| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Genetics and Molecular Biology |
2Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS, UK; 3School of Biological Sciences, Queen Mary College, University of London, E1 4NS, UK; 4Institute of Biophysics, Academy of Sciences of the Czech Republic, Kralovopolska 135, CS-61265 Brno, Czech Republic; 5Florida Museum of Natural History and the Genetics Institute, University of Florida, Gainesville, Florida 32611 USA; 6Department of Botany and the Genetics Institute, University of Florida, Gainesville, Florida 32611 USA
Received for publication September 2, 2003. Accepted for publication March 11, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: centromere • chromosomal evolution • fluorescent in situ hybridization (FISH) • genome size • polyploidy • rDNA • subtelomere • Tragopogon
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Up to 70% of all species of flowering plants may have experienced one or more episodes of polyploidy in their evolutionary past, and many crops are of polyploid origin including maize, wheat, and cotton (e.g., Stebbins, 1971
; Grant, 1981
; Masterson, 1994
). Polyploids often have novel phenotypes that are not present in their diploid progenitors or exceed the range of the parental species (reviewed in Levin, 1983
, 2002
; Ramsey and Schemske, 2002
). The mechanisms by which polyploidy contributes to novel variation are not well understood, in part because polyploidization involves a number of genetic and epigenetic phenomena (reviewed in Soltis and Soltis, 1999
, 2000
; Matzke et al., 1999
; Wendel, 2000
; Liu and Wendel, 2002
; Osborn et al., 2003
; Soltis et al., 2004
). The degree of genomic change in recent natural and synthetic allopolyploids seems to vary across taxa. For example, synthetic polyploids in Gossypium (Liu et al., 2001
; Adams and Wendel, in press
) and recently formed natural polyploids in Spartina (Baumel et al., 2001
, 2002
; Ainouche et al., in press
) had few changes in overall genome structure. In contrast, Arabidopsis (Madlung et al., 2002
; Chen et al., in press
), Brassica (Song et al., 1995
; Pires et al., in press
), Triticum (Ozkan et al., 2001
; Shaked et al., 2001
; Levy and Feldman, in press
), and Nicotiana (Skalicka et al., 2003
; Kovarík et al., in press
) demonstrate genomic change in association with allopolyploidy.
Another phenomenon that can accompany polyploidization is a change in genome size. One might expect a general pattern of additivity in genome size between progenitor diploids and derivative polyploids, as is seen in Gossypium (Liu and Wendel, 2002
). However, there are also examples of non-additivity in polyploid genome size, as seen in polyploids reported to show decreased DNA amounts relative to diploids (e.g., Vigna, Parida et al., 1990
), as well as polyploids with increased amounts of DNA (e.g., Linum, Cullis, 1979
). In fact, the overall trend in polyploid evolution across the angiosperms is for "genome downsizing" (reviewed in Leitch and Bennett, in press
). These and other recent studies have drawn attention to the potentially dynamic nature of polyploids (e.g., Soltis and Soltis, 1999
). However, many aspects of the molecular consequences of polyploid formation and subsequent genome evolution remain to be elucidated.
The immediate consequences of polyploidization can best be studied in either synthetic polyploids or natural polyploids of recent ancestry. While evolutionary biologists can create synthetic polyploids from any number of taxa, only four plant genera have been documented to have generated natural polyploid species in the last 150 yr: Cardamine (Urbanska et al., 1997
), Spartina (Baumel et al., 2001
, 2002
; reviewed in Ainouche et al., in press
), Senecio (Ashton and Abbott, 1992
; reviewed in Abbott and Lowe, in press
), and Tragopogon (Ownbey, 1950
; reviewed in Soltis et al., in press
). Tragopogon (Asteraceae) has approximately 150 species native to Eurasia. Three diploid (2n = 2x =12) species (T. dubius, T. pratensis, and T. porrifolius) that were introduced into eastern Washington State, USA, and adjacent Idaho in the early 1900s have recently and recurrently formed two allopolyploid (2n = 4x = 24) species, T. mirus (T. dubius x T. porrifolius) and T. miscellus (T. dubius x T. pratensis) in western North America (Ownbey, 1950
; Ownbey and McCollum, 1954
; Brown and Schaak, 1972
; Soltis and Soltis, 1989
, 1991
, 1999
; Soltis et al., 1995
; Cook et al., 1998
). Ownbey and McCollum (1954)
used traditional cytogenetic methods to karyotype the six pairs of chromosomes in the introduced diploid species of Tragopogon. Along with morphological characters, Ownbey and McCollum (1954)
observed sufficient chromosomal variation (e.g., terminal knobs and secondary constrictions) among the different populations of diploid species to infer the multiple origins of the two recently formed allopolyploids, T. mirus and T. miscellus. These recurrent formations were later confirmed with molecular methods (Soltis et al., 1995
).
Our goal is to determine whether chromosomal rearrangements or changes in genome size have occurred in the Tragopogon allotetraploids since their recent formation. One way to survey for genomic rearrangements in polyploids is to locate repetitive DNA physically on chromosomes using the tools of molecular cytogenetics such as fluorescent in situ hybridization (FISH), a technique that has provided insights into genome and chromosome evolution (reviewed in Heslop-Harrison, 1991
, 2000
; Jiang and Gill, 1994
, 1996
; Leitch and Bennett, 1997
; Schwarzacher and Heslop-Harrison, 2000
; Singh, 2003
). While some researchers have localized rDNA loci onto chromosomes of members of Asteraceae (reviewed in Torrell et al., 2003
), few studies have isolated tandem repeats (satDNA) from representatives of the family for use in molecular cytological studies (Crepis, Jamilena et al., 1993
, 1995
; Brachycome, Houben et al., 2000
). This is surprising given the historical interest in chromosome evolution and polyploidy in this large plant family (e.g., Tahara, 1915
; Lawrence, 1929
; Gustafsson, 1932
; Babcock and Stebbins, 1938
; Clausen et al., 1945
; Carr, 1985
; Baldwin, 1993
). Because tandem repeats have provided insights into genome and chromosome evolution (Jiang and Gill, 1994
, 1996
; Heslop-Harrison, 2000
), we characterized four tandem repeats in Tragopogon: two novel tandem repeats (TPRMBO and TGP7) that we isolated specifically for this study and two rDNA loci (18S-5.8S-26S and 5S). We used molecular cytogenetics (FISH) to determine the number and distribution of these four tandem repeats to investigate the genetic consequences of allopolyploidy in Tragopogon. The FISH technique was carried out on multiple populations of the three diploid species (T. dubius, T. pratensis, and T. porrifolius), and karyotypes were constructed. The same probes were then hybridized to multiple populations of the recently formed allotetraploid species (T. mirus and T. miscellus) to determine if chromosomal rearrangements had occurred subsequent to polyploidization. We also measured DNA C values for the same diploid and tetraploid Tragopogon species to determine the dynamics of genome size evolution in this polyploid complex.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
using Pisum sativum cv. Minerva Maple (4C = 19.46 pg) as a calibration standard. In brief, root tips from newly germinated Tragopogon seeds were fixed in a freshly prepared solution of 3 : 1 ethanol : glacial acetic acid at –20°C overnight. Roots were hydrolyzed in 5 mol/L hydrochloric acid at 25°C for 35 min, stained in the dark for 2 h in Feulgen reagent at 23°C, then washed three times each for 10 min in SO2 water (per Hanson et al., 2001
). Root tips were squashed in 45% acetic acid, and cover slips were temporarily sealed with rubber solution. Slides were kept in a refrigerator overnight, and readings were taken the following day on a Vickers M85a scanning microdensitometer (York, Yorkshire, UK). Thirty nuclei (10 nuclei judged to be at mid-prophase in mitosis on each of three slides) were measured for each Tragopogon species and for the calibration standard (Pisum). For each slide, the mean and standard deviation of readings were calculated and used to obtain the overall mean and standard deviation. Arbitrary values were converted to picograms from the ratio of the mean absorbency of the Tragopogon test species to that of the calibration standard. For T. miscellus populations 2604 and 2605, for which we also sampled populations representing their diploid ancestors (T. dubius 2613 and T. pratensis 2608), we compared the genome size value predicted by the sum of the parental genomes with the observed C values using a modified t test (Sokal and Rohlf, 1980
, p. 231).
DNA extraction, Southern blotting, and DNA sequencing
Total genomic DNA was extracted from 1 g of fresh young leaves of each Tragopogon species using a cetyltrimethylammonium bromide (CTAB) method and ethanol and ammonium acetate precipitations (Doyle and Doyle, 1987
) with minor modifications (Smith et al., 1991
). Total genomic DNA was purified using spin columns in the QIAquick PCR purification kit (Qiagen, Valencia, California, USA). DNA was diluted to approximately 50–200 ng/mL. Methods for Southern blotting and methylation analysis follow standard procedures as described in Kovarík et al. (2000)
. DNA sequencing was carried out using the dideoxytermination method (Sequenase Version II DNA sequencing kit, USB, Amersham Biosciences, Cleveland, Ohio, USA). Sequence data were deposited, and the GenEMBL database was searched for similar sequences.
Cloning of tandem repeats from Tragopogon
Isolation of TPRMBO tandem repeat
DNA of Tragopogon pratensis was digested with an excess of MboI restriction endonuclease (10 units/g DNA) overnight and size-separated on a 7% polyacrylamide gel. A band of approximately 160 bp was excised, purified, and ligated into the BamHI site of pZero vector (Invitrogen, Carlsbad, California, USA). The ligation mix was used to transform E. coli DH5 
cells, and clones were screened for recombinants that labeled strongly in Southern blot hybridization against the 32P-labeled genomic DNA of T. pratensis and contained 160-bp inserts. Inserts from 14 candidate colonies were sequenced (GenEMBL #AJ307715-28) and aligned using Wisconsin GCG software. A consensus sequence was constructed because all colonies contained essentially the same sequence with only a few mutations.
Isolation of TGP7 tandem repeat
Because only low amounts of genomic DNA were obtained from T. porrifolius, we employed a novel strategy for the isolation of a repeat from this species. The method is based on the original AFLP methodology (Vos et al., 1995
) and involved the ligation of an adapter to the Sau3AI-restricted genomic DNA from T. porrifolius. The Sau3AI adapter was prepared by annealing two single-stranded complementary oligonucleotides (5'-GACGATGAGTCCTGAG-3', 5'-GATCCTCAGGACTCAT-3')] and 50 pmols was used in a 50-µL ligation mixture containing 0.5 µg restricted genomic DNA, reaction buffer, and 1 unit T4 ligase (Roche, Mannheim, Germany). After ligation (2 h, 20°C), the DNA was purified using the QIAquick polymerase chain reaction (PCR) purification kit (Qiagen), and approximately 1/10 volume was amplified by PCR (0.5 µmol/L primer [5'-GATGAGTCCTGAGGATC-3'] in 0.2 mmol/L dNTPs, 2.5 mmol/L MgCl2, enzyme buffer, and 1 unit of Taq polymerase; initial denaturation [3 min, 92°C] and 25 cycles of denaturation, 15 s at 92°C; annealing, 30 s at 50°C; extension, 30 s at 72°C). The PCR conditions were empirically adjusted to obtain 
100–200 ng of PCR products of <600 base pairs (bp) in length. Amplified DNA was purified using the QIAquick PCR purification kit (Qiagen), digested with Sau3AI, and ligated into the BamHI site of pBluescript vector (Stratagene, La Jolla, California, USA). The ligation mix was used to transform E. coli DH5 cells. Twenty randomly selected recombinant plasmids were digested with enzymes (BstXI + PstI) and screened by Southern hybridization with 32P-labeled genomic DNA of T. porrifolius. The clone with the strongest signal was recovered (TGP7delta) and sequenced. On Southern blots, the cloned sequence appeared shorter than the monomeric unit of the satellite. The PCR primers were designed with a 2-bp overlap to obtain the full-length clone. The PCR products were cloned into the EcoRV site of pBluescript KS + vector (Stratagene, La Jolla, California, USA) and sequenced (GenEMBL #AF493256).
Analysis of chromatin structure
Nuclei of T. pratensis were prepared for micrococcal nuclease (MNase) treatments using a protocol modified from Espinás and Carballo (1993)
. All operations were performed at 0–4°C. Approximately 1.5 g of fresh leaves were ground in liquid nitrogen and transferred to 50 ml of buffer A (10 mmol/L NaCl, 10 mmol/L MES pH 6.0, 5 mmol/L EDTA, 0.15 mmol/L spermidine, 20 mmol/L beta-mercaptoethanol, 0.6% TRITON X-100, 0.2 mol/L sucrose, and 0.1 mmol/L PMSF) and quickly homogenized. The homogenate was filtered through silk and centrifuged at 2620 x g for 10 min. The pellet was resuspended in 15 mL of buffer A, centrifuged again, resuspended in buffer B (88% mass/mass PERCOL in buffer A), and centrifuged at 5150 x g for 10 min. Floating nuclei (0.4 mL) were removed, resuspended in 1 mL of MNase digestion buffer (10 mmol/L Tris-HCl pH 7.5, 100 mmol/L, ammonium sulfate, 5 mmol/L MgCl2, 3 mmol/L CaCl2, 5 mmol/L beta-mercaptoethanol), and treated with 30 units/mL of MNase at 37°C. Aliquots were removed from the reaction mixture at various times and the digestion stopped with equal volumes of stop buffer (1% sarcosyl, 0.25 mol/L EDTA pH 8, 5 mmol/L EGTA, and 0.5 mol/L NaCl). The samples were treated with proteinase K for several hours at 52°C. DNA was extracted with phenol-chloroform-isoamyl alcohol and ethanol precipitated. Five micrograms of DNA were loaded and fractionated on a 2% agarose gel. After subsequent hybridization of blots with TPRMBO, TGP7, and rDNA, the signal in the oligonucleosomal fraction (mononucleosome, dinucleosome, etc.) was quantitatively evaluated (using a Storm phosphorimager; Storm, Molecular Dynamics, Sunnyvale, California, USA). The average length of DNA was expressed in nucleosome multiples for each selected time interval of MNase digestion according to the method described in Fulnecek et al. (2002)
. The initial rate of MNase digestion was calculated and expressed as a decrease of fragment size per time interval. Only the nearly linear regions of curves (0–3 min) were considered for calculation of digestion rates.
Chromosome preparation and fluorescent in situ hybridization
Chromosome preparations and fluorescent in situ hybridization (FISH) were performed for multiple populations of the three parental diploid species (T. dubius, T. pratensis, and T. porrifolius) and two allopolyploid derivatives (T. mirus and T. miscellus) following Leitch et al. (1994)
with slight modifications as in Lim et al. (2000b)
. Healthy root tips of germinating seeds were pretreated at 4°C for 18 h and fixed at room temperature in 3 : 1 ethanol : glacial acetic acid overnight. Root tips were digested in 0.2% (mass/volume) cellulase R10, 0.2% (mass/volume) pectolyase Y23, and 0.2% (mass/volume) drieselase for 30 min and then transferred to 1% citrate buffer for 3–4 h. The meristematic cells behind the root cap were dissected, macerated, and squashed on a glass slide in a drop of 60% acetic acid. Five probes were used in the FISH experiments: (1) TPRMBO, (2) TGP7, (3) TRS (telomere repeat sequence, Cox et al., 1993
), (4) pTa71, a clone containing a 9-kb EcoRI fragment of 18S-5.8S-26S rDNA and the intergenic spacer from Triticum aestivum (Gerlach and Bedbrook, 1979
); and (5) pTZ19-R, a 120-bp fragment of the 5S rDNA unit isolated from Nicotiana rustica (Venkateswarlu et al., 1991
). Probes cut from plasmids were labeled using nick translation as follows: TPRMBO, TRS, and 5S rDNA were labeled with digoxigenin-11-dUTP (Roche Biochemicals, Sussex, UK) detected with antidigoxigenin-FITC; TGP7 and 18S-5.8S-26S rDNA were labeled with biotin-16-dUTP (Sigma-Aldrich Company Ltd, Poole, Dorset, UK) detected with avidin-Cy3 (Amersham Pharmacia Biotech, Little Chalfont, UK). The probe mixture was prepared as in Lim et al. (2000b)
. After overnight hybridization at 37°C, the slides were given a stringent wash in 20% (volume/volume) formamide in 0.1% SSC at 40–42°C. Chromosomes were counterstained for DNA with 4,6-diamidine-2-phenylindole, dihydrochloride (DAPI). Images were taken with a Leica Aristoplan epifluorescent microscope, and photographs were scanned and processed using Adobe Photoshop (Adobe Systems, Edinburgh, UK). Only those functions that apply equally to all pixels in the image were used. For every set of observations, between five and 10 metaphases were examined from each of 1–3 slides.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The DNA C values of T. mirus are not substantially different from the sum of the diploid parents. The genome size of T. mirus (population 2601 from Pullman, Washington, USA; mean 4C DNA value of 24.33) is additive of the genome sizes of its putative diploid progenitor populations, T. dubius (population 2613 from Pullman, Washington USA; mean 4C DNA value of 11.76) and T. porrifolius (population 2611 Pullman, Washington, USA; mean 4C DNA value of 12.5). Although we measured 4C DNA values for another population of T. mirus (2603 from Rosalia, Washington, USA), we cannot comment unequivocally on the additivity of its genome size because we were able to measure only the genome size of the T. dubius parent; plants of the second parent, T. porrifolius, are no longer present in Rosalia and are presumed extinct. Nonetheless, the two 4C DNA measurements for plants of T. porrifolius from other localities are very similar.
In contrast, the genome size of at least some populations of T. miscellus appears to have undergone downsizing when compared to its diploid progenitors. Populations 2604 (Moscow, Idaho, USA) and 2605 (Pullman, Washington, USA) have the same diploid parental genotypes (T. pratensis [2608 Moscow, Idaho, USA] and T. dubius [2613 Pullman, Washington, USA]), but are the result of reciprocal parentage (Soltis and Soltis, 1989
). The genome size values for these two populations are 20.30 pg and 20.99 pg, respectively, which are lower than the value predicted (24.20 or 23.27 pg) by adding the values of the diploid parents (t = 18.79958, P < 0.05; t = 13.88154, P < 0.05, for predicted values of 24.20 and 23.27, respectively). This level of downsizing—approximately 15%—is similar to that reported in the allopolyploids Brassica napus, B. juncea, and B. carinata (Naryan, 1998
). The genome size of a third population of T. miscellus (2606 from Spangle, Washington, USA) is 21.76 pg; the value for the parental T. pratensis from Spangle is 11.09 pg, but we lack genome size data for T. dubius from Spangle.
Detection, isolation, and characterization of the TPRMBO repeat
Total genomic DNA of T. pratensis digested with MboI yielded 160-bp bands after size fractionation on PAGE and ethidium-bromide staining (not shown). The 160-bp fragments were excised and cloned (GenEMBL #AJ307715-28). One clone, named TPRMBO4, was used as the probe for Southern hybridization. The ladder of bands generated using different restriction endonucleases in the Southern hybridization experiments indicates that TPRMBO is a tandem repetitive sequence (Fig. 1a). Regular ladders were found with Sau3AI and MboI restriction enzyme digestions and irregular ladders in higher multimer fractions using RsaI, AluI, MboII, SspI, and DraI (Fig. 1a). Database searches revealed no sequence similarities to other known families of repetitive sequences. The sequenced TPRMBO monomers are 160 ± 2 bp in length and moderately A + T rich (67%). A search for internal structural motifs revealed no direct repeats, palindromes, or inverted repeats, although the sequence does have several oligoA and oligoT motifs in an arrangement typical for tandemly arranged and curved satellite sequences (Fitzerald et al., 1994
).
|
). Thus, the data show that TPRBMO contains some methylation at asymmetrical nucleotide sites. To examine the occurrence of TPRMBO repeats in species of Tragopogon, we digested genomic DNA from T. dubius, T. pratensis, T. porrifolius, T. mirus, and T. miscellus with HindIII and hybridized the blot with the TPRMBO4 probe (Fig. 1b). A hybridization signal of comparable intensity was found in all species. Band polymorphisms were revealed in low-molecular-mass fractions. The species-specific hybridization bands (indicated by asterisks, Fig. 1b) were additive in lanes loaded with DNAs from the allotetraploids T. mirus and T. miscellus. It is likely that Tragopogon allotetraploids inherited the genomic arrangement of TPRMBO repeats from their respective diploid ancestors.
Isolation and characterization of the TGP7 repeat
A novel isolation strategy was used to obtain a repeated sequence from T. porrifolius. The main advantage of this AFLP-based cloning strategy is the ability to isolate a sequence even when only low amounts of DNA or degraded DNA are available. While conventional band excision cloning requires at least 5000 ng of high-molecular-mass DNA, the described method enabled the successful cloning of a repeat from only 50 ng of genomic DNA. The only limitation of this method appears to be the requirement for a GATC site in the monomeric unit of a repeat. This limitation could be overcome by using multiple adapters to target sequences of frequent-cutting restriction enzymes (e.g., MseI, RsaI, TaqI).
We recovered a clone carrying a 532-bp insert that hybridized strongly with 32P-labeled genomic DNA from T. porrifolius. This clone, TGP7, was used as a probe for Southern hybridization (Fig. 2). Highly regular ladders of bands were obtained using Sau3AI, HaeIII, RsaI, BstNI (Fig. 2a), and HindIII (Fig. 2b). These restriction enzymes indicated a tandem arrangement of units. Most of the repeats were digested with HindIII into monomeric units of 530 bp, suggesting a conserved HindIII site. MboI digested DNA more efficiently than its methylation-sensitive Sau3AI isoschisomer, suggesting cytosine methylation of some GATC motifs. Similar results were obtained with MspI/HpaII and BstNI/EcoRII isoschisomeric pairs.
|
To determine the occurrence of TGP7 repeats in other Tragopogon species, genomic DNA from each species was digested with HindIII and probed with TGP7 (Fig. 2b). The ladder of signals indicates that TGP7 is a tandem repetitive sequence found in both diploid and allotetraploid Tragopogon species. However, the signal in the lane loaded with DNA from T. dubius was significantly less (three- to four-fold), suggesting a lower content of TGP7 repeats in this species compared to the others.
Analysis of chromatin structure
There may be spatial control of DNA condensation as seen for repetitive subtelomeric and intercalary sequences of wheat and related species (Vershinin and Heslop-Harrison, 1998
). Therefore, we compared the chromatin structure of TGP7 and TPRMBO in Tragopogon by using a micrococcal nuclease sensitivity assay. As expected, both probes revealed a pattern of bands with nucleosomal periodicity (Fig. 3a). However, the two probes differed in some aspects. First, the length of the nucleosomal unit containing TPRMBO sequence appeared shorter (by 5–7 bp) than that of TGP7—based on length differences between corresponding neighboring bands (mononucleosome–dinucleosome, dinucleosome–trinucleosome, etc.). This feature of TPRMBO chromatin could be related to the relatively short (160-bp) TPRMBO monomeric units. Second, the kinetics (rate) of digestion of the sequences were different (Fig. 3b). The rates of DNA shortening calculated from slopes of kinetic curves were –2.25 nucleosome/min for 18S rDNA (Southern blot data not shown), –2.21 nucleosome/min for TGP7, and –1.25 nucleosome/min for TPRMBO. Therefore, the TPRMBO chromatin appears substantially less accessible to digestion than TGP7 and 18S rDNA loci and is probably more compact. Third, while the basic MNase digestion profile for TPRMBO reveals a mononucleosomal periodicity, in TGP7 every third band is more prominent (Fig. 3c). It is probable that the higher order structure of TGP7 is of a basic unit with a trinucleosomal length (530–540 bp).
|
|
|
Chromosomal location of rDNA loci
The FISH technique was conducted on T. dubius (Fig. 5a), T. pratensis (Fig. 5b), T. porrifolius (Fig. 5c), T. miscellus (Fig. 5d), and T. mirus (Fig. 5e–f) using 18S-5.8S-26S rDNA and 5S rDNA (Fig. 5a– f) as probes. The distributions of the rDNA loci on the diploid species of Tragopogon are shown on the karyotypes (Fig. 5g). Tragopogon dubius has one pair of 18S-5.8S-26S rDNA loci and one pair of 5S rDNA loci on the largest pair of chromosomes (Fig. 5a). This largest pair of chromosomes corresponds to chromosome pair A following Ownbey and McCollum's karyotype nomenclature (1954)
. Tragopogon pratensis also has one pair of 18S-5.8S-26S rDNA loci and one pair of 5S rDNA loci on chromosome pair A (Fig. 5b). In contrast, T. porrifolius has two pairs of 18S–5.8S–26S rDNA loci and two pairs of 5S rDNA loci (Fig. 5c). One pair of each type of rDNA locus is on chromosome A like the other two diploids; the extra 18S–5.8S–26S rDNA locus is on chromosome D, and the extra 5S rDNA locus is on chromosome F (Fig. 5g). The observation of an extra 18S–5.8S–26S locus on chromosome D in T. porrifolius is consistent with the additional satellite seen on Ownbey and McCollum's (1954)
karyotypes and with restriction site studies of the 18S-5.8S-26S rDNA region (Soltis and Soltis, 1991
). The rDNA hybridization signals in the polyploids T. miscellus (Fig. 5d) and T. mirus (Fig. 5e–f) were additive of those observed in the diploids (Fig. 5a–c, g). In sum, no changes in the distribution or abundance of either rDNA locus were detected in the polyploids relative to their diploid progenitors.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, 1971
; Grant, 1981
; Levin, 1983
), it is important to understand the range of molecular evolutionary phenomena associated with genome doubling. Genome doubling causes a range of responses that vary across taxa (reviewed in Soltis and Soltis, 1999
; Matzke et al., 1999
; Wendel, 2000
; Liu and Wendel, 2002
; Osborn et al., 2003
; Soltis et al., 2004
). With respect to chromosomal rearrangements, some polyploid genomes appear to be the sum of their parental diploid parts, such as Gossypium (Liu et al., 2001
; Adams and Wendel, in press
) and Spartina (Baumel et al., 2001
, 2002
; Ainouche et al., in press
). Other polyploids such as Brassica (Song et al., 1995
; Pires et al., in press
), Triticum (Ozkan et al., 2001
; Shaked et al., 2001
; Levy and Feldman, in press
), and Nicotiana (Lim et al., 2000a
; Skalicka et al., 2003
; Kovarík et al., in press
) undergo rapid and dynamic genomic changes. Similarly, polyploid taxa can vary in genome size because some polyploids have a general pattern of additivity in genome size when compared to their progenitor diploids, while other polyploids do not (reviewed in Leitch and Bennett, in press
). Here, we compare diploid and polyploid Tragopogon to determine whether rapid changes in genome size and chromosomal rearrangement may have accompanied recent polyploidization.
Genome size in diploid and polyploid Tragopogon
Tragopogon diploids vary in genome size, both within and among populations. Although measuring intrapopulational variation in genome size is technically difficult (Greilhuber, 1998
), variation up to 40% has been confirmed in some species (e.g., Zea mays). In Zea mays, intrapopulation variation has been associated with changes in the heterochromatin knob number (Greilhuber, 1998
). In addition, there are now reports correlating intraspecific variation in genome size of Hordeum with changes in copy number of retrotransposons (Vicient et al., 1999
).
Tragopogon mirus does not show large departures from additivity of the genome sizes of its diploid progenitors (Table 1). This is consistent with genome size results for other polyploid taxa in the Asteraceae. Specifically, polyploids in Artemisia (Torrell and Valles, 2001
), Leucanthemum (Asteraceae; Marchi et al., 1983
), Hypochaeris (Cerbah et al., 1999
), and Microseridinae (Asteraceae; Price and Bachmann, 1975
) all show additivity of diploid genome sizes.
However, at least two populations of T. miscellus show evidence of genome downsizing. Thus, T. miscellus allopolyploids may be similar to polyploids in Brassica and Poaceae, and in angiosperms generally, in which both recent and ancient allopolyploids show signs of genome reduction (reviewed in Levy and Feldman, 2002
; Leitch and Bennett, in press
).
Repetitive DNA
To assess whether large-scale chromosomal rearrangements may have occurred, we used four tandem repetitive sequences (TPRMBO, TGP7, 18S–5.8S–26S rDNA, 5S rDNA) as molecular cytogenetic markers. Repetitive DNAs are useful molecular cytogenetic markers (Jiang and Gill, 1994
, 1996
; Heslop-Harrison, 2000
) and can be used to detect chromosomal rearrangements in polyploids (Lim et al., 2000b
; Mishima et al., 2002
; Kotseruba et al., 2003
). Repetitive DNA includes dispersed and tandem repeats, which can be further categorized into rDNA and satellite DNA (satDNA). Blocks of satDNA often form heterochromatin predominantly in the centromeric and/or subtelomeric regions of the chromosome (Ganal et al., 1991
; Maluszynska and Heslop-Harrison, 1991
; Hennig, 1999
; Kishii et al., 1999
, 2001
; Guerra, 2000
). The association of satDNA tandem repeats with heterochromatic regions, as well as their unit length similarity (150–180 bp) and nucleosomal periodicity, suggests a role in the maintenance of chromosomal structure (Martinez-Zapater et al., 1986
; Heslop-Harrison and Bennett, 1990
; Heslop-Harrison, 1991
; Schmidt and Heslop-Harrison, 1998
). Only a few tandem repeats have been isolated from species of Asteraceae (e.g., Jamilena et al., 1993
, 1995
; Houben et al., 2000
), including those reported here. The Tragopogon genome has at least two kinds of satDNA repeats, TPRMBO and TGP7.
TPRMBO is a 160-bp centromeric tandem repeat
TPRMBO is a typical tandem repeat that has a predominantly centromeric localization to all chromosomes of all diploid Tragopogon species studied. It probably represents the main component of centromeric heterochromatin because the FISH signal corresponds to the DAPI-positive bands at the centromere (Fig. 4). TPRMBO has several characteristics typical of a centromeric repeat. First, the size of its monomeric unit (160 bp) appears to be similar to other satellites forming centromeric heterochromatin in plants (Martinez-Zapater et al., 1986
; Crowhurst and Gardner, 1991
; Rajagopal et al., 1999
; Pedrosa et al., 2001
) and animals (Mestrovic et al., 2000
). Second, irregular ladders revealed by some enzymes (e.g., RsaI, AluI, MboII, SspI, DraI in Fig. 1a) indicate considerable heterogeneity in sequence and arrangement of the TPRMBO repeats. The arrays of regular tandem repeats at the centromere could be intermingled with unrelated sequences (e.g., retrotransposon-related DNA sequences, Miller et al., 1998
; Presting et al., 1998
; Wiens and Sorger, 1998
; Francki, 2001
; Saunders and Houben, 2001
; Zhang et al., 2001
). Third, in most restriction enzyme digests, considerable undigested relic DNA was observed in the upper part of the gel, suggesting mutation of the restriction site in a number of units (Fig. 1a). Indeed, sequencing of 14 clones revealed relatively wide divergence among the TPRMBO units, with sequence similarities ranging from 82 to 93%. It is apparent that the relatively conserved regions (e.g., at the 5' and 3' terminus) were flanked by stretches of higher sequence divergence, suggesting functional constraints imposed on only some subregions. Chromosome-specific polymorphisms have been described in centromeric satellites of Arabidopsis thaliana (Heslop-Harrison et al., 1999
) and Ornithogalum longibracteatum (Pedrosa et al., 2001
). We speculate that TPRMBO sequence heterogeneity could reflect locus-specific differences in the Tragopogon genome. TPRMBO also showed some dispersion at pericentromeric locations (Fig. 4d, g), similar to satellite sequences in Ornithogalum longibracteatum (Pedrosa et al., 2001
).
Mammalian centromeres contain an histone H3 variant called CENP-A (Palmer et al., 1991
). Although it remains to be determined whether CENP-A is actually present in the form of a canonical nucleosome structure in vivo, it may be possible that binding of CENP-A could account for some of the distinct features of TPRMBO chromatin, including its relatively high degree of condensation and short nucleosomal periodicity.
TGP7 is a 532-bp subtelomeric tandem repeat
The FISH technique revealed that TGP7 is a typical subtelomeric tandem repeat in the chromosomes of the three diploid species of Tragopogon examined (Figs. 4, 5). The 532-bp monomeric length of TGP7 was conserved across the Tragopogon species studied (Fig. 2b). Monomers of similar lengths have been found for subtelomeric repeats in other genera (e.g., 570 bp in Leymus, Kishii et al., 1999
, and Secale, Vershinin et al., 1996
). The sequence is methylated, which is also typical of tandem repeats (Fig. 2a). In other satellite repeats, short sequences, such as oligo dA and CAAAA motifs, are over-represented (Macas et al., 2002
). TGP7 also has several oligo dA tracts, and these can account for conformation polymorphisms observed in the satellite DNA (Fitzerald et al., 1994
). In addition, six CAAAA motifs were found (five on the bottom strand) at a frequency that greatly exceeded expectation. It has been proposed that CAAAA pentanucleotides in heterochromatin might be involved in the breakage–reunion properties of repeated sequences (Macas et al., 2002
) and can lead to sequence translocations and chromosomal rearrangements (Langdon et al., 2000
). However, our analysis by FISH suggests no such evolution in the allopolyploids investigated. The restriction enzyme digestions of TGP7 repeats, using several enzymes, showed highly regular ladders in both diploids and allotetraploids (Fig. 2). Perhaps the TGP7 sequences in subtelomeric regions have experienced recent expansion and/or homogenization.
In MNase digests, the TGP7 bands revealed a periodicity of multiples of three basic TGP7 (3, 6, 9 ... n) units (Fig. 3), suggesting that every third linker of a nucleosome fiber is more sensitive to nuclease cleavage. If this is the case, nucleosomal linker regions along the long, tandemly repeated TGP7 sequence might not be structurally equivalent. Such anisotropy could have several explanations. (1) There might be binding sites for proteins that prevent or enhance DNA cleavage. For example, HMGY protein binds to AT-rich tracts in mouse alphoid satellite DNA (Solomon et al., 1986
). (2) There might be differential epigenetic modification of chromatin by methylation and/or acetylation occurring at a three-nucleosomal periodicity. Certainly heterogeneity in cytosine methylation occurs among units of the tobacco subtelomeric satellite repeat sequence HRS60 (Kovarík et al., 2000
) and perhaps in other simple satellites as well (Saunders and Houben, 2001
). (3) Nucleosome binding to a preferential structural motif could induce a conformation change in DNA, resulting in altered sensitivity of DNA toward nuclease. The prominent structural motifs in satellites are oligo dA tracts (Macas et al., 2002
), known to induce DNA curvature (Goodsell et al., 1994
). The dA4–6 tracts are clustered in approximately one half of the TGP7 monomer, while the second half is relatively devoid of these motifs. Such uneven distribution of oligo (dA) tracts has been observed in a 1.2-kb satellite of Secale cereale (Japanese rye, Nagaki et al., 1999
) and could contribute to higher-order packing of chromatin.
Organization and distribution of TPRMBO and TGP7 in Tragopogon allopolyploids
The repetitive sequences TGP7 and TPRMBO appear to be directly inherited in the allopolyploids without structural, organizational, or distributional changes. This inference is supported by additivity of both Southern hybridization bands (Figs. 1b, 2b) and FISH signals at metaphase (Fig. 4). The subtelomeric repeat TGP7 showed this more clearly because there were only eight subtelomeric FISH signals (corresponding to four genetic loci) in T. dubius and 18 in T. porrifolius and T. pratensis. The allotetraploids each had 26 subtelomeric signals, suggesting that the parental chromosome subtelomeric regions remained intact. In allotetraploid cotton, the two parental diploid genomes (A and D) have distinct repetitive elements. Following polyploidy, there has been substantial colonization of the D-genome by A-genome repetitive elements (Hanson et al., 1998
; Zhao et al., 1998
).
Organization and distribution of rDNA in Tragopogon diploids and allopolyploids
The distributions of the rDNA loci on the diploid species of Tragopogon were consistent with previous cytogenetic observations and restriction site studies (Ownbey and McCollum, 1954
; Soltis and Soltis, 1991
). Tragopogon dubius and T. pratensis each have one pair of 18S-5.8S-26S rDNA loci and one pair of 5S rDNA loci on the largest pair of chromosomes (Fig. 5a, b), while T. porrifolius has two pairs of 18S-5.8S-26S rDNA loci and two pairs of 5S rDNA loci (Fig. 5g). The number and distribution of the rDNA loci in the polyploids T. mirus (Fig. 5d) and T. miscellus (Fig. 5e–f) were additive of those observed in the diploids (Fig. 5a– c, g). This is similar to the situation in Nicotiana tabacum, where most cultivars have a number of rDNA loci that reflect the sum of the diploid progenitors (Kenton et al., 1993
; Lim et al., 2000b
; Kovarík et al., in press
). This pattern contrasts with other allopolyploid systems, such as Zingeria (Poaceae; Kotseruba et al., 2003
) and Sanguisorba (Rosaceae; Mishima et al., 2002
), where rDNA loci were evidently lost after polyploidization.
Chromosome evolution in diploid and polyploid Tragopogon
Collectively, the distribution of the four tandem repetitive DNA loci (TPRMBO, TGP7, 18S-5.8S-26S rDNA, and 5SrDNA) among the chromosome pairs allowed the construction of molecular cytogenetic karyotypes for the three diploid species of Tragopogon (Fig. 5g). Thus, the number, location, and intensity of the FISH signals for all the mapped loci allowed for the identification of several of the diploid parental chromosomes in the polyploids. If rearrangements had taken place upon or immediately following polyploidization, we would expect to observe nonadditive patterns in the polyploids. For example, the number of rDNA loci in a polyploid could be greater or fewer than that found separately in the two diploid progenitors. Alternatively, rearrangements could move subtelomeric repeats found in the diploids to interstitial locations in the polyploids. Some synthetic allopolyploid species, including members of Triticeae (Shaked et al., 2001
) and Brassica (Song et al., 1995
), display rapid (within a few generations) evolution of polymorphic markers. However, other polyploid species show additive patterns, as we observed for Tragopogon; these include synthetic polyploids in Gossypium (Liu et al., 2001
) and the natural Nicotiana polyploids (N. rustica and N. arentsii) compared with their putative diploid parents (Lim et al., 2004
).
We found no evidence for major genomic rearrangements in the allopolyploids T. mirus and T. miscellus. The number and location of the tandem repetitive sequences TGP7 and TPRMBO appear to be directly inherited in the allopolyploids from their corresponding diploid ancestors without organizational or distributional changes. Similarly, the 18S-5.8S-26S and 5S rDNA loci in T. miscellus and T. mirus were exactly as predicted from the number and location of these loci in their diploid progenitors. There are several possible explanations for our inability to detect alterations in parental subgenomes in the allotetraploid species of Tragopogon. An obvious possibility is that perhaps no such rearrangements have occurred in T. mirus and T. miscellus. Alternatively, allopolyploid-induced changes may have occurred in Tragopogon, but are located in the middle- and low-repetitive fractions of the genome that have not been analyzed in this study. That is, we used a coarse approach; there may be microrearrangements that these methods were not able to detect. Similarly, ribosomal RNA genes, whose sequence and copy numbers are supposed to be under greater selective constraints than the non-transcribed satellite repeats, can undergo homogenization in allopolyploids (Wendel et al., 1995
; Volkov et al., 1999
; Lim et al., 2000a
). In future cytogenetic studies and fine-scale mapping studies, these unexplored areas of genome structure in Tragopogon can be examined. However, based on the tandem repetitive sequences in the allopolyploids studied, we have been unable to detect major changes in overall genome structure that may have occurred within the time frame of 80 years since polyploidization.
While polyploid genomes may be quiescent in some respects, they can simultaneously be dynamic in other aspects. For example, polyploids in Gossypium and Nicotiana have an additive number of rDNA loci, but still exhibit concerted evolution between loci where repeats at the different loci become homogenized to the same sequence (reviewed in Adams and Wendel, in press
; Kovarík et al., in press
). In addition, rDNA loci may be additive in number, but then exhibit differences in gene expression. Interspecific hybrids often have rRNA genes of one parent functionally dominant over the rRNA of the other parent (nucleolar dominance, Navashin, 1934
), and there are many examples of such regulation of rRNA gene activity in allopolyploids (reviewed in Pikaard, 2000
). Ongoing studies suggest that concerted evolution and nucleolar dominance may be occurring in the recently formed polyploids Tragopogon mirus and T. miscellus (A. Kovarík et al., unpublished data). These phenomena could have profound impacts on attempts at phylogeny reconstruction based on rDNA loci such as ITS. Future studies in Tragopogon will put chromosome evolution into a broader phylogenetic context as is being done in other genera of the Asteraceae (Hypochaeris, Cerbah et al., 1999
; Samuel et al., 2003
; Weiss et al., 2003
) and in other groups (Sanguisorba, Rosaceae, Nicotiana, Solanaceae, Lim et al., 2000b
, 2004
; Mishima et al., 2002
).
|
FOOTNOTES |
|---|
7 Current address: Department of Agronomy, University of Wisconsin, Madison, Wisconsin 53706 USA (phone 608-262-8228; FAX 608-262-5217; e-mail: jcpires{at}facstaff.wisc.edu
)
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Adams K. L. J. F. Wendel In press Exploring the genomic mysteries of polyploidy in cotton. Biological Journal of the Linnean Society.
Ainouche M. L. A. Baumel A. Salmon In press Spartina anglica Schreb.: a natural model system for analysing early evolutionary changes that affect allopolyploid genomes. Biological Journal of the Linnean Society.
Ashton P. A. R. J. Abbott 1992 Multiple origins and genetic diversity in the newly arisen allopolyploid species, Senecio cambrensis Rosser (Compositae). Heredity 68: 25-32
Babcock E. B. G. L. Stebbins 1938 The American species of Crepis: their relationships and distribution as affected by polyploidy and apomixis. Carnegie Institute of Washington, Publication No. 504.
Baldwin B. G. 1993 Molecular phylogenetics of Calycadenia (Compositae) based on ITS sequencing of nuclear ribosomal DNA—chromosomal and morphological evolution reeaximined. American Journal of Botany 80: 222-238[CrossRef][ISI]
Baumel A. M. L. Ainouche R. Kalendar A. H. Schulman 2002 Retrotransposons and genomic stability in populations of the young allopolyploid species Spartina anglica C. E. Hubbard (Poaceae). Molecular Biology and Evolution 19: 1218-1227
Baumel A. M. L. Ainouche J. E. Levasseur 2001 Molecular investigations in populations of Spartina anglica C. E. Hubbard (Poaceae) invading coastal Brittany (France). Molecular Ecology 10: 1689-1701[CrossRef][Medline]
