Construction of cytogenetic map of Gossypium herbaceum chromosome 1 and its integration with genetic maps

Background Cytogenetic map can provide not only information of the genome structure, but also can build a solid foundation for genetic research. With the developments of molecular and cytogenetic studies in cotton (Gossypium), the construction of cytogenetic map is becoming more and more imperative. Results A cytogenetic map of chromosome 1 (A101) of Gossypium herbaceum (A1) which includes 10 bacterial artificial chromosome (BAC) clones was constructed by using fluorescent in situ hybridization (FISH). Meanwhile, comparison and analysis were made for the cytogenetic map of chromosome 1 (A101) of G. herbaceum with four genetic linkage maps of chromosome 1 (Ah01) of G. hirsutum ((AD)1) and one genetic linkage map of chromosome 1 of (A101) G. arboreum (A2). The 10 BAC clones were also used to be localized on G. raimondii (D5) chromosome 1 (D501), and 2 of them showed clear unique hybridized signals. Furthermore, these 2 BAC clones were also shown localized on chromosome 1 of both A sub-genome and D sub-genome of G. hirsutum. Conclusion The comparison of the cytogenetic map with genetic linkage maps showed that most of the identified marker-tagged BAC clones appearing same orders in different maps except three markers showing different positions, which might indicate chromosomal segmental rearrangements. The positions of the 2 BAC clones which were localized on Ah01 and Dh01 chromosomes were almost the same as that on A101 and D501 chromosomes. The corresponding anchored SSR markers of these 2 BAC clones were firstly found to be localized on chromosome D501 (Dh01) as they were not seen mapped like this in any genetic map reported.


Background
Allopolyploid formation in plants simply reflects promiscuity of plants or provides a selective advantage in survival has been long debated [1]. Most angiosperm (approximately 70 percent) was thought to have incurred one or more polyploidization events in their history [2]. The genus Gossypium, which comprises of 52 species had been classified into 8 diploid (2n = 2× = 26) genomes, i.e. A, B, C, D, E, F, G, and K, and as well one allotetraploid (2n = 4× = 52) genome, i.e. AD [3,4]. In the early stages of the genus evolution, A genome diploids and D genome diploids diverged, acquiring a 2-fold difference in genome size subsequently. The genome-size difference is thought probably associated with the expansion and contraction of repetitive elements including transposons, whereas the homoeologous sequences flanking the genes are highly conserved [5][6][7]. These two divergent genomes later became reunited with allopolyploid formation approximately 1-2 million years ago (MYA) [8], in the New World by showing hybridization between a maternal Old World "A" genome taxon resembling Gossypium herbaceum (2n = 2× = 26) and a paternal New World "D" genome taxon resembling Gossypium raimondii or Gossypium gossypioides (both 2n = 2× = 26) [9,10]. It is old enough for sequence divergence but relatively recent to maintain genome stability [11]. This proves that cotton (Gossypium spp) is not only an important economic plant, but also an excellent system for study on genomic organization, genome-size variation, genome evolution and polyploidization in plants. Therefore, due to its importance in scientific research, sequencing the cotton genome and carrying out genomic research are highly necessary.
The genetic linkage map, which can indicate the orders of markers on chromosomes, is a powerful molecular tool to dissect the genome. Nevertheless, the exact cytogenetic positions of the genetic loci and genomic sequences on the chromosomes shall be not accurately identified by genetic linkage map, due to the unequally distributed crossovers on chromosome arms. Thus loci physically far apart on chromosomes can be linked tightly on genetic linkage maps and vice versa [12]. Cytogenetic map, which accurately represent direct inspection of distinctive loci on chromosomes, can compensate for the disadvantage of genetic linkage map. It not only can provide information on the structure and evolution of genomes but also is useful in the synteny comparison between relative genomes, especially for complex-genome organisms which has large amounts of repetitive DNA, such as maize and wheat [13]. So integrating genetic linkage map with cytogenetic map can provide new insight in chromosome structure. However, cytogenetic maps are nascent and relatively underdeveloped in many plants especially in cotton, despite the long history of cytology [14]. Thus, the majority of the genetic linkage maps of cotton were not integrated with any type of physical map.
Though the application of FISH in cotton has lagged behind other crops, the development of target DNA has greatly improved its resolution and promoted its application in cytogenetic study [36,37]. Chromosomes of many cotton species were identified and many researches of genome structure of cotton were achieved by using FISH [38][39][40]. Cytogenetic maps of cotton, in which A h 12 and D h 12 homologous chromosomes including 15 and 21 SSR-derived BACs, respectively, had also been developed, and the integration between cytogenetic maps and genetic linkage maps has been comprehensively analyzed [11]. However the cytogenetic maps of cotton are far from completed. In this paper, a cytogenetic map of chromosome 1(A 1 01) of G. herbaceum including 10 BAC clones was constructed by using BAC-FISH mapping method and as well the relationship between the cytogenetic map and genetic linkage maps has been analyzed. Furthermore 2 of these 10 clones were apparently localized on chromosome 1 (D 5 01) of G. raimondii as well as on chromosomes D h 01 and A h 01of G. hirsutum. And also, the relationship of the 2 BAC clones positions in different chromosomes has been analyzed subsequently.

Results
Construction of cytogenetic map of G. herbaceum chromosome 1(A 1 01) To construct a cytogenetic map of G. herbaceum chromosome 1(A 1 01), Pima 90-53 BAC library was screened using sixteen SSR markers. The SSR markers were selected from five genetic linkage maps and used to screen the BAC library. A total of 47 positive BAC clones were identified ( Table 1). The chromosome-specific BAC clone 52D06 was used to identify chromosome 1(A 1 01) of G. herbaceum [41,42], and all positive BAC clones of 16 SSR markers were selected for FISH mapping. Clones of 6 SSR markers which showed repetitive FISH signals in mitotic metaphase were discarded and clones of one SSR marker which showed strong signals on other chromosomes but not on chromosome 1 were also discarded (data not shown). The remaining clones of nine SSR markers with unique clear hybridization signals ( Figure 1) were used for FISH mapping. Seven of them were localized on the long arm and the other two were localized on the short arm. The FISH signals of each BAC clone from more than 10 cells with clear chromosome spreads were measured and the relative positions of FISH signals were computed. More than 10 cells with clear chromosome spreads were chosen to distinguish the position of the centromere and to compute the exact cytogenetic position of the centromere. The data was analyzed to construct the cytogenetic map of G. herbaceum chromosome 1(A 1 01) ( Figure 2).

Integration and analysis of clone positions across maps
In order to analyze the relationship between genetic linkage maps and the constructed cytogenetic map, the composite alignment was constructed to compare the FISH map directly to the genetic linkage maps of G. hirsutum ((AD) 1 ) and G. arboreum (A 2 ) chromosome 1 ( Figure 2). The alignment provided a global view of the relationship between the genetic positions of the SSR markers and the cytogenetic positions of the BAC clones anchored by corresponding SSR markers.
The comparative analysis showed that the order of most selected marker-anchored BAC clones was the same as in the linkage map, except three BAC clones anchored by markers NAU3135, BNL3580 and NAU4044 appeared different positions. These three BAC clones were very tight on the cytogenetic map, but the positions of corresponding SSR markers of them were visibly different from those on the genetic map C. Two BAC clones 305A19 and 260J3 anchored by SSR markers NAU2015 and BNL2921 respectively had good corresponding locations between genetic maps and the cytogenetic map. But the locations of other BAC clones and their corresponding markers had obvious discrepancies, especially BAC clone 216B15 anchored by NAU4891 showed a maximum difference of 32.7 RMP units. BAC clone 85P13 anchored by NAU3135 and BAC clone 426C18 anchored by TMB0062  The BAC clone anchored by SSR marker NAU2285 showed two clear signals on chromosome 1(A 1 01), and the location of one signal was concordant with marker position in the corresponding genetic map. SSR marker HAU076 is in a short linkage group which only has two markers on chromosome 1(A h 01) according to genetic map E [47], here its corresponding BAC clone was localized on chromosome 1(A 1 01) and its exact physical position was obtained afterwards.
Two clones on chromosome A 1 01 were localized on chromosome D 5

01, A h 01 and D h 01
Chromosome-specific BAC clone 389K13 and 48F11 were used to identify chromosome 1(D 5 01) of G. raimondii and chromosome 15(D h 01) of G. hirsutum, respectively [41,48]. Chromosome-specific BAC clone 52D06 was used to identify chromosome 1(A h 01) of G. hirsutum [41]. All BAC clones of nine SSR markers distributed on chromosome 1(A 1 01) of G. herbaceum were selected for FISH mapping. Clones of seven SSR markers either showed repetitive FISH signals or no FISH signal in mitotic metaphase (data not shown). The remaining clones of two SSR markers showed unique clear hybridization signals on chromosome D 5 01 (Figure 3). Clone 305A19 anchored by NAU2015 was localized in the long arm and clone 216B15 anchored by NAU4891 was localized in the short arm. We measured the FISH signals of each BAC clone from more than 10

Comparison and analysis of clone positions across different chromosomes
By analyzing the relationship between different maps, a whole insight of chromosome structure can be displayed. And by comparing maps of homoeologous chromosomes, well conserved regions between homoeologous chromosomes can be detected. Here the positions of the two clones were compared across different chromosomes ( Figure 4).
The comparative analysis showed that the order of the three marker-anchored (NAU2015, NAU4891, BNL3902) BAC clones on chromosome D 5 01 was same as in the whole-genome DNA marker map, while the distance between the markers has little discrepancies. The positions of BAC clones tagged SSR markers NAU2015 and The acquisition of the locations of centromeres in different chromosomes will be helpful to the study of their structure. In our research more than 10 cells were chosen with clear chromosome spreads to distinguish the positions of the centromeres of different chromosomes. And of these ten SSR markers used in this research, the centromere position was expressed as follows: G. herbaceum chromosome 1(A 1 01) was located between SSR marker NAU2474 and BNL2921, BNL2921 is the nearest SSR marker to the centromere. As crossovers are always low at the region near the centromere, the near markers between the two markers in genetic maps may be considered physically far apart.

Construction of cytogenetic map for G. herbaceum chromosome 1 (A 1 01)
Cytogenetic map can provide information on the structure and evolution of genomes [49]. It can compensate for the disadvantages of genetic map which based on recombination frequencies that vary widely in relation to physical distances. However, few researches on cytogenetic mapping of cotton have been reported. To our knowledge, the cytogenetic map of G. herbaceum chromosome 1 (A 1 01) reported here is the first case. The cytogenetic map including the ten BAC clones are all anchored by SSR markers. Among the ten BAC clones, eight of them were localized on the long arm, and two of them were localized on the short arm. As reported in the documentation, using conserved genes or clones in related species have proved a successful strategy in examining genome structures and relationships and predicting genes locations and DNA markers [21]. So the ten BAC clones in this article also can be used in the related species, Theobroma cacao even from which Gossypium was diverged 18-58 million years ago [50], in order to help study the relationship between these species. As the resolution of metaphase FISH is limited and highly repetitive regions which may not be represented in the BAC library exist, the coverage of the BAC map is still thought uncompleted. However, this will not influence on the cytogenetic map be used to provide a solid theoretical foundation and as an useful method to analyze the structure of G. herbaceum chromosome 1(A 1 01). At the same time, a reliable backbone was built up to guide in sequencing the G. herbaceum chromosome 1(A 1 01).

Integration of genetic linkage maps and cytogenetic map
Genetic linkage map can provide some valuable information of the genome, but they can only provide little information about the exact cytogenetic locations of markers and the distance between them. In the present work, a cytogenetic map of G. herbaceum chromosome 1(A 1 01) was constructed and was shown integrated with three genetic maps of G. hirsutum and one genetic map of G. arboretum, using a standardized map unit-relative map position (RMP). It is thought exciting that this integration of maps can provide new insights into structure and organization of G. herbaceum.
The comparative map alignments showed that the distance of marker NAU2474 and BNL2921 on genetic map C was 12.3 RMP (16.3 cM) but on the cytogenetic map was 42.3 RMP. This phenomenon also happened between other markers (BNL2921 and TMB0062, BNL2921 and NAU4891). The discrepancies between genetic and cytogenetic distances may be because of the suppression of genetic recombination. The same observations were reported in many other plant including maize [51], wheat [52], barley [53], and sorghum [54,55]. In particular chromosomes, the distance between two markers which is more than 50% in genetic map can be very near in cytogenetic map because of the suppression of genetic recombination [53][54][55]. The order of positions of Marker NAU3135 BNL3580 and NAU4044 in the cytogenetic FISH map are different from the order in the genetic map C. Marker NAU3135 and NAU4044 on genetic map C were very tightly linked. The discrepancies in markers order might indicate small chromosome rearrangements, but it may also be due to the insufficient resolution of the FISH cytogenetic map in specific regions of the chromosome. The BAC clone tagged SSR marker NAU2285 showed 2 clear signals on chromosome 1 (A 1 01), the location of one of the signals was well corresponding with its location on genetic maps. This may indicate that the conversed region has a translocation within chromosome 1 (A 1 01). Based on the research work of Zhang [47], HAU076 is in a short linkage group that has only two markers on chromosome 1(A h 01). In our study, its corresponding BAC clone 378J7 was localized cytogenetically on chromosome 1 (A 1 01). The location of the SSR marker will be very useful for cotton genome organization architecture.
Two SSR markers first localized on chromosome D 5 01 and D h 01 Although more than 30 genetic maps of cotton had been constructed [56], most of them used different mapping populations with different population sizes. As a result, the genetic markers were often mapped at different genetic positions, even at different chromosomes, in different maps. So this makes it very difficult to study gene distribution, chromosome evolution, and map-based cloning between different populations [57]. The cytogenetic map is known inherently informative as it can represent direct positions of location chromosomes. It can help to resolve the locations of markers that couldn't be resolved on genetic linkage maps, especially markers linked closely with very low rate of recombination chromosome sections. The location of two SSR markers which had never been included on chromosome D 5 01 or D h 01 in any genetic map by using BAC-FISH method displayed here demonstrated to be a good example. In another word, the location of the two SSR markers can not only consummate genetic maps of D 5 01 and D h 01, but also means the chromosome section which linked closely to the two SSR markers, may infer a very low rate of recombination. The first location of the two SSR markers on chromosome D 5 01 and D h 01 will facilitate the whole-genome physical alignment, sequencing, and mapping of genes for cotton improvement.

Two conserved regions detected between chromosomes D 5 01 (D h 01) and A 1 01 (A h 01)
It is known that the allotetraploid cotton contains two sub-genomes originated from the related ancestor species with nearly two fold sizes difference. The comparison analysis between homoeologous loci and chromosomes in tetraploid cotton showed that most duplicated genes in allopolyploid cotton evolved independently of each other [58], and some regions were highly conserved [5,59]. In the research of Wang et al. [11], ten of the eleven BAC clones which were localized on chromosome A h 12 also showed signals on chromosome D h 12; fourteen of the twenty BAC clones which were localized on chromosome D h 12 also showed signals on chromosome A h 12, suggesting that the majority of this pair of homoeologous chromosomes remains conserved and homologous after polyploidization occurrence. In this paper, nine BAC clones from cytogenetic map of G. herbaceum chromosome A 1 01 were used to hybrid chromosomes of G. raimondii, as a result two of them were localized on chromosome D 5 01. In addition, the two clones were also tested on chromosome A h 01 and D h 01 of G. hirsutum, showing the same positions as that on chromosome A 1 01 and D 5 01. But other seven BAC clones were not localized on chromosome D 5 01 or D h 01. This may suggest that most parts of the pair of homologous chromosomes were not conserved.

Conclusion
Cotton is an excellent system for the study of genome evolution and polyploidization. However, the cytology study on cotton is far behind other leading crops. Using BAC-FISH presented here, individual BAC clones all anchored by SSR markers were accurately localized on chromosomes. Two markers perhaps were first mapped on chromosome D 5 01 and D h 01, and the conserved regions tagged by the two markers were detected between D 5 01 (D h 01) and A 1 01 (A h 01) chromosomes. Development of a clone-based cytogenetic map seen in the present work may also offer a resource to accelerate Integration of genetic and cytogenetic maps not only verifies the quality of the four genetic maps but also provides important information for cotton breeding and evolution.

Plant materials and BAC library
The plant materials (G. herbaceum, accession name is Zhongcao-1; G. raimondii, accession name is D5-2; and G. hirsutum, accession name is CCRI-12) were obtained from National Wild Cotton Nursery in Hainan Island, China, sponsored and owned by the Institute of Cotton Research of Chinese Academy of Chinese Academy of Agricultural Sciences (CRI-CAAS). They are also conserved in the greenhouse at CRI-CAAS' headquarter in Anyang City, Henan Province, China.
Pima 90-53 BAC library which was screened in this paper was kindly provided by Prof. Zhiying Ma (Hebei Agricultural University, China).

DNA probes preparation
The probes BAC DNA were isolated by using a standard alkaline extraction [60]. The chromosome-specific BAC clones were labeled by standard DIG-nick translation reactions, whereas the screened chromosome-specific BAC clones were labeled with Biotin-nick translation reactions, according to the instructions of the manufacturer (Roche Diagnostics, USA).

Chromosome preparation and FISH
Mitotic chromosomes preparation and FISH procedure were conducted using a modified protocol [61]. Biotinlabeled and digoxigenin-labeled probes were detected by avidin-fluorescein (green) and anti-digoxigenin-rhodamine (red) (Roche Diagnostics, USA), respectively. Chromosomes were counterstained by 4′,6-diamidino-2-phenylindole (DAPI) in the antifade VECTASHIELD solutions (Vector Laboratories, Burlingame, CA). For the probecocktail mixture, gDNA was used as block DNA. The dose of block DNA was 200 times of the chromosome-specific BAC DNA. The hybridization signals were observed using a fluorescence microscope (Leica MRA2) with a chargecoupled device (CCD) camera (Zeiss Axioskop2 plus). Final image adjustments were performed by using Adobe Photoshop CS3 software.

Comparison of maps using standardized map unit
Different kinds of maps are constructed in different method, and their units are also different. For example, the unit of genetic maps is cM (centimorgan) while the unit of cytogenetic maps is FL (FL: the percentage of the distance from the FISH site to the end of the short arm relative to the total length of the chromosome). In order to integrate different type of maps with shared markers for a comprehensive view of genome structure, the relative map position (RMP) units were used in the present study. The RMP unit of the cytogenetic map is the percentage of the distance (μm) from the FISH signal site to the end of the short arm showed relative to the total length of the chromosome (μm) and the RMP value of the genetic map is the percentage from the genetic location (cM) of each locus along the total length (cM) of the corresponding linkage group [12]. In order to establish the exact position of each clone, hybridization signal of each BAC clone was measured in more than 10 cells and the average position was computed.