Introduction

Cotton, cultivated in nearly 100 countries, is the most important natural textile fiber and the second largest oilseed crop. The genus Gossypium consists of ∼50 species, including 40-45 diploids (2n = 2x = 26) and five allotetraploids (2n = 4x = 52) (Fryxell 1992). Two groups of diploid species, designated A- and D-genome, diverged from a common ancestor about 5-10 million years ago (MYA). These diploid species were combined to form new allotetraploid species through a polyploidization event that occurred 1–2 MYA, giving rise to five divergent allotetraploid species (Cronn et al. 2002; Seelanan et al. 1997; Wendel 1989). Two of the allotetraploid species, Gossypium hirsutum and Gossypium barbadense, account for ∼95% and ∼5% of worldwide cotton cultivation, respectively. The genus Gossypium not only has its economic importance, but also provides an excellent system for the study of genome evolution and polyploidization in plants.

Polyploidy or whole genome duplication is common in many plants and some animals (Otto and Whitton 2000). It is estimated that all angiosperms underwent at least one event of polyploidy in their evolutionary history, and ∼70% of flowering plants are polyploids (Masterson 1994). After polyploidization, the genomes may undergo genetic and epigenetic changes, leading to gene expression and phenotypic variation (Chen 2007). In cotton, duplicate genes originating from the progenitors evolve independently at the same rate as those of their diploid progenitors (Cronn et al. 1999) and display few genomic changes in the allotetraploids (Liu et al. 2001). In the analysis of AdhA and CesA genomic regions and random whole-genome shotgun trace reads, the genome-size difference in cotton is probably associated with the expansion and contraction of repetitive elements including transposons, whereas the homoeologous sequences flanking the genes are highly conserved (Grover et al. 2004, 2008; Hawkins et al. 2006). However, a generally high level of sequence conservation between homoeologous genomic regions makes it difficult to annotate and assemble whole-genome shotgun sequences in allotetraploid species including cotton and wheat.

To dissect the genomic complexity in allotetraploid cotton, extensive efforts have been placed on developing high-density genetic maps. To date, over 1,000 DNA markers have been mapped in the populations of allotetraploid cotton (Guo et al. 2008; Rong et al. 2004). In addition, over 350,000 expressed gene sequence tags (ESTs) have been analyzed (Udall et al. 2006; Yang et al. 2006), which revealed an enrichment of genome-specific ESTs during early stages of fiber development (Yang et al. 2006). Moreover, several large insert bacterial artificial chromosome (BAC) libraries have been developed (Hu et al. 2009; Tomkins et al. 2001; Yin et al. 2006). Although linkage maps provide useful genetic information including colinearity of DNA markers, physical distance, and organization of the homoeologous chromosomes in the allopolyploid species are largely unknown. In addition, many DNA markers are monomorphic and cannot be mapped in both homoeologous linkage groups. Integration of DNA markers into cytogenetic maps provides unique insights into genome organization at the chromosomal level (Wang et al. 2006a). Direct localization of genetically mapped sequences onto chromosomes using fluorescence in situ hybridization (FISH) is a powerful method for mapping plant chromosomes (Jiang and Gill 2006; Kato et al. 2005; Lakatošová and Holečková 2007; Schwarzacher 2003). Meiotic pachytene chromosomes are often used for developing high-resolution FISH maps in plants because they are 10-25 times longer than somatic metaphase chromosomes and the morphology of the chromosomes and cytological landmarks are well preserved (De Jong et al. 1999; Fransz et al. 1998; Harper and Cande 2000; Xu and Earle 1996).

Previous cotton cytogenetic maps were limited to mitotic or meiotic metaphase chromosomes using repetitive DNA (Hanson et al. 1996; Ji et al. 2007; Zhao et al. 1998), BACs (Hanson et al. 1995), and DNA marker-derived BACs (Wang et al. 2006b, 2007b, 2008). This is partly because tetraploid cotton contains many chromosomes (2n = 4x = 52), and it is difficult to prepare pachytene chromosomes (Jack Mursal and Endrizzi 1976). To date, integrated genetic and cytogenetic maps using pachytene chromosomes were not available in cotton (Ji et al. 2007). Here, we report the first high-resolution integrated genetic and cytogenetic maps of homoeologous chromosomes 12A and 12D, which contain important genes related to fiber fuzz, gland development, and male sterility in allotetraploid cotton (Chen et al. 2009; Dong et al. 2007; Rong et al. 2005). The advantages of using chromosomes 12A and 12D also includes that no chromosome translocations (Gerstel 1953) and inversions (Rong et al. 2004; Guo et al. 2008) compared with their diploid ancestors have been found on these two chromosomes. The pachytene chromosomes of 12A and 12D were mapped using monomorphic and polymorphic BACs that were anchored by DNA markers (Wang et al. 2007a). Integration of genetic loci into physical localizations showed considerable variation of genome organization, structure, and size between 12A and 12D homoeologous chromosomes.

Methods

Materials

G. hirsutum acc. TM-1 was used for cytological studies. All BACs used for FISH mapping were identified by screening two genomic BAC libraries derived from TM-1 (Hu et al. 2009) and a restorer line, 0-613-2R (G. hirsutum), for cytoplasmic male sterile line of Gautieria harknessii (Yin et al. 2006). The BACs from these libraries were identified by the prefixes “t” for TM-1 BACs and “r” for restorer line BACs. The simple sequence repeat (SSR) markers used for BAC screening were selected from high-density genetic maps derived from the populations of tetraploid Gossypium species (Guo et al. 2008).

Chromosome preparation

Meiotic pachytene chromosome spreads were prepared as previously described (Wang et al. 2006b) with several modifications. Briefly, young buds containing pachytene-stage meiocytes were incubated at 37°C for 5 h in an enzyme mixture containing 2% cellulose and 2% pectolyase in 10 mM citrate buffer, pH 4.5. After washing in deionized water, several anthers were carefully transferred onto an ethanol-washed glass slide with a drop of 60% acetic acid (v/v), and the pollen mother cells were dissected out by teasing apart the anthers using forceps. After covering with a coverslip, the material was immediately squashed on a hot plate at 80°C. The slides were stored at −70°C. After removing the coverslips, slides were dehydrated through an ethanol series (70, 90, and 100%; 5 min each) and used in FISH.

FISH

The FISH procedure was similar to the previously published protocol (Wang et al. 2007c). In order to reduce the background, all BAC probes for each chromosome were separated into two or three cocktails and then applied in two or three rounds of hybridization using high-quality pachytene chromosome preparations. Biotin- and digoxigenin-labeled probes were detected using rhodamine-conjugated anti-digoxigenin and fluorescein-conjugated avidin (Roche Diagnostics), respectively. Chromosomes were counterstained with DAPI (4′, 6-diamidino-2-phenylindole; Sigma, USA) and antifade (Vector, USA) under a coverslip.

Image analysis

Slides were examined under an Olympus BX51 fluorescence microscope. The gray images of chromosome and FISH signal channels were captured and merged using Image-Pro Express software V5.0 with an Evolution VF CCD camera (Media Cybernetics, USA). To determine physical positions of signals, only chromosomes without apparent morphological distortion were analyzed, and their physical positions of signals were measured using Image-Pro Express. Final image adjustments were performed using Adobe Photoshop 6.0 software. The path of the chromosomes was computationally traced and straightened according to the manual provided by the Image J software V1.41 (http://rsb.info.nih.gov/ij).

Results

High-resolution cytogenetic maps of 12A and 12D homoeologous chromosomes

To construct a cytogenetic map with a high resolution, we screened as many BACs as possible using the SSR markers anchored on homoeologous chromosomes 12A and 12D in tetraploid cotton (Guo et al. 2008). For example, SSR primer pair NAU1237 produced two polymerase chain reaction fragments in G. barbadense cv. Hai7124 and two in G. hirsutum acc. TM-1 (Fig. S1a), one pair of which was co-dominant and polymorphic, NAU1237-160 in Hai7124 and NAU1237-155 in TM-1, and mapped on chromosome 12A (Guo et al. 2008). However, another fragment NAU1237-140 was monomorphic between Hai7124 and TM-1 and could not be mapped. This unmapped locus is probably located in its homoeologous chromosome 12D. Using the SSR primer pair NAU1237, we identified two positive BAC clones, r081K08 associated with the polymorphic locus NAU1237-155 and designated polymorphic BAC, and t215O23 related to the monomorphic locus NAU1237-140 and designated monomorphic BAC. Using our previously described method (Wang et al. 2007a, b), we performed FISH analysis using r081K08 BAC simultaneously with another BAC r043C02 that was known to be on chromosomes 12A (Wang et al. 2007a), and both BACs (r081K08 BAC as same as r043C02) were co-localized in the same chromosome 12A (Fig. S1b). As expected, another BAC t215O23 was mapped on 12D in FISH analysis. This approach has been proved to be useful for identification and mapping of more BACs onto homoeologous chromosomes in allotetraploid cotton (Wang et al. 2007a).

We selected 92 SSR markers that dispersed ∼3-cM intervals along the entire linkage maps of 12A and 12D and used them to screen for the corresponding BACs. A total of 85 positive BACs, 26 derived from 19 monomorphic SSR markers and 59 from 54 polymorphic SSRs were identified using 73 SSR markers. Out of 85 BACs, 53 (62.4%) were not mappable because they produced high hybridization background or light signals that were scattered on the chromosomes other than 12A and 12D. Consistent with our previous results, these BACs may contain one or more dispersed repetitive elements, suggesting a high amount of repetitive DNA dispersed in tetraploid cotton (Wang et al. 2006b, 2007b). The wide distribution of repetitive DNA leads to a relatively low percentage (37.6%) of available BACs that can be used for FISH analysis in tetraploid cotton.

The remaining 32 BACs that consistently produced unambiguous bright FISH signals on chromosome 12A or 12D were used in the construction of cytogenetic maps on pachytene chromosomes (Table 1). Among them, ten BACs were derived from monomorphic SSR markers, and three, t283I16, t208C15, and r070L23, were homoeologous-specific BACs because they generated bright signals on both 12A and 12D, indicating sequence homology between these BACs retained in 12A and 12D (Wang et al. 2007a; Fig. S2a-c). BAC t097G20 was mapped to the centromeric region of all 52 chromosomes not only in mitosis, but also in meiosis (Fig. S2d, e) in tetraploid cotton and probably contained centromere-specific repeats. Nevertheless, the centromere location can be accurately defined using this centromere-specific BAC t097G20. Together, 32 BACs generated a total of 36 cytogenetic markers (FISH signals), which were used to construct integrated genetic and cytogenetic maps of homoeologous chromosomes 12A and 12D (Fig. 1).

Table 1 List of BACs and their corresponding DNA markers used in the study
Full size table
Fig. 1
figure 1

BAC FISH mapping of pachytene chromosomes 12A and 12D in allotetraploid cotton. a, b Chromosomes 12A (a) and 12D (b) from different pachytene cells of TM-1 were hybridized with 15 and 21 BAC clones, respectively. The centromere-specific BAC t097G20 hybridized to all 52 cotton centromeres (red signals). Arrowheads indicate centromeres in homoeologous chromosomes 12A and 12D. c, d Chromosomes 12A (c) and 12D (d) were computationally straightened from images shown in a and b, respectively. Mapped BACs are shown on the right. Bars 10 µm

Full size image

Cytogenetic map of chromosome 12A

A total of 15 SSR-derived BACs (Table 1) including six monomorphic, five polymorphic, three homoeologous-specific (t283I16, t208C15, and r070L23), and one centromeric BAC (t097G20) were mapped on chromosome 12A (Figs. 1 and 2a). Five of them were localized in the short arm and nine in the long arm. To compare genetic and cytogenetic maps, we measured the FISH signals of each BAC from ten cells with clear chromosome spreads and calculated the relative position of FISH signals in a cytogenetic map (Fig. 2b). The genetic distances of SSR markers associated with the corresponding BACs were also converted into the relative positions in the corresponding linkage map (Fig. 2a). We found that the orders and positions of polymorphic and monomorphic BACs on chromosome 12A were concordant with marker positions in the corresponding linkage group.

Fig. 2
figure 2

An integrated cytogenetic map of homoeologous chromosomes 12A and 12D showing a high correlation between genetic positions of SSR markers and their corresponding physical locations of BACs. a, c The partial genetic maps of homoeologous chromosomes 12A (a) and 12D (c; Guo et al. 2008), in which the marker positions (cM) were converted to the positions relative to the total length of the linkage group (%). b, d Ideograms of pachytene chromosome 12A (b) and 12D (d) based on the FISH mapping results shown in Fig. 1. Heterochromatic regions with bright DAPI staining are shown in dark blue. Green and red circles represent FITC and rhodamine signals, respectively, of FISH generated by BAC probes. The positions of BACs on the chromosome were drawn based on the data from ten samples and were converted to the distances relative to the length of the chromosome (%). The polymorphic and monomorphic BACs were linked to corresponding SSR markers by black and black dotted lines, respectively. The dot-dash lines indicate the BACs with homoeologous-specific markers. Bar represents 10 µm or 10 cM in the cytogenetic and linkage maps, respectively

Full size image

Mapping monomorphic BACs can increase the number of mapped homoeologous loci. Of six monomorphic BACs, three (r038P15, r051E21, and t259M16) were mapped with relative distances of 78.0, 86.6, and 100%, respectively, in the long arm of 12A. BAC r051E21 was localized between two homoeologous-specific BACs t208C15 and r070L23, overlapping with r070L23. The other three monomorphic BACs, t200M10, r078J02, and r075A11, were mapped between the centromere and BAC t336G12 or SSR marker NAU3561-370 in the short arm of 12A (Fig. 2b). Intriguingly, the monomorphic BAC r075A11 was located in the short arm of chromosome 12A, whereas its corresponding genetic locus NAU2356-170 was mapped in the long arm of 12D (Fig. 2b, c). This suggests a rearrangement between 12A and 12D homoeologous chromosomes.

To test the coverage of genetic markers in the distal ends of 12A, all DNA markers near the ends of both arms were used to screen for BAC libraries. NAU4047-400, one of the four markers tested, detected a BAC for FISH. The marker was located in the short arm of the linkage map, and its corresponding BAC was located near the end of short arm (Fig. 1c), ∼1.8% of the length from the top of the chromosome (Fig. 2b). Another monomorphic BAC t259M16 was mapped near the end of the long arm (Fig. 1c). No corresponding BAC clone was found for the end-marker JESPR300-210. BAC t115A01, corresponding to JESPR300-205 in the linkage map 12D, was located within the chromosomal arm proximal to t259M16 (see below). The analysis of integrated cytogenetic and linkage maps indicated that a good coverage of genetic markers in the linkage map and a complete colinearity between genetic loci and physical positions of corresponding BACs in 12A.

Cytogenetic map of chromosome 12D

We developed a cytogenetic map of chromosome 12D using 21 BACs, including two monomorphic (t215O23 and r012M10), 15 polymorphic, three homoeologous-specific (t283I16, t208C15, and r070L23), and one centromeric BAC (t097G20; Table 1). The homoeologous-specific and centromeric BACs were common markers for both 12A and 12D. Similar to what was found in chromosome 12A, there was an obvious difference between the relative physical positions detected by polymorphic BACs and the genetic distances detected by linkage analysis near the end of two arms (Fig. 2c, d). Within a chromosomal arm such as the pericentric region of the long arm, the physical and genetic distances were relatively consistent. The order of polymorphic and monomorphic BACs in chromosome 12D was generally colinear with that of corresponding genetic markers in the linkage map, and only a few closely linked loci, between NAU4912-160 and NAU2372-265, JESPR300-205 and NAU3896-230, and NAU2715-250 and NAU3291-290, displayed changes in the order between the genetic markers and BAC locations (Fig. 2c, d). For example, BACs t107F20 and t246C12 nearly overlapped, although the genetic distance between their corresponding genetic markers BNL386-195 and NAU3006-220 was 4.6 cM.

Both polymorphic BAC r067L14 and monomorphic BAC t259M16 of NAU3896-230 were located near the end of chromosomes 12A and 12D (Fig. 1c, d), consistent with the location of SSR marker NAU3896-230 near the end of the linkage map. To evaluate the coverage of markers and BACs in the short arm of the linkage map 12D, we screened for BACs using 11 SSR markers located at 0 ∼ 41.2 cM in the distal region of short arm (Guo et al. 2008). Two SSR markers NAU4089-190 and NAU877-205 (25.6% and 31.5% from the distal end of short arm, respectively) detected polymorphic BACs that produced unambiguous FISH signals. The physical locations of corresponding BACs were 9.4% and 13.0% from the distal end of short arm (Figs. 1d and 2d).

The first integrated cytogenetic and linkage maps of homoeologous chromosomes 12A and 12D indicate that the orders of most genetic markers tested are colinear with corresponding BAC FISH signals. However, some discrepancies exist; r078J02 and t331N02 derived from NAU877 were anchored in the short arm of both chromosomes. However, physical positions between NAU1237-associated t215023 BAC and NAU3006-associated BAC t246C12 or NAU2356-associated r075A11 in 12D were reversed in 12A. This may be caused by a chromosome rearrangement possibly an inversion between NAU1237 and NAU2356 in 12D in the pericentric region relative to that in 12A.

Structural differences between homoeologous chromosome 12A and 12D

We analyzed the distribution of euchromatic and heterochromatic regions in two homoeologous chromosomes using DAPI-staining method (Kapuscinski 1995). The DAPI staining patterns were averaged from pachytene chromosomes in 20 cells. The length and intensity of blue staining regions corresponded to the relative length and intensity of euchromatin and heterochromatin along the chromosome (Fig. 2b, d). The majority of heterochromatin was distributed in the pericentric regions of 12A and 12D. A larger DAPI-staining block in the pericentric region of 12A than that of 12D indicated more heterochromatin in 12A. BAC r075A11 in 12A and t215O23 and r010G17 in 12D were localized in the pericentric regions of heterochromatin. Without Cot-1 DNA blocking, the FISH signals of these BACs in mitotic metaphase chromosomes were visible on all chromosomes or one homoeologous group of chromosomes (Fig. S3) probably because these BACs contain high amounts of repetitive DNA sequences that are dispersed in the genome. The overall similar patterns and distributions of euchromatic and heterochromatic regions suggest stable maintenance of chromatin structures between these two homoeologous chromosomes.

Detection of three homoeologous-specific BACs t283I16, t208C15, and r070L23 in this study indicated a homologous relationship between these three regions of two homoeologous chromosomes (Fig. S2). We further analyzed the homology and diversity of 12A and 12D homoeologous relationships using BAC-FISH intensities and patterns. To detect weak signals, we increased the probe DNA to 300 ng or more in each FISH assay using the protocol developed in maize (Kato et al. 2006). Only the signals that were found in three or more cells were further analyzed. The BACs mapped on chromosome 12D were divided into two groups. The first group included 13 BACs spanning from 61.2% to 100% in the distal region of the long arm, and the remaining seven BACs, including five being in the pericentric region of the long arm and two in the distal region of the short arm, belonged to the second group. Interestingly, we detected all 13 FISH signals of the first-group BACs in 12D, as well as in the distal regions of the long arm in 12A, although the FISH signals in 12A were weaker than that in 12D (Fig. 3a), except for three homoeologous-specific BACs and r067L14. Their orders were completely colinear between these two homoeologous chromosomes, indicating conservation of genome structure in the long arm of 12A. To our surprise, among the second group of seven BACs derived from 12D, only one BAC (t163D05) in the distal regions of short arm generated a weak homoeologous signal in 12A (Fig. 3b), and BACs r010G17 and t081H09 produced very weak and unstable signals in 12A. This suggests genomic changes in this pericentric region between 12D and 12A. The changes might occur in the progenitor species prior to polyploidization or in the homoeologous chromosomes evolution after polyploidization. To test this, we hybridized five 12D-derived BACs (t215O23, t081H09, t246C12, t107F20, and r010G17) that generated no signals in 12A (Fig. 3b) with A-genome diploid cotton, Gossypium arboreum. No hybridization was found in the diploid cotton either (Fig. S4). Together with integrated cytogenetic maps, we suggest a chromosomal rearrangement that is most likely driven by illegitimate recombination in this region prior to polyploidization.

Fig. 3
figure 3

FISH mapping showing variation between homoeologous chromosomes 12A and 12D. a One set of 13 BACs anchored on the long arm of 12D generated signals on both long arms of chromosomes 12A and 12D with the same order in one cell. BAC clones are indicated in chromosome 12A. b Seven BACs anchored on the short and long arms of 12D were hybridized, whereas only one BAC t163D05 produced signals in both chromosomes 12A and 12D. Another BAC r105F05 (red signals, arrowhead) was used as a control for the long arm. BAC clones are indicated in chromosome 12D. c With the exception of r043C02, all BAC probes anchored on the long arm of chromosome 12A generated signals in chromosome 12A and 12D as well. BACs r079A04 and t174A01 produced weaker signals than others. BAC clones are indicated in both 12A and 12D. d Five BACs anchored on the short arm of chromosome 12A were hybridized with both chromosomes 12A and 12D with a consistent order. Note that the different distance between BACs t200M10 and r078J02. BAC clones are indicated in chromosome 12D. Bars 10 µm

Full size image

Excluding three homoeologous-specific BACs and one centromere-specific BACs, we divided 11 BACs in 12A into two groups: six in the long arm and five in the short arm. Six long-arm BACs, except for r043C02, generated bright FISH signals in 12D (Fig. 3c), although r079A04 and t174A01 produced relatively weak signals in 12D. Five short-arm BACs generated bright FISH signals in 12D, and t263K12 and t336G12 produced brighter signals than other three BACs (Fig. 3d). The physical order of all BACs tested was consistent between two homoeologous chromosomes. Like in 12D, the BACs in the distal regions of 12A produced brighter FISH signals than those in pericentric regions, suggesting that the sequence identity is generally higher in the distal regions than in the pericentric regions of the two homoeologous chromosomes.

Genome-size variation between homoeologous chromosomes 12A and 12D

To test genome-size variation between 12A and 12D homoeologous chromosomes, we measured the lengths of both mitotic-metaphase and meiosis-pachytene chromosomes in 50 and 20 well-spread cells, respectively. The centromere location was accurately defined using the centromere-specific BAC t097G20. The total lengths of 12A and 12D were 141.91 and 107.76 µm, respectively, in the pachytene stage and 3.05 and 2.30 µm, respectively, in the metaphase stage. The arm ratios (short/long), 0.54 and 0.55 in the metaphase, and 0.41 and 0.40 in the pachytene were very similar in 12A and 12D. Two homoeologous chromosomes showed the same stretching ratio (∼47 times) between pachytene and metaphase stages. The total length of chromosome 12A was 1.3-fold longer than that of 12D (Table 2). To find out the locations of size differences between the two chromosomes, we calculated the ratios of physical distances between the same pair of BACs using 13 D-subgenome- and 5 A-subgenome-specific BACs that detected FISH signals on both chromosomes as presented above (see Fig. 3a, d). As expected, the distance ratios (A/D BACs) between 13 out of 19 pairs of BAC tested were larger than 1 (Fig. 4), indicating a longer chromosomal segments in 12A than in 12D. Two regions displaying largest values with ratios of 2.8 between t336G12 and t200M10 and 1.83 between t097G20 to t112E02 may be responsible for genome-size variation between 12A and 12D. The more expansion between t097G20 and t112E02 in 12A than that in 12D indicated a high amount of heterochromatin in the pericentric region of 12A, which is consistent with the DAPI-staining results. The largest ratio of 2.8 between t336G12 and t200M10 indicated either an expansion in 12A or a deletion in 12D. Our comprehensive analysis of integrated genetic and cytogenetic maps suggests that uneven expansion or contraction between different regions of homoeologous chromosomes leads to genome-size variation in A and D homoeologous chromosomes.

Table 2 Lengths and arm ratios of pachytene chromosomes 12A and 12D in Gossypium hirsutum L.
Full size table
Fig. 4
figure 4

Genome-size variation between homoeologous chromosomes 12A and 12D. Ideograms of pachytene chromosomes 12A and 12D based on the FISH mapping results shown in Fig. 2a, c. The segments separated by the same BAC pairs in chromosomes 12A and 12D are linked by lines, and the distance ratios are indicated. Values on the left of 12A and right of 12D are the cytological distances separating the BACs. The data are averaged from seven measurements

Full size image

Discussion

Integrated genetic and cytogenetic maps of homoeologous chromosomes 12A and 12D

Several cotton genetic maps have been developed and refined through linkage analysis using molecular markers (Guo et al. 2008; Nguyen et al. 2004; Rong et al. 2004). Although they are valuable genomic resources, the linkage maps provide little information about physical locations, distributions, distances, and sometimes orientations of genetic markers. Using a large number of marker-associated BACs, we developed the first comprehensive cytogenetic maps of homoeologous chromosomes 12A and 12D in tetraploid cotton. The integration between genetic positions of DNA markers into physical locations of BACs provides new insights into structure and organization of homoeologous chromosomes in tetraploid cotton, which will ultimately facilitate the assembly of cotton genomic sequences (Chen et al. 2007). Meiotic pachytene chromosomes are generally 10-25 times longer than metaphase chromosomes and provide excellent cytogenetic maps with a resolution between metaphase chromosomes and DNA fibers (Harper and Cande 2000). The morphology and cytological landmarks of pachytene chromosomes are well preserved, making it suitable for FISH mapping. Although cotton has relatively small chromosomes, pachytene chromosomes are notably long, ∼47 times longer than their corresponding mitotic metaphase chromosomes. FISH mapping using marker-associated BACs enables us to effectively integrate cytogenetic maps with corresponding genetic loci. The integrated cytogenetic maps of homoeologous chromosomes 12A and 12D indicated that the orders of most genetic markers are colinear with corresponding BAC locations although several closely linked loci in 12D display inconsistent orders or locations compared with those in BAC FISH maps. It is valueless for resolving small genetic distances in linkage maps. For example, two pairs of homoeologous BACs, t246C12 and t107F20 and r105F05 and t112E02, which were located on euchromatic and heterochromatic regions, respectively, and separated by 0.6 and 1.9 cM in linkage maps, were unambiguously resolved in pachytene FISH maps. The physical distance can be as close as 50 kb between two adjacent BACs t246C12 and t107F20. Although these two BACs did not overlap, after library screening using SSR markers NAU3006-220 and BNL840-165, we found BAC t246F19 with a 50-kb insert that bridged the two BACs, suggesting that the physical distance between BACs t246C12 and t107F20 is close to 50 kb. The HindIII-fingerprinting results indicated no overlap between these two BACs (data not shown).

Suppression of genetic recombination in the pericentromeric regions often leads to the discrepancies between genetic and physical distances in many plant species, including maize (Wang et al. 2006a), wheat (Gill et al. 1996), tomato (Tanksley et al. 1992), barley (Kunzel et al. 2000), and sorghum (Islam-Faridi et al. 2002; Kim et al. 2005). In particular, recombination in the pericentromeric regions, which may account for up to 50% of the length of each chromosome, is highly suppressed in sorghum, wheat, and barley (Gill et al. 1993; Islam-Faridi et al. 2002; Kim et al. 2005; Kunzel et al. 2000). In this study, the genetic distance between markers NAU877-205 and BNL3816-195 is only 3.2 cM, representing 2.5% of total linkage distance of chromosome 12D in the linkage map (Guo et al. 2008), whereas the physical distance between these two markers corresponds to 20.8% of total length of the chromosome (Fig. 2d). This implies very low recombination rates in this pericentromeric region. However, recombination between markers BNL1669-170 and NAU2372-265 in the long arm of 12D was also suppressed although there is no obvious heterochromatin in this region according to the DAPI-staining results (Fig. 2d). The two BACs are separated by 20.9% of the total chromosome length, which corresponds to only 6.6% of the total genetic distance. If the recombination suppression in this region is related with its chromosome rearrangement remains to be investigated.

In chromosome 12A, the recombination suppression in the pericentromeric region was not detected because the current physical map consisted of a limited number of polymorphic BACs. Compared to the physical map (43.2% physical length), the recombination frequency (34.5%) was slightly reduced (1.25-fold) between markers NAU2096-200 and NAU3561-370, which covered the entire pericentromeric region extending to the short arm. However, marker NAU2096-200 was close to the pericentromeric heterochromatin block (Fig. 2b), and the genetic and physical distances between them were equal (46.0%). This suggests that 1.25-fold recombination reduction occurs in the pericentromeric region of the short arm, which is consistent with absence of recombination suppression in the pericentromeric region of 12D.

Homologous relationships between chromosomes 12A and 12D

G. hirsutum L. is an allotetraploid that contains two subgenomes originating from related ancestor species with different genome sizes. Genomic changes in homoeologous chromosomes have been studied in tetraploid cotton, suggesting that the homoeologous loci are mainly additive (Liu et al. 2001), evolve independently (Cronn et al. 1999), and well conserved in the regions containing AdhA and CesA genes (Grover et al. 2004, 2007). These results were obtained from analyzing random genomic sequences and a few pair of homoeologous BAC sequences. These views have been updated by our compressive analysis of two homoeologous chromosomes in tetraploid cotton. Among 31 BACs, 25 (81%) including three homoeologous-specific BACs, except for the centromere-specific BAC, produced consistent orders of FISH signals in both homoeologous chromosomes (Fig. 3). This suggests that the majority of homoeologous chromosomes 12A and 12D remains conserved and homologous after polyploidization that occurred ∼2 MYA (Cronn et al. 2002; Seelanan et al. 1997; Wendel 1989). However, five BACs located in the pericentromeric region from centromere to 41.1% in 12D do not have obvious FISH signals on its corresponding homoeologous chromosome 12A. The integrated cytogenetic and linkage maps indicated that this region has active recombination events in 12D, although the heterochromatic region is relatively small. All homoeologous-specific BACs were dispersed in the distal region of long arms. Likewise, BACs in the distal region of short arms generated bright homoeologous FISH signals. Several previous studies have shown that high recombination rates are usually correlated to high gene densities (Barakat et al. 1997; Broman et al. 1998; Collins et al. 1996; Gill et al. 1996; Thuriaux 1977). Our data imply that pericentromeric regions of long arms in homoeologous chromosomes 12A and 12D may also be gene-rich. In general, the distal regions of homoeologous chromosomes 12A and 12D are more similar than the pericentromic regions.

Potential mechanisms for genome-size evolution in chromosomes 12A and 12D

Conservation of genic and intergenic regions suggests that genome-size evolution between A- and D-subgenomes occurs at regional levels (Grover et al. 2004). This observation has not been tested in a genome-wide scale. Here, we provide a detailed chromosomal view of genome-size variation between homoeologous chromosomes 12A and 12D that show 1.3-fold size variation. Each alignment between homoeologous regions exhibits a 1.23-fold size variation, which is approximately the overall size difference (∼1.3-fold) between the two chromosomes. But uneven expansion or contraction between different regions of homoeologous chromosomes leads to genome-size variation in A and D homoeologous chromosomes. The distal regions of chromosomes 12A and 12D are highly conserved; however, either an expansion in 12A or a deletion in 12D occurred in this region between t336G12 and t200M10 result in 2.8 time difference in genome size between 12A and 12D. DAPI-staining and BAC-FISH all revealed that the more expansion with a high amount of heterochromatin in the pericentric region of 12A than that in 12D leads to more genome size in 12A.

In summary, the first set of integrated genetic and cytogenetic maps in two homoeologous chromosomes in cotton presents a comprehensive and comparative analysis of conservation and changes in genetic and genomic footprints between these two homoeologous chromosomes. These maps have not only provided insights into genome evolution of homoeologous chromosomes but also serve as a template to facilitate genome sequencing and sequence assembly in tetraploid cotton.