Copyright Ó 2007 by the Genetics Society of America DOI: 10.1534/genetics.107.072751 Concerted Evolution of rDNA in Recently Formed Tragopogon Allotetraploids Is Typically Associated With an Inverse Correlation Between Gene Copy Number and Expression Roman Matya´sˇek,* Jennifer A. Tate,† Yoong K. Lim,‡ Hana Sˇrubarˇova´,* Jin Koh,† Andrew R. Leitch,‡ Douglas E. Soltis,† Pamela S. Soltis§ and Alesˇ Kovarˇ´ık*,1 *Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i, Laboratory of Molecular Epigenetics, Kra´lovopolska´ 135, CZ-61265 Brno, Czech Republic, †Department of Botany, University of Florida, Gainesville, Florida 32611, ‡School of Biological Sciences, Queen Mary, University of London, E1 4NS, UK and §Florida Museum of Natural History, University of Florida, Gainesville, Florida 32611 Manuscript received March 1, 2007 Accepted for publication May 23, 2007 ABSTRACT We analyzed nuclear ribosomal DNA (rDNA) transcription and chromatin condensation in individuals from several populations of Tragopogon mirus and T. miscellus, allotetraploids that have formed repeatedly within only the last 80 years from T. dubius and T. porrifolius and T. dubius and T. pratensis, respectively. We identified populations with no (2), partial (2), and complete (4) nucleolar dominance. It is probable that epigenetic regulation following allopolyploidization varies between populations, with a tendency toward nucleolar dominance by one parental homeologue. Dominant rDNA loci are largely decondensed at interphase while silent loci formed condensed heterochromatic regions excluded from nucleoli. Those populations where nucleolar dominance is fixed are epigenetically more stable than those with partial or incomplete dominance. Previous studies indicated that concerted evolution has partially homogenized thousands of parental rDNA units typically reducing the copy numbers of those derived from the T. dubius diploid parent. Paradoxically, despite their low copy number, repeats of T. dubius origin dominate rDNA transcription in most populations studied, i.e., rDNA units that are genetic losers (copy numbers) are epigenetic winners (high expression). PIGENETIC silencing of parental genes is an im- DNA methylation and histone acetylation (for review E portant feature of newly established allopolyploids see Preuss and Pikaard 2007). Other factors may also and is believed to reconcile regulatory incompatibilities play a role in ND, e.g., the position of rRNA genes in the between parental subgenomes (Adams and Wendel nucleus (Shaw and Jordan 1995), the relative amount 2005; Birchler et al., 2005). Nucleolus organizer re- of heterochromatin (Viegas et al. 2002), and the activity gions (NORs) contain tandemly arranged highly reit- of unlinked genes (Lewis et al. 2004). Despite consid- erated ribosomal rRNA genes coding for 18S-5.8S-26S erable progress in our understanding of ND, little is rRNA whose expression is under epigenetic control known about the relationship between ND and sequence (Pikaard 2000; for review see Neves et al. 2005). homogenization, particularly in young populations of Indeed nucleolar dominance (ND) was one of the first newly formed polyploidy species. The goal of this article examples of differential gene expression discovered in is to study this relationship and the inheritance patterns plant hybrids nearly a century ago (Navashin 1934). of ND and sequence homogenization at individual Numerous studies in both plants and animals support and population levels in recently formed allotetraploid the view that ND/rRNA gene silencing may be stable species. or unstable, partial, or reversible (Volkov et al. 2007). Besides epigenetic silencing homogenization of re- In Arabidopsis and Brassica biochemical (Chen and peats seems to be another interesting aspect of rDNA Pikaard 1997a) and genetic (Lawrence et al. 2004; biology. The consequence of homogenization is little Probst et al. 2004) studies provided evidence that rRNA sequence variability among the units within and across silencing depends on epigenetic factors, including the arrays in genome. In many allopolyploids’ rDNA one particular unit overwrites preexisting units, probably mediated by recombination or copy number expansion/ Sequence data from this article have been deposited with the EMBL/ contraction mechanism that leads to the fast evolu- GenBank Data Libraries under accession nos. AM493990 and AM493995. tion of novel families, a process called concerted evo- 1Corresponding author: Institute of Biophysics, Academy of Sciences of lvarez endel the Czech Republic, v.v.i., Laboratory of Molecular Epigenetics, Kra´lovo- lution (for review see A and W , 2003; polska´ 135, CZ-612 65 Brno, Czech Republic. E-mail: [email protected] Volkov et al. 2007). However, the frequency/occurrence Genetics 176: 2509–2519 (August 2007) 2510 R. Matya´sˇek et al. of homogenization does vary between species. Allopo- TABLE 1 lyploids of Arabidopsis and Brassica apparently have no Summary of collections of Tragopogon used rDNA unit loss or concerted evolution, despite an evolutionary history lasting several thousands of gen- No. of erations (Bennett and Smith 1991; O’kane et al. 1996). Collection individuals In contrast studies in some recently formed allopoly- Species no. Locality analyzed ovarik ploids (K et al. 2005), synthetic interspecific T. dubius Population 2613 Pullman, WA 3 in luster hybrids and polyploids (L et al. 1985; C et al. Population 2614 Rosalia, WA 3 1996; Weiss and Maluszynska, 2000; Skalicka´ et al. Population 2615 Spokane, WA 3 2003; Pontes et al. 2004; Lim et al. 2006) reveal evidence T. porrifolius Population 2607 Troy, ID 3 of astonishingly fast genetic change influencing thou- Population 2611 Pullman, WA 3 sands of rDNA units, perhaps in just a few generations Population 2612 Troy, ID 3 T. pratensis Population 2598 Colton, WA 3 following allopolyploidy. The reasons for interspecies Population 2608 Moscow, ID 3 differences remain enigmatic, but evolutionary change Population 2609 Spangle, WA 3 probably occurs through a range of processes including T. mirus Population 2601 Pullman, WA 17 locus loss/gain, amplification/reduction in repeat copy Population 2602 Palouse, WA 29 number, gene conversion, and recombination. Population 2603 Rosalia, WA 10 In nature there are several examples of allopolyploids Population 2690 Tekoa, WA 6 of recent (,150 years) origin, including in the genera Population 2673 Oaksdalle, WA 6 T. miscellus Population 2604 Moscow, ID 10 Senecio (Abbott and Lowe 2004), Spartina (Ainouche oltis Population 2605 Pullman, WA 7 et al. 2004), and Tragopogon (S et al. 2004), en- Population 2606 Spangle, WA 14 abling direct comparison of expression patterns be- tween parents and polyploids. However, in none of these systems has the question of parental rRNA genes expres- these questions we have analyzed rDNA copy numbers, sion been addressed so far. Tragopogon mirus Ownbey rRNA gene expression, and rDNA chromatin conden- and T. miscellus Ownbey are allotetraploids (2n ¼ 24) sation at interphase and metaphase in several popula- that formed during the past 80 years following the tions each of T. mirus and T. miscellus. introduction of three diploids ½T. dubius, T. pratensis, T. porrifolius (2n ¼ 12) from Europe to western North America (Ownbey 1950; Soltis et al. 2004). T. mirus and MATERIALS AND METHODS T. miscellus originated by hybridization between T. dubius Plant material: Seeds of the three diploid and two allote- and T. porrifolius, and T. dubius and T. pratensis, respec- traploid species of Tragopogon were collected from natural tively. Both allotetraploids have formed repeatedly— populations in Idaho and Washington (Table 1). For each there may be as many as 21 lineages of separate origin of population sampled, we collected seeds from 10 to 20 in- dividual plants; the seeds from each plant were kept separate. T. miscellus and 11 of T. mirus in the Palouse region of These seeds were subsequently planted in the greenhouses at Washington and Idaho (Soltis and Soltis 1989; Cook the University of Florida; plants were grown to maturity and et al. 1998). We previously reported little or no deviation allowed to self-pollinate. The resultant selfed seeds (S1) were from additivity for chromosome numbers, satellite se- collected from individual plants (Kovarik et al. 2005). Some of quences, and genome size (Pires et al. 2004). However, a these seeds, representing one generation of selfing (S1) in the greenhouse, were later germinated in a greenhouse in the quantitative study of the inheritance of rRNA genes Institute of Biophysics, Brno for use in this investigation. among several populations of T. mirus and T. miscellus Collected leaf samples were dipped into RNAlater solution revealed up to 70–95% reduction in the number of (Ambion, Austin, TX), stored at À20°, and analyzed within 12 T. dubius repeats, likely by deletion or recombination months after harvest. Root tips for cytogenetic studies and leaf events. The genetic lability of the T. dubius genome tissue for DNA isolation were harvested from 5- to 8-week-old plants. Plants were sampled in duplicate. In some cases, plants component in the allopolyploids was recently demon- were sampled during both the first and second year of the life strated by Tate et al. (2006), who found stochastic gene span of these biennials. All individuals were genotyped using loss at several low-copy loci in T. miscellus individuals PCR-CAPS analysis of the PP2C gene (Tate et al. 2006) and from two populations. their diploid parentage