Introduction

Wheat (Triticum aestivum L.) is grown on more areas of land than any other crop plant and is a staple food for 40% of the world’s population, providing 20% of the calories consumed (Gill et al. 2004). It is an allohexaploid species (2n=6x=42, genome formula AABBDD) that resulted from two independent hybridization events. The first lead to the allotetraploid species T. turgidum, in which the A genome of the diploid species T. urartu and the B genome of a yet unidentified diploid species were combined. Aegilops speltoides is the living species believed to be most closely related to the B genome of hexaploid wheat (Daud and Gustafson 1996). The second hybridization occurred only about 8,000 years ago and involved T. turgidum and Ae. tauschii, a diploid species carrying the D genome, and resulted in hexaploid wheat (Feldman et al. 1995). Due to the presence of three closely related (homoeologous) subgenomes (A, B and D), sequence redundancy is a characteristic feature of the wheat genome. This, together with a genome size of 16,974 Mb/1C (Bennett and Smith 1991)—which is fivefold the human genome, and more than 80% of repetitive DNA (Smith and Flavell 1975)—represents a great challenge for gene discovery and genome sequencing efforts.

One way to simplify the molecular genetic analyses is to create resources specific for the three subgenomes of wheat. Bacterial artificial chromosome (BAC) libraries, which represent one of the most useful genome resources (Zhang and Wu 2001), were constructed for the D genome of T. tauschii (Moullet et al. 1999) and for the Am genome of domesticated diploid wheat T. monococcum (Lijavetzky et al. 1999), which is related to T. urartu [known to be the putative donor of the A genome of wheat (Dvorak et al. 1993)]. More recently, Cenci et al. (2003) constructed a BAC library for tetraploid durum wheat (T. turgidum var. durum, genome formula AABB). These BAC resources turned out to be very useful and facilitated gene cloning (Faris et al. 2003; Ling et al. 2003; Yan et al. 2003) and discovery of organization of the wheat genome into gene-rich islands, where gene densities are similar to that of Arabidopsis, and gene-poor regions loaded with retroelements (Wicker et al. 2001; Brooks et al. 2002). These observations showed that despite the initial scepticism, map-based cloning is viable in wheat.

Unfortunately, the diploid BAC libraries have limitations stemming mainly from a dynamic and rapid genome evolution in wheat species and changes induced by polyploidization (Ozkan et al. 2001). For example, a comparison of a T. tauschii BAC clone sequence with orthologous regions of the wheat A and B genomes revealed rapid sequence divergence (Anderson et al. 2003). Sequencing and annotation of two physical contigs covering orthologous glutenin loci on chromosome 1AS of T. monococcum and T. turgidum var. durum showed that sequence conservation was limited to the regions containing glutenin genes. The rest of the two contigs, representing 90% of the total sequence, was not conserved (Wicker et al. 2003). These data imply that BAC libraries created from wheat progenitors and related species cannot fully replace BAC resources made from hexaploid wheat. The recent progress in BAC cloning methodology facilitated preparation of several BAC libraries for hexaploid wheat (Allouis et al. 2003; Liu et al. 2000; Nilmalgoda et al. 2003; Ma et al. 2000). Although they represent invaluable tools, due to the large number of clones needed to achieve a desired genome coverage (0.5–1×106 clones), their maintenance, handling and screening represents a formidable challenge.

A strategy is needed to reduce the complexity of the huge genome of hexaploid wheat, without sacrificing the access to large genomic DNA clones. The creation of chromosome and chromosome-arms-specific BAC libraries is an attractive possibility. Representing only a few per cent of the total genome, they would open completely new horizons for wheat genome analysis and gene cloning. With this in mind, Vrána et al. (2000) developed protocols that could be used to purify wheat chromosomes in large numbers and high purity, using laser flow cytometry. Subsequently, Kubaláková et al. (2002) demonstrated that any chromosome arm of wheat could be sorted from telosomic lines, and Šimková et al. (2003) developed a protocol for preparation of high-molecular-weight (HMW) DNA suitable for cloning from sorted chromosomes. These advances facilitated the development of protocols for the cloning of individual flow-sorted chromosomes into BAC libraries and the construction of the first chromosome-specific BAC library in plants: the chromosome 3B library (Šafář et al. 2004). Consisting of only 67,968 clones, with an average insert size of 103 kb, the TA-3B library represents an easily manageable resource with over a sixfold coverage of the chromosome (Šafář et al. 2004).

In this work, we continued cloning of hexaploid wheat chromosomes and report the construction of the second BAC library ever produced from flow-sorted wheat chromosomes. The library was produced for chromosomes 1D, 4D and 6D that were sorted from cv. Chinese Spring. Creation of the subgenomic library confirms a possibility to prepare BAC libraries from DNA of sorted wheat chromosomes. The availability of the subgenomic BAC library will aid in comparative genome analysis as well as gene cloning from the three wheat D-genome chromosomes.

Materials and methods

Plant material

Seeds of hexaploid wheat T. aestivum L. cv. Chinese Spring (2n=6x=42) were germinated in the dark at 25±0.5°C on moistened filter paper in glass petri dishes for 3 days to achieve optimal root length (2–3 cm). In total, 4,000 seeds were germinated and used in batches of 20 for preparation of chromosome suspensions.

Chromosome analysis and sorting

Suspensions of intact mitotic chromosomes were prepared from synchronized root tip meristems as described by Vrána et al. (2000). Mitotic chromosomes were released into the modified isolation buffer (IB) of Šimková et al. (2003), stained by 2 μg/ml 4′,6-diamidino-2-phenylindole (DAPI) and analysed using the FACSVantage flow cytometer (Becton Dickinson, San José, Calif., USA). In order to sort chromosomes 1D, 4D and 6D, a sort window was set on a dot plot of fluorescence pulse area versus fluorescence pulse width. Chromosomes were sorted in aliquots of 1.0×105 into 160 μl 1.5× IB buffer. In order to check the purity in sorted fractions, 2,000 chromosomes were sorted at regular intervals onto a microscope slide into a 15-μl drop of PRINS buffer supplemented with 5% sucrose and air-dried. Sorted chromosomes were identified after fluorescence in situ hybridization (FISH), with a probe for the Afa repeat (Kubaláková et al. 2002).

BAC library construction

Preparation of HMW DNA and BAC library construction were performed as described previously (Šafář et al. 2004). Briefly, each batch of 105 flow-sorted chromosomes was pelleted, and the chromosomes were embedded in 12 μl of low-melting-point agarose. In total, 100 agarose miniplugs, representing 1×107 sorted chromosomes, were prepared and used. Chromosomal DNA was partially digested with HindIII and size-selected by pulsed field gel electrophoresis (PFGE). Four regions of the gel corresponding to approximately 50–100 kb (H0 fraction), 100–150 kb (H1 fraction), 150–200 kb (H2 fraction) and 200–250 kb (H3 fraction) were excised. HMW DNA was electroeluted and ligated into a dephosphorylated vector pIndigoBAC (Caltech, Pasadena, Calif., USA), prepared according to Chalhoub et al. (2004). Escherichia coli ElectroMAX DH10B competent cells (GIBCO–BRL, Gaithersburg, Md., USA) were transformed by electroporation. The library was ordered in 384-well plates in FM medium (Woo et al. 1994). The plates were incubated at 37°C overnight, replicated and stored at −80°C.

BAC library characterization

Insert size of BAC clones selected randomly from all four ligations was determined as described previously (Šafář et al. 2004). The library was pooled and screened using microsatellite (SSR) markers specific for 1D, 4D and 6D chromosomes as well as markers specific for the remaining 18 wheat chromosomes (Tables 1, 2). Screening was done also with mit3 and mit4 markers specific for mitochondrial DNA (B. Chalhoub et al., unpublished data) and markers specific for chloroplast genes rbcL and psbA (Ogihara et al. 2000). BAC clones from each of the 384-well plates were pooled, pelleted and resuspended in 4 ml TE (10:1 mM) buffer. Bacterial suspensions were lysed at 95°C for 30 min and pelleted at 3,000 g for 60 min. Supernatant was diluted 20-fold by deionized water and used for PCR. The reaction mix (10 μl) consisted of 2 μl template DNA, 1× PCR buffer, 1.5 mM MgCl2, 0.2 mM dNTPs, 1.25 μM primers and 0.5 U AmpliTaq DNA polymerase (Roche, Mannheim, Germany). PCR reaction was performed as follows: nine cycles consisting of 5 min at 94°C, 12 s at 65°C and 18 s at 72°C; 35 cycles of 12 s at 94°C, 12 s at 55°C and 18 s min at 72°C; and final extension at 72°C for 7 min. Presence of PCR products was checked by electrophoresis on 4% MetaPhor agarose gel (FMC Bioproducts, Philadelphia, Pa., USA).

Table 1 Library screening with SSR markers specific for chromosomes 1D, 4D and 6D
Full size table
Table 2 Library screening with SSR markers specific for potentially contaminating chromosomes
Full size table

FISH with ‘low copy’ BAC clones

A total of 1,536 BAC clones of the H0 sublibrary (four 384-well plates) were spotted onto two 8×12-cm Hybond-N+ filters (AP Biotech, Piscataway, N.J., USA). The filters were hybridized with digoxigenin-labelled C0t-1 fraction of genomic DNA (Zwick et al. 1997) to reveal clones with low proportions of repetitive sequences. Twenty-one ‘low-copy’ BAC clones (Table 3) showing no or weak hybridization signals were chosen for FISH. Preparations of digoxigenin-labelled probes from BAC inserts and FISH on flow-sorted chromosomes were performed as described by Šafář et al. (2004). When needed, dispersed repetitive DNA was blocked by C0t-1 fraction at various ratios of probe/blocking DNA. In order to identify individual chromosomes, a biotin-labelled probe for the Afa repeat was included in the hybridization mix. The sites of probe hybridization were detected using anti-digoxigenin-FITC and streptavidin-Cy3, and the chromosomes were counterstained by DAPI. The preparations were evaluated using an Olympus BX60 microscope (Olympus, Tokyo, Japan) equipped with optical filter sets appropriate for DAPI, fluorescein and Cy3 fluorescence. The images of DAPI, fluorescein and Cy3 fluorescence were acquired separately with a b/w CCD camera, which was interfaced to a PC running the ISIS software (Metasystems, Altlussheim, Germany).

Table 3 Cytogenetic mapping of putatively ‘low copy’ clones selected from the subgenomic bacterial artificial chromosome (BAC) library specific for chromosomes 1D, 4D and 6D. Clones determined after hybridization with C0t-1 fraction of genomic DNA
Full size table

Results

Flow cytometric analysis of relative fluorescence intensity of DAPI-stained chromosomes isolated from cv. Chinese Spring resulted in a distribution (flow karyotype) on which four peaks (I, II, III and 3B) could be discriminated (Fig. 1). While the peak 3B represented solely the chromosome 3B, peaks I, II, and III represented groups of 3, 7 and 10 chromosomes, respectively (Kubaláková et al. 2002). Chromosomes 1D, 4D and 6D from peak I (Fig. 1, insert) were sorted to prepare the subgenomic BAC library. In total, 107 chromosomes were sorted and used to prepare 100 miniplugs (105 chromosomes each), representing approximately 13.4 μg of DNA. The sorting took 18 working days, and 4,000 seeds were needed to prepare 200 samples of chromosome suspensions. Purity, in sorted fractions as determined microscopically after FISH with the Afa repeat, was 91%. It was found that sorted chromosomes were contaminated at an average frequency of 9% by various chromosomes and chromosome arms, without an apparent prevalence of certain chromosome types.

Fig. 1
figure 1

Histogram of relative fluorescence intensity (flow karyotype) obtained after flow cytometric analysis of 4′,6-diamidino-2-phenylindole (DAPI)-stained suspension of chromosomes prepared from hexaploid wheat cv. Chinese Spring. The flow karyotype consists of three composite peaks (I, II and III) representing groups of chromosomes and a clearly discriminated peak representing chromosome 3B. Examples of chromosomes 1D, 4D and 6D, which were sorted from peak I, are shown in the insert after fluorescence in situ hybridization (FISH) with a probe for Afa repeat (black signals) that facilitated their identification

Full size image

Four BAC sublibraries were created from four independent ligation reactions with different size classes of partially digested chromosomal DNA, namely the sublibraries H0, H1, H2 and H3. Ninety-six BAC clones were selected randomly from the four sublibraries and used to estimate the average insert size using PFGE (Fig. 2a). The most abundant sublibrary, H0, contained 42,240 clones and its average insert size was 70 kb. The H1 sublibrary had 33,024 clones and an average insert size of 100 kb. The least abundant sublibrary with only 3,456 clones was H2, with 90 kb of average insert size, while the H3 sublibrary had 8,448 clones, and the average insert size was 92 kb. The complete BAC library (designated as TA-subD) consisted of 87,168 clones ordered in 227 384-well plates, with an average insert size of 85 kb. Of the clones, 53% showed inserts larger than 100 kb; only 29% of BAC clones have inserts shorter than 75 kb (Fig. 2b).

Fig. 2
figure 2

Insert-size analysis of randomly chosen clones from the TA-subD bacterial artificial chromosome (BAC) library. a BAC DNA was digested with NotI enzyme to release the insert, separated by pulsed field gel electrophoresis (PFGE) and stained with ethidium bromide. Lanes M contain λPFGE marker. The 7.5-kb band represents the BAC vector. b Insert-size distribution of 96 randomly chosen BAC clones from the BAC library. Fifty-three per cent of BAC clones contain inserts larger than 100 kb; only 29% of BAC clones have inserts shorter than 75 kb

Full size image

Considering the wheat genome size of 16,974 Mb/1C (Bennett and Smith 1991) and relative chromosome lengths given by Gill et al. (1991), molecular sizes of the three chromosomes were estimated as 606 Mb for 1D, 649 Mb for 4D and 714 Mb for 6D. Together they represent 1,969 Mb, or 11.6%, of the hexaploid wheat genome. If a contamination with other chromosomes of 9% is considered, the subgenomic BAC library represents 3.4 size equivalents of the three chromosomes. Using the formula of Clarke and Carbon (1976), a probability of recovering any DNA sequence present on 1D, 4D and 6D chromosomes in the library was estimated to be 96.5%. Conversely, a probability of finding DNA sequences from chromosomes that contaminated the sorted fraction was calculated to be 4.8%.

With the aim to confirm the chromosome specificity and verify the genome coverage of 3.4-fold, the subgenomic BAC library was screened by PCR with SSR markers. The library was pooled into 227 pools, each pool representing one 384-well plate. The screening was done using 19 SSR markers specific for chromosome 1D, 11 SSR markers specific for chromosome 4D, and 20 SSR markers specific for chromosome 6D. The number of pools that gave PCR products of expected length with each SSR marker is listed in Table 1. Based on these results, the coverage for chromosomes 1D, 4D, and 6D was estimated to be 5.95-fold, 6.34-fold and 5.7-fold, respectively, indicating equal representation of the three chromosomes in the library. The coverage of the complete BAC library estimated using the SSR markers was thus sixfold. The extent of contamination with DNA of other 18 chromosomes was estimated using 36 SSR markers. Only a scarce contamination by DNA of chromosomes 1A, 2D, 5B and 7B was detected (Table 2).

While no positive pools were found after screening with mitochondria-specific primers, few positive pools were found after screening with markers specific for chloroplast DNA. However, simultaneous occurrence of products with both chloroplast primer pairs in one pool was observed only in 10% of the positive pools. Considering the fact that the rbcL and psbA genes are separated only by 55 kb in the wheat chloroplast genome (Ogihara et al. 2000), our observation indicates that sequences integrated in nuclear genome were detected rather than DNA coming from organelles.

One of the uses of the library will be the construction of detailed cytogenetic maps. To test this approach, 1,536 clones from the H0 sublibrary were hybridized with C0t-1 fraction of genomic DNA to reveal clones with the lowest amount of repetitive DNA and potentially containing low- or single-copy sequences. Twenty-one of these were chosen for cytogenetic mapping using FISH. Although the clones were selected for low proportion of repeats, FISH with 13 of them resulted in complete labelling of all wheat chromosomes (Table 3). FISH localization of the eight remaining BAC clones on chromosomes 1D, 4D and 6D is shown in Fig. 3 and summarized in Table 3. Two of the BAC clones (110/C2 and 109/O19) localized in centromeres of all wheat chromosomes. It is interesting to note that these clones localized specifically to centromeres of all barley chromosomes, too (not shown). Two other BAC clones (214/J1 and 214/O2) gave specific signals on only one of the D-genome chromosomes. Four clones (214/C8, 109/F2, 110/K6 and 109/C22) localized to discrete loci on two of the three D chromosomes from which the library was created. BAC clone 110/K6 hybridized to chromosome 4A and BAC clone 109/F2 to chromosome 7D. While the BAC clone 214/C8 gave no signal on the other 19 wheat chromosomes, BAC clone 109/C22 hybridized also to six other wheat chromosomes (Table 3).

Fig. 3
figure 3

Physical mapping of selected BAC clones, using FISH on wheat chromosomes. The addresses of individual BAC clones are given above the lines. For each BAC probe, two representative examples of FISH are given. In each pair of chromosomes on the figure, the chromosome on the left shows hybridization with a BAC probe detected by fluorescein (yellow signals); the chromosomes were counterstained by propidium iodide (red). The chromosomes on the right show simultaneous hybridization with the BAC probe detected by fluorescein (yellow signals) and a probe for the Afa repeat, detected with Cy3 (red), which facilitated chromosome identification; the chromosomes were counterstained by DAPI (blue)

Full size image

Discussion

This paper reports on the construction of a subgenomic BAC library (TA-subD) specific for chromosomes 1D, 4D and 6D of hexaploid wheat. This is the second BAC library ever produced from particular plant chromosomes, after the chromosome 3B-specific BAC library created by Šafář et al. (2004). Its construction proves that chromosomes purified using flow cytometry may be used for preparation of chromosome-specific BAC libraries in wheat and other plant species as the methods for chromosome sorting are available for other cereal and legume species (Doležel et al. 2004).

The library screening with SSR markers proved a representative coverage of the three chromosomes and their equal representation in the library. The genomic coverage of sixfold estimated after screening with SSR markers was higher compared to that estimated with average insert size (3.4-fold). A similar phenomenon was observed for the TA-3B BAC library specific for wheat chromosome 3B (Šafář et al. 2004). As only PCR products with expected length were considered, the sixfold coverage could not be due to the presence of duplicated loci. Therefore, as with the TA-3B BAC library (Šafář et al. 2004), the discrepancy could either be due to overestimation of molecular size of the three chromosomes or to underestimation of the average insert size. The subgenomic library compares well with large-insert DNA libraries created using genomic DNA of hexaploid wheat. Liu et al. (2000) reported on the construction of a transformation-competent artificial chromosome (TAC) library with average insert size of 54 kb and 3.07-fold genomic coverage. A BAC library created by Nilmalgoda et al. (2003) with average insert size of 79 kb, has a similar coverage of 3.1-fold. Recently, we achieved a much better coverage (9.3-fold) by constructing a BAC library from cv. Chinese Spring, consisting of 1.2×106 clones with average insert sizes of 130 kb (Allouis et al. 2003) as well as from cv. Renan (unpublished results). However, maintenance and handling of these two libraries with a huge number of BAC clones remains a non-trivial task.

Specificity of the TA-subD BAC library for chromosomes 1D, 4D and 6D was independently confirmed by two approaches. First, the sorted fractions were periodically checked microscopically for the presence of contaminating chromosomes. On average, 9% of chromosomes other than 1D, 4D and 6D were detected after FISH with the Afa repeat. This approach facilitates unambiguous identification of the D-genome chromosomes (Kubaláková et al. 2002). The second approach involved library screening with 36 SSR markers specific to chromosomes other than 1D, 4D and 6D (one marker per each chromosome arm). Although only a small number of SSR markers were used, in principle the results confirmed low contamination as determined by FISH. Only markers specific for chromosomes 1A, 2D, 5B and 7B were detected, demonstrating the high specificity of the library (Table 2).

One of the uses of the BAC library will be the construction of a detailed molecular cytogenetic map. Cytogenetic maps consisting of a series of molecular landmarks aid in construction of physical contig maps by resolving the order of local contigs and closely spaced clones, and confirming the precise physical position of centromere and markers at the end of linkage groups (Islam-Faridi et al. 2002; Kim et al. 2002). Cytogenetic maps are also very useful to study chromosome organization and karyotype evolution (Jackson et al. 2000; Ziolkowski and Sadowski 2002). Our results with a limited number of BAC clones indicate that the subgenomic BAC library will be a useful source of D-genome cytogenetic markers. Cytogenetically mapped low-copy BAC clones could serve as anchor makers and facilitate construction of physical contig maps (Stein et al. 2000; Faris et al. 2003). The reason that most of the ‘low-copy BAC clones’ as identified by no or weak hybridization with C0t-1 fraction of genomic DNA were not chromosome specific is probably due the fact that these clones still contain few but highly repeated sequences which is not surprising considering the wheat genome structure (Akhunov et al. 2003; Chantret et al. 2004).

Our study indicates a possibility to isolate BAC clones specific to single-wheat chromosomes. This observation is surprising when considering the presence of three homoeologous chromosome sets in hexaploid wheat, and may be explained by a rapid and dynamic evolution of the Triticum genomes and divergence of homologous regions (Anderson et al. 2003; Wicker et al. 2003). Recently, Zhang et al. (2004) identified several BAC clones whose fragments hybridized only to specific chromosome pairs of wheat, thus supporting our observations. Localization of some BACs to non-homologous chromosomes could be due to interchromosomal duplications. Alternatively, these observations may reflect the structural changes accompanying the evolution of the hexaploid wheat karyotype (Gill et al. 1991). As none of the BAC clones has been sequenced, the nature of the presumably homologous sequences is not known. Identification of BAC clones containing specific repeats, such as the two clones containing centromere-specific repeats, can facilitate the analysis of molecular organization of functional domains of wheat chromosomes.

We expect that the most attractive application of the library will be gene cloning, as a number of agronomically interesting genes are known to be localized on one of the three D-genome chromosomes. These include seed storage protein genes (glutenins and gliadins) on chromosomes 1D and 6D (Anderson and Greene 1997; Sabelli and Shewry 1991; Spielmeyer et al. 2000) and genes related to disease resistance on chromosome 1D (Brooks et al. 2002). The Rht-D1 dwarfing gene was mapped to the short arm of chromosome 4D (King et al. 1996; Cadalen et al. 1998; Sourdille et al. 1998). A major gene conferring aluminium tolerance as well as a locus controlling the K+/Na+ discrimination in saline environments were localized on the long arm of chromosome 4D (Rodriguez Milla and Gustafson 2001). Although the interesting genes may be cloned using the existing genomic BAC libraries (Moullet et al. 1999; Cenci et al. 2003; Allouis et al. 2003), the newly created BAC library specific for chromosomes 1D, 4D and 6D, which represents only about 11% of the whole genome, offers a targeted and hence more efficient and faster approaches for map-based cloning.

Targeting the gene-rich regions of the D genome is another attractive use of the library. This could be achieved by screening the library with appropriate probes such as ESTs to reveal gene-rich BAC clones. Alternatively, the library could be pooled and the chromosome DNA used for Cot analysis (Peterson et al. 2002; Yuan et al. 2003) to generate a shotgun library enriched in low-copy sequences for targeted sequencing of the genic fraction. Although the analysis could also be performed using genomic DNA, working with only about one tenth of the wheat genome should simplify the sequence annotation and assembly.

Rapid production of complete chromosome contigs is now possible using the subgenomic TA-subD BAC library. Their production would be greatly simplified by already existing genetic (Boyko et al. 1999, 2002) and physical (http://wheat.pw.usda.gov/PhysicalMapping/) maps for the diploid D genome of Ae. tauschii, which could serve as a framework and also by genetic/physical wheat map relationships that were recently published (Sourdille et al. 2004). The construction of complete BAC contigs for the three D-genome chromosomes has a potential to transform the positional cloning of their genes into a trivial bioinformatic exercise (Lagudah et al. 2001). The TA-subD library will allow appreciation of the D-genome evolution in the hexaploid context and facilitate studies to explain the lack of polymorphism and difficulties associated with the development of genetic maps for the D genome. Is it not clear whether this is because the D genome is more conserved (if yes, why?): because it is less prone to evolution or because the recombination in the D genome is inhibited? A physical approach should allow answering these questions.

In conclusion, this study confirmed that the large genome of hexaploid wheat may be dissected into small and defined fractions corresponding to one or few chromosomes. While this paper reports on a BAC library specific for chromosomes 1D, 4D and 6D, we previously reported the preparation of the TA-3B BAC library specific for chromosome 3B (Šafář et al. 2004). These unique resources can greatly facilitate genome sequencing and gene cloning as well as comparative analyses aimed at revealing the genome changes accompanying the evolution of polyploid wheat. The work is in progress to construct BAC libraries specific for other chromosomes of hexaploid wheat.