Introduction

Papaya, Carica papaya L., is a nutritious fruit crop grown in all tropical and many subtropical countries of the world. This small tree is primarily grown as a garden crop and also in commercial fields for local and export markets. Papaya is trioecious having three sex types (hermaphrodite, female, and male) and it has several properties that facilitate its use in genetic studies. It has a short generation time of 9 to 15 months from seed-to-seed and it produces numerous fruit containing thousands of seeds. Papaya produces a fruit at approximately 4 to 5 day intervals 12 months of the year over a period of 2 to 4 years. Its large flowers allow for individual hybridization to guarantee parentage for genetic studies. It has a small genome (372 Mbp) that has been sequenced (Ming et al. 2008). Several high-density genetic maps have been produced (Blas et al. 2009; Chen et al. 2007; Ma et al. 2004) and a BAC-based physical map has been integrated with genome sequence and high-density genetic map (Yu et al. 2009).

High-density genetic mapping using amplified fragment length polymorphism (AFLP) and microsatellite markers, fine mapping, and physical mapping of the sex determination genes led to the conclusion of a nascent Y chromosome with a small non-recombining region controlling sex determination in papaya (Liu et al. 2004; Ma et al. 2004). The male or hermaphrodite-specific region of the Y chromosome (MSY or HSY) is about 10–15 % of the Y chromosome (Liu et al. 2004; Zhang et al. 2008). Further sequence analysis of divergence of four X/Y gene pairs located at HSY and the homologous region on the X chromosome revealed that the papaya Y chromosome evolved 2–3 million years ago (MYA) (Yu et al. 2008).

In addition to genetic linkage group identification, chromosome nomenclature was developed for papaya based on fluorescence in situ hybridization (FISH) (Wai et al. 2010; Zhang et al. 2010). The sex chromosome has been designated as chromosome 1 (Zhang et al. 2010) and linkage group (LG) 1 in both the AFLP and microsatellite marker (or simple sequence repeat, SSR) genetic maps (Blas et al. 2009; Chen et al. 2007; Ma et al. 2004). Chromosome 1 is the second largest chromosome in papaya, but the size difference with the largest chromosome is within the range of standard deviation (Zhang et al. 2010). Chromosome 1 has five heterochromatic knobs in the proximal region and one heterochromatic knob close to the distal end (Zhang et al. 2008, 2010). The rest of chromosome 1 is stained lightly by DAPI, but one of the chromosome arms is always stained more intensely than the other (Zhang et al. 2010). The five proximal knobs are named as K1 to K5 and their location on the HSY region was confirmed by FISH (Zhang et al. 2008). Both chromosomes X and Yh (the hermaphrodite Y chromosome) contain the knob K1, while the remaining four knobs appear only on the Yh chromosome (Zhang et al. 2008). The greater DNA content of Yh than of X chromosome may be responsible for the twist of the Y chromosome between K4 and K5 (Zhang et al. 2008). The DNA associated with K2 to K5 is highly methylated compared to the homologous region on X (Zhang et al. 2008). However, the reason for differential methylation in this region remains unknown. The centromere of the chromosome 1 was shown to be located on or close to K4 (Zhang et al. 2008).

Although the genomic resources of papaya are fairly extensive, the cytological features of papaya chromosomes have not been linked with the existing genome sequence or with the genetic and physical maps. The most direct way to integrate the chromosomal features of papaya with the genetic and physical maps would be to generate a cytogenetic map for localizing microsatellite-tagged BACs directly on pachytene chromosomes by FISH. Meiotic pachytene chromosomes are more relaxed than somatic metaphase chromosomes giving them much higher levels of resolution (Cheng et al. 2002; Danilova and Birchler 2008; Jiang and Gill 2006). Localization of labeled BACs on pachytene chromosomes to integrate genetics and cytology has been applied in sorghum, tomato, potato, cucumber, rice, maize, and Arabidopsis (Cheng et al. 2001; Fransz et al. 1996; Iovene et al. 2008; Kim et al. 2005b; Koo et al. 2008; Koumbaris and Bass 2003; Pedrosa-Harand et al. 2009; Peters et al. 2009; Ren et al. 2009; Wang et al. 2006). A molecular cytogenetic map combines chromosome structures with recombination rate and physical distance, thus providing integrated biological information on genome organization. This information allows for an estimation of physical distances of trait loci for use in map-based cloning (Budiman et al. 2004; Kim et al. 2005b), enables comparative genomic studies between closely related species (Iovene et al. 2008; Kim et al. 2005a), and contributes to development of genome sequencing strategies (Szinay et al. 2008).

In our previous study, we demonstrated the feasibility of BAC-FISH on high-resolution papaya pachytene chromosome spreads that helped resolve problems about the orientation of linkage groups (Wai et al. 2010). Here, we report the construction of a molecular cytogenetic map for chromosome 1 by labeling 16 genetically mapped microsatellite markers and two cytological feature-associated markers on pachytene chromosomes. The recombination frequency along chromosome 1 was analyzed, which allowed us to analyze differences of chromosome condensation in selected regions and the relationships between chromosomal morphology, transposable elements and gene content.

Materials and methods

Plant materials and chromosome preparation

C. papaya varieties SunUp and AU9 were planted and maintained at the Hawaii Agriculture Research Center substation located at Kunia, Hawaii. Hermaphrodite and male flowers (5–7 mm in length) were collected and fixed in Carnoy’s solution (100 % ethanol/glacial acetic acid = 3:1) with 15 min vacuum infiltration. Meiotic pachytene chromosomes were prepared as described in Wai et al. (2010). Microscopic slides holding chromosome preparations were pretreated by incubating on 60 °C slide warmer for 30 min, then fixed in 4 % paraformaldehyde (in phosphate-buffered saline) for 10 min followed by 2× SSC for three times at intervals of 5 min. The slides were then dehydrated in 70 and 95 % ethanol for 2 min each.

Probes and fluorescence in situ hybridization

Microsatellite markers previously mapped on LG 1 of papaya SSR genetic maps were chosen randomly based on the criterion of being an approximate distance of 10 cM apart. BAC clones that harbor corresponding microsatellite markers were chosen based on the assembled papaya genome sequence database (version 2007-04-27) available at the Advanced Studies of Genomics, Proteomics and Bioinformatics (ASGPB) homepage (http://asgpb.mhpcc.hawaii.edu). Corresponding BAC clones for knob 2 and knob 4 at the HSY region were selected according to Zhang et al. (2008). All BAC clones used in FISH were obtained from the SunUp hermaphrodite BAC library (Ming et al. 2001). The specificity of BAC clone signals on chromosomes was tested on pachytene chromosome spreads before making cytological measurements. Only BAC clones with single clear signals were used for cytological measurement. BAC DNA extraction, probe labeling and fluorescence in situ hybridization methods were described by Wai et al. (2010), except that C0t-1 DNA was not added to the hybridization mixture.

Cytological measurement

The distribution of euchromatin and heterochromatin at the short and long arms were measured in 18 pachytene chromosome spreads. Short arm euchromatin refers to the region from the telomeric end carrying the microsatellite marker P6K72 to knob 5 of the HSY. Long arm euchromatin refers to the region from knob 1 of the HSY to the telomeric end carrying the microsatellite marker CPM1737.

For BAC clones giving unambiguous signals, two to four clones were labeled simultaneously on meiotic pachytene chromosomes for measuring their cytological positions. FISH images were captured using a 2.0-megapixel monochrome QuantiFire® camera (Optronics) attached to an Olympus BX51 epifluorescence microscope. Grayscale images on each color channel were captured and merged by PictureFrame™ 2.2 image system (Optronics) with final adjustment by Adobe Photoshop 7. The length of chromosome and position of signals from chromosome ends carrying the microsatellite marker P6K72 were measured on merged images by Rincon®. The relative cytological position (CP) was calculated as:

$$ {\mathrm{CP}}\left( {{\mathrm{in}}\,{\mathrm{percentage}}} \right){ =}\left( {S/T} \right) \times 100\% $$

where S is the distance between signals and telomeric end carrying marker P6K72 and T is the total length of the chromosome. The fraction length (FL) was calculated as:

$$ {\rm{FL = }}\left( {S/T} \right) \times {\rm{total genetic distance of LG1,}} $$

where the total genetic distance of LG 1 is 145 cM. For each marker, a minimum of ten pachytene chromosome spreads were measured. The measurement data were analyzed using Microsoft Excel 2003.

The average cytological distance between the telomere of the short arm and the probe signal was determined on eight pairs of markers using linear regression analysis. The average chromosome length of 45.35 μm was used to calculate the degree of condensation of the euchromatin and heterochromatin on chromosome 1.

Repetitive sequences and gene density analysis

Scaffold sequences used for analysis of chromosome 1 were based on the physical map published by Yu et al. (2009) and only those sequences larger than 100 kb were included in this analysis. Transposable elements and repetitive sequences in 18 scaffolds were identified using RepeatMasker version 3.2.8 (Smit, A.F.A and Hubley, R., http://www.repeatmasker.org) and cross match version 0.990329 (http://www.phrap.org) with custom library composed RepBase (Jurka et al. 2005) (http://www.girinst.org/repbase, 2009-06-04 version), TIGR plant repeat database (Ouyang and Buell 2004) and C. papaya specific transposable element database (ftp:/ftp.cbcb.umd.edu/pub/data/CPR-DB) (Nagarajan et al. 2008).

For putative gene content, the presence of genes on each scaffold was either predicted by a combination of Augustus, Fgenesh, Genscan, GlimmerHMM, SNAP, and TWINSCAN or supported by presence of EST sequences. The predicted gene location and number were accessed through ASGPB homepage (http://asgpb.mhpcc.hawaii.edu/papaya/).

Results

Morphology of chromosome 1

We observed that papaya DAPI-stained chromosome 1 was mainly euchromatic, except for the central region of the chromosome. Five darkly stained knobs were observed at the center of this chromosome. The first knob (K1) was the largest in size and appeared on both chromatids. The other four knobs (K2, K3, K4, K5) alternated with euchromatin and were found only on the Y chromatid. In addition to these five knobs, one heterochromatic knob was also observed on the distal end of euchromatic arm.

Based on the papaya sequence-tagged genetic map, 16 microsatellite markers randomly distributed along the lengths of LG1 were selected for cytogenetic mapping (Table 1). BAC clones harboring the selected microsatellite markers were identified by searching for microsatellite sequences and papaya BAC end sequences into the papaya genome sequence. BAC clones were labeled on pachytene chromosome spreads by FISH and the distances between specific marker and the telomeric end were measured. The relative CP of each marker was calculated and the result is shown in Table 1. We measured the length of the two arms of chromosome 1 referencing K4 as the centromere of the chromosome 1 (Zhang et al. 2008). The distance between the telomere end close to marker P6K72 and K4 occupied 42 % of the chromosome length and the distance between the other telomere end and K4 was about 58 % of the chromosome length. Thus, the chromosome arm containing marker P6K72 and K5 was defined as the short arm and the other arm was defined as the long arm. Chromosome 1 was sub-metacentric with an arm ratio of 0.71. Based on the measurement of 18 pachytene chromosome spreads, 84.71 ± 2.35 % of the Y chromosome was euchromatic where 51 ± 4.6 % of the euchromatin was on the short arm and 49 ± 4.6 % on the long arm (Table S1). The remaining 15.29 % of the Y chromosome was heterochromatic and mainly located in the central region of the chromosome.

Table 1 Genetic and cytological positions of microsatellite and knob-associated markers on chromosome 1
Full size table

The condensation degree of euchromatin and heterochromatin can vary at the pachytene stages and thus affect the precision of our cytological measurements. We measured 18 chromosome spreads to record the lengths of the euchromatin and centromeric heterochromatin (Table S1). The correlation between the length of the euchromatin and heterochromatin was positive and significant (r = 0.703, p ≤ 0.01) and suggested both regions change proportionately when the chromosome length varies in different stages of relaxation. We concluded that pachytene chromosome 1 spreads of different lengths were suitable to construct the cytogenetic map.

Cytogenetic map of chromosome 1

We selected 16 BACs harboring microsatellite markers from non-clustering regions on LG 1 (Table 1) to construct a molecular cytogenetic map. Two morphological feature-associated markers, BAC 99O03 on K2 and BAC 52H15 on K4, were also included to allow us to characterize the HSY region. The total genetic length of papaya chromosome 1 (LG 1 of the genetic map) was 145 cM (Chen et al. 2007) and the 16 selected markers were located between 1 and 145 cM of LG 1. The genetic distance between markers ranged from 5 to 24 cM with an average distance of 9 cM.

The BAC clone signals of 16 markers and 2 knobs were measured on a minimum of 10 pachytene chromosome spreads (Fig. 1). By comparing the FISH signals and DAPI-staining pattern, BAC clone 19A10 containing the marker P6K1066 (25 cM) was labeled on the short arm of chromosome 1 (Fig. 1c). The BAC clone that contained the other distal end marker on LG 1, Ctg566 (145 cM), was mapped on the long arm by FISH. FISH mapping results revealed that all 16 markers except P3K2981 were located in euchromatic region, four on the short arm, one around the euchromatin–heterochromatin boundary of the short arm, and ten on the long arm. The order of BAC clone signals on chromosome 1 was concordant with the order of microsatellite markers on the genetic map except for the three markers, P3K2981, P6K1460 and Ctg566 (Fig. 2). FISH results showed the positions of markers P3K2981 and P6K1460 were reversed on the cytological map compared with their locations on the genetic map (Fig. 2). The marker Ctg566 (145 cM) is likely mapped wrongly on the genetic map and showed a signal between the markers P3K629 (101 cM) and CPM1737 (113 cM) on the long arm of chromosome 1 (Fig. 1d). The second distal end SSR genetic marker of LG 1, P3K7101 (122 cM) was located closer to the telomeric end of the chromosome than marker Ctg566 (145 cM). About 96 % of the length of chromosome 1 was covered by markers P6K72 (1 cM) through P3K7101 (122 cM) (Fig. 1e and Table 1), indicating a high coverage of the genetic map of chromosome 1.

Fig. 1
figure 1

FISH mapping of microsatellite marker-anchored BAC clones on papaya chromosome 1. a The DAPI-stained papaya pachytene chromosome spread was converted into a black-and-white image. Five heterochromatic knobs on chromosome 1 are labeled K1 through K5. b The same chromosome spread as in 1A was hybridized with BAC clone 19B15 (46 cM) and 19I15 (48 cM), which spans knobs K2 to K5 of the MSY region. c Simultaneous hybridization of four BAC clones, 19A10 (25 cM), 73P20 (30 cM), 14L22 (35 cM) and 4E12 (89 cM) revealed the orientation of the genetic map of chromosome 1. d BAC clone 19E20 harboring the most distal microsatellite marker ctg566 (145 cM) is positioned between BAC clones 26 N04 (101 cM) and 96C17 (113 cM). e Simultaneous hybridization of three BAC clones, 55 G05 (1 cM), 1 M23 (60 cM) and 99B02 (122 cM), on pachytene chromosome 1. BAC clones 55 G05 and 99B02 span the majority of chromosome 1, except for the telomeres. f BAC clones 99O03 and 52 H15, associated with knobs K2 and K4, respectively, was co-mapped with BAC clone 4E12 (89 cM). Bars, 5 μm

Full size image
Fig. 2
figure 2

Genetic map, cytogenetic map and recombination rate of chromosome 1. The chromosomal positions of 15 microsatellite markers (except Ctg566) were measured by FISH mapping and diagrammed in scale. The genetic positions of markers are based on the microsatellite marker genetic map published by Chen et al. (2007). The dark circles on cytogenetic map represent the six heterochromatic knobs on chromosome 1. On the right panel, the recombination frequency was calculated as genetic to cytological distance (cM/FL, x-axis) and plotted against cytological distance (in fraction length, FL) in y-axis. The average recombination frequency (0.87 cM/FL) of chromosome 1 is marked as the red dashed line

Full size image

Zhang et al. (2008) showed that a BAC clone harboring the marker CPM1055 located at 50 cM in the euchromatin of the long arm adjacent to K1. Our cytogenetic results showed that the FISH signals of P3K2981 and P6K1460 flanked K2 to K5 of the centromeric heterochromatin (Fig. 1a and b). Thus, the five proximal knobs were located between markers P3K2981 (46 cM) and CPM1055 (50 cM) of the genetic map.

Recombination frequency along chromosome 1

Although a physical map of chromosome 1 is available (Yu et al. 2009), gaps remained between several scaffolds so that the actual physical distances between some of the genetic markers are unknown. The BAC clones used in FISH are tagged with microsatellite markers, so that the cytological map can be directly integrated with the genetic map for comparing recombination frequencies of different regions along the chromosome.

The ratio of genetic to cytological distance was in a narrow range along the euchromatin regions of chromosome 1, ranging from 0.61 to 2.12 cM/FL (0.47 to 1.64 FL/cM) (Table 2, Fig. 2). The cytological distance between markers P6K72 (1 cM) and P3K7101 (122 cM) was 138.62 FL. Thus, the average ratio of genetic to cytological distance between these two markers was 0.87 cM/FL. The mean ratio of the short arm between markers P6K72 and P6K1460 was 0.87 ± 0.07 (47 cM/53.86 FL), while the mean ratio of the long arm between markers P3K2388 and P3K7101 was 1.39 ± 0.14 (62 cM/44.77 FL). Overall, the long arm had a higher ratio of genetic to cytological distance than the short arm.

Table 2 Recombination frequency along papaya chromosome 1
Full size table

The centromeric heterochromatic region between markers P6K1460 and P3K2981 showed a severe suppression of recombination. Although this region was approximately 12.2 % (16.97 FL/145 FL) of the chromosome in cytological length, it contributed only 1.3 % (2 cM/145 cM) of the genetic distance. The ratio of genetic to cytological distance in this region was 0.12 ± 0.02 cM/FL (2 cM/16.97 FL), 6 to 18 folds lower than that of the euchromatic regions (Table 2).

Condensation degree of euchromatic and heterochromatic regions of chromosome 1

Aligning the cytogenetic map with the physical map allowed us to calculate the condensation degree along the chromosome 1. We selected eight pairs of markers along the chromosome 1, two pairs on the short arm, five on the long arm, and one pair in the centromeric/pericentromeric heterochromatin region (Table 3). We measured the cytological distances on 108 pachytene spreads. The average cytological length of the chromosome 1 at pachytene stage was 45.35 ± 9.55 μm. The mean absolute cytological distance of each marker pair (in micrometers) was determined using linear regression analysis against the chromosome length (p < 0.05). The condensation in euchromatin regions showed significant variation ranging from 273 to 806 kb/μm with an average at 474 kb/μm (Table 3). The euchromatin region covered by the marker pair of P3K418 and P3K7101, located close to the telomere of the long arm, was more condensed (806 kb/μm) than any other euchromatic regions examined. The condensation degree between K2 and K4 in the centromeric/pericentromeric heterochromatin region was estimated at 1,711 kb/μm, nearly fourfolds higher than was the average degree condensation of the euchromatin regions. The two pairs of markers from the short arm (112 and 132 kb/cM) showed higher ratios of physical distance/genetic distance than all but one pair of markers from the long arm. One should note that this pair of markers on the long arm were located close to the telomere and showed the highest degree condensation.

Table 3 Physical, genetic and cytological relationships of eight pairs of markers
Full size table

Abundance of repetitive sequences and genes on chromosome 1

The genome of a female plant of transgenic papaya, ‘SunUp’, has been sequenced (Ming et al. 2008) and 25 scaffolds in total length of 21,968,080 bp were anchored on chromosome 1(Yu et al. 2009). Among the 25 anchored scaffolds, 18 of them with size greater than 100 kb were included in this cytogenetic analysis. Based on their chromosomal locations determined by the genetic markers on chromosome 1, we divided the 18 scaffolds into three groups, short arm euchromatin, centromeric/pericentromeric heterochromatin, and long arm euchromatin (Table S2). There are four scaffolds (scaffolds 39, 66, 177 and 215) located at the centromeric/pericentromeric heterochromatic region of X chromosome, five scaffolds (scaffolds 14, 99, 116, 123, and 140) at the short arm, and nine (scaffolds 21, 26, 36, 49, 64, 142, 160, 249 and 286) at the long arm. After removing the gaps within scaffolds, the remaining sequence was analyzed with RepeatMasker, including 4,333,240 bp on the short arm, 2,869,702 bp on the centromeric/pericentromeric heterochromatin, and 8,096,849 bp on the long arm (Table S2).

The percentages of repetitive sequences in the short arm, centromeric/pericentromeric heterochromatin of X chromosome, and long arm were 39, 49, and 26 %, respectively (Table 4). The abundance of repetitive sequences increased toward the centromeric/pericentromeric region (Fig. 3; Supporting Information, File S2). Transposable elements were the most abundant repetitive sequences. Class I retrotransposons and Class II DNA transposons contributed 33 % of the genomic sequence in the short arm, 45 % of the centromeric/pericentromeric heterochromatin of X chromosome and 20 % of long arm (Table 4). Among all transposable elements, Ty3/Gypsy retrotransposons were the most abundant repeats in all three regions and occupied 14 to 30 % of the genomic sequence of the X chromosome. The Ty1/Copia retrotransposons were the second most abundant transposable elements, ranging from 2.74 to 8.28 % among the three regions.

Table 4 Repeat content of chromosome 1
Full size table
Fig. 3
figure 3

Distribution of transposable elements and gene density along the papaya X chromosome. The 18 scaffolds were joined together as a pseudomolecule and subjected to transposable element distribution and gene density analysis. The percentage of total transposable element (green diamond) and number of genes per 100 kb (red square) in each scaffold were plotted against the physical distance from the telomere of the short arm (in base pair) to the telomere of the long arm. The X chromosome was divided into three regions: the short arm euchromatin, the long arm euchromatin, and the centromeric/pericentromeric heterochromatin region. The large blue dot represents the heterochromatin knob 1 in the centromeric/pericentromeric region, and the small blue dot represents the heterochromatic knob on the long arm. The gray bar highlights the centromeric/pericentromeric heterochromatin region

Full size image

The total amount of DNA transposons in centromeric/pericentromeric region was slightly higher than the euchromatin of short and long arms (Table 4). The majority of the DNA transposons on the X chromosome remained to be classified. A relatively higher number of CACTA elements were found in the euchromatin region on the long arm, a slightly lower number on the short arm, and the lowest number in the centromeric/pericentromeric heterochromatin region.

The number of predicted genes per 100 kb was calculated to compare the gene density along chromosome 1. We found 526 genes on the five scaffolds located on the short arm, 1,229 genes on the nine scaffolds located on the long arm, and 284 genes on the centromeric/pericentromeric heterochromatin region (Table 4). The long arm had the highest gene density at 15.18 genes per 100 kb, the short arm had 12.14 genes per 100 kb, and the centromeric/pericentromeric region showed the lowest density at 9.90 genes per 100 kb. Along the chromosome 1, the gene density increased toward the telomeres and decreased toward the centromere (Fig. 3; Table S2).

Discussion

Plant sex chromosomes, in contrast to sex chromosomes of mammals, evolved more recently and provide a particularly informative system to study the origin and process of sex chromosome evolution. It has been widely accepted that sex chromosomes may have evolved from a pair of autosomes by selective pressure to link two sex determination mutations, one that promoted maleness and the other suppressing femaleness (Charlesworth and Charlesworth 1978). These two sex determination loci would have become permanently linked when suppression of recombination occurred in this region. The absence of recombination would then lead to accumulation of deleterious mutations, degeneration of the Y chromosome, and the formation of heteromorphic sex chromosome pairs. The suppression of recombination between the sex determination genes is therefore the pivotal event in sex chromosome evolution (Ming et al. 2007).

Chromosomal recombination frequency is related with chromosome structures and can be altered by the proximity of certain architectural elements, such as the centromere or large blocks of heterochromatin (Roberts 1965). Furthermore, recombination suppression is a common characteristic of all eukaryotic centromeres (Roberts 1965; Choo 1998; Lambie and Roeder 1986; Roberts 1965; Tanksley et al. 1992). For example, artificially repositioning of the yeast Cen3 resulted in reduced recombination at new sites and increased recombination at former sites (Lambie and Roeder 1986). Thus, the centromere location might have provided initial reduction of recombination between sex determination genes, which facilitated the evolution of the papaya sex chromosomes.

A molecular cytogenetic map integrated with genetic linkage map and genomic sequence of sex chromosomes could potentially reveal features of genome structure and evolution that cannot be obtained from either linkage maps or genome sequence. The genome size of the sex determination region of the papaya sex chromosome is sufficiently small that has been genetically mapped so that if it were mapped cytogenetically, it could provide evidence for probable mechanisms in the evolution of plant sex chromosomes.

Three high-density genetic maps have been constructed for papaya using different marker systems and mapping populations (Blas et al. 2009; Chen et al. 2007; Ma et al. 2004). Severe recombination suppression was observed in the HSY region of all three genetic maps (Blas et al. 2009; Chen et al. 2007; Ma et al. 2004). Our molecular cytogenetic map revealed that the HSY region spanned about 12.2 % of the chromosome in cytological length and was covered by five knob-like heterochromatin structures. Low recombination or the lack of recombination has been frequently observed in heterochromatic centromere of plant and animal chromosomes (Roberts 1965; Choo 1998; Lambie and Roeder 1986; Roberts 1965; Tanksley et al. 1992). Recombination was completely suppressed in the centromere and pericentromere of the papaya chromosome 1, and resulted in the clustering of a large number of markers on the genetic maps. The order of the clustered markers cannot be determined solely by genetic mapping. We mapped 16 microsatellite marker-tagged BAC clones on pachytene chromosome spreads and a majority of these markers showed a good correlation between their order on the genetic and cytogenetic maps, with only three genetic markers changed in their relative positions. These three mis-mapped markers were cytogenetically mapped to either heterochromatic or pericentromeric/centromeric regions. The discrepancy between the genetic map and our cytogenetic map was likely caused by the skewed segregation in the highly heterochromatic or centromeric regions.

Cytogenetic analyses of chromosomes revealed the genetic and physical relationships in specific chromosome regions. Integration of the chromosome cytological features and molecular DNA sequence would facilitate the use of the genetic and physical maps for comparative genomics and may accelerate the cloning of economically important genes. We analyzed the recombination frequency along the papaya sex chromosomes. The ratio between genetic and physical distance across chromosome 1 was not constant, primarily due to variation in level of genetic recombination in different regions. As expected, the lowest recombination frequency was found in the centromeric heterochromatin. It was interesting that the long arm of chromosome 1 showed 60 % higher recombination than in the short arm. The long arm was genetically mapped as about twice the length of the short arm. However, our cytogenetic analysis indicated that chromosome 1 was sub-metacentric and the p/q ratio was 0.71. The difference between cytological and genetic length may be caused by different sequence composition of genes and repetitive sequences on the two arms.

Both the short and long arms of chromosome 1 were mainly euchromatic. However, one of the euchromatic arms was noted as more densely stained by DAPI than the other (Ming et al. 2008; Zhang et al. 2008). Our present analysis indicated that the long arm contained 32 % less repetitive sequences and 25 % higher gene density than the short arm. Different repetitive sequence contents found in the two arms may explain why the two arms were not evenly stained by DAPI.

Genetic recombination shapes an organism’s genome structure and plays an important role in the evolution of plant genomes (Gaut et al. 2007). Unequal crossing over and non-homologous recombination can generate mutation (Sturtevant 1925), change gene copy number (Zhang and Gaut 2003), and cause chromosomal rearrangements (McClintock 1929). Thus, genes in high recombination regions would evolve more rapidly than genes in low-recombination regions. Under natural selection, beneficial mutations would rise rapidly, and deleterious mutations would be eliminated more effectively in high recombination regions. Consistent with this evolutionary advantages of high recombination regions, gene density and recombination frequency are positively correlated in plant species (Anderson et al. 2006; Wright et al. 2003), moreover, the accumulation of transposable elements and organellar DNA is often found in low-recombination heterochromatic regions (Matsuo et al. 2005).

Papaya chromosome 1 conforms to the above generalities in that its gene density and recombination were positively correlated. The lowest recombination frequency was found in the centromeric/pericentromeric region where the gene density was the lowest as well. The gene density of the centromeric/pericentromeric region was 35 % lower and the repetitive sequences were 85 % higher compared with the long arm euchromatin region. In addition, recombination frequency and gene density increased towards the distal ends of both arms and reached maximum at the sub-telomere and telomeric ends. Although the trend is not universal, it has been reported in several plant species that recombination tends to be the highest in the distal chromosome regions (Drouaud et al. 2006; See et al. 2006; Wu et al. 2003). The high recombination in the telomeric regions might result in rapid evolution and make telomeric regions highly dynamic. Thus, insertions of transposable elements or organellar DNA in telomeric regions would not be stable and would likely be removed effectively. Consistent with this hypothesis, the telomeres of the papaya chromosome 1 contained the lowest repetitive sequences and the highest gene density. In an opposite trend, the pericentromeric/centromeric region contained the highest repetitive sequences and the lowest gene density.

It has been reported that retrotransposons are enriched in gene-poor and recombination-inactive regions (IRGSP 2005; Schnable et al. 2009). Our results showed a similar distribution of retrotransposons in the papaya chromosome 1. Detailed analysis of transposable elements along the papaya chromosome 1 revealed differential accumulation of different types of transposable elements. Ty3/Gypsy was the most abundant retrotransposable element and did not show obvious preferential accumulation across the entire chromosome. However, Ty1/Copia was more abundant in the centromeric/pericentromeric heterochromatin region than in the euchromatic arms. The overall ratio of Ty1/Copia to Ty3/Gypsy was about 1:5 at the genome level (Nagarajan et al. 2008), similar to the ratio observed in the long arm euchromatin region. The Ty1/Copia to Ty3/Gypsy ratio increased to 1:4 in the short arm euchromatin and to 1:3.6 in the centromeric/pericentromeric heterochromatin region. Although different patterns of Ty1/Copia accumulation have been observed in other plant species (Baucom et al. 2009; Peters et al. 2009), Ty1/Copia is more abundant in the pericentromeric heterochromatin than in the chromosome arms in Arabidopsis (Pereira 2004). Interestingly, Ty1/Copia is also mainly accumulates at Y chromosome of Silene latifolia, compared to autosomes (Cermak et al. 2008). Pereira (2004) reported that Ty1/Copia insertions are younger than the host species and that the uneven distribution of Ty1/Copia in these species has been caused by the rapid elimination of these elements from the genome, rather than having an insertional bias. As heterochromatic and MSY regions are recombinationally suppressed, the Ty1/Copia element may have slower evolutionary rate and thus retained in the genome. It would be interesting to test this hypothesis in papaya by genome-wide dating of its retrotransposons.

The papaya chromosome 1 centromeric/pericentromeric heterochromatin region contained relatively more DNA transposons than in the euchromatin arms. CACTA elements showed significant differential distribution among centromere/pericentromere and euchromatin arms. The euchromatin arms contained about twice the number of CACTA elements than did the centromeric/pericentromeric region. This unequal distribution of CACTA in chromosome 1 is similar to the situation in papaya chromosome 6, where we found three times more CACTA in each euchromatin arm than in the centromeric heterochromatin (Wai, C.M. and Yu, Q., unpublished data). Although the mechanism for differential distribution of CACTA is unknown, cytogenetic studies of transposable elements in Triticeae species revealed a significant reduction of En/Spm elements around centromeres (Altinkut et al. 2006). Preferential accumulation of CACTA in pericentromeric region is also observed in Arabidopsis (Miura et al. 2004).

The papaya chromosome 1 short arm euchromatin contained about twice the amount of DNA transposon MuDr than the centromeric/pericentromere heterochromatin and the long arm euchromatin. The centromeric heterochromatin contained similar amounts of MuDr as did the long arm euchromatin, suggesting the uneven distribution of this transposable element may not be related to recombination rate. Uneven distribution of specific retrotransposons and DNA transposons can be caused by their preferential insertions along the chromosome or by positive selection for their retention in the genome during evolution. It was also possible that the distribution patterns of transposable elements could be species-, or chromosome-specific. It has been reported that each transposable element has its own specific methylation environment which affect its distribution in the rice genome (Takata et al. 2007). This specificity may contribute to the diverse location of individual transposable elements among different species.