Genome polymorphisms and gene differential expression in a ‘back-and-forth’ ploidy-altered series of weeping lovegrass (Eragrostis curvula)

ARTICLE IN PRESS Journal of Plant Physiology 164 (2007) 1051—1061

www.elsevier.de/jplph

Genome polymorphisms and gene differential expression in a ‘back-and-forth’ ploidy-altered series of weeping lovegrass (Eragrostis curvula) Martı´n A. Mecchiaa, Ana Ochogavı´ab, Juan Pablo Selvab, Natalia Laspinaa, ´n Spangenbergc, Silvina Felittic, Luciano G. Martelottoa, Germa Viviana Echeniqueb, Silvina C. Pessinoa, a

Laboratorio Central de Investigaciones, Facultad de Ciencias Agrarias de la Universidad Nacional de Rosario, Parque Villarino, S2125ZAA Zavalla, Santa Fe, Argentina b Departamento de Agronomı´a, Universidad Nacional del Sur, CERZOS CONICET, San Andre´s 800, Bahı´a Blanca, Argentina c Primary Industries Research Victoria, Victorian AgriBiosciences Centre, 1 Park Drive, Bundoora, VIC 3083, Australia Received 18 May 2006; accepted 5 July 2006

KEYWORDS Apomixis; Gene expression; Genome evolution; Non-mendelian inheritance; Polyploidy

Summary Molecular markers were used to analyze the genomic structure of an euploid series of Eragrostis curvula, obtained after a tetraploid dihaploidization procedure followed by chromosome re-doubling with colchicine. Considerable levels of genome polymorphisms were detected between lines. Curiously, a significant number of molecular markers showed a revertant behavior following the successive changes of ploidy, suggesting that genome alterations were specific and conferred genetic structures characteristic of a given ploidy level. Genuine reversion was confirmed by sequencing. Cluster analysis demonstrated grouping of tetraploids while the diploid was more distantly related with respect to the rest of the plants. Polymorphic revertant sequences involved mostly non-coding regions, although some of them displayed sequence homology to known genes. A revertant sequence corresponding to a P-type adenosine triphosphatase was found to be differentially represented in cDNA libraries obtained from the diploid and a colchiploid, but was not found expressed in the original tetraploid. Transcriptome profiling of inflorescence followed by real-time polymerase chain reaction validation showed 0.34% polymorphic bands between apomictic tetraploid and sexual diploid plants. Several of the polymorphic sequences corresponded to known genes. Possible correlation

Abbreviations: AFLP: amplified fragment-length polymorphisms; ATPase: adenosine triphosphatase; cDNA: complementary DNA; CoA: coenzyme A; dNTP: deoxynucleotide triphosphate; EST: expressed sequence tag; PCR: polymerase chain reaction; RAPD: random amplified polymorphic DNA Corresponding author. Tel.: +54 341 4970085x154; fax: +54 341 4970085x115. E-mail address: [email protected] (S.C. Pessino). 0176-1617/$ - see front matter & 2006 Elsevier GmbH. All rights reserved. doi:10.1016/j.jplph.2006.07.002

ARTICLE IN PRESS 1052

M.A. Mecchia et al. between the results observed here and a recently reported genome-wide nonMendelian inheritance mechanism in Arabidopsis thaliana are discussed. & 2006 Elsevier GmbH. All rights reserved.

Introduction Polyploidization is considered to be a major factor in the evolutionary history of angiosperms and one of the most important contributor mechanisms to speciation and adaptation in the plant kingdom (Wendel, 2000). Either autopolyploidy or allopolyploidy has influenced the genome structures of many important cultivated plant species such as alfalfa, potato, wheat, oat, cotton, coffee and canola. Furthermore, other major crops like maize, soybean and cabbage and even the model plant Arabidopsis thaliana have been recognized as ancestral polyploids (paleopolyploids), although the evidence for this has been often obscured by genomic rearrangements (Osborn et al., 2003). Polyploidization may involve major genetic and epigenetic alterations of progenitor genomes, which can be established either rapidly or as a long-term response (revised in Ma and Gustafson, 2005). This more or less extensive genome modification apparently involves both coding and noncoding regions and is immediately followed by alterations of gene expression (Osborn et al., 2003). The altered patterns of gene expression are probably significant for adaptation, as whole gene complement re-duplication requires modification of regulatory networks in order to maintain homeostasis (Kellogg, 2003). As polyploidization is both an ancestral and a contemporary process, it is relevant to consider the molecular evolutionary events that characterize the early stages of polyploid formation as distinct from those that are responsible for long-term genomic changes (Wendel, 2000). Newly synthesized polyploids provide useful systems to reveal genetic alterations occurring immediately following a modification at the ploidy level. Rapid genetic changes were observed in autopolyploid Paspalum (Martelotto et al., 2005a) and allopolyploid Brassica (Song et al., 1995), Aegilops triticum (revised in Ma and Gustafson, 2005) and triticale (Ma et al., 2004). However, species like cotton (Liu et al., 2001), Brassica juncea (Axelsson et al., 2000) and Spartina anglica (Baumel et al., 2002) showed almost complete additivity of gene complements in full genome surveys of parental loci. Moreover, phylogenetic analysis suggests that the cotton genome has been quiescent and conserved over

1–2 million years (Liu and Wendel, 2002). The observation of genomic stasis in at least some species suggests that the response to allopolyploidization is highly variable and can result in divergent genome behavior under different environments. Major mechanisms leading to novel forms of gene expression in polyploids generally relate to an increased variation in dosage-regulated gene expression, alteration of regulatory networks and/or rapid genetic and epigenetic changes. Gene expression alterations were detected in a recently formed autopolyploid of Paspalum notatum (Martelotto et al., 2005b). Similarly, autotetraploid Arabidopsis thaliana lines generated by colchicine doubling also showed differential expression of several genes when compared to the original diploid (Wang et al., 2004). Gene expression repatterning associated with genetic and epigenetic modifications was reported for allopolyploid wheat (Kashkush et al., 2002, 2003), Arabidopsis (Comai et al., 2000; Wang et al., 2004) and cotton (Adams et al., 2003). The quiescent genome behavior observed in cotton, compared to the detection of major gene expression re-patterning suggests that even minimal changes in sequence structure and/or methylation patterns can lead to dramatic changes in gene expression. Eragrostis curvula (Schrad.) Nees is a polymorphic grass native to Southern Africa. Several of its forms, known as lovegrasses, were introduced to Australia, USA, and Argentina as soil conservation cultivars or forage grasses (Voigt et al., 2004). Most Eragrostis curvula cultivars are autopolyploid and reproduce by obligate apomixis (asexual reproduction by seeds) of the pseudogamous diplosporous type (Voigt et al., 2004). Diploid forms are very infrequent and do not occur in all forms of the species (Voigt et al., 2004). In vitro culture assays aimed at the development of related sexual and apomictic materials of different ploidy levels to be used for breeding and research purposes were recently conducted (Cardone et al., in press). As a consequence of these experiments, an euploid series of the species consisting of an apomictic tetraploid natural genotype (Tanganyika) (T, 2n ¼ 4x ¼ 40), a sexual dihaploid derivative obtained from T by in vitro culture (D, 2n ¼ 2x ¼ 20) and two tetraploid sexual plants originated from D

ARTICLE IN PRESS Genome polymorphisms and gene differential expression by colchicine treatment (C and M, 2n ¼ 4x ¼ 40) was produced (Cardone et al., in press). We report here the results of a general comparison of the genetic structure and gene expression levels in this series. The particular ‘back-and-forth’ ploidy-altered system used here provides a unique perspective to appreciate the molecular changes related to ploidy level modifications. These results may contribute to a general understanding of the mechanisms involved in rapid genome variation and gene expression re-patterning associated with ploidy changes.

Materials and methods Plant material Plants used in this study were obtained as previously reported (Cardone et al., in press). Briefly, emerging inflorescences from an Eragrostis curvula (Schrad.) Nees (weeping lovegrass) Tanganyika genotype (T, 2n ¼ 4x ¼ 40) were cultured on Murashige and Skoog’s (1962) medium supplemented with 2,4-dichlorophenoxyacetic acid and 6benzylaminopurine. Cultures were kept in the dark at 25 1C for 4 weeks and were then transferred to a growth chamber with a photoperiod of 16 h (66 mmol m2 s1). After 4 months, regenerated plants ðR0 Þ were transferred to pots containing soil and were grown in a glasshouse. Three months later, plants were transplanted to a field. One of a total of 23 R0 plants presented a distinctive phenotype. Chromosome counting and flow cytometry studies showed that it was euploid and had 20 chromosomes in somatic cells. This plant ð2n ¼ 2x ¼ 20Þ was named D (UNST1122 in Cardone et al., in press). Twenty diploid clones obtained from D by dividing/rooting of tillers were allowed to crosspollinate in the glasshouse in order to produce an R1 generation. No other diploid was present in the glasshouse, except for the D-derived clones. The diploid chromosome number in the R1 generation was confirmed by chromosome counting and flow cytometry. In order to double the chromosome number, 500 seeds of one diploid R1 plant were treated with a solution of 0.05% (w/v) colchicine and 2% (v/v) dimethylsulfoxide for 6, 12, 24 or 48 h in the dark at 25 1C. The treated seeds were sown in pots of soil, germinated, and grown in the glasshouse for 2 months. Chromosome numbers were assessed using a Partec II Flow Cytometry DNA Analyzer and also by counting in roots tips. Two tetraploid plants were identified. These plants ð2n ¼ 4x ¼ 40Þ were named M and C (formerly

1053

UNST1112 and UNST1131, respectively) (Cardone et al., in press). Cytoembryological analysis and progeny tests showed that all D, M and C plants reproduced sexually, although residual apomixis could not be completely excluded (Cardone et al., in press). Comparisons of 15 propagules of D and 15 plants of Tanganyika indicated the presence of highly significant differences in leaf width and length and plant height between the dihaploid and its progenitor (Cardone et al., in press). Other unusual characteristics of D were a typical pubescence on the base of the adaxial surface of the leaves, a clear tendency to produce aerial tillers, and a very small ligule compared with the original cultivar (Cardone et al., in press). Panicles exhibited variation in size and appearance, with secondary panicle branches held rigidly away from the primary branches in the dihaploid. Anthers of D were pale yellow. Seeds were ellipsoidal with a light brown coat, and were smaller and lighter than those produced by Tanganyika. In the glasshouse D produced flowers throughout the year, except in the Southern-hemisphere cold winter months of June and July (Cardone et al., in press). The two colchiploid plants (M and C) were morphologically very similar to and produced flowers in the same period as the original cultivar Tanganyika (from October to early March) (Cardone et al., in press; Fig. 1).

RAPD analysis Genomic DNA was extracted by using the protocol described at the CIMMYT Applied Molecular Genetics Laboratory Protocols, CIMMYT, Mexico (www.cimmyt.org). Sample quality was checked by measuring the Abs. 260 nm/Abs. 280 nm index and electrophoresis of agarose 1% (w/v) gels, to confirm DNA integrity and absence of RNA contamination. Random amplified polymorphic DNA (RAPD) experiments (Williams et al., 1990) were performed on duplicated samples by using 21 primers from the British Columbia University RAPD Primer Synthesis Project (set 3). Amplifications were prepared following the technique described in the CIMMYT Applied Molecular Genetics Laboratory Protocols, CIMMYT, Mexico (www.cimmyt.org). Briefly, each amplification reaction was performed in a volume of 25 mL containing 25 ng primer, 20 ng genomic DNA, 1X Platinum Taq polymerase reaction buffer (Invitrogen), 10 mM each deoxynucleotide triphosphates (dNTPs), 1.5 mM magnesium chloride and 1 U of Platinum Taq polymerase (Invitrogen). Amplifications were carried out in a MJ Research

ARTICLE IN PRESS 1054

M.A. Mecchia et al.

A T

D

C

M D

Tanganyika 2n=4x=40 T

In-vitro culture

Dihaploid 2n=2x=20

B

D T

D

C

M Vegetative propagation

C

Colchicine T

D

C

M

treatment

colchiploid

colchiploid

2n=4x=40

2n=4x=40

C

M

Figure 1. Exomorphology of the Eragrostis curvula genotypes used for genetic and expression comparisons. Panel A: Leaf blades originated from genotypes T, D, C and M (as indicated at the top). Panel B: base of the adaxial leaf surface for genotypes T, D, C and M (as indicated at the top). Panel C: seeds obtained from genotypes T, D, C and M (as indicated at the top). Panel D: schematic representation for the generation of the ‘back-and-forth’ ploidy series.

thermocycler programmed as follows: an initial denaturation step at 931 for 1 min, 45 cycles of 1 min at 921, 1 min at 361 and 1.5 min at 721 and a

final extension step of 5 min at 721. Negative controls were performed by eliminating genomic DNA from the assays. Reactions were

ARTICLE IN PRESS Genome polymorphisms and gene differential expression electrophoresed in 5% (w/v) polyacrylamide denaturing gels and silver-stained.

AFLP analysis Aliquots of 900 ng genomic DNA were simultaneously digested with EcoRI and MseI. Fragments were ligated to EcoRI and MseI adaptors and constituted the template for further amplification. Pre-amplification primers contained one selective nucleotide. Pre-amplification products were diluted (1:10) in Tris 10 mM/EDTA 0.1 mM and were used as templates for second-round amplification, where eighteen combinations of the EcoRI and MseI amplified fragment-length polymorphisms (AFLP) primers (containing three selective nucleotides) were employed. An MJ Research thermocycler programmed for the cycling profile indicated in Vos et al. (1995) was used for all amplifications. Reliability was assessed by the use of duplicates. Following amplification, the polymerase chain reaction (PCR) products were mixed at the right proportion with 3X loading dye (98% (v/v) formamide, 10 mM EDTA, 0.025% (w/v) bromophenol blue and 0.025% (w/v) xylene cyanol), denatured at 951 for 5 min and immediately placed on ice. Five microliter of the denatured samples were loaded onto denaturing 5% (w/v) polyacrylamide gels. Amplification products were visualized by silver staining and digitized for scoring analysis.

Dendrogram construction Bands were counted and data was converted into a binary matrix, in which 1 and 0 indicated band presence and absence, respectively. Failed amplification or equivocal results were coded as missing data. Matrixes were analyzed for the determination of similarity coefficients between pairs of individuals and group clustering. The Jaccard’s coefficient (J) was used: J ¼ a=ða þ b þ cÞ, in which a is the number of bands common to both individuals, b the number of bands present in the first individual and absent in the second one and c the number of bands present in the second individual and absent in the first one (Jaccard, 1908). Group clustering analysis was performed through the unweighted pair group method with arithmetic mean method using the Infostat computational pack (Facultad de Ciencias Agropecuarias, National University of Cordoba, Argentina).

Cloning of fragments RAPD and AFLP bands and polymorphic features from differential display analysis were excised from

1055

gels. The gel section was carefully chopped into small pieces and eluted in buffer 0.5 M ammonium acetate/1 mM EDTA pH 8. The DNA was precipitated in ethanol, and re-amplified using the same PCR conditions as described above. All fragments were cloned with the aid of the pGEMs-T Easy Vector System (Promega).

Sequencing and BLAST analysis Sequencing of the RAPD/AFLP fragments and differential display complementary DNA (cDNA) clones was performed by Macrogen Inc. (Korea). Analysis of DNA similarity was carried out using the BLAST package 2.0 on the National Center of Biotechnology Information server (http://www. ncbi.nlm.nih.gov/BLAST/).

Library construction and EST sequencing Four cDNA libraries were constructed from sexual dihaploid flowers (EC01), apomictic tetraploid flowers (EC02), apomictic tetraploid leaves (EC03) and sexual tetraploid flowers (EC04), respectively. Total RNA was extracted from flowers or leaves using the RNAeasys total RNA isolation kit (Promega). The cDNAs were obtained using a SMARTs PCR synthesis kit, cloned into the pGEMs-T easy vector and used for transformation of XL10-Golds Ultracompetent Escherichia coli cells. Two thousand and four hundred randomly selected cDNA clones from each library were sequenced using a MegaBACETM DNA sequence analyzer. Raw expressed sequence tag (EST) sequence traces were processed and annotated after the analysis using the XGI pipeline (http://www.ncgr.org/xgi) for automated EST clustering. Processed ESTs were clustered into consensus sequences that were compared using BLASTX. Functional categories were assigned by means of gene ontology annotation.

Differential display analysis and validation Total RNA was obtained from flowers by using the SV Total RNA Isolation kit from Promega. Reverse transcription was performed using Superscript II reverse transcriptase (Gibco-Life Technologies) as indicated by the manufacturers. Differential display experiments were conducted under the general protocol reported by Liang and Pardee (1992) with minor modifications. The anchored oligonucleotide used corresponded to the sequence 50 T(12)(ACg)A30 . Twenty decamers from the British Columbia University RAPD Primer Synthesis Project (set 3) were used in combination with the anchored

ARTICLE IN PRESS 1056

M.A. Mecchia et al.

oligonucleotide to create primer pairs. PCR reactions were prepared in duplicate in a final volume of 25 mL containing 1  Taq activity buffer (Promega), 1.5 mM MgCl2, 50 mM dNTPs, 0.70 mM arbitrary primer, 2.5 mM anchored primer, 2 U of Taq DNA polymerase enzyme (Promega) and 2.5 mL of the reverse transcription reaction (previously diluted 1/20). Negative control reactions were performed using total non reverse-transcribed RNA, to verify the absence of chromosomal DNA in the RNA preparations. The cycle program consisted of an initial step of 3 min at 94 1C, 40 cycles of 20 s at 94 1C, 20 s at 38 1C and 30 s at 72 1C, followed by a final step of 5 min at 72 1C. Samples were mixed with denaturing loading buffer, treated for 3 min at 95 1C and separated in a 5% (w/v) polyacrylamide gel. Amplification products were silver-stained. Bands were scored only in the middle portion of the gel, where resolution was maximal and profiles were fully reproducible. Differential fragments were excised, eluted in a buffer of 0.5 M ammonium acetate and 1 mM EDTA pH 8, precipitated in ethanol, and re-amplified using the same PCR conditions described above. All fragments were cloned with the aid of the pGEMsT Easy Vector System (Promega). Sequencing of the differential display cDNA clones was performed by Macrogen Inc. (Korea). Analysis of DNA similarity was carried out using the BLAST package 2.0 on the National Center of Biotechnology Information server (http://www.ncbi.nlm.nih.gov/ BLAST/). In some cases, specific primers were designed for target fragments and used in realtime PCR amplification to validate differential expression.

Table 1.

Results RAPD analysis RAPD analysis was initially performed to compare the genetic structure of the T, D and C genotypes. Two hundred and eighteen molecular markers were obtained using 21 random decamers. There was in theory seven possible patterns of band incidence distribution across the three genotypes: 1/1/1, 1/ 1/0, 0/1/1, 1/0/1, 1/0/0, 0/1/0 and 0/0/1. Each was observed in different proportions (Table 1). Surprisingly, most of the polymorphisms observed corresponded to the revertant patterns 1/0/1 and 0/1/0, which together involved a 22.47% of the total counted bands. Other polymorphic patterns accounted for a 6.43% of the total bands, while monomorphic bands represented a 71.10% (Fig. 2).

AFLP analysis The genetic structure of the T, D and C lines was further analyzed with AFLP markers. Amplification patterns obtained with 18 primer combinations were compared. As a whole, 790 molecular markers were scored, of which 254 (32.15%) were polymorphic (Table 1). AFLP results were congruent with those from RAPD amplification, although the percentage of polymorphisms detected was rather larger. Again, the majority of polymorphisms corresponded to the revertant patterns 1/0/1 and 0/1/0, which together involved a 30.25% of the total counted bands. Other polymorphic patterns

Type, number and percentage of bands counted in the RAPD and AFLP gels

Pattern

Score RAPD

Typea

T (4x)a

D (2x)a

C (4x)a

No. of bands

1/1/1 1/0/0 0/1/1 1/1/0 0/0/1 1/0/1 0/1/0

1 1 0 1 0 1 0

1 0 1 1 0 0 1

1 0 1 0 1 1 0

155 3 2 6 3 34 15

Total a

218

0 indicates lack of a given band; 1 indicates presence of a given band.

AFLP P (%) 71.10 1.38 0.92 2.75 1.38 15.59 6.88 100.0

No. of bands 536 4 10 1 0 173 66 790

P (%) 67.85 0.51 1.26 0.13 0.00 21.90 8.35 100.0

ARTICLE IN PRESS Genome polymorphisms and gene differential expression

haploid genome. However, in the AFLP experiments P values differed significantly, which might reflect incomplete detection of epigenetic modification (see discussion).

A Monomorphic 71.10 %

Polymorphic (others) 6.43 %

Confirmation of polymorphism reversion in a second colchiploid

Polymorphic (revertant) 22.47 %

B Monomorphic 67.85 %

Polymorphic (revertant) 30.25 %

Polymorphic (others) 1.90 %

Figure 2. Percentage of monomorphic and polymorphic bands observed during the genetic profiling. Panel A: results obtained from the RAPD experiments: Panel B: results obtained from the AFLP experiments.

Table 2. Percentage of polymorphic markers characQ teristics of a given ploidy level per haploid genome ( ) Technique

RAPD AFLP

1057

P 4x

2x

3.90 5.47

3.44 4.17

accounted for a 1.90% of the total bands, while monomorphic bands represented a 67.85% (Fig. 2).

Ploidy-related polymorphisms proportional to haploid genome content A parameter termed ‘percentage of polymorphic markers characteristic of a given ploidy level per Q haploid genome’ ( ), was calculated as the ratio of the percentage of polymorphic bands present only in a given ploidy level to the number of genome complements in the nucleus (two for diploids, four Q for tetraploids). The values for tetraploid and diploid genotypes obtained with RAPDs were very similar (Table 2), such that approximately the same number of bands that disappeared in T appeared in D during the haploidization process, as a proportion of haploid genome content. Moreover, the same number of RAPD bands reappeared in C during tetraploid formation as disappeared from D, per

A total of 695 (33 RAPD+662 AFLP) markers were used to profile a second colchiploid obtained from an independent autopolyploidization event from a different seed of the same R1 plant (see Materials and methods). These 695 markers were included within the set of 1008 markers (218 RAPD+790 AFLP) previously used to differentiate T, D and C (Table 3). The genetic structure of genotype M was almost identical to that of genotype C. The only polymorphisms detected between C and M consisted of two bands of the 1/0/0/1 and the 0/0/0/1 type, respectively. These observations indicated that the reversion pattern of bands previously detected for T, D and C was consistently maintained in M, so the pattern of reversion was almost identical for both colchiploids. Therefore, reversion appeared to represent a non-random, specific process generating a structure typical of the tetraploid ploidy level. RAPD and AFLP results were summarized in Table 3. Sectors of RAPD and AFLP gels showing band reversion were shown in Fig. 3.

Clustering analysis Cluster analysis revealed that all tetraploid individuals appeared included in a single group (Fig. 4). The only individual excluded from clustering was the diploid. The T, C and M genotypes grouped together at a genetic distance of 0.03, while D appeared apart, at a distance of 0.3 with respect to the tetraploids. The coefficient of cophenetic correlation for the dendrogram was 1.000.

Cloning and sequencing of polymorphic fragments Twenty-seven polymorphic bands amplified with the same AFLP primer pair (E33M40) were cloned and sequenced in order to discriminate against comigrating fragments of similar molecular weight and address the presence of allelic forms in the tetraploid and dihaploid plants. These 27 bands included eight different groups of revertant bands that appeared in T, C and M but were absent in D (eight bands/plant  three plants ¼ 24 bands) and three bands that were present only in D, but were

ARTICLE IN PRESS 1058

M.A. Mecchia et al.

Table 3. Type, number and percentage of bands counted in the molecular marker profiling of T, D, C and a second colchiploid (M), obtained in an independent autotetraploidization event Pattern

Score

Typea

T (4x)a

D (2x)a

C (4x)a

M (4x)a

No. of bands

1/1/1/1 0/1/1/1 1/1/0/0 0/0/1/1 1/0/1/1 0/1/0/0 1/0/0/1 0/0/0/1 1/0/0/0

1 0 1 0 1 0 1 0 1

1 1 1 0 0 1 0 0 0

1 1 0 1 1 0 0 0 0

1 1 0 1 1 0 1 1 0

471 11 1 2 152 53 1 1 3

67.77 1.58 0.14 0.29 21.87 7.63 0.14 0.14 0.44

695

100.00

Total a

P (%)

0 indicates lack of a given band; 1 indicates presence of a given band.

A T

B T

D

D

C

C

M M

T

D D

C

M

T

D

C

M

M

C

T

0.4

0.3

0.2

0.1

0.0

Figure 4. Dendrogram showing the genetic distance between the Eragrostis curvula genotypes. The T, C and M genotypes group together, while D appears apart with respect to the tetraploids.

Figure 3. Band reversion in the back-and-forth ploidy series. Panel A: RAPD gel showing amplicons obtained from genotypes T, D, C and M (as indicated at the top). Amplification reactions were conducted by duplicate. Panel B: AFLP gel showing amplicons obtained from genotypes T, D, C and M (as indicated at the top). Arrows point to revertant bands of the 1/0/1/1 or 0/1/0/0 type.

absent in T, C and M (three bands/plant  one plant ¼ 3). From the eight groups originated from the tetraploids, revertant bands were confirmed to correspond to the same sequence in five cases (groups 1, 3, 6, 7 and 8). In one group we detected identical sequences for two of the bands (originated from T and C) but efforts to sequence the third type failed. The two remaining groups showed crossed contamination between them (group 4 and 5). Three of the groups originated in the tetraploids (groups 6, 7 and 8) were composed by homologous

sequences, revealing the possible presence of allelic or multigenic isoforms that were simultaneously modified with the ploidy change. None of the bands originated from the dihaploid were homologous (allelic) to bands originated from the tetraploids, initially suggesting that alleles of the 1/0/1/1 type polymorphic bands were not present in the dihaploid genome and alleles of the 0/1/0/0 type polymorphic bands were not present in the tetraploid genome. Furthermore, preliminary DNA blot hybridization analysis revealed that at least some of the clones originated from tetraploids hybridized showing additional bands in the tetraploids with respect to the diploids. On the contrary, clones obtained from the diploid presented extra bands in the diploid with respect to the tetraploids (data not shown). BLAST searches revealed that although the majority (10) of the unique sequences (14) did

ARTICLE IN PRESS Genome polymorphisms and gene differential expression UBC216 T

T

D

UBC241 D

T

T

D

D

1059

that at least some polymorphic revertant sequences may be effectively expressed in plant tissues.

Differential display

Figure 5. Differential display experiments showing polymorphic bands between T and D. Sectors of gels displaying amplicons obtained from two primer pair combinations. Decamers used for amplification are indicated at the top. Amplification reactions were conducted by duplicate.

not show similarity to known genes, a few (4) corresponded to coding sequences, including a retrotransposon protein (Gen identity: ABA95071.1, E value: 3e34, BLASTX), a putative sucrose-phosphate synthase 1 (Gen identity: XM_481429.1, E value: 7e27, BLASTN), a methylase (Gen identity: AAP21237.1, E value: 8e9, BLASTX) and an haloacid dehalogenase-type hydrolase/P type adenosine triphosphatase (ATPase, Gen identity: CAE05846.2, E value: 6e17, BLASTX). A total of 31 sequences were used as queries, including the 27 sequences mentioned in the former paragraph and other four additional revertant bands (derived from RAPDs) of the 1/0/1/1 or the 0/1/0/0 type, that had been obtained from T or D, respectively.

Differential display banding patterns from floral samples of T and D were compared in duplicated tests (Fig. 5). From of a total of 20 oligonucleotide combinations originally assayed, nine were selected for further analysis based on generation of coherent profiles. Candidate transcripts were distinguished by a clear differential signal between T and D samples, compared to equivalent signals from others. Bands showing a clear presenceabsence pattern were isolated and derived fragments were cloned. Thirteen polymorphic bands (0.65% of the analyzed bands) were detected, but only eight were successfully obtained for sequencing, ranging in size from 221 to 1127 bp. BLAST searches revealed homologies for some of the fragments with retrotransposons (Gene identity: XP_463537.1, E value: 2e15 , BLASTX) , a yippee family protein like protein (Gene identity: NM_1118913.2, E value: 3e12, BLASTX), an isovaleryl coenzyme A (CoA) dehydrogenase (Gene identity: AK119504.1, E value: 3e55, BLASTN) and a succinate dehydrogenase (Gene identity: AY104384, 1e56, BLASTN). Sequences corresponding to an isovaleryl CoA dehydrogenase were confirmed to be differentially represented in the ECO1 (D) and ECO4 (T) libraries. We were unable to detect the rest of the sequences in the ESTs libraries, presumably because the small number of clones sequenced did not allow the revealing of low-abundance transcripts.

Discussion Expression of polymorphic sequences Coding sequences detected as polymorphic in the genomic tests were compared against the EST database obtained from cDNA libraries produced from T, D and M. Each one of the libraries was represented in the database by 2400 clones. No sequence corresponding to the clone enconding a putative methylase was found in any of the libraries. However, clones homologous to a different region of the same P type ATPase gene as that identified from the RAPD experiments (Gene identity: XP_482103.1, E value: 4e30, BLASTX) were found to be represented at least once in ECO1 (D) and three times in ECO4 (M). This demonstrated

Results reported here demonstrated the occurrence of rapid genetic changes in Eragrostis curvula associated with successive modifications of the ploidy level. The generation of polymorphisms was large-scale, involving 31.45% of the genetic loci randomly distributed across the genome. As half of the tetraploid genome was lost during dihaploidization, a relatively high rate of genetic polymorphism was already anticipated, but it should be directly related to the heterozygosity level. Surprisingly, the majority of the total modifications (90.85%) were fully reproducible and reversible. This implied that the progenitor allelic forms were lost during dihaploidization and then recovered when the original ploidy level was restored. The low proportion

ARTICLE IN PRESS 1060 of bands that were irreversibly lost during dihaploidization suggested that low levels of heterozygosity were present in the autotetraploid Eragrostis curvula genome. In the presence of these unexpected observations, every possible consideration or measure were taken to minimize the risk of genetic contamination: diploid clones originated from D by vegetative propagation were the only material with this ploidy level kept at the greenhouse; the ploidy levels of plants and progenies were carefully checked by flow cytometry and chromosome counting; pollen contamination of materials was systematically avoided; and analysis of a second independently obtained colchiploid permitted elimination of the effects of stochastic changes and contamination. As the changes were detected by both RAPD and AFLP analysis, they must necessarily be genetic in nature. However, in the case of AFLPs, the partial sensitivity of EcoRI to methylation of its recognition site may allow codetection of accompanying epigenetic changes. This effect might explain the slight differences in the rate of polymorphisms detected through RAPD and AFLP experiments and in the P parameter values (see Tables 1 and 2). Although the effect of reversion was not easily explained, it was obvious that: (1) the progenitor genetic structure could be recovered when the original ploidy was restored, (2) therefore, the genetic structures of the lines were typical of a given ploidy level and (3) there must be a mechanism of retention of sequence information eliminated from the genome, to permit re-establishment. However, this is not the first report of reversible genomic changes occurring in synthetic poyploids. Molecular marker variants absent from several generations of a population of synthetic Brassica polyploid plants reappeared subsequently in the population (Song et al., 1995). Moreover, an analogous mechanism of genome-wide non-Mendelian gene inheritance was recently detected for recessive mutant alleles of the organ fusion gene hothead (hth) in Arabidopsis (Lolle et al., 2005). It was suggested that the genetic restoration events were the result of a template-directed process that makes use of an ancestral RNA-sequence cache (Lolle et al., 2005). Genome-cached information could be reinstated with the genome by an uncharacterized mechanism, possibly activated by stress. The demonstration of transcription from some of the polymorphic regions (P-type ATPase) in cDNA libraries of the diploid and colchiploid genotypes of Eragrostis curvula might support this mechanism. However, several of the revertant polymorphic sequences were either non-coding or

M.A. Mecchia et al. were not detected in the libraries (probably due to the limited depth of cDNA sampling to date). Therefore, an exhaustive analysis including fully representative libraries and full-length sequencing of alleles detected in each library would be necessary prior to any further speculation. The detection of genomic polymorphisms during autotetraploidization in Eragrostis curvula was particularly high (29.77%). Rates found in synthetic autotetraploid Paspalum notatum were rather lower (9.20%), but still considerable (Martelotto et al., 2005a). The rate of reversion reported here was large in comparison with that reported in Brassica (Song et al., 1995). It would be relevant to consider if apomictic species like Eragrostis curvula and Paspalum notatum might be more sensitive to ploidy changes than those reproducing exclusively by sexuality. Gametophytic apomixis is always associated with increased ploidy levels, and apomictic species often organize as agamic complexes in which diploid individuals are sexual and polyploid individuals are apomictic. An increased sensitivity of the genome to ploidy changes in apomictic species compared to sexual ones could represent an evolutionary advantage and should be systematically analyzed in the future.

Acknowledgements We are in debt with Dr. Pablo Rabinowicz and Dr. Luca Comai for valuable discussion. Thanks are also due to Dr. Juan Pablo A. Ortiz for critically reading the manuscript, and to Dr. Francisco Espinoza and Lic. Cesar Olsina for allowing access to methods and equipment. This work was funded by ANPCyT (Agencia Nacional de Promocio ´n Cientı´fica y Tecnolo ´gica, Argentina, BID 1201-OC-AR PICT03 14624, PAV 137), Convenio Argentino-Brasilen ˜o de Biotecnologı´a (CABBIO 2004 Project 012), CONICET (PIP 6509) and the National University of Rosario (SPU 026). Luciano G. Martelotto and Juan Pablo Selva doctoral fellows of CONICET (Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas). Natalia Laspina is a doctoral fellow of ANPCyT (Agencia Nacional de Promocio ´n Cientı´fica y Tecnolo ´gica). Silvina Felitti, Viviana Echenique and Silvina C. Pessino are career members of CONICET (Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas).

References Adams KL, Cronn RC, Percifield R, Wendel JF. Genes duplicated by polyploidy show unequal contributions

ARTICLE IN PRESS Genome polymorphisms and gene differential expression to the transcriptome and organ-specific reciprocal silencing. Proc Natl Acad Sci USA 2003;100:4649–54. Axelsson T, Bowman CM, Sharpe AG, Lydiate DJ, Lagercrantz U. Amphidiploid Brassica juncea contains conserved progenitor genomes. Genome 2000;43: 679–88. Baumel A, Ainouche M, Kalendar R, Schulman AH. Retrotransposons and genomic stability in populations of the young allopolyploid species Spartina anglica CE Hubbard (Poaceae). Mol Biol Evol 2002;19:1218–27. Cardone S, Polci P, Selva JP, Mecchia M, Pessino SC, Herrmann P, et al. Novel genotypes of the subtropical grass Eragrostis curvula for the analysis of apomixis (diplospory). Euphytica [in press]. Comai L, Tyagi AP, Winter K, Holmes-Davis R, Reynolds SH, Stevens Y, et al. Phenotypic instability and rapid gene silencing in newly formed Arabidopsis allotetraploids. Plant Cell 2000;12:1551–67. Jaccard P. Nouvelles recherches sur la distribution florale. Bull Soc Vaud Sci Nat 1908;44:223–70. Kashkush K, Feldman M, Levy AA. Gene loss, silencing and activation in a newly synthesized wheat allotetraploid. Genetics 2002;160:1651–9. Kashkush K, Feldman M, Levy AA. Transcriptional activation of retrotransposons alters the expression of adjacent genes in wheat. Nat Genet 2003; 33:102–6. Kellogg EA. What happens to genes in duplicated genomes. Proc Acad Natl Sci USA 2003;100:4369–71. Liang P, Pardee AB. Differential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science 1992;257:967–71. Liu B, Wendel JF. Non-Mendelian phenomena in allopolyploid genome evolution. Curr Genom 2002;3(6): 1–17. Liu B, Brubaker CL, Mergeai G, Cronn RC, Wendel JF. Polyploid formation in cotton is not accompanied by rapid genomic changes. Genome 2001;44:321–30. Lolle SJ, Victor JL, Young JM, Pruitt R. Genome-wide nonMendelian inheritance of extra-genomic information in Arabidopsis. Nature 2005;434:505–9. Ma X-F, Gustafson JP. Genome evolution of allopolyploids: a process of cytological and genetic

1061

diploidization. Cytogenet Genome Res 2005;109: 236–49. Ma X-F, Fang P, Gustafson JP. Polyploidization-induced genome variation in triticale. Genome 2004;47: 839–48. Martelotto LG, Ortiz JPA, Espinoza F, Quarin CL, Pessino SC. A comprehensive analysis of gene expression and genomic alterations in a newly formed autotetraploid of Paspalum notatum. In: Proceedings of the fourth international symposium on the molecular breeding of forage and turf, Aberystwyth, Wales, UK, 2005a. p. 178. Martelotto LG, Ortiz JPA, Espinoza F, Quarin CL, Pessino SC. A comprehensive analysis of gene expression alterations in a newly synthesized Paspalum notatum autotetraploid. Plant Sci 2005;169:211–20. Murashige T, Skoog F. A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol Plant 1962;15:473–97. Osborn TC, Pires JC, Birchler JA, Auger DL, Chen ZJ, Lee H-S, et al. Understanding mechanisms of novel gene expression in polyploids. Trends Genet 2003;19:141–7. Song K, Lu P, Tang K, Osborn TC. Rapid genome change in synthetic polyploids of Brassica and its implications for polyploid evolution. Proc Natl Acad Sci USA 1995;92: 7719–23. Voigt P, Rethman N, Poverene M. Lovegrasses. In: American Society of Agronomy, Crop Science Society of America, Soil Science Society of America, WarmSeason (C4) Grasses, Agronomy monograph no. 45, 2004. p. 1027–56 [Chapter 32]. Vos P, Hogers R, Bleeker M, Reijans M, Van de Lee T, Hornes M, et al. AFLP: a new concept for DNA fingerprinting. Nucl Acids Res 1995;23:4407–14. Wang J, Tian L, Madlung A, Lee HS, Chen M, Lee JJ, et al. Stochastic and epigenetic changes of gene expression in Arabidopsis polyploids. Genetics 2004;167:1961–73. Wendel JF. Genome evolution in polyploids. Plant Mol Biol 2000;42:225–49. Williams JGK, Kubelik AR, Livak KJ, Rafalski JA, Tingey SV. DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucl Acids Res 1990; 18:6531–5.