Polymorphic Nuclear Gene Sequences Indicate a Novel Genome Donor In

Hereditas 148: 8–27 (2011)

Polymorphic nuclear gene sequences indicate a novel genome donor in the polyploid genus Thinopyrum MATT ARTERBURN1 , ANDRIS KLEINHOFS2 , TIMOTHY MURRAY3 and STEPHEN JONES2 1 Department of Biology, Washburn University, Topeka, KS, USA 2 Department of Crop and Soil Sciences, Washington State University, Pullman, WA, USA 3 Department of Plant Pathology, Washington State University, Pullman, WA, USA

Arterburn, M., Kleinhofs, A., Murray, T. and Jones, S. 2010. Polymorphic nuclear gene sequences indicate a novel genome donor in the polyploid genus Thinopyrum . – Hereditas 148: 8–27. Lund, Sweden. eISSN 1601-5223. Received 16 June 2008. Accepted 15 November 2010.

For decades, the wheatgrass genus Thinopyrum has been of interest to plant breeders as a source of genes that confer competitive traits. This genus has been a considerable challenge to plant systematists because of the impacts of polyploidization on the evolution of this group. This study was aimed to augment existing cytogenetic data with a sequence-based investigation of the genomes of these species. Sequences of the internal transcribed spacer 1 (ITS1), introns 9 through 11 of the granule-bound starch synthase (GBSSI) gene and intron III of the beta-amylase gene (Bmy1 ) were isolated from the genomes of polyploid Thinopyrum species by PCR, cloning and sequencing and the evolutionary distances between these species and putative diploid ancestors were estimated with Kimura ’ s two-parameter method. Phylogenetic analysis of these sequences largely agrees with what has been established through cytogenetic means for the Th. caespitosum (Koch) Liu & Wang and Ps geniculata (Trin.) Á . Lö ve, and suggests a contribu- tion of the St genome of Ps. spicata (Pursh) Á . L ö ve to the tetraploids Th. scirpeum (Presl) Dewey and Th. junceiforme ( Á . L ö ve & D. L ö ve) Á . L ö ve. A unique Bmy1 allele, divergent from other Triticeae but shared between Th. caespitosum , Th. intermedium (Host) Barkworth & Dewey, Th. junceum (L.) Á . L ö ve and Th. ponticum Barkworth & Dewey, implies a connection between these species. Distinct oligonucleotide polymorphisms and distance calculations based on the three loci implicate Crithopsis delileana (Schult.) Roshev. and Taeniatherum caput-medusae (L.) Nevski in the evolution of the hexaploid Th. intermedium and the decaploid Th. ponticum and also suggest a potential connection of these polyploids with Elytrigia repens (L.) Desv. ex Nevski. None of these species have been previously associated with the Thinopyrum genus. Allele-speciﬁ c PCR was employed to detect the putative Crithopsis allele of ITS1 in a number of accessions. Real-time PCR indicates that two of six genomes of the hexaploid Th. intermedium have the Crithopsis -type ITS1 allele and that all ITS1 loci in the decaploid Th. ponticum are of this type. These results are supportive of the hypothesis that concerted evolution has homogenized the rDNA of Th. ponticum to the allele derived from the Crithopsis or Taeniatherum ancestor. Discovery of these novel alleles, with close homology to Ta. caput-medusae , may represent a fundamental change in the view of the evolution of Th. intermedium and Th. ponticum .

Matt Arterburn, Department of Biology, Washburn University, 1700 SW College Ave, Topeka, 66621 KS, USA. E-mail: matt. [email protected]

The wheatgrass genus Thinopyrum harbors species Members of the genus Thinopyrum have, at different which range in ploidy from diploid to decaploid and times, been included in the genera Elymus , Elytrigia , thus exhibit considerable evolutionary complexity at the Lophopyrum, Pseudoroegneria and Agropyron . Because genome level. By contrast, the anatomical features of their robust capacity for wide hybridization, classifi ca- of these species are often so similar that historically col- tion of these species has been diffi cult, although a consid- lectors of these wild grasses have sometimes grouped erable amount of effort has been invested in doing so. accessions of different ploidy levels into the same spe- Current classifi cation of members in the Thinopyrum cies. Although all these species are perennial, some genus is largely based on anatomical and karyotypic are rhizomatous while others have caespitose, bunch- homology, and the pairing behaviors of their chromo- like growth habit. Although frequently employed by somes in interspecifi c hybrids and/or amphiploids. The plant breeders in wide hybridizations, and the object polyploid members of this group are believed to have of a considerable amount of cytogenetic scrutiny, the originated from polyploidization events involving three Thinopyrum group is less well-characterized in terms of putative progenitors of the genus: Th. elongatum Dewey molecular sequence data. A growing body of sequence (2n ϭ 2x ϭ 14, Ee Ee ), Th. bessarabicum (Savul & Rayss) data should help illuminate new aspects of the evolution- Á . L ö ve. (2n ϭ 2x ϭ 14, E b E b ), and either Pseudoroegne- ary history and the dynamics of polyploidization that ria strigosa (Bieb) Á . L ö ve (2n ϭ 2x ϭ14, StSt) or the have occurred in this genus, an understanding which will closely-related Ps. spicata (Pursh) Á . L ö ve (2n ϭ 2x ϭ 14, aid crop improvement. StSt) (D EWEY 1984). The genomes of Th. elongatum and

Th. bessarabicum have previously been denoted E and J hybrids (2n ϭ 4x ϭ Ee E e E b Eb ) ( LIU and WANG 1992, respectively, although these are now frequently referred to 1993). On the basis of GISH analysis, R EFOUFI et al. (2001) as the Ee and E b genomes, based on the work of WANG suggested that Th. junceiforme is instead an autotetraploid (1992) which suggests that chromosome pairing patterns of the Ee genome. Two other species, Th. nodosum (Boiss. in interspecifi c hybrids of these two is such that a single & Heldr.) Á . L ö ve and Th. caespitosum (Koch) Liu & letter designation should be used to indicate their close Wang are thought to be allotetraploids of the Eb and St evolutionary relationship. We will use the Ee and Eb desig- genomes (2n ϭ 4x ϭ E b Eb StSt) ( LIU and WANG 1989, nation for these genomes throughout this study, for clarity. 1992). Th. scirpeum (Presl) Dewey is considered an auto- Designation of the various genomes present in Thinopy- tetraploid of the E e genome ( LIU and WANG 1993). rum polyploids has been based largely on the ability to The higher ploidy Thinopyrum species are of particular generate interspecifi c hybrids, the ability of chromosomes interest because they are frequently used in wide crosses to pair in these hybrids or in amphiploids and on the rela- to introgress agronomically useful genes into cereal crops tive intensity of signal produced via genomic in situ and for production of perennial grains for sustainable hybridization (GISH) when using genomic DNA of puta- agricultural systems (C HEN et al. 1998a; COX et al. 2002; tive ancestors as a probe ( WANG 1989; ZHANG et al 1996a; Z HANG et al 1996b; S CHEINHOST et al. 2001). Th. junceum C HEN et al. 1998b). (L.) Á . L ö ve is described as an allohexaploid combining The current evolutionary view of the genus Thinopy- the Ee and Eb genomes (L IU and WANG 1993). LIU rum defi nes the polyploid members as either autopoly- and W ANG (1993) described Th. intermedium (Host) ploids or allopolyploids of the Ee , E b and St genomes, in Barkworth & Dewey as an allohexaploid of the Ee and various combinations (Table 1). Of the eight tetraploids in St genomes (2n ϭ 6x ϭ 42, E e Ee E e E e StSt) on the basis this genus, four, Th. distichum (Thunb.) L ö ve, Th. curvi- of chromosome pairing and C-banding in interspecifi c folium (Lange) D.R. Dewey, Th. sartorii (Boiss. 7 Heldr.) hybrids. C HEN et al. (1998b), based on GISH analysis, Á . L ö ve and Th. junceiforme ( Á . L ö ve & D. L ö ve) suggested that this species is an allohexaploid of the Á . L ö ve, are considered allotetraploids of the Ee and Eb Eb genome, St genome and J s genomes (2n ϭ 6x ϭ 42, genomes based on chromosome pairing in interspecifi c Eb E b J S J S StSt). The JS designation is based on their

Table 1. Summary of ITS1, Bmy1 and GBSSI sequence analysis. Sequenced alleles are matched to the closest putative diploid ancestor. Evolutionary distance values from closest ancestor alleles are given in parentheses; multipliers indicate number of clones isolated of each allelic type. Hypothetical genome constitutions are given in the rightmost column .

Sequence Current genome based Species (accesion) designations ITS results Bmy1 results GBSSI results constitution

Th. elongatum (PI547313) E e E e E e E e E e Th. bessarabicum (PI531711) Eb Eb Eb E b Eb Ps. strigosa (PI499493) St St St St St Cr. delileana (01C4200003) K K K K K Ta. caput-medusae (PI222048) Ta Ta Ta Ta Ta Th. scirpeum (PI531749) Ee E e * E e (0.034 x5) St (0.008 x2) ND E e St Th. junceiforme (PI531731) Ee E b * Ee (0.021 x5) E e (0.000 x1) ND E e St E e E e ∗ ∗ ∗ St (0.011 x1) Ps. geniculata (PI565009 ) Ee St ∗ ∗ St (0.021 x5) E e (0.019 x2) ND E e St St (0.008 x2) Th. caespitosum (PI228276) Ee St ∗ ∗ St (0.021 x4) Eb (0.014 x3) ND Eb X or E b St X (0.272 x1) Th. junceum (PI414667) Ee E b E b ∗ Ee (0.026 x7) E e (0.060 x3) ND Ee Eb X Eb (0.030 x3) X (0.251 x2) Eb (0.008 x1) Th. intermedium (PI172688) E eEeSt ∗ K (0.043 x4) St (0.008 x4) Ta (0.016 x9) Ee StTa Eb J sSt † Eb (0.035 x4) E e (0.021 x3) Ee (0.036 x1) Ee StX St (0.071 x4) X (0.263 x3) Th. ponticum (PI547313) Ee E e E e StSt † † K (0.065 x16) E e (0.022 x6) Ta (0.011 x10) Ee E e StTaX Eb E b E b E bEb ‡ St (0.013 x3) Ee (0.030 x2) Eb E b E b J sJs † X (0.253 x1)

∗ ∗ ∗ ∗ ∗ ∗ (LIU and WANG 1993); (LIU and WANG 1992); (REFOUFI et al. 2001); †(CHEN et al. 1998); † † (ZHANG et al. 1996) ; ‡ (MURAMATSU 1990) . 10 M. Arterburn et al. Hereditas 148 (2011) observation that these chromosomes demonstrate increased estimates, sequence motifs, such as insertions and dele- signal at the telomeres and centromeres when hybridized tions, shared only between Thinopyrum members, can with St genomic DNA probes, and they suggest that indicate an evolutionary link between them. Sequences this genome is the product of recombination between Eb unique to a single polyploid species should indicate muta- (aka J) and St chromosomes (C HEN 2005). The deca- tions that occurred after a polyploidization event. However, ploid Th. ponticum Barkworth & Dewey has indepen- as with any method, using sequence data to determine dently been described as an autodecaploid of the Eb genome origin in polyploids comes with its own set of genome (2n ϭ 10x ϭ Eb E b Eb E b E b E b E b Eb E b E b , M URAMATSU challenges and caveats. S HAKED et al. (2001) demonstrated 1990; W ANG et al. 1991), an allodecaploid with three that considerable genome editing may occur upon poly- copies of the Eb genome and two copies of the J s genome ploidization; sequences are silenced, some lost and others (2n ϭ 10x ϭ 70, E b E b Eb E b E b Eb J S J S JS JS ) ( CHEN et al. 1998b) form pseudogenes. Evidence suggests that sequence diver- and an allohexaploid with three copies of the Ee genome gence in eukaryotes is correlated with expression level, and two copies of the St genome (2n ϭ 10x ϭ 70, possibly because unexpressed sequences are heterochro- E e Ee Ee Ee Ee E e StStStSt) ( ZHANG et al. 1996b). The divergent matinized, inert and thus less mutable (K RYLOV et al. 2003). viewpoints on genome origins of these species, each Thus, selection pressures and sequence changes could eas- supported by one or more forms of evidence, emphasize ily differ even amongst homologous loci in polyploid the challenges in discerning genome constitutions within genomes. These phenomena add further dimensions to the Thinopyrum genus. sequence-based analysis of polyploid evolution compared The study described here is a sequence-based approach with groups of solely diploid descent. These diffi culties are to investigating Thinopyrum evolution, to augment the amplifi ed in the Thinopyrum genus, as some species, such existing data gathered on this group, and help further as Th. intermedium , are obligate outcrossers and so segre- defi ne the events that shaped this genus. In situ hybridiza- gation may be the source of further variation, even in estab- tion is informative, but the results obtained can be subject lished collections. With these cautions in mind, sequence to the chemical conditions and stringencies employed in analysis, based on discovery of unique Thinopyrum the process, and meiotic pairing analysis, while a highly sequence motifs held in common between polyploids and useful analytical tool, is dependent upon the genetic con- on overall sequence homology, both with each other and text in which such observations are made. J AUHAR et al. their putative ancestors, should help supplement existing (2004), examining chromosome pairing in a trigeneric data and aid in reconstructing the history of this genus. hybrids between Thinopyrum species and durum wheat In this study, the 222 nt internal transcribed spacer 1 (2n ϭ 4x, ABEe E b), reported decreased pairing between (ITS1) region, a 420 nt region of beta-amylase ( Bmy1 ) homoeologous E e and E b chromosomes when the Ph1 intron III, a 672 nt region of the granule-bound starch syn- pairing regulator was active, compared with hybrids in thase (GBSSI) gene, and the 301 nt spacer region between which Ph1 was absent or mutated. These fi ndings indicate the chloroplast genes trnL , which encodes the tRNA that that genome relatedness might be less than what is seen in binds leucine, and trnF, which encodes the tRNA that other studies lacking the scrutiny of the Ph1 system. Con- binds phenylalanine, were sequenced and evaluated in the versely, reliance on pairing studies exclusively in wheat genomes of available Thinopyrum species as well as those amphiploids could result in underestimation of genome- Pseudoroegneria species believed to have been involved relatedness. It is unknown what level of pairing regulation in hybridizations with Thinopyrum species (Ps. strigosa , is employed by Thinopyrum species, or indeed in most of Ps. spicata and Ps geniculata (Trin.) Á . L ö ve). For diploid the wild Triticeae, and the system employed by wheat species and chloroplast sequences this was accom- might be excessively stringent by comparison. These plished by direct sequencing of PCR products while for viewpoints, each supported by quantitative evidence, have polyploid species both direct-sequencing and TA cloning led to debate over the relatedness of the Ee and Eb genomes. of PCR products was carried out and an adequate number These complex dynamics highlight the need for continued of clones were selected to be confi dent of covering all studies with a variety of methods to create a cogent narra- potential genomes. Evolutionary distance was estimated tive for Thinopyrum evolution. The intent of this study is between species using Kimura ’ s two-parameter distance to add molecular sequence information to the data acquired method and phylogenetic trees were created using FastMe from the many studies performed to date and help refi ne (K IMURA 1980, DESPER and GASCUEL 2002). Several shared that narrative a further step. oligonucleotide polymorphisms were evident and the Sequence data has the advantage of an increased ability blastn algorithm was employed to detect these in motifs to estimate the degree of evolution that has occurred after other sequenced Triticeae species in GenBank ( Ͻ http:// each Thinopyrum species underwent polyploidization, ncbi.nlm.nih.govϾ ). Allele-specifi c primers were used while also providing tangible links between polyploids to detect unique ITS sequences in a larger sample of and reputed ancestors. In addition to overall homology Thinopyrum, Taeniatherum and Elytrigia accessions and Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 11

real-time PCR was used to detect relative copy numbers 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 2.5 mM MgCl2 , of this sequence in Th. intermedium and Th. ponticum . 200 μ M dNTP, 1.0 U Taq polymerase (Invitrogen) and 25 pmol of each primer in a 20 μ l reaction. The reactions were carried out over 35 cycles of 94° C for 30 s, 56– 58 ° C MATERIAL AND METHODS (depending on primer set used) for 30 s and 72 ° C for 90 PCR and DNA sequencing seconds. Five μ l of the PCR product was loaded into a 1.0% agarose gel and run at 100V; the resulting PCR frag- DNA was isolated from seeds of wild Triticeae species, ments were excised and purifi ed using the QiaQuick Gel selected at random from bulked seed of each accession pro- Extraction Kit (Qiagen). vided by the Western Regional Plant Introduction State in Purifi ed PCR products of amplifi ed nuclear sequences Pullman, Washington, USA. Accessions used for analysis from polyploid samples were ligated into pGEMT-EZ vec- are indicated in the left-hand column of Table 1. For each tor with the pGEMT-EZ Cloning Kit (Promega), trans- accession studied, fi ve seeds from the bulked collection formed into E. coli DH5α cells and plated, following kit were pooled together and extracted; this was done to obtain protocols. Colonies identifi ed through blue/white screen- a random sampling from the population rather than rely on ing were cultured in LB broth with 100 μ g ml Ϫ1 ampicillin sequence information from a single plant. Pooling of greater at 37 ° C for 16 h. Plasmids were isolated from the broth than fi ve seeds or extractions from multiple individual seeds cultures using the Wizard SV Miniprep Kit (Promega) and of each accession would necessitate a prohibitive amount of μ plasmids were eluted in 50 l of ddH 2O. Nuclear sequences cloning and sequencing and increased the risk of “missing” from diploids and chloroplast sequences were sequenced alleles present in lower copy numbers. Germplasm of directly from PCR products. Th. nodosum , Th. sartorii , Th. distichum and Th. curvifo- The number of clones isolated and sequenced for each lium was not available for distribution, due to poor fertility species was determined using an equation based on the and thus subsequent lack of distributable seed from collec- Product Rule of probability. For clarity and uniformity it ϭ tions. Crithopsis delileana (Schult.) Roshev (2n 14, KK) was assumed that a single gene copy (or region, in the case was provided by the Genebank of the Research Institute of the rDNA) is present in each genome of a polyploid for Crop Production (Prague). Chromosome counts were (even if it is not expressed), and therefore that each is rep- conducted for each accession obtained before PCR and resented in proportion to genome dosage (i.e. one allele sequencing were attempted, to confi rm the identity of the per diploid genome) in the PCR product. The possibility of germplasm. After glumes were removed, seed tissue from heterozygosity was not considered in the equation, due to each accession was ground in a DNA extraction buffer con- time and cost, although this was compensated for by peak- sisting of 100 mM NaCl, 100 mM Tris-HCl, 10 mM EDTA, by-peak analysis of chromatograms obtained from direct- 2% polyvinyl-polypyrrolidone and 1% sarkosyl, then sequenced products. Multiple fl uorescent peaks at a given purifi ed with 25:24:1 phenol:chloroform:iso-amyl alcohol, position, which indicate heterozygosity, were not observed precipitated with isoproponal, dried, and resuspended in outside of those originating from the alleles presented in μ ϫ 100 l of 1 TE buffer. Sequences were amplifi ed from this study. Given the above assumptions, the chance of fail- 30 ng of genomic DNA from samples representing species ing to obtain one of the two sequences from cloned PCR of the genus Thinopyrum , related species of the genus products of a tetraploid is q ϭ (1/2) n where n is the number Pseudoroegneria and twenty accessions of Taeniatherum of colonies selected and subsequently sequenced. To be ϭ ϭ caput-medusae (2n 2x 14, TaTa). certain within q Ͻ 0.05 that both sequences present in the Amplifi cation of the ITS regions was carried out using tetraploid have been observed, at least fi ve colonies must ′ ′ primers ITSL (5 - TCGTAACAAGGTTTCCGTAGGTG -3 ) be picked, so that q ϭ (1/2) 5 ϭ 0.03125. For hexaploid spe- ′ ′ and ITS4 (5 -TCCTCCGCCTTATTGATATGC-3 ), as cies the equation q ϭ (2/3) n ϩ (1/2) n must be satisfi ed, so described by HSIAO et al. (1994). Amplifi cation of the eight colonies must be picked to achieve a q Ͻ 0.05. For ′ GBSSI regions was carried out using primers 5 -TGC the decaploid species Th. ponticum , at least 16 colonies ′ ′ GAGCTCGACAACATCATGCG-3 and 5 -CGCTGAGGCG must be picked to satisfy the equation q ϭ (4/5)n ϩ (3/4)n ϩ ′ GCCCATGTGG-3 , as described in MASON-GAMER (2001). (2/3)n ϩ (1/2) n . In cases where examination of direct- Amplifi cation of the Bmy1 intron III was accomplished sequenced PCR products of tetraploids clearly indicated ′ ′ with primers 5 -ATGAATCTCCRAYGCCTGG-3 and no additional alleles present, this equation was relaxed ′ ′ 5 -CTGCTGCTGCTTTGAARTCTG-3 , as described by (this occurred for Th. scirpeum and Th. junceiforme). M ASON-GAMER (2005). The gene spacer region between the chloroplast genes trnL , and trnF was amplifi ed using Sequence analysis and phylogenetic tree construction the primers 5′ -TCCGTCGACTTTATAAGTCGTG-3′ and 5 ′ -TGCCAGGAACCAGATTTGAACT-3′ , as described Forward and reverse sequencing was carried out on each by N ISHIKAWA et al. (2002). The PCR solution consisted of template using dideoxy-terminator sequencing with the 12 M. Arterburn et al. Hereditas 148 (2011) same primers used for amplifi cation. The 10 μl reaction ITS1 and Bmy1 allele sequence together in tandem to form consisted of 4 μ l BigDye Terminator 3.1 Ready Reaction a ∼ 662 nt sequence, to represent the loci present in a single Mix (Applied Biosystems), 75 ng DNA and 5 pmol of genome of the polyploid. When the ITS and Bmy1 alleles primer. Sequence reactions were purifi ed using CentriSep isolated from a polyploid had closest calculated similarity spin columns (Princeton Separations) and dessicated in a to the same diploid (ex: TINT-ITS1-St allele and TINT- vacuum centrifuge. Sequencing was carried out on an Bmy1 -St allele), these were assumed to share origins and ABI Prism 373 DNA Analyzer by the Laboratory for thus combined together to represent loci in a single Bioanalysis and Biotechnology at Washington State genome. When a sample exhibited only one allele at a University. Sequences were aligned using the program locus, then this allele was combined with each allele of the BioEdit and the ClustalW algorithm (ver. 1.81) with the other locus. The second tree, the ‘ species tree ’ , was con- IUB DNA weight matrix ( THOMPSON et al. 1994, T IPPMANN structed by combining all alleles of the ITS1 and Bmy1 2004). Nucleotide substitution analysis was performed loci together to provide a sequential representation of the and sequence comparisons were made using Kimura ’ s alleles present in all genomes of the species. This was two-parameter distance method with the Data Analysis accomplished by combining six ITS1 and six Bmy1 in Molecular Biology and Evolution software package sequences together in tandem, for total average sequence (X IA and X IE 2001). Phylogenetic trees were constructed length of approximately 3850 nt, with the proportion of using both FastMe and ClustalW (ver. 1.81) (D ESPER alleles in the sequence weighted by the number of clones and G ASCUEL 2002). Both analyses produced trees with recovered of each allele. For example, for Th. intermedium the same groupings and nearly-identical distances and ITS1, four clones each were isolated with closest distance thus only the trees generated by FastMe are presented in to the Ee , Eb and K alleles respectively, thus two copies of this report. Oligonucleotide polymorphisms (insertions/ each of these allele sequences were combined for the ITS1 deletions) of four nt or greater were noted and sequences portion of the sequence (an equal proportional representa- were submitted to GenBank ( Ͻ www.ncbi.nlm.nih.govϾ ) tion of alleles). Allele sequence order was preserved as for BLAST analysis to identify other sequences contain- closely as possible so that comparison would be made ing these motifs. On this basis, the ITS1 and Bmy1 sequences between sequences of closest origin: Ee alleles were placed of Ta. caput-medusae reported by BLATTNER (2004), and fi rst, Eb second, St third, K fourth and X last. The third tree the GBSSI sequences of Ta. caput-medusae , Ps. spicata was based solely on individual alleles of GBSSI sequence; and Ta. caput-medusae reported by MASON-GAMER (2005) this was done because we did not acquire GBSSI data for were downloaded from GenBank and added to sequence every specimen, and because, unlike the ITS1 and Bmy1 alignments for distance calculations and tree construc- intron III, the GBSSI sequences contained both coding tion. The ITS1 sequence submitted by H SIAO et al. (1994), and non-coding portions (also, some alleles were frame- the Bmy1 sequence submitted by M ASON-GAMER (2005) shifted pseudogenes) and thus were likely infl uenced by and the GBSSI sequence of M ASON-GAMER et al. (1998) different selection pressures. of Triticum monococcum were download from GenBank and added to each sequence alignment to serve as an out- Allele-specifi c PCR of ITS regions group. Sequences generated in this study were submitted to GenBank. PCR was performed on 42 accessions of Thinopyrum/ To make sorting, alignment and tree construction sim- Pseudoroegneria, twenty accessions of Taeniatherum pler, alleles identifi ed were given genome-designations caput-medusae (L.) Nevski (PI598389, PI314697, based on the sequence of the putative diploid ancestor PI217615, PI220590, PI561094, PI208075, PI220589, with which they shared closest sequence identity and low- PI227665, PI317475, PI561110, PI577708, PI204557, est evolutionary distance value. For example, the sequence PI204557, PI561108, PI577710, PI251387, PI317476, TINT-ITS1-Eb is used to denote the Eb -allele, one of the PI270591, PI361110 and PI208075) and 23 accessions three ITS1 alleles isolated from Th. intermedium , which of Elytrigia repens (L.) Desv. ex Nevski (PI207451, according to distance calculations is most closely-related PI317410, PI222960, PI565006, PI502361, PI229920, to ITS1 in the Eb genome of Th. bessarabicum . Full PI440076, PI595136, PI499490, PI380624, PI173628, description and rationale for these designations is pro- PI547340, PI172359, PI621812, PI240128, PI311333, vided in the Results section. PI439989, PI180407, PI429780, PI838582, PI531747, Construction of sequence-based phylogenetic trees is PI240129 and PI618807), using the non-selective reverse problematic when analyzing groups of mixed ploidy lev- primer ITS4 and either of the selective forward primers els, as existing algorithms are not designed with these ITSK1 (5 ′ -GTCCGTGGCGAGGGTTTTCC-3′ ) or ITSK2 complex groups in mind. Three trees were constructed (5′ -GTCCGTGGCGACGGTTTACC-3′ ), to detect the using nuclear gene sequence alignments. The fi rst tree, the presence of the ITS1 K-allele based on the polymorphism “ genome tree,” was constructed by combining a single at nucleotide 56 of the ITS sequence in these species. Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 13

The former primer is denoted ITSK1 because it is identi- Th. elongatum (Ee) GACGGCATCGTCTGTCGCTCGA cal to the Cr. delileana sequence and the latter is denoted ITSK2 because it has a single T→ A transversion com- Th. bessarabicum (Eb) GACGGCACCGTCCGTCGCTTGA pared to the Cr. delileana sequence, but is the type found in Th. intermedium and Th. ponticum. PCR conditions Ps. strigosa (St) GACGGCACCATCCGTCGCTTGG were identical to those described earlier, with an anneal- (K) GACGGTTTACACCGTCGCTCGG ing temperature of 56° C. Cr. delileana Th. intermedium Eb(x4) GACGGCACCGTCCGTCCGTCCG Real-time PCR detection of ITS regions St (x4) GATGGCACCATCCATTGCTCGG Real-time PCR analysis was performed on a BioRad iCy- Th. intermedium cler device, using 30 ng of genomic DNA from accessions Th. intermedium K (x4) GACGGTTTTCACCGTCGCTCGG of Th. intermedium and Th. ponticum . Genomic DNA of Chinese Spring wheat was used as a negative control, in Th. ponticum K (x16) GACGGTTTTCACCGTCGCTTGG addition to an H2 O control. Primers ITSK2 and ITS2 were used at 10 pmol per reaction to produce a 250 bp product. Fig. 1. Alignment of ITS1 sequences, nucleotides 50-71. Shaded Positive controls consisted of a cloned 250 bp fragment of boxes indicate sequence differences, including a motif possibly ITS1 K-allele DNA from Th. intermedium at 1.5 ng, 0.15 ng, displaced by the poly-T insertion at nucleotide 56 (bracketed). 0.015 ng, 0.0015 ng and 0.00015 ng, respectively. PCR was carried out in a solution of 50 mM KCl, 20 mM Tris-HCl (pH 8.4), 0.2 mM of each dNTP, 0.025 U iTaq species. Sequencing of Cr. delileana ITS1 in this study DNA polymerase and 10 nM SYBR Green I fl uorescein. confi rmed the presence of this allele in that species. How- Reactions were carried out with an initial denaturation ever, twenty Ta. caput-medusae accessions analyzed in phase for 10 min at 95° C, 40 cycles of 95° C for 20 s, 59° C this study all lacked this motif; of these, six (PI598389, for 15 s and 72 ° C for 30 s and a fi nal denaturation phase PI314697, PI217615, PI220590, PI561094 and PI208075) at 95 ° C for 3 min. Dissociation values of the resulting were assessed through direct sequencing and all twenty PCR products was established using a melt curve by via allele-specifi c PCR. This suggests that the Ta. caput- increasing temperature from 55° C to 95° C over 80 cycles. medusae specimen examined by BLATTNER (2004) may be Comparison of genomic samples to standards using the a unique ecotype. Because of its homology to Cr. delile- BioRad iCycler software determined the relative abundance ana , a diploid of the K genome, this ITS1 allele was des- of K-allele ITS sequences per sample. ignated the K-allele in this report. Of all reported sequences, only Cr. delileana, a single accession of Ta. caput-medusae, Th. intermedium and RESULTS Th. ponticum appear to possess the K-allele of ITS1. Other polymorphisms in the ITS1 region between the organisms Sequence motifs and allele designations examined in this study consisted of single nucleotide The ITS1 sequences exhibit a fairly narrow range of polymorphisms, mainly Py→ Py transitions. The only con- sequence polymorphism, with the most divergent sistent polymorphisms seen both in putative diploid ances- sequences differing by sixteen nucleotides and the most tors and in one or more of the Thinopyrum polyploids are similar differing by only three. The most distinct feature is a C→ T transition at nt 63 in Th. elongatum and a GT dinu- an oligonucleotide polymorphism, detectable at nt 56 in cleotide at nt 78 in Ps. spicata . Sequences with these two Th. ponticum and Th. intermedium (Fig. 1). This allele features had closest overall sequence identity with appears to have formed via a four-thymidine insertion and Th. elongatum and Ps. spicata respectively, and were des- an accompanying deletion of the downstream CCGT, ignated the E e-allele and the St-allele, although it should which is present as a direct repeat in nearly all other Trit- be noted that single nucleotide polymorphisms such as iceae species, as reported on GenBank. This hypothetical these could have arisen independently. mechanism is supported by the displaced CACCGT motif Sequences of intron III of the Bmy1 gene have been in this polymorphic region. BLAST analysis revealed successfully used for genotyping and prediction of evolu- that only Cr. delileana ( HSIAO et al. 1994) and Ta. caput- tionary distance in barley and its relatives ( SJAKSTE medusae ( BLATTNER 2004) have been reported with a and ZHUK 2006). Analysis of a 420 nt region of this intron similar ITS1 allele, hence the inclusion of these species in in Thinopyrum reveals a broader range of divergence this investigation as potential ancestors to the Thinopyrum than observed for ITS1, with the most similar sequences polyploids. It should be noted that the fourth thymidine being completely identical and the most divergent having of this motif is instead an adenosine in both of these two 63 total transitions and transversions. The most distinct 14 M. Arterburn et al. Hereditas 148 (2011) oligonucleotide polymorphism occurs at nucleotide 179 A 672 nt region of the granule bound starch synthase in some species, involving the insertion of the eight nucle- (GBSSI) gene, beginning in intron 9 and terminating in otide motif TTTTGAGT (Fig. 2, shaded box). This motif intron 11, was compared between Th. ponticum, Th. inter- is found in Th. intermedium , Th. ponticum, Th. scirpeum , medium and likely ancestors. Five polymorphic motifs, Th. caespitosum and Ps. geniculata . GenBank searches present in a single allele, are of particular interest (Fig. 3). revealed that the only other Triticeae species reported to BLAST analysis demonstrated that all fi ve of these have a homologous allele is Ps. spicata (accession unique sequences have identity with Ta. caput-medusae PI232117), as described by MASON-GAMER (2005), and accession PI208075 and select E. repens accessions, so this sequence was designated the St-allele. A 15- as described by MASON-GAMER (2001). Thus, this allele nucleotide deletion at position 123 in Cr. delileana and a was designated the Ta-allele because of its similarity to fi ve-nucleotide insertion at position 363 in one of the Ta. caput-medusae and was found in Th. intermedium and three Th. junceum alleles are the only other polymor- Th. ponticum alone amongst the sequences examined in phisms observed involving more than three residues. this study. This allele was not observed in Cr. delileana . Additionally, this sequence exhibited a number of single, The intron polymorphisms at nucleotides 73, 96, 151 and di- and tri-nucleotide polymorphisms shared between 655 of the Ta-allele are likely oligonucleotide additions, polyploids and putative ancestor species, and these are possibly caused by polymerase slippage events that refl ected in evolutionary distance estimates. formed direct repeats. At nucleotide 73 the sequence Evolutionary distance estimates for the Bmy1 sequences CACC is repeated three times in tandem in other alleles, are summarized in Table 2. Alleles found in the polyploids but only the sequence CCCC is present, in a single copy, that demonstrate closest similarity to Th. elongatum were in the Ta-allele. At position 96, the sequence TTC is pres- designated the Ee -allele of Bmy1 and those closest to ent only once in the middle of a C:T rich region in other Th. bessarabicum were referred to as the Eb -allele. Alleles alleles, but is present as a tandem repeat in the Ta-allele. in the polyploids that had closest identity with Ps. spicata In Th. bessarabicum , nine nucleotides are deleted at all exhibited the exact eight-nucleotide insertion at posi- this position compared to other alleles. At nucleotide 151, tion 179, as well as combinations of other common point the sequence CAGG is present as a tandem repeat in the mutations; these alleles were all designated St-alleles. A Ta-allele and in Ps. spicata , but only once in the other fourth sequence recovered in four of the polyploids shows sequences assessed. At position 655, the sequence considerable variation from all other Triticeae Bmy1 CAGTCCTTCTT is present in Ps. spicata, while the sequences, as evidenced by the evolutionary distances sequence CCTTCCT is present in T. monococcum and reported in Table 2. Aside from a 78-nt stretch that has the tri-nucleotide motif CCT is present in the Ta-allele. 87% identity with the Bmy1 intron of several Hordeum The other sequences examined in this study lacked any of vulgare cultivars, the remainder of the sequence has no these motifs at this position. While the other four poly- appreciable similarity to other sequences that have been morphisms are located in introns, the oligonucleotide submitted to GenBank. The origins of this sequence change at nucleotide 572 is a 10 nucleotide deletion in remain enigmatic and so this allele was designated the the exon 11 that results in a non-functional frameshift, ren- X-allele because its sequence could not be ascribed to dering subsequent coding regions of the gene unusable in any putative ancestor. The X-allele is very divergent from the Ta-allele. While many Triticeae accessions have non- the putative ancestors, with only 56 – 57% identify with the functional GBSSI genes, only Ta. caput-medusae , E. repens, putative ancestors and an average evolutionary distance and, in our study, Th. intermedium and Th. ponticum have value from these of 0.243, an approximately 2.5-fold been reported to have this specifi c frameshift mutation, increase over the next most divergent Bmy1 sequence, according to BLAST analysis. Th. ponticum and Th. inter- among the Thinopyrum accessions examined. BLAST medium exhibit all fi ve of the Ta-allele oligonucleotide analysis indicated that the fl anking exons in the X-allele motifs, and estimated evolutionary distances of this allele sequence do indeed code for beta amylase, and since it in these two polyploids are closest to E. repens (0.005 and was observed in samples from Th. intermedium , Th. pon- 0.013, respectively) but also very close to Ta. caput-medusae ticum Th. caespitosum, and Th. junceum it is unlikely to (0.011 and 0.016). be the result of contamination or similar PCR error. Direct sequencing of GBSSI PCR products from Unusual sequence motifs are not entirely unexpected as Th. ponticum and Th. intermedium revealed heteroge- the Triticeae Bmy1 intron III has been shown to be prone neous chromatograms in which the two GBSSI alleles, to rearrangements of modular sequences, including micro- the non-functional and functional genotypes, are visible. satellites, transposable elements and transcription factor In both of these species the non-functional allele pro- binding sites, which could involve sequence inversions duced the more intense signal and it is therefore not sur- and other mechanisms that could have greatly changed the prising that only 1 out of 10 clones from Th. intermedium nature of this sequence ( SJAKSTE et al. 2003). and 2 out of 12 clones sequenced from Th. ponticum were Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 15 l alleles with

nt 190 indicates sequence * TACGGCT--- G -----TTTGA TATGGCT--- G -----TTTGA TATGGCTTTT TGAGTTTTGA G --TGGCT--- G -----TTTGA TATGGCT--- G -----TTTGA TATGGCT--- T -----TTTGA TTTGGCT--- G -----TCCGA TTTGGCT--- G -----CCCGA arrant separate display. arrant separate display. nt 175 quences were necessarily identical). Multipliers ...... nt 95 ATGAAA CGGTAA ------CGGTAA ----GCCCAG ------TGGTAA ----ACCCAA ------ACCCAA ------TGGTAA ----ACCCAA ------TGGTAA ----ACCCAA ------CGATAA ----ACCCAA AAGTTTT TTCAACAAAT CAATACCCAA CCAAGTGACG ATCTGCATTG AAAAGGGTTC CCAA------CCAAGAGACG ATTTGCACTG AAAAAGGTTC CCAAGTGACG ATCTGCACTG AAAAAGGTTC CCAAGAGACA ATTTGCACTG AAAAAGGTTC CCAAGAGACG ATTTGCACTG AAAAAGGTTC TCAAGTGATG ATCTGCACTG AAAAAGGTTC CCAAGTGACG ATCTACATTG GAAAAGTTTT TTCAACAAAC CAATACCCAA ATGAAA nt 40 (x3) e allele allele b e X allele St allele E E CDEL (K) TMED (Ta) TINT X (x3) TJUN E intron III allele sequences, nucleotides 40-95 and 175-190. Shaded boxes indicate sequence changes. Left-hand column lists al Bmy1 ) ) e b (St) (E (x6) (xE) (x1) (x1) (E (x1) (x1) e e e e b b X (x1) X (x3) X (x2) X (x1) E E E E St (x2) St (x2) E TINT TJUN TPON TELO TINT PGEN TPON TINT St (x4) TPON St (x3) PGEN TJUN E TCAES TJUNF PSPIC* TSCIR TJUNF St (x1) TCAES TBESS Abrreviated alignment of downloaded from GenBank (cited in text). downloaded from GenBank (cited in text). Fig. 2. Fig. matching oligonucleotide motifs and closest evolutionary distances to the sequences displayed (this does not mean that these se indicate number of clones recovered of each sequence type from polyploids. Alleles TJUN-Ee and TINT-X were distinct enough to w TINT-X TJUN-Ee and Alleles indicate number of clones recovered each sequence type from polyploids. 16 M. Arterburn et al. Hereditas 148 (2011) X TPON St TPON

e E TPON X TIN ransversions between the two the two between ransversions St TIN volutionary distance value(s) for volutionary distance value(s) for

e E TIN X TJUN

b E TJUN

e E TJUN St TJUF

e E

TJUF

St PGEN

e E St-allele.] Th. Ponticum Th. junceum Elytrigia repens Elytrigia

PGEN Th. intermedium

ϭ ϭ ϭ ϭ X TCAS tponticum .

b Th E TCAS TPON TPON EREP EREP TJUN TINT St TSCP TSCP Th. scirpeum Th. caespitosum

Ps. geniculata

Th. junceiforme Th. junceiforme ϭ ϭ

ϭ ϭ TSCIRP

PGEN TJUF ELO TBESS PSPIC CDEL TMED

TCAES 0.114 0.089 0.014 0.089 0.100 0.100 0.0800.118 0.093 0.008 30/37 0.093 20/12 0.104 21/7 0.102 20/11 0.090 22/7 0.019 15/14 0.261 0.096 4/3 0.09 28/38 21/12 0.093 21/7 0.093 0.099 26/36 21/12 23/7 26/35 28/41 21/14 13/11 25/39 21/14 15/11 28/38 0.039 0.0190.030 0.091 0.000 0.0740.087 0.088 0.026 0.060 0.079 0.028 0.094 0.017 0.0710.042 0.092 0.077 0.014 0.022 0.255 0.072 0.075 0.094 0.089 0.0680.042 0.260 0.077 0.069 0.022 0.019 0.090 0.029 0.094 18/7 0.075 0.252 0.030 0.066 0.077 0.074 6/1 0.069 0.096 0.029 19/7 0.060 0.259 0.030 21/6 0.071 10/12 0.003 0.074 20/13 10/11 0.074 0.096 23/43 20/12 0.022 0.259 25/42 1/0 0.078 0.003 17/15 7/1 0.066 18/7 21/46 0.074 0.099 11/12 20/6 21/40 0.022 0.242 23/8 23/39 0.078 1/0 19/44 0.066 7/1 11/12 0.099 20/7 0.242 22/6 25/8 19/7 22/40 0.005 22/40 24/39 0.074 19/44 0.240 2/0 22/7 23/40 22/7 22/40 0.099 0.0780.039 0.088 0.014 0.097 0.091 21/9 0.026 21/9 2/1 21/6 23/8 21/13 30/35 30/41 18/8 7/3 2/1 21/6 20/7 3/2 3/1 21/8 23/9 21/14 15/12 11/13 26/36 26/44 19/8 8/3 2/1 21/6 24/34 23/41 19/8 8/3 4/1 23/8 24/33 25/41 Triticum monococcum Triticum TMONO T

Th. bessarabicum Ta. caput-medusae Ta.

Ps. spicata

ϭ Cr. delileana Cr. Th. elongatum

) 0.113 0.088 22/8 20/12 20/10 24/7 3/2 30/39 19/12 24/7 19/11 23/7 16/16 2/1 28/40 20/12 22/7 26/38 20/12 24/7 26/36

) 0.030 19/11 20/7 4/2 3/2 20/6 20/11 29/41 6/1 22/6 6/1 21/6 10/11 20/12 25/42 7/1 20/6 23/39 7/1 22/6 24/39 Evolutionary distance estimates between all Bmy1 alleles observed in this study. Values above the diagonal are of transition/t are the diagonal above Values alleles observed in this study. all Bmy1 Evolutionary distance estimates between b e ϭ e e ϭ b e b ϭ ϭ ϭ e e CDEL CDEL TMED PSPIC TBESS TELO TMONOTELO (E 7/4 23/15 23/11 6/5 8/6 23/10 24/15 31/45 9/5 23/10 7/4 24/10 24/16 15/15 27/46 10/5 23/10 25/43 10/5 25/10 26/43 species and values below the diagonal are distance values based on two-parameter distance. Shaded cells indicates the closest e Shaded cells indicates the closest distance. distance values based on two-parameter are the diagonal species and values below no distance. identical and show cases, equences are an allele on either axis. In four Table Table 2. TPON St is the [Alleles sequenced from a polyploid are indicated: KEY DIPLOIDS TMONO TETRAPLOIDS HIGHER POLYPLOIDS PGEN StTJUF E 0.096TJUF StTJUN E 0.075TJUN E 0.099 0.085 0.078 0.008 0.088 0.085TIN St 0.084 0.008TIN X 0.000 0.091TPON E 0.096 0.080TPON St 0.087 0.242 0.268 0.075TPON X 0.003 0.071 0.102 0.083 0.240 0.085 0.247 0.258 0.080 0.255 0.008 0.074 0.231 0.091 0.088 0.235 0.003 20/6 0.237 0.261 0.084 0.013 0.078 0.000 0.249 0.095 1/0 0.207 0.080 0.226 0.250 0.090 13/11 0.242 0.240 0.005 21/8 0.245 24/8 0.071 0.036 0.086 26/37 0.210 14/11 0.000 0.236 0.252 0.223 19/7 27/37 0.075 0.226 0.078 0.041 0.003 20/7 0/0 0.240 0.005 0.227 0.069 0.231 1/0 24/35 0.080 0.090 0.210 0.245 0.008 19/7 25/35 0.225 0.231 0.245 0.074 0.074 0.215 20/7 0.039 0.096 2/0 0.246 0.240 0.233 0.237 0.227 3/0 24/34 0.080 0.034 0.005 25/34 24/35 0.231 0.236 0.211 19/7 0.081 0.018 22/40 0.232 2/0 26/45 0.219 24/34 5/1 26/32 PSPIC (St) CRDE (k)TMED (Ta) 0.029 StTSCP 0.017TCAS E 0.096 0.099 0.094 0.080 0.085 0.008 0.091 5/4 0.084 21/8 21/12 29/42 21/7 6/3 30/36 18/7 21/8 4/2 0/0 22/8 20/6 21/13 1/0 11/13 25/41 13/11 7/3 23/8 26/37 21/8 19/7 23/40 0/0 7/3 26/37 23/8 19/7 24/40 2/0 24/44 TBESS E TCAS XPGEN E 0.287 0.260 0.262 0.239 0.282TJUN X 0.273TIN E 0.242 0.268 0.247 0.242 0.230 0.221 27/42 0.253 30/36 0.255 29/41 21/36 0.225 23/45 0.238 30/40 0.049 6/12 0.238 28/42 0.225 30/36 0.242 6/7 0.230 0.244 28/42 0.252 32/40 5/6 24/43 26/37 5/8 24/43 28/37 5/7 Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 17 nce changes.

nt 669 nt nt 159 was distinct enough to warrant separate display. was distinct enough to warrant separate display. e played (this does not mean that these sequences were and Intron 11). Shaded boxes indicate seque and Intron 11). AGGCAAAGTGA----CAGGCG TTG AGGCAAAGTGA----CAGGCG TTG AG-CAAAGTGACAGGCAGGTG TTG AGGCAAAGTGACAGGCAGGCG TTA AGGCAAAGTGACAGGCAGGCG TTA AGGCAAAGTGA----CAGGCG TTG AGGCAAAGTGA----CAGGCG TTG nt 140 ...... Start of intron11 of Start nt 111 570-669 (Exon 11 570-669 (Exon 11 and closest evolutionary distances to the sequences dis GTGAGCACCCACCCACCAGC ACAAAGATTT CTTC---CTC TTCTTCCGGT GTGAGCACCCACCCACCCAC ACAAAGATTT CTTC------CGGT GTGAGCACCCACCCACCCAC AAAAAGATTT CTTC---CTC TTCTTCCGGT GTGAGCACCC C------C ACAAAGATTT CTTCTTCCTC TTCTTCGGGT GTGAGCACCCACCCGCCCAC ACAAAGATTT CTTC---CTC TTCTTCCGGT GTGAGCACC- ACCCACCAGC ACAAAGATTT CTTC---CTC TTCTCCCGGT CACCAGCCGC TTCGAGCCCT GCGGCCTCAT CCAGCTCCAG GGAATGCGCT ACGGAACGGT AGACTTTTCC TTCTT------GCCA CACCAGCCGC TTCGAGCCCT GCGGCCTCAT CCAGCTCCAG GGGATGCGCT ACGGAACGGT AAACTTTTCC TTCTT------GCCA CACCAGCCGCTTCGAGCCCT GTGGCCTCAT CCAGCTCCAG GGAATGCGCT ACGGAACGGT AAACGCCTCC TCCTTCAGTC CTTCTTGCCA CA------CGAGCCCT GCGGCCTCAT CCAGCTCCAG GGGATGCGCT ACGGAACGGT AAACGCCTCC TCCTT------CCTGCCA CACCAGCCGC TTTGAGCCCT GCGGCCTCAT CCAGCTCCAG GGGATGCGCT ACGGAACGGT AAACTTTTCC TTCTT------GCAA CACCAGCCGCTTTGAGCCCT GCGGCCTCAT CCAGCTCCAG GGGATGCGCT ACGGAACGGT AAACGCCTCC TCCT-----C CTTCCTGCCA GTGAGCACCC C------C ACAAAGATTT CTTCTTCCTC TTCTTCGGGT CACCAGCCGC TTCGAGCCTT GCGGCCTCAT CCAGCTCCAG GGGATGCGCT ACGGAACGGT AAACTTTTCC TTCTT------GCCA nt 62 nt 570 nt sequence, nucleotides 62-111 (Intron 9), 140-159 10) and sequence, nucleotides 62-111 ) ) b b (x1) (x1) e e GBSSI Ee allele Ee allele TMONO* TMONO* CDEL (K) CDEL (K) Ta allele* Ta allele* Ta PSPIC (St)* PSPIC (St)* PSPIC TBESS (E TBESS (E TINT E TINT E ) ) e e (Ta) (Ta) (Ta) (Ta) (E (E (x4) (x4) e b Ta (x9) Ta (x5) Ta (x9) Ta (x5) Ta E E EREP EREP TELO TELO TMED TMED TINT TINT TPON TPON TPON TPON Abbreviated alignment of indicates sequences downloaded from GenBank (cited in text). indicates sequences downloaded from GenBank (cited in text). * Fig. 3. Fig. Left-hand column lists all alleles with matching oligonucleotide motifs

necessarily identical). Multipliers indicated number of clones recovered of each sequence type from polyploids. Allele TINT-E Allele necessarily identical). Multipliers indicated number of clones recovered each sequence type from polyploids. 18 M. Arterburn et al. Hereditas 148 (2011) of the functional GBSSI allele. However it should be calculations used in ‘ species tree ’ construction because mentioned that the PCR was not attempted with addi- existing phylogenetic algorithms are not designed for tional primer sets, so primer bias cannot be ruled out as polyploid analysis. The method used, while believed to be the cause of this skew, unlike in our investigation of the the most accurate possible for our study, biased diploids, ITS1 K-allele. Regardless, this suggests that Th. ponti- with their single unique allele each (copied in tandem six cum and Th. intermedium both possess a functional and times), to be calculated as more distinct from each other, non-functional allele of the GBSSI gene, and also sug- than with tetraploids which have a combination of alleles gests a possible evolutionary link between these two poly- similar to the putative ancestors. This bias is less intense ploids, Ta. caput-medusae and E. repens . in the ‘ genome tree ’ , and its associated calculations, The sequence of the 306-nucleotide spacer region because equal proportions are being compared; in this between the chlorosplast genes trnL and trnF were found sense comparison of diploids to tetraploids is similar to to be nearly identical between all samples examined. Only comparing sequences of homozygotes and heterozy- ﬁ ve single nucleotide polymorphisms were evident: an gotes. These methods should still provide estimation of A →G transition shared by Th. bessarabicum , Ps. genicu- relatedness, but with a degree of skew when comparing lata and Th. ponticum , a G→ A transition at residue 25, groups of different ploidy levels. Evolutionary distance shared by Th. elongatum and Th. scirpeum, a G→ T trans- estimates between single alleles, or between species of version at residue 25 in Th. intermedium , an A → G transi- the same ploidy, are also reported and are not subject to tion at nucleotide 179 in Ps. strigosa and Th. junceiforme , these biases. For these reasons, we chose to examine the and a G→ T transversion at position 235 in Th. scirpeum . evolutionary distance between the putative diploid ances- From a phylogenetic standpoint this sequence was not tors in an independent calculation, before comparing especially informative, although it does suggest a level of them with the polyploids. common cytoplasmic inheritance in some species. Overall sequence homology, based on comparing the combination of ITS1, Bmy1 and GBSSI sequences between diploid species, places Th. elongatum close to Cr. delile- Evolutionary distance between diploids ana (0.032 distance value), with a more distant relation- Examination of conserved sequence motifs, evolutionary ship to Th. bessarabicum (0.045) and Ta. caput-medusae , distance estimates and positions on the constructed phy- and most divergent from Ps. spicata (0.077). The ITS1 logenetic trees indicates the relatedness of Thinopyrum and GBSSI sequences of Th. elongatum are most closely polyploids and their genomes to each other and to their related to those of Th. bessarabicum (0.030 and 0.028 putative ancestors. These results are summarized in Table 1. respectively), and most distant from Ps. spicata (0.078 Figure 4 depicts the ‘ species tree ’ , based on phylogenetic and 0.094 respectively). When comparing Bmy1 sequences analysis of all combined allelic sequences from the ITS1 alone, Th. elongatum is again closest to Ta. caput-medusa and Bmy1 alleles. Figure 5 depicts the ‘ genome tree ’ in (0.014) and Cr. delileana (0.017), and distant from which single ITS1 and Bmy1 alleles were paired to repre- Th. bessarabicum (0.088) and Ps. spicata (0.078). sent putative genomes from each polyploid. Figure 6 Sequence comparison based on all three loci indi- provides evolutionary distances and a phylogenetic tree cates that Th. bessarabicum is most closely related to based on GBSSI sequences alone. It should be noted that Th. elongatum (0.045) and most divergent from Ps. spicata there is unavoidable bias in the evolutionary distance (0.082). The ITS1 sequence was closest to Th. elongatum

TELO(Ee) TJUNF(EeEeEeEeEeEe / EeEeEeStStSt) PGEN(StStStStStSt/ EeEeEeStStSt) PSPIC(St) (EeEeEeEeEeEe / EeEeEeStStSt) TSCIRP (EeEeEeEeEeEe / StStStStStSt) TBESS (Eb) TCAES TJUN (EeEeEeEeEbEb / EeEeEbEbXX) TINT (EbEbStStKK / EeEeStStXX)

TPON (KKKKKK / EeEeStStXX) TMED (Ta) CDEL (K) TMONO (A) Fig. 4. Phylogenetic “ species tree ” based on combining all alleles of ITS1 and Bmy1 intron III into a single sequence set. Putative diploid ancestors are shaded grey, while abbreviations of polyploid species are followed in parenthesis by notation indicating the six ITS and six Bmy1 alleles, respectively, that were combined into a single sequence. Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 19

TBESS (Eb) TCAES1 (St/Eb x7) TJUN2 (Eb/Eb x4) PSPIC (St) TSCIRP(Ee/St x7) TJUNF2 (E/St x6) PGEN1 (St/St x7) TINT1(K/St x8) TPON1(K/St x19) TCAES2 (St/X x5) TINT3b (K/X x7) TPON3 (K/X x17) TJUN3 (Ee/X x9) TJUN1 (Ee/Ee x10) TELO(Ee) TJUNF1 (Ee/Ee x6) TMED (Ta) PGEN2(St/Ee x7) TINT2 (Eb/Ee x7) TINT3a (K/Ee x7) TPON2 (K/Ee x22) CDEL (K) TMONO (A) Fig. 5. Phylogenetic “ genome tree” for individual polyploid genomes. Each data point represents a combination of a single ITS allele and a single Bmy1 allele that have been combined to represent the allele combinations present in a single genome from each polyploid species. TINT3a and TINT3b represent two possible combinations of the K-allele in Th. intermedium .

(0.030) and most divergent from Cr. delileana (0.074), The Bmy1 allele of Ps. spicata is divergent from the other while the GBSSI sequence was most divergent from three diploids in the range 0.078 to 0.094, with Th. elon- Ps. spicata (0.102) and closest to Th. elongatum (0.028). gatum the most closely related. The GBSSI sequence of Th. bessarabicum exhibited the most distinct Bmy1 allele Ps. spicata is the most divergent of the diploids, ranging of the diploids, diverging from the other diploids with from 0.094 to 0.113 in distance. The ITS allele is similar distance values ranging from 0.088 to 0.099. to Th. elongatum (0.034) and Th. bessarabicum (0.030). Sequence comparison based on combined loci indicates Based on combined sequence analysis, Cr. delileana is that Ps. spicata is the most signiﬁ cantly diverged from the most closely related to Th. elongatum (0.032) and Ta. caput- other diploids. It is most closely related to Th. elongatum medusae (0.052) and most divergent from Ps. spicata (0.077) and most distant from Cr. delileana (0.091). (0.091). The ITS1 sequence is closest to Ta. caput-medusae

PSPIC (St) PSTRIG(St) TELO (Ee) 0.032 CRDE(K) TBESS (Eb) 0.034 0.027 TPON Ee (x1) 0.013 TINT Ee (x2) EREP(Ta) 0.005 TPON Ta (x9) TINT Ta (x8) 0.014 TAEN (Ta) 0.016 TMONO (A) Fig. 6. Phylogenetic tree based on GBSSI intron 9-11 sequence. Putative diploid ancestors are shaded grey. Multipliers indicate the number of clones recovered of each allele from polyploids. Arrows indicate Kimura ’ s two-parameter distance value between the two bracketed sequences. Putative diploid ancestors are shown in grey. 20 M. Arterburn et al. Hereditas 148 (2011)

(0.047), the Bmy1 is close to Ta. caput medusae (0.026) polyploids with Bmy1 St-alleles. These data indicate that and the GBSSI sequence is closest to Th. elongatum Th. junceiforme could be an allotetraploid combining the (0.032) and Th. bessarabicum (0.034). Ee genome and St genome. The combined ITS1/ Bmy1 sequences of Ta. caput- Ps. geniculata has two Bmy1 alleles: an St-allele and an medusae are most closely related to Th. elongatum (0.045) Ee -allele. The Ee -allele is of equal similarity to Th. elonga- and Cr. delileana (0.052). The GBSSI sequence is greatly tum (0.019) and the Ee -allele of Th. junceiforme (0.019). diverged from Ps. spicata (0.113) and differs from the The St-allele is completely identical to those found in other three diploids in a narrow range of 0.062 to 0.068. Th. scirpeum (0.000) and Th. intermedium (0.000) and The Bmy1 allele is very similar to Th. elongatum (0.014) nearly identical with those found in Th. junceiforme and Cr. delileana (0.026), which accounts for their over- (0.003), Th. ponticum (0.005) and Ps. spicata (0.008). all close distance. In summary, the closest sequence The presence of the X-allele in Th. intermedium and relationships, when considering combined ITS1 and Th. ponticum prevents them having a close distance with Bmy1 sequences, appear to exist between Th. elongatum , Ps. geniculata when considering combined sequences, Cr. delileana and Ta. caput-medusae , with Th. bessarabicum and thus Ps. geniculata is estimated to be closest overall more distantly related and Ps. spicata the most divergent. to Th. junceiforme (0.015) and also close to Ps. spicata (0.031), Th. scirpeum (0.036) and Th. elongatum (0.038) Evolutionary distance between tetraploids and as seen on the ‘ species tree’ . On the ‘ genome tree’ , one Ps. geniculata e putative ancestors allele combination of clusters with E -allele combinations from Th. intermedium and Th. ponticum , Combined ITS1/ Bmy1 sequence data indicates that while the other clusters with St-allele combinations. These Th. scirpeum is closely related to fellow tetraploids data support the established identity of Ps. geniculata Ps. geniculata (0.036) and Th. junceiforme (0.038) and to as an allotetraploid in which the Ee and St genomes are the diploid Ps. spicata (0.026). This is not surprising because combined (L IU and WANG 1992; JENSEN et al. 1995). the sole Bmy1 allele observed in Th. scirpeum , designated Analysis of overall sequence relatedness to Th. caespito- an St-allele, is completely identical to the St-alleles of sum is complicated by the presence of the Bmy1 X-allele, Ps. geniculata (0.000) and nearly identical to the St-alleles which causes infl ated distance values and clusters it on of Th. junceiforme (0.003) and Ps. spicata (0.008). The the ‘ species tree ’ with other species bearing X-alleles: Th. scirpeum St-allele possesses the identical oligonucle- Th. junceum (0.075), Th. intermedium (0.085) and Th. pon- otide insertion and multiple point mutations found in the ticum (0.091), although overall distances are not particu- other St-alleles. This St-allele is also nearly identical larly close between these. A comparison of individual (0.003) to that found in Th. intermedium, but the presence alleles is more illuminating. Th. caespitosum has a single of the other alleles, particularly the X-allele, causes a ITS1 St-allele with closest homology to Ps. spicata combined alignment to throw these two farther apart on (0.021) and Ps. geniculata (0.013). The Bmy1 X-allele the ‘ species tree ’ . The Ee -allele of Th. scirpeum is closest predictably has lower distances values only when com- to the E e-alleles found in Th. elongatum and Th. juncei- pared with other X-alleles such as that found in Th. ponti- forme (both 0.034). These data support Th. scirpeum ’ s cum (0.032), Th. intermedium (0.032) and Th. junceum position in the ‘ species tree ’ , grouped with Ps. spicata and (0.051), and values in excess of 0.252 when compared sharing common ancestry with Ps. geniculata and Th. jun- to any other sequences. The other Bmy1 sequence is an ceiforme, and its clustering with the other genomes bear- Eb -allele that matches closely with Th. bessarabicum ing St-alleles in the ‘ genome tree ’ . (0.014) and the Eb -allele of Th. junceum (0.019) but is Th. junceiforme is most closely related to the diploid divergent from other sequences (values range from 0.090 to ancestors Th. elongatum (0.031) and Ps. spicata (0.038) 0.114). The presence of the X-allele in this species and to fellow tetraploid Ps. geniculata (0.031). The complicates classifi cation, but the ITS1-St allele and the St-allele of Th. junceiforme possesses the oligonucle- strong association of the Bmy1 Eb -allele with Th. bessara- otide motif and other commonalities with other St-allele bicum and other Eb -alleles suggests a role for this ances- sequences, and is particularly similar to the St-allele tor. Thus the data support a view of Th. caespitosum as an of Th. scirpeum (0.003) and Ps. geniculata (0.003). allotetraploid of the Eb and St genomes. The other Bmy1 sequence observed is an E e -allele whose sequence is completely identical to Th. elongatum (0.000). Evolutionary distance between higher polyploids, The ITS allele of Th. junceiforme is an E e -allele most tetraploids and diploids closely related to Th. elongatum (0.021). On the ‘ genome tree ’ , one allele combination of Th. junceiforme is linked Because the higher Thinopyrum polyploids all exhib- closest to Th. elongatum and Ta. caput-medusae , while ited the divergent and enigmatic X-allele, and because the other is linked closely to Th. scirpeum and other they have a more heterogenous combination of alleles, Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 21 evolutionary distance values between sequences of these (0.030) and divergent from all other sequences (0.067 – species are infl ated when compared to lower ploidy spec- 0.094 range, excluding X-alleles). The third Bmy1 allele is imens that lack the X-allele. Overall sequence homology, an X-allele. Th. intermedium exhibited two GBSSI alleles. based on combined ITS1 and Bmy1 allele sequences, The fi rst is the Ta-allele, with the exon 11 frameshift and indicates that Th. junceum is most closely related to the four other oligonucleotide motifs, and exhibited clos- Th. intermedium (0.064) and Th. ponticum (0.072) and est distance to Th. ponticum (0.014), Ta. caput-medusae expectedly distant from species lacking the X-allele. (0.016) and E. repens (0.013) and was divergent from all Th. junceum exhibited two ITS1 alleles: one E e -allele other sequences (0.065– 0.122 range). The second is an with closest homology to Th. elongatum (0.026) and the apparently functional sequence, found in one of ten clones E e -allele of Th. junceiforme (0.013) and one Eb -allele only, with close similarity to Th. ponticum (0.013), roughly with closest homology to Th. bessarabicum (0.030), the equal similarities to Cr. delieana (0.037), Th. elongatum E b -allele of Th. intermedium (0.026), and, curiously, the (0.036) and Th. bessarabicum (0.038), and divergent from St-allele of Ps. geniculata (0.021). Th. junceum exhibited Ps. spicata (0.112). Interestingly, while distance calcula- three Bmy1 alleles: an Eb -allele that is very close to tions indicate that this functional allele is closest to Th. bessarabicum (0.008) and the Eb -allele found in Th. elongatum , thus it was designated an Ee-allele, it Th. caespitosum (0.019), a more divergent E e -allele that shares all four intron polymorphisms with the Ta-alleles, is closest to Th. elongatum (0.061) and Th. junceiforme but not the exon frameshift, perhaps because this is the (0.061), and the X-allele which is similar only to other sole functional GBSSI allele in Th. intermedium . These X-alleles, most closely to that of Th. ponticum (0.034). data strongly indicate a role for Ps. spicata in Th. interme- There is a notable lack of St-alleles at the loci in Th. jun- dium ’ s history, indicates a close relationship with Th. pon- ceum , and overall sequence homology is closest to ticum , and a connection with Th. caespitosum and Th. bessarabicum (0.090) of the diploids and, expectedly Th. junceum due to the common X-allele. due to the X-allele, Th. caespitosum (0.075) from among Based on combined sequence calculations, Th. ponti- the tetraploids. Th. junceum is most divergent from cum is most closely related to Th. intermedium (0.029), Cr. delileana (0.120) and Ta. caput-medusae (0.118), due with the presence of the X-allele causing it to appear diver- to lack of the Ta-allele and K-allele. These data support gent from other species. Th. ponticum has the single ITS1 the established view of Th. junceum as an allohexaploid K-allele, evidenced by homogeneity of sequence chro- combining the E e genomes and Eb genomes, with a strong matograms from direct-sequenced PCR products with two indication of a relationship with Th. caespitosum, and distinct primer sets and by the recovery of 16 clones of this possibly Th. intermedium and Th. ponticum, due to the type only. Th. ponticum has three Bmy1 alleles. The fi rst is shared X-allele (L IU and WANG 1993). an St-allele with close homology with Ps. spicata (0.013) Based on combined sequences, Th. intermedium dem- and very close homology to the St-alleles of tetraploid onstrates, by far, closest distance to Th. ponticum (0.029), Th. scirpeum (0.005), Ps. geniculata (0.005) and Th. jun- with the next closest similarity to Th. junceum (0.064), ceiforme (0.008) and to the St-allele of Th. intermedium although the shared X-allele in these species weighs heav- (0.005), and divergent from the other diploids (0.083– 0.097 ily in this calculation. Th. intermedium has three ITS1 range). The second is an Ee -allele with very close homol- alleles, found in an equal number of clones: a divergent ogy to the Ee -alleles of Th. intermedium (0.005), Ps. genic- St-allele with closest similarity to Ps. spicata (0.071) and ulata (0.003) and Th. junceum (0.007), similarity to Ps. geniculata (0.071), a K-allele with closest distance to Th. elongatum (0.022), Cr. delileana (0.028) and Ta. caput- Cr. delileana (0.043) and a third allele which shows clos- medusae (0.030) and divergent from Th. bessarabicum est identity with both Ps. spicata (0.026) and the E b -allele (0.094) and Ps. spicata (0.080). The third Bmy1 allele is an of Th. junceum (0.026); this was tentatively designated an X-allele. Th. ponticum has the non-functional Ta-allele E b -allele because of the closer overall sequence relation- of GBSSI with close similarity to E. repens (0.005), ship between Th. intermedium and Th. junceum, though it Ta. caput-medusae (0.011) and Th. intermedium (0.014). could be an St-allele. The fi rst of three Bmy1 alleles in The functional GBSSI in Th. ponticum, recovered in two of Th. intermedium are an St-allele that is completely identi- 12 clones, is similar to Th. elongatum (0.030) and to the cal to alleles found in Th. scirpeum (0.000) and P. geniculata functional allele in Th. intermedium (0.013). It is important (0.000), highly homologous to alleles in Th. ponticum to note that the functional Ee -allele of Th. ponticum does (0.005), Th. junceiforme (0.003) and Ps. spicata (0.008) not share any of the fi ve oligonucleotide polymorphisms and divergent from the other diploids (0.076 – 0.090 range). with the Ta-alleles or the Ee -allele of Th. intermedium . The second Bmy1 allele is an Ee -allele that is highly These data suggest strongly suggest a role for Ps. spicata homologous to alleles in Th. ponticum (0.005), Ps. genic- and Th. elongatum in Th. ponticum evolution, as well as ulata (0.003), similar to Th. junceiforme (0.021), Th. elon- Ta. caput-medusae and possibly Cr. delileana, based on the gatum (0.021), Cr. delileana (0.028) and Ta. caput-medusae ubiquity of ITS1 K-allele and the GBSSI Ta-allele. 22 M. Arterburn et al. Hereditas 148 (2011)

Allele-specifi c and real-time PCR copies per genome; barley has a comparable genome size to a Thinopyrum diploid and has an estimated 1880 copies Allele-specifi c primers were developed for the ITS1 of ITS1 (Z HANG et al. 1990). Thinopyrum polyploids K-allele and PCR was performed on control samples to have multiple rDNA loci; thus determining the abun- distinguish the K-allele from other genotypes (Fig. 7). dance of this allele may help trace the origins of these Even the single nucleotide difference between primers loci. Real-time analysis is shown in Fig. 8 and indicates ITSK1 and ITSK2 had a considerable impact on ampli- that Th. ponticum has approximately 6.67 times as many fi cation specifi city in the controls. Subsequently, 76 acces- copies of the K-allele compared to Th. intermedium . This sions of Thinopyrum , Pseudoroegneria , Taeniatherum fi nding is supportive of the proportions of cloned caput-medusae and Elytrigia repens were tested with sequences, suggesting that the K-allele is likely present these primers. Results for the Thinopyrum and Pseudor- in at most two genomes of the hexaploid Th. interme- oegneria samples are shown in Table 3. Only accessions dium, but is ubiquitous in Th. ponticum. However, it of Th. intermedium and Th. ponticum generated PCR must be noted that the precise number of rDNA loci product with these primers, with the exception of six acces- in the decaploid Th. ponticum is uncertain; between sions of Th. elongatum and two accessions of E. repens . seventeen and twenty loci have been identifi ed through Successful K-allele amplifi cation in Th. elongatum fl uorescent in situ hybridization (FISH) (L I and Z HANG accessions PI142012, PI255149, PI205279, PI206622, 2002; B RASILIERO-VIDAL et al. 2003). In their FISH study PI109452 and W98526 prompted chromosome counts with Th. distichum and Th. elongatum , KOSINA and to confi rm their identities; all possess either 42 or 70 H ESLOP-HARRISON (1996) noted differences in numbers chromosomes and likely represent Th. ponticum or of rDNA loci between and within plants. It must be noted Th. intermedium accessions. Both E. repens accessions that amplifi cation and sequencing of the Th. ponticum (PI380624 and PI229920) have the expected 42 chro- ITS1 region was repeated using different combinations mosomes; both represent accessions of subspecies elon- of forward and reverse primers to ensure that a bias in gatiformis and were collected from Iran. Because GBSSI primer annealing was not responsible for the recovery of analysis revealed close homology between Th. interme- clones with exclusively K-allele sequences. dium and accessions of E. repens, it is not entirely surprising that some accessions of E. repens may possess the K-allele and that these two hexaploids might have DISCUSSION an evolutionary connection. Sequencing of ITS1 in six randomly chosen Ta. caput-medusae accessions revealed From analyzing evolutionary distance calculations, based that all have closest homology to the Eb -alleles. on combined sequences of all three loci examined, it Allele-specifi c primers were used in a real-time PCR appears that Ps. spicata is the most distant from the other assay to determine the relative abundance of the K-allele putative diploid ancestors, not a surprising result since it in Th. intermedium and Th. ponticum samples. As part is a member of a different genus from Th. elongatum and of the rDNA cluster, ITS1 is generally present in many Th. bessarabicum. Less expected is the close homology

abPRIMER ITSL PRIMER ITSK2

1 2 3 4 5 6 H2O 1 2 3 4 5 6 H2O

c PRIMER ITSK1

1 2 3 4 5 6 H2O

Fig. 7a-c. Results of PCR using the non-selective forward primer ITSL (a ) and allele speciﬁ c forward primers ITSK2 (b ) and ITSK1 ( c). Lanes 1-3 represent PCR of Th. elongatum , Th. bessarabicum and Ps. strigosa respectively. Lanes 4 and 5 are Th. ponticum and Th. intermedium respectively and Lane 6 is Cr.delileana . Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 23

Table 3. Results of ITS allele-specifi c PCR in Thinopyrum values ranging from 0.032 to 0.045, when excluding the and Pseudoroegneria samples. Chromosome counts of more divergent Ps. spicata. By contrast, the calculated samples which deviated from expectations (noted with* ) distances of other diploids with each other range from are indicated. 0.056 to 0.082 for Th. bessarabicum, 0.082 to 0.091 for Ps. spicata, 0.052 to 0.088 for Cr. delileana and 0.052 to # ACCESSION SPECIES ITS-K? CHROM. 0.066 for Ta. caput-medusae. This might suggest that PI222958 TELO NO Th. elongatum is the least diverged from a common ances- PI255149 TELO YES* 42 tor of these diploids. These data do little to aid the debate PI204383 TELO NO regarding the evolutionary closeness of Th. bessarabicum PI142012 TELO YES* 70 and Th. elongatum, save to indicate that the two are signifi - PI109452 TELO YES* 70 cantly closer in relationship than either is with Ps. spicata, PI206622 TELO YES * 42 while Th. elongatum appears to be more closely related to PI206623 TELO NO Cr. delileana than it is to Th. bessarabicum. * W98526 TELO YES 70 Analysis of both individual and combined ITS1 and Bmy1 PI531750 TELO NO sequences indicates connections between Th. elongatum PI547313 TELO NO PI531711 TBESS NO and Ps. spicata with the three tetraploids Th. scirpeum , PI531712 TBESS NO Th. junceiforme and Ps. geniculata . Total evolutionary W610232 TBESS NO distance calculations ranged from 0.025 to 0.038 between PI449493 PSTRIG NO Ps. spicata and these three, while Th. elongatum showed PI595172 PSTRIG NO values within that range compared with Ps. geniculata PI531752 PSTRIG NO and Th. junceiform but higher with Th. scirpeum (0.059), W614049 PSTRIG NO likely because a single St-allele for Bmy1 was observed in PI531749 TSCIRP NO the latter specimen. Evolutionary distance estimates PI531750 TSCIRP NO between individual alleles also support the connection PI228276 TCAES NO between Th. elongatum , Ps. spicata and the tetraploids PI531731 TJUNF NO Th. junceiforme and Ps. geniculata, with distance values PI502271 PGEN NO for Bmy1 ranging from 0.000 (identical) to 0.021. These PI565009 PGEN NO PI632554 PGEN NO data suggest that both Ps. geniculata and Th. junceiforme e PI227184 TJUN NO are allotetraploids with the E and St genomes. The pres- PI439604 TJUN NO ence of the Bmy1 St-allele in Th. scirpeum, with the char- PI516555 TJUN NO acteristic 8-nt insertion and a very close distance value PI531729 TJUN NO of 0.008 with Ps. spicata, indicates that the St genome PI547327 TJUN NO is present in this tetraploid. However, the ITS1 E e -allele PI578702 TJUN NO data, and a signifi cant body of cytogenetic work, indicates PI255150 TINT YES a role for Th. elongatum in the history of Th. scirpeum PI266118 TINT YES ( LIU and WANG 1993). Thus systematists might consider PI273737 TINT YES the possibility that Th. scirpeum is an allotetraploid with PI316122 TINT YES contributions of the Ee and St genomes. PI318923 TINT YES The Bmy1 E b -allele observed in Th. caespitosum exhib- PI380630 TINT YES PI383370 TINT YES its a close homology (0.014) to Th. bessarabicum , while PI383271 TINT YES the ITS1 data suggests a potential role for Ps. spicata . PI3400066 TPON YES These data support the established view of Th. caespito- PI531737 TPON YES sum as an allotetraploid with a combination of the Eb and PI508561 TPON YES St genomes (L IU and WANG 1992). Since no Bmy1 PI547312 TPON YES St-allele was detected, it is possibly, though entirely hypothetical, that a mutational event such as a sequence between sequences of Th. elongatum, Cr. delileana and inversion or transposition, mutated the St-allele of Ta. caput-medusae ; the two most closely related of the Th. caespitosum into the X-allele. The presence of the diploids we analyzed are Th. elongatum and Cr. delileana unique X-allele, similar to those found in the three higher (0.032 distance value). Th. elongatum is equally close to polyploids (0.031 to 0.051 range), indicates a connection Th. bessarabicum and Ta. caput-medusae, but Th. bessar- between these four species, as exhibited on the ‘ species abicum is less similar to Ta. caput-medusae and Cr. deli- tree’ . This could indicate that Th. caespitosum is the leana than is Th. elongatum . Th. elongatum has closest tetraploid species that participated in polyploidization overall similarity with all the other diploids, with distance events to form these high polyploids; sequence data would 24 M. Arterburn et al. Hereditas 148 (2011)

Th. ponticum Th. intermedium

Chinese Spring

H2O Control

Fig. 8. Real-time PCR results using primers ITSK2 and ITS2. Unlabeled curves represent positive control samples of pure K-allele sequence at 1.5 ng, 0.15 ng, 0.015 ng, 0.0015 ng and 0.00015 ng respectively (from left to right). Although six accessions each of Th. ponticum and Th. intermedium were analyzed, only one of each are shown here for clarity. suggest that a hybridization event with Th. bessarabicum between Th. caespitosum and Th. elongatum or Th. bessar- occurred to form Th. junceum and with Ps. spicata to abicum could have resulted in the hexaploid Th. junceum . form Th. intermedium and Th. ponticum . This hypothesis, It seems unlikely that Th. junceum resulted from genetic however, does not account for the introgression of the isolation and speciation in a population of Th. interme- K-allele and Ta-allele in these latter two species. This dium, although both possess the X-allele, because Th. jun- hypothetical model also assumes that the X-allele is indeed ceum is lacking any St-alleles, the K-allele of ITS1 a mutated St-allele. This idea is supported by the sequence and the Ta-allele of GBSSI, all of which are present in data, because the St-allele has the closest distance (0.252) Th. intermedium . to the X-allele, save for other X-alleles. Based on sequence homology of all three loci exam- The sequence data indicates that Th. junceum has the ined, Th. intermedium is most similar to Th. ponticum strongest connection to Th. bessarabicum , based on the (0.029) and somewhat similar to Th. junceum (0.064). very close distance of the Bmy1 E b-allele (0.008) and also Sequence of the St-allele of Bmy1 in Th.intermedium the presence of the ITS1 Eb -allele. There is also evidence shares close homology to Ps. spicata (0.008) and shares for the presence of an Ee genome in this hexaploid, based the characteristic oligonucleotide motif, a fi nding strongly on alleles at both loci examined. These data support the suggestive of the presence of the St-genome in this current view of this species as an allohexaploid of the hexaploid. The Ee -alleles of Bmy1 and GBSSI, both closely E e and E b genomes (L IU and WANG 1993). However, estab- related to Th. elongatum , suggests that the Ee genome is lishing the origins of the third genome in Th. junceum is also present. These two fi ndings agree with established complicated by the presence of the X-allele. Distance esti- views of this species (L IU and WANG 1993). The origins mates suggest that the X-allele is a mutated St-allele, of the third genome is the most diffi cult question, and although Th. intermedium and Th. ponticum have both an has been the subject of debate around this species, with St-allele and X-allele at the Bmy1 locus, so unless the theories including another Ee genome and the J s genome. specimens examined were segregating for this allele, or The unexpected observation of the Bmy1 X-allele, ITS1 Th. intermedium has multiple contributions of the St K-allele, and GBSSI Ta-allele adds further questions. genome, this may not prove an accurate concept. The The complexity and potential new evolutionary players X-allele sequence is unique enough that it seems unlikely that these sequences represent might help explain the that the event that generated it could happen indepen- challenges that many researchers have faced in reaching a dently in these species, so it is hard to imagine an evolu- defi nitive conclusion on the genome constitution of tionary scenario in which there is no connection between this species, and support the hypotheses of K ISHII et al. Th. junceum and Th. caespitosum, especially when the (2005), who suggest a novel, as yet unconfi rmed, genome E b-alleles in these two lines are so very close to each donor involved in Th. intermedium genome evolution. other (0.019). A hybridization and polyploidization event The homology of these two alleles to Cr. delileana and Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 25

Ta. caput-medusae respectively, and the six precise oligo- (0.082), despite lacking the X-allele. A hypothetical his- nucleotide polymorphisms shared amongst the three, are tory could have seen unreduced gametes of Th. caespito- unlikely to have evolved independently and suggest that sum and Ta. caput-medusae to hybridize to form the one of these species might have been involved in a poly- hexaploid Th. intermedium. Hybridization of unreduced ploidization event that ultimately resulted in Th. interme- gametes of Th. intermedium with Ps. geniculata might dium . Because the Ta-allele has such close homology then have resulted in the decaploid Th. ponticum . This (0.011) to Ta. caput-medusae and because of the report of model is speculative, but is supported by both sequence the K-allele in this species by BLATTNER (2004), Ta. caput- analysis and in part by cytogenetic studies performed to medusae seems the more likely to have been involved than date that implicate an Ee and St genome in Th. ponticum Cr. delileana. The close sequence homology between ( ZHANG et al. 1996a). Th. elongatum , Cr. delileana and Ta. caput-medusae that Real-time PCR results support the hypothesis that we report here indicates a relationship between these three Th. intermedium has the K-allele present in a single and might help explain why cytogenetic evidence has led genome and is present in high copy numbers in Th. ponti- researchers to suspect the presence of a second Ee genome cum. The ITS regions of Th. ponticum are thus unique in in Th. intermedium, if indeed one of these two is instead a that they indicate homogeneity of the rDNA, despite this genome donor (L IU and WANG 1993). It is also possible decaploid species having as many as twenty rDNA loci, as that an as-yet uncharacterized relative of these three was established in the study of L I and Z HANG (2002). These involved. The very close homology of the Th. intermedium researchers suggested that concerted evolution had favored Ta-allele to the hexaploid E. repens (0.013) adds another a single type of rDNA in Th. ponticum . The heterogeneous layer of complexity to the issue. Many described acces- alleles of GBSSI and Bmy1 indicate that Th. ponticum is sions of E. repens are hexaploids and may contain one or not an autodecaploid, yet the evidence in this study cor- more genomes derived from Ta. caput-medusae . The pres- roborates the idea that a single rDNA allele has become ence of the X-allele further complicates analysis, and the ubiquitous in the genomes of this species, verifi ed by PCR presence of both the St-allele and X-allele in Th. interme- with multiple primers to reduce the possibility of primer dium is evidence against the hypothesis that the X-allele is bias. L I and Z HANG conjectured that the rDNA regions of a mutated St-allele, unless these alleles were segregating Th. ponticum had undergone concerted evolution via gene in the group we examined. The data provided from this conversion, similar to a process described in tetraploid investigation is unlikely to be suffi cient to defi nitively peonies and soybeans (S ANG et al. 1995; J OLY et al. 2004). characterize the genomes of Th. intermedium , but they do The rationale for this phenomenon is the conversion of support the existing concept of the Ee and St genomes all rDNA in the polyploid genome to the type that is presence in this hexaploid and provide a new set of spe- preferentially expressed. cies for investigators to consider as potential genome Polyploid genera such as Thinopyrum tend to evade donors for this species. phylogenetic characterization owing to their complexity. Analysis of the Th. ponticum sequence data is beset by Polyploidy-induced genome-wide sequence alterations, the same complications as Th. intermedium, and overall karyotypic changes, multiple gene copies and pseudogene sequence homology indicates a very close relationship formation are considerable obstacles to sequence analysis between these two. Th. ponticum has the same set of and may be sources of some of the unique sequences alleles as Th. intermedium, save for the ubiquity of the observed in this study ( SHAKED et al. 2001). The sequence ITS1 K-allele and the absence of any other ITS1 alleles. data presented in this study has indicated potential new Among the allele sequences that these two share, each participants in the evolution of the Thinopyrum grasses, had closest homology to each other over any other has supported existing views of the genome constitutions sequence, as indicated by the pairing of these two of some species, and has raised further intriguing ques- together in every instance on both the ‘ species tree’ and tions about this group. These data have not conclusively ‘ genome tree ’ . This suggests that Th. intermedium and delineated the evolutionary history of this complex genus, Th. ponticum are close relatives and diverged relatively nor is it likely that any one type of evidence will do so. recently from each other. The most straightforward Indeed this evidence highlights the challenge of assign- explanation is that the hexaploid Th. intermedium ing genome designations, as more circuitous routes than engaged in a hybridization event, involving unreduced single polyploidization events may have shaped the gametes, with a tetraploid species to form the decaploid genomes of many of these species, possibly including Th. ponticum . The isolation of only a single X-allele aneuploidy and chromosomal rearrangement. Hopefully clone from Th. ponticum suggests that this tetraploid was further studies, capitalizing upon a union of cytogenetic not Th. caespitosum. Overall sequence homology would analysis and sequenced-based efforts, and as-yet unex- indicate Ps. geniculata as the most likely candidate, plored methods, can further elucidate the history of this since it is closest to Th. ponticum of all the tetraploids important group. 26 M. Arterburn et al. Hereditas 148 (2011)

Acknowledgements – The authors gratefully acknowledge the interactivity are correlated in eukaryoytic evolution. – Genome partial funding of this work by a USDA-CSREES Special Grant Res. 13: 2229 – 2235. and the Washington State Wheat Commission. Li, D. and Zhang, X. 2002. Physical localization of the 18S-5. 8S-26S rDNA and sequence analysis of ITS regions in Thinopyrum ponticum (Poaceae: Triticeae): implications for REFERENCES concerted evolution. – Ann. Bot. 90: 445 – 452. Liu, Z. and Wang, R. 1989. Genome analysis of Thinopyrum Blattner, F. 2004. Phylogenetic analysis of Hordeum (Poaceae) caespitosum . – Genome 32: 141 – 145. as inferred by nuclear rDNA ITS sequences. – Mol. Liu, Z. and Wang, R. 1992. Genome analysis of Thinopyrum Phylogenet. Evol. 33: 289 – 299. junceiforme and Th. sartorii . – Genome 35: 758 – 764. Brasiliero-Vidal, A. C., Cuadrado, A., Brammer, S. et al. 2003. Liu, Z. and Wang, R. 1993. Genome constitutions of Thinopyrum Chromosome characterization in Thinopyrum ponticum curvifolium , T. scirpeum , T. distichum and T. junceum (Triticeae, Poaceae) using in situ hybridization with different (Triticeae: Gramineae). – Genome 36: 641 – 651. DNA sequences. – Genet. Mol. Biol. 26: 505 – 510. Mason-Gamer, R. 2001. Origin of North American Elymus Chen, Q. 2005. Detection of alien chromatin introgression (Paeceae: Triticeae) allotetraploids based on granule-bound from Thinopyrum into wheat using S genomic DNA as probe starch synthase genes. – Syst. Bot. 26: 757 – 758. – a landmark approach for Thinopyrum genome research. Mason-Gamer, R. 2005. The ß -amylase genes of grasses and a – Cytogenet. – Genome Res. 109: 350 – 359. phylogenetic analysis of the Triticeae (Poaceae). – Am. J. Chen, Q., Connor, R., Ahmad, F. et al. 1998a. Molecular Bot. . 92: 1045 – 1058. characterization of the genome composition of partial Mason-Gamer, R., Weil, C. and Kellogg, E. 1998. Granule- amphiploids derived from Triticum aestivum ϫ Thinopyrum bound starch synthase: structure, function and phylogenetic ponticum and T. aestivum ϫ Th. intermedium as sources of utility. – Mol. Biol. Evol. 15: 1658 – 1673. resistance to wheat streak mosaic virus and its vector, Aceria Muramatsu, M. 1990. Cytogenetics of decaploid Agropyron tosichella . – Theor. Appl. Genet. 97: 1 – 8. elongatum ( Elytrigia elongata ) (2n ϭ 70). I. Frequency of Chen, Q., Conner, R., Laroche, A. et al. 1998b. Genome analysis decavalent formation. – Genome 33: 811 – 817. of Thinopyrum intermedium and Thinopyrum ponticum using Nishikawa, T., Salomon, B., Komatsuda, T. et al. 2002. Molecular genomic in situ hybridization. – Genome 41: 580 – 586. phylogeny of the genus Hordeum using three chloroplast Cox, T. S., Bender, M. and Picone, C. 2002. Breeding perennial DNA sequences. – Genome 45: 1157 – 1166. grain crops. – Critical Rev. Plant Sci. 21: 59 – 91. Refoufi , A., Jahier, J. and Esnault, M. 2001. Genome analysis Desper, R. and Gascuel, O. 2002. Fast and accurate phylogeny of Elytrigia pycnantha and Thinopyrum junceiforme and reconstruction algorithms based on the minimum-evolution of their putative natural hybrid using the GISH technique. principle. – J. Comput. Biol. 9: 687 – 805. – Genome 44: 708 – 715. Dewey, D. R. 1984. The genomic system of classifi cation as a guide Sang, T., Crawford, D. and Stuessy, T. 1995. Documentation of to intergeneric hybridization with the perennial Triticeae. – In: reticulate evolution in peonies (Paeonia ) using internal Gustafson, J. P. (ed.), Gene manipulation in plant improvement. transcribed spacer sequences of nuclear ribosomal DNA: Proc. 16th Stadler Genet. Symp. N. Y., pp. 209 – 279. implications for biogeography and concerted evolution. – Proc. Hsiao, C., Chatterton, N., Asay, K. et al. 1994. Phylogenetic Natl Acad. Sci. 92: 6813 – 6817. relationships of the monogenomic species of the wheat Scheinost, P., Lammer, D. and Cai, X. 2001. Perennial wheat: a tribe, Triticeae (Poaceae), inferred from nuclear rDNA sustainable cropping system for the Pacifi c Northwest. – Am. (internal transcribed spacer) sequences. – Genome 38: J. Alternative Agric. 16: 146 – 150. 211 – 223. Shaked, H., Kashkush, K., Ozkan, H. et al. 2001. Sequence Jauhar, P. P., Dogramaci, M. and Peterson, T. 2004. Synthesis and elimination and cytosine methylation are rapid and cytological characterization of trigeneric hybrids of durum reproducible responses of the genome to wide hybridization wheat with and without Ph1. – Genome 47: 1173 – 1181. and allopolyploidy in wheat. – Plant Cell 13: 1749 – 1759. Jensen, K., Curto, M. and Asay, K. 1995. Cytogenetics of Eurasian Sjakste, T. G. and Zhuk, A. F. 2006. Novel haplotype description bluebunch wheatgrass and their relationship to North American and structural background of the eventual functional bluebunch and thickspike wheatgrasses. – Crop Sci. 35: signifi cance of the barley β -amylase gene intron III segments. 1157 – 1162. – Theor. Appl. Genet. 113: 1063 – 1079. Joly, S., Rauscher, J., Sherman-Broyles, S. et al. 2004. Evolutionary Sjakste, T. G., Tashal, I. and Roder, M. S. 2003. Inheritance of dynamics and preferential expression of homeologous microsatellite alleles in pedigrees of Latvian barley varieties 18S-5.8S-26S nuclear ribosomal genes in natural and artifi cial and related European ancestors. – Theor. Appl. Genet. 106: Glycine allopolyploids. – Mol. Biol. Evol. 21: 1409 – 1421. 539 – 549. Kimura, M. 1980. A simple method for estimating evolutionary Thompson, J. D., Higgins, D. G. and Gibson, T. J. 1994. rates of base substitutions through comparative studies of CLUSTAL W: improving the sensitivity of progressive nucleotide sequences. – J. Mol. Evol. 16: 111 – 120. multiple sequence alignment through sequence weighting, Kishii, M., Wang, R. and Tsujimoto, H. 2005. GISH analysis position specifi c gap penalties and weight matrix choice. revealed new aspect of genomic constitution of Thinopyrum – Nucleic Acids Res. 22: 4673 – 4680. intermedium . – Book of Abstracts, 5th Int. Triticeae Symp. Tippmann, H. F. 2004. Analysis for free: comparing programs Prague, Czech Republic, p. 21. for sequence analysis. – Briefi ngs Bioinf. 5: 82 – 87. Kosina, R. and Heslop-Harrison, J. 1996. Molecular cytogenetics Wang, R. 1989. An assessment of genome analysis based on of an amphiploid trigeneric hybrid between Triticum durum , chromosome pairing in hybrids of perennial Triticeae. – Thinopyrum distichum and Lophopyrum elongatum. – Ann. Genome 32: 179 – 189. Bot. 78: 583 – 589. Wang, R. R. C. 1992. Genome relationships in the perennial Krylov, D., Wolf, Y., Rogozin, I. et al. 2003. Gene loss, protein Triticeae based on diploid hybrids and beyond. – Hereditas sequence divergence, gene dispensability, expression level and 116: 133 – 136. Hereditas 148 (2011) Polymorphic sequences indicate a novel ancestor of Thinopyrum 27

Wang, R., Marburger, J. and Hu, C. 1991. Tissue culture facilitated Zhang, X. Y., Koul, A., Petroski, R. et al. 1996a. Molecular production of aneupolyhaploid Thinopyrum ponticum and veriﬁ cation and characterization of BYDV-resistant germ amphidiploids of Hordeum violaceum X H. bogdanii and their plasms derived from hybrids of wheat with Thinopyrum uses in phylogenetic studies. – Theor. Appl. Genet. 81:151– 156. ponticum and Th. intermedium. – Theor. Appl. Genet. 93: Xia, X. and Xie, Z. 2001. DAMBE: Data analysis in molecular 1033 – 1039. biology and evoluiton. – J. Hered. 92: 371 – 373. Zhang, X. Y., Dong, Y. and Wang, R. 1996b. Characterization of Zhang, Q., Saghai-Maroof, M. and Allard, R. 1990. Effects genomes and chromosomes in partial amphiploids of the on adaptedness of variations in ribosomal DNA copy number hybrid Triticum aestivum ϫ Thinopyrum ponticum by in situ in populations of wild barley (Hordeum vulgare ssp. hybridization, isozyme analysis and RAPD. – Genome 39: spontaneum ). – Proc. Natl Acad. Sci. 87: 8741 – 8745. 1062 – 1071.