CHROMOSOME NUMBERS AND KARYOTYPE EVOLUTION

Transcription

1 American Journal of Botany 91(3): CHROMOSOME NUMBERS AND KARYOTYPE EVOLUTION IN HOLOPARASITIC OROBANCHE (OROBANCHACEAE) AND RELATED GENERA 1 GERALD M. SCHNEEWEISS, 2 TERESA PALOMEQUE, 3 ALISON E. COLWELL, 4 AND HANNA WEISS-SCHNEEWEISS 2,5 2 Department of Higher Plant Systematics and Evolution, Institute of Botany, University of Vienna, Rennweg 14, 1030 Vienna, Austria; 3 Departamento de Biologı a Experimental, A rea de Genética, Universidad de Jaén, Jaén, Spain; and 4 Western Fisheries Research Center, United States Geological Survey, Seattle, Washington, USA Chromosome numbers and karyotypes of species of Orobanche, Cistanche, and Diphelypaea (Orobanchaceae) were investigated, and 108 chromosome counts of 53 taxa, counted for the first time, are presented with a thorough compilation of previously published data. Additionally, karyotypes of representatives of these genera, including Orobanche sects. Orobanche and Trionychon, are reported. Cistanche (x 20) has large meta- to submetacentric chromosomes, while those of Diphelypaea (x ) are medium-sized submetato acrocentrics. Within three analyzed sections of Orobanche, sects. Myzorrhiza (x ) and Trionychon (x ) possess mediumsized submeta- to acrocentrics, while sect. Orobanche (x ) has small, mostly meta- to submetacentric, chromosomes. Polyploidy is unevenly distributed in Orobanche and restricted to a few lineages, e.g., O. sect. Myzorrhiza or Orobanche gracilis and its relatives (sect. Orobanche). The distribution of basic chromosome numbers supports the groups found by molecular phylogenetic analyses: Cistanche has x 20, the Orobanche-group (Orobanche sect. Orobanche, Diphelypaea) has x, and the Phelipanche-group (Orobanche sects. Gymnocaulis, Myzorrhiza, Trionychon) has x,. A model of chromosome number evolution in Orobanche and related genera is presented: from two ancestral base numbers, x h 5 and x h 6, independent polyploidizations led to x 20 (Cistanche) and (after dysploidization) x (Orobanche-group) and to x and x (Phelipanche-group), respectively. Key words: chromosome number evolution; Cistanche; Diphelypaea; karyotype evolution; Orobanche; Phelipanche. Orobanche is the largest genus among the holoparasitic members of Orobanchaceae (sensu Young et al., 99) and comprises approximately 170 species distributed predominantly in the Northern Hemisphere. Following the treatment of Beck-Mannagetta (30), most authors divide Orobanche into four sections. The two larger sects. Orobanche and Trionychon occur in the Old World, while the smaller sects. Gymnocaulis and Myzorrhiza are confined to the New World. Other authors treat these sections as the separate genera Orobanche, Phelipanche, Aphyllon, and Myzorrhiza, respectively (Soják, 72; Holub, 77, 90; Teryokhin et al., 93). Within Orobanche, three basic chromosome numbers are present, x, x, and x. The first is confined to members of sect. Orobanche, the second occurs in sects. Myzorrhiza and Trionychon, and the third in sect. Gymnocaulis. The pronounced karyological split between O. sect. Orobanche and the other sections (x vs. x and x ) indicates that this section might be phylogenetically independent from the remaining sections, thus supporting a narrower generic concept. Recent molecular phylogenetic analyses provide contradictory results. While nuclear ITS-data corroborate this karyological split (Schneeweiss et al., in press), plastid markers suggest a close relationship of O. sect. Trionychon to sect. Orobanche and not to the American sections (Wolfe and depamphilis, 98; Young et al., 99). Chromosome studies, employing both classical and molec- 1 Manuscript received June 2003; revision accepted 7 October We thank J. Fernandez-Casas, W. Gutermann, C.-G. Jang, O. Jimenez, L. Lofquist, J. Fernandez Piqueras, J. Pueche, M. Ruı z Rejón, P. Schönswetter, B. E. Smythies, D. Specht, M. Staudinger, and A. Tribsch for help in collecting plant material. Financial support by the Austrian Science Foundation (FWF; P14352-Bio) is highly appreciated. 5 hanna.weiss@univie.ac.at. 439 ular cytogenetic methods, have proven to be an important source of information for analyzing relationships and evolution of taxa. Hence, this study aims to contribute towards a better understanding of the cytological evolution of Orobanche and closely related genera. We report chromosome numbers for 53 taxa (several from more than one accession), of these counted for the first time, plus karyotypes for representatives of most of the major lineages. The new and previously reported karyological data are synthesized and used together with available molecular phylogenetic information (Schneeweiss et al., in press) to formulate hypotheses concerning the evolution of basic chromosome numbers and the karyotypes within Orobanche and related genera. MATERIALS AND METHODS One hundred accessions of 49 taxa of Orobanche sects. Orobanche, Trionychon, and Myzorrhiza and eight accessions of two taxa each of Cistanche and Diphelypaea were investigated cytologically (Appendix 1; see supplemental data accompanying the online version of this article). From the other genera belonging to the clade including Orobanche (Boschniakia, Conopholis, and Epifagus; Young et al., 99) no material was available. Due to the lack of typical root meristems, chromosome numbers were determined from meiotic divisions in pollen mother cells (PMCs) and first mitoses in developing microspores. Young flower buds were fixed in the field in 3 : 1 ethanol : glacial acetic acid for at least h at room temperature and stored at 20 C until use. Feulgen staining with Schiff s reagent was done following the standard method. Briefly, material was hydrolyzed in 5N HCl (MERCK, Vienna, Austria) for 20 min at room temperature, washed briefly with tap water, and stained with Schiff s reagent (SIGMA, Munich, Germany) in darkness for 1 h. Squash preparations were made in a drop of 45% acetic acid. Preparations with a minimum of 10 good quality chromosome spreads were analyzed for each individual and at least two individuals were scored for each accession. Chromosome spreads were analyzed with a light microscope (Polyvar, Reichert-Jung, Vienna, Austria) and photographed with Technical Pan 100 ASA

2 440 AMERICAN JOURNAL OF BOTANY [Vol. 91 TABLE 1. New chromosome counts for species of Cistanche, Diphelypaea, and Orobanche (an asterisk indicates counts published by Palomeque and Sañudo in Löve [81] without images). For collection details see Appendix 1 (see supplemental data accompanying the online version of this article). Taxon n 2n Figure Cistanche phelypaea (L.) Coutinho subsp. lutea (Desj.) F. Casas & M. Lainz phelypaea (L.) Coutinho subsp. lutea (Desj.) F. Casas & M. Lainz phelypaea (L.) Coutinho subsp. lutea (Desj.) F. Casas & M. Lainz phelypaea (L.) Coutinho subsp. phelypaea phelypaea (L.) Coutinho subsp. phelypaea Diphelypaea coccinea (M. Bieb.) Nicolson coccinea (M. Bieb). Nicolson tournefortii (Desf.) Nicolson Orobanche sect. Myzorrhiza californica Cham. & Schlecht. subsp. californica californica Cham. & Schlecht. subsp. grandis Heckard pinorum Geyer ex Hook. sect. Orobanche alba Steph. ex Willd. alba Steph. ex Willd. alba Steph. ex Willd. alsatica Kirschl. subsp. alsatica alsatica Kirschl. subsp. libanotidis (Rupr.) Pusch amethystea Thuill. subsp. amethystea amethystea Thuill. subsp. amethystea amethystea Thuill. subsp. amethystea* anatolica Reut. anatolica Reut. austrohispanica M. J. Y. Foley caryophyllacea Sm. caryophyllacea Sm. cernua Loefl. var. cernua cernua Loefl. var. cernua cernua Loefl. var. cernua cernua Loefl. var. cumana (Wallr.) G. Beck cernua Loefl. var. cumana (Wallr.) G. Beck* colorata C. Koch crenata Forsk. crenata Forsk. crenata Forsk. densiflora Salzm. ex Reut. densiflora Salzm. ex Reut. elatior Sutt. flava Mart. ex F. Schultz foetida Poir. foetida Poir.* grossheimii Novopokr. grossheimii Novopokr. hederae Duby hederae Duby lutea Baumg. lutea Baumg. macrolepis (Coss.) Coss. ex Nyman macrolepis (Coss.) Coss. ex Nyman macrolepis (Coss.) Coss. ex Nyman macrolepis (Coss.) Coss. ex Nyman macrolepis (Coss.) Coss. ex Nyman minor Sutt. owerinii (G. Beck) G. Beck owerinii (G. Beck) G. Beck owerinii (G. Beck) G. Beck owerinii (G. Beck) G. Beck A, C, E 1B 1D 1F, G 2A, B 2C 3A 3B 3C 3D 3E 3F 3G 3I 3J 3K 3L 57 3M 3N 3O 3P Q 76 3R 3S

3 March 2004] SCHNEEWEISS ET AL. KARYOLOGY OF OROBANCHE AND RELATED GENERA 441 TABLE 1. Continued. Taxon n 2n Figure owerinii (G. Beck) G. Beck picridis F. Schultz cf. raddeana G. Beck raddeana G. Beck rapum-genistae Thuill. subsp. benthamii (Timb.-Lagr.) P. Fourn. rapum-genistae Thuill. subsp. rapum-genistae rapum-genistae Thuill. subsp. rapum-genistae rapum-genistae Thuill. subsp. rapum-genistae rapum-genistae Thuill. subsp. rapum-genistae teucrii Holandre sect. Trionychon cf. aegyptiaca Pers. cf. aegyptiaca Pers. arenaria Borkh. bungeana G. Beck bungeana G. Beck cf. coelestis (Reut.) G. Beck coelestis (Reut.) G. Beck coelestis (Reut.) G. Beck cf. graciosa G. Beck lavandulacea Rchb. lavandulacea Rchb. nana (Noe ex Reut.) Nyman nana (Noe ex Reut.) Nyman nana (Noe ex Reut.) Nyman nana (Noe ex Reut.) Nyman* nowackiana Markgr. olbiensis Nym. pulchella (C. A. Mey.) Novopokr. purpurea Jacq. s.l.* purpurea Jacq. subsp. purpurea ramosa L. rosmarina G. Beck tunetana G. Beck tunetana G. Beck 3T 3U 3V 3W 2D 2E, F, G 2H 2J 2K 2L 2I 2M 2N 2O 2P 2Q 2R 2S 2T 2U film (Kodak, Vienna, Austria). For measurements of chromosome length, metaphase of first mitosis in microspores was chosen to allow comparison of chromosome length between genera. In order to allow a sounder evaluation of chromosome number distribution, previously published data were included as exhaustively as possible. Because chromosome numbers in Orobanche and other analyzed genera are very likely of palaeopolyploid origin (see Discussion), we differentiate terminologically between the basic chromosome number x, which refers to the lowest haploid chromosome number known in each lineage, and the hypothetical basic chromosome number x h, which indicates the putative ancient basic chromosome number not found in any current species. RESULTS Chromosome numbers The newly obtained chromosome numbers (Table 1) are presented together with previously reported ones (Appendix 2; see supplemental data accompanying the online version of this article). Apart from a few exceptions discussed, our data and those from literature agree in the inferred distribution of chromosome numbers among taxonomic groups. Cistanche (Fig. 1A E) All accessions investigated were 2n 40. Aneuploid numbers, if present, are restricted to single microspores (Fig. 1C). Pazy (98) reports n 30 for one individual from one population of Cistanche salsa in Israel. Kadry (52) reports exclusively the aneuploid chromosome number of n 21 for all analyzed individuals of Cistanche phelypaea. Meiotic irregularities resulting in uneven distribution of chromosomes to the microspores (see Fig. 1C) or asynaptic meiosis (Pazy and Plitman, 96; Pazy, 98) are known for closely related Cistanche taxa and might also be

4 442 AMERICAN JOURNAL OF BOTANY [Vol. 91 Fig. 1. Chromosomes and karyotypes of Cistanche (A E) and Diphelypaea (F G): (A) C. phelypaea subsp. lutea, n 20 (MitI); (B) C. phelypaea subsp. lutea, 2n 40 (diakinesis); (C) C. phelypaea subsp. lutea, n 21 (MitI); (D) C. phelypaea subsp. phelypaea, 2n 40 (AI/TI; only one pole of PMC shown); (E) karyotype of C. phelypaea subsp. lutea n 20 based on (A); (F) D. coccinea, n (MitI; the same cell in two focal plains); (G) karyotype of Diphelypaea coccinea based on (F). Arrows indicate the bivalents associated with the nucleolus and thus carrying NORs; arrowheads indicate overlapping chromosomes. Bar 10 m. Abbreviations: AI, anaphase I; MI, metaphase I; MitI, metaphase of first mitotic division in microspore; NOR, nucleolus organizer region; PMC, pollen mother cell; TI, telophase I. present in the population investigated by Kadry (52), thus accounting for the deviant number. Diphelypaea (Fig. 1F, G) Members of this genus have been counted here for the first time. Both species investigated, D. coccinea and D. tournefortii, have 2n, a number also found in many members of Orobanche sect. Orobanche (see next; Appendix 2). Orobanche (Figs. 2, 3) Within Orobanche, two groups can be distinguished. The first is characterized by chromosome numbers based on x (sects. Myzorrhiza and Trionychon; Fig. 2) or x (sect. Gymnocaulis). A few deviant chromosome numbers have been reported for species belonging to this group. Löve (54) indicated 2n for O. ludoviciana, which has otherwise been reported to have 2n 48. Srivastava (39) counted 2n for O. aegyptiaca, which normally has 2n (Appendix 2). In both cases, the deviant numbers are likely due to species misidentification. Because no vouchers are cited in either publication, their identification cannot be checked. Pillai (77) reports 2n 44 for O. aegyptiaca under its synonym O. indica Hain-Buch. (sic). Assuming that the determination of the investigated plants is correct, this would be the first and only report of a polyploid and, at the same time, aneuploid number in sect. Trionychon. However, this unexpected number within Orobanche could be due to confusion of Orobanche aegyptica with a Striga species. This may be the case for the following reasons: Orobanche indica, albeit with different nomenclatural authority, is also a synonym of Striga gesnerioides; the description of the investigated plant given by Pillai as having bluish-pink flowers agrees with Striga rather than with Orobanche; additionally, 2n 44 fits into the range of chromosome numbers reported for Striga (Iwo et al., 93; Aigbokhan et al., 98). Because no voucher specimen is indicated by Pillai, this explanation remains hypothetical. Jensen (51) reports 2n 36 for the sexual form and 2n ca. 70 for the parthenogenetic form of O. uniflora of sect. Gymnocaulis. Heckard and Chuang (75), however, found only 2n 48 and suspected that in the study of Jensen (51) some of the smaller chromosomes were undetected, resulting in the deviant numbers. The second group includes Orobanche sect. Orobanche (Fig. 3) and is characterized by chromosome numbers based on x. A single deviant count of 2n for O. cumana (Žukov, 39) is probably due to confusion with a species from O. sect. Trionychon. Polyploidy Polyploidy is unevenly distributed among species of Orobanche and related genera. Cistanche (x 20), Diphelypaea (x ), Orobanche sects. Gymnocaulis (x Fig. 2. Chromosomes and karyotypes of Orobanche sect. Myzorrhiza (A C) and sect. Trionychon (D U): (A) californica subsp. grandis, 2n 48 (TI); (B) californica subsp. grandis, 2n 48 (diakinesis); (C) pinorum, 2n 48 (diakinesis); (D) cf. aegyptiaca, n (MitI); (E) arenaria, 2n (diakinesis); (F) arenaria (pachytene with DNA fragmentation; see text for details); (G) arenaria, 2n (AI), arrowheads indicate unseparated bivalents; (H) bungeana, 2n (AI); (I) coelestis, 2n (diakinesis); (J) lavandulacea, 2n (diakinesis); (K) mutelii, n (MitI); (L) mutelii, n (MitI); (M) mutelii, 2n (diakinesis); (N) nana, 2n (diakinesis); (O) nana, n (MitI); (P) nowackiana, n (MitI); (Q) olbiensis, 2n (AI/TI); (R) pulchella, 2n (diakinesis); (S) purpurea, n (MitI); (T) rosmarina, 2n (MI/AI); (U) tunetana, 2n (AI). Arrows indicate the bivalents associated with nucleolus and thus carrying NORs. Bar 10 m. Abbreviations as in Fig. 1.

5 March 2004] SCHNEEWEISS ET AL. KARYOLOGY OF OROBANCHE AND RELATED GENERA 443

6 444 AMERICAN JOURNAL OF BOTANY [Vol. 91 ) and Trionychon (x ) are devoid of polyploids or, if such are present, they are very rare and their presence coincides with meiotic irregularities of one or all analyzed individuals of single populations, as seen in Cistanche spp. (Pazy, 98) and Orobanche uniflora (Jensen, 51). Orobanche sect. Myzorrhiza (x ) consists nearly exclusively of polyploids (Appendix 2). In Orobanche sect. Orobanche (x ), polyploidy is restricted to three lineages (Fig. 4). The first is the tetraploid O. macrolepis, a phylogenetically distinct species of the Iberian Peninsula and adjacent Africa (Schneeweiss et al., in press). The second comprises species with shiny red coloration on at least the interior of the corolla (O. austrohispanica, O. foetida, and O. gracilis of grex Cruentae, Beck-Mannagetta, 1890, 30) and the phylogenetically close O. densiflora. Apart from one diploid accession of O. gracilis reported from Slovakia (Uhrı ková et al. in Löve, 83), this clade comprises exclusively tetra- and hexaploids (Appendix 2). The third isolated occurrence of polyploidy is found in O. transcaucasica. This species belongs to the O. minor aggregate, which otherwise comprises diploid taxa such as O. minor and O. crenata (Appendix 2). Only a few species are known to exhibit more than one ploidy level. The broadest range of ploidy levels is exhibited by O. cooperi (sect. Myzorrhiza; x ), where di-, tetra-, hexa-, and octoploids have been reported (Appendix 2). Di-, tetra-, and hexaploids are also known from O. gracilis (sect. Orobanche; x ), while in O. transcaucasica (sect. Orobanche; x ) di- and tetraploids are present (Appendix 2). Of those, the cytotype mixture (that is, the presence of more than one cytotype within a population) is known from O. cooperi (tetra-, hexa-, and octoploids) and O. gracilis (tetraand hexaploids). Karyotypes and chromosome pairing during meiosis Cistanche The chromosomes (2n 2x 40) are relatively large (5 10 m; Fig. 1A C), and are all meta- to submetacentric. During meiosis, all chromosome pairs form regular bivalents, each usually possessing two chiasmata (Fig. 1B). In diakinesis, usually two bivalents are associated with the nucleolus, suggesting the presence of at least two pairs of active 45S-rDNA loci (Fig. 1B; data not shown). No satellites have been detected in mitotic chromosomes in developing microspores. No meiotic irregularities such as laggards or bridges have been found (Fig. 1D), and only in one microspore of more than 30 analyzed in a single anther a deviating chromosome number (n 21; Fig. 1C) has been observed. Diphelypaea The chromosomes (2n 2x ) are medium sized (2.5 5 m; Fig. 1F, G). Because of the restricted amount of material available, only microspores undergoing first mitotic division were analyzed for each accession. Most chromosomes are submetacentric and acrocentric, with some telocentrics (1 3) also present. At least one chromosome (number 3) in a microspore was observed to possess a satellite (Fig. 1G). Orobanche Species of sect. Myzorrhiza analyzed in the present study are tetraploids (2n 4x 48). Their chromosomes are small (2 3 m; Fig. 2A C) and submetacentric to acrocentric. During meiosis all homologous chromosome pairs form regular bivalents with one or two chiasmata each (Fig. 2B, C). In both species analyzed, two bivalents are associated with the nucleolus in diakinesis suggesting the presence of at least two pairs of active 45S-rDNA loci (Fig. 2B). No meiotic irregularities have been found. However, in early prophase I (leptotene to pachytene) DNA-fragmentation has been observed (not shown). Species of Orobanche sect. Orobanche are diploids (2n 2x ; Fig. 3A D, F I, K, O P, S U), tetraploids (2n 4x 76; Fig. 3E, J, L, N, Q R, V W), or hexaploids (2n 6x 114; Fig. 3M). The chromosomes are small (1 3 m), differing slightly in length from species to species, e.g., 1 2 m in O. gracilis (6x and 4x; Fig. 3M N) or m ino. alsatica subsp. libanotidis (Fig. 3B). Most of the chromosomes are meta- and submetacentric, and 5 7 acrocentric pairs are also present in some species (e.g., O. cernua, Fig. 3H). During meiosis, all homologous chromosome pairs form regular bivalents that usually possess two chiasmata (Fig. 3O, R). The pairing in meiosis is regular, regardless of the ploidy level of the plant as seen in diploid O. grossheimii (Fig. 3O), tetraploid O. macrolepis (Fig. 3R), and hexaploid O. gracilis (not shown). During diakinesis in diploid species, one bivalent is associated with the nucleolus suggesting the presence of at least one pair of active 45S-rDNA loci (Fig. 3O; data not shown). In tetraploids, two bivalents are associated with the nucleolus (Fig. 3R). Only a few cells ( 1%) in individual anthers show meiotic irregularities such as univalents or uneven distribution of chromosomes to the cell poles in anaphase I. As in sect. Myzorrhiza, DNA fragmentation has been observed during early prophase I (not shown). Species of Orobanche sect. Trionychon are diploids (2n 2x ; Fig. 2D U). Their chromosomes are of medium size (3 4 m) and submetacentric to acrocentric (Fig. 2D, K). During meiosis, all homologous chromosome pairs form regular bivalents with one, two, or three chiasmata (e.g., Fig. 2H J). In most species during diakinesis, one bivalent is associated with the nucleolus suggesting the presence of at least one pair of active 45S-rDNA loci (Fig. 2E, M, N). During metaphase in the first mitotic division, one satellite is visible (Fig. 2D). As in the former group, the number of cells with irregularities in meiosis, such as bridges and laggards (Fig. 2G) or uneven distribution of chromosomes to the cell poles in anaphase I (not shown), is very low. As with members of sects. Myzorrhiza and Orobanche, DNA fragmentation was observed during early prophase I (Fig. 2F). Fig. 3. Chromosomes and karyotypes of Orobanche sect. Orobanche: (A) alba, n (MitI); (B) alsatica subsp. libanotidis, n (MitI); (C) amethystea subsp. amethystea, 2n (AI); (D) anatolica, 2n (MI); (E) austrohispanica, 2n 76 (TI); (F) cernua var. cernua, 2n (MI); (G) cernua var. cumana, 2n (MI); (H) cernua, 2n (karyotype of somatic cell); (I) colorata, n (MitI); (J) densiflora, 2n 76 (AI); (K) elatior, 2n (MI/ AI); (L) foetida, 2n 76 (MI/AI); (M) gracilis, n 57 (MitI); (N) gracilis, n (MitI); (O) grossheimii, 2n (diakinesis); (P) hederae, 2n (AI/ TI); (Q) macrolepis, 2n 76 (TI); (R) macrolepis, 2n 76 (diakinesis); (S) owerinii, 2n (AI/TI); (T) rapum-genistae, n (MitI); (U) transcaucasica, 2n (AI/TI); (V) transcaucasica, 2n 76 (AI/TI); (W) transcaucasica, 2n 76 (AI/TI; only one pole of the PMC). Arrows indicate the bivalents associated with nucleolus and thus carrying NORs. Bar 5 m. Abbreviations as in Fig. 1.

7 March 2004] SCHNEEWEISS ET AL. KARYOLOGY OF OROBANCHE AND RELATED GENERA 445

8 446 AMERICAN JOURNAL OF BOTANY [Vol. 91 Fig. 5. Phylogenetic relationships of Boschniakia, Cistanche, Conopholis, Diphelypaea, Epifagus, and Orobanche inferred from nuclear ITS sequences (Schneeweiss et al., in press). The basic chromosome numbers of the phylogenetic lineages are indicated (missing information on chromosome numbers is marked with a question mark). Fig. 4. Phylogenetic relationships within Orobanche sect. Orobanche inferred from nuclear ITS sequences (Schneeweiss et al., in press). Polyploid taxa are indicated by bold branches, and taxa and supporting branches with lacking information on the chromosome number are shaded in gray. DISCUSSION Distribution of basic chromosome numbers and karyotypes The distribution of basic chromosome numbers correlates well with taxonomic groups (Fig. 5). The Old World Cistanche and the small American genus Conopholis have the basic chromosome number x 20 (Appendix 2). The chromosomes of Cistanche are large meta- to submetacentrics. No information is available on morphology of the chromosomes of Conopholis. Within Orobanche, two cytological groups congruent with two clades inferred from phylogenetic analyses of nuclear ITS-sequences (Fig. 5; Schneeweiss et al., in press) are found: (1) the Phelipanche-group, containing sects. Gymnocaulis, Myzorrhiza, and Trionychon, characterized by basic chromosome numbers of x (n,, 36, 48) and, very likely derived from that, x (n, 48), and small to medium-sized submetacentric to acrocentric chromosomes; (2) the Orobanche-group, containing sect. Orobanche, with basic chromosome number of x (n,, 57) and small mostly meta- and submetacentric and few acrocentric chromosomes. Diphelypaea, a small West Asian genus with morphologically unique large, single, terminal, bright red flowers, has the same basic chromosome number x as Orobanche sect. Orobanche, but larger chromosomes albeit of similar morphology (Fig. 1F, G). This is in good congruence with recent molecular phylogenetic analyses, which have shown that Diphelypaea belongs to the Orobanche group (Fig. 5). Evolution of basic chromosome numbers and karyotype relationships Few hypotheses have been proposed concerning the evolution of the basic chromosome numbers found today in Orobanche and related genera, and all agree that palaeopolyploidization played a major role in their evolution. Gardé (52) suggested that all three base numbers found in Orobanche and Cistanche, x,, and 20, respectively, arose from an ancestral number of x h 6. Polyploidization led to x 2x h in O. sect. Trionychon (Fig. 6IIb), while polyploidization and subsequent aneuploidy led to x 3x h 1 in O. sect. Orobanche and similarly to x 3x h 2 20 in Cistanche (Fig. 6V). Because of the absence of any extant species of Orobanche having 2n 2x h, Heckard and Chuang (75) proposed x 2x h as found in O. sect. Trionychon for the basic chromosome number within Orobanche, from which 2x 4x h as found in sects. Gymnocaulis and Myzorrhiza was derived (Fig. 6III). This idea was later corroborated by the detection of diploid O. cooperi (2n 4x h ; Reveal and Moran, 77). Heckard and Chuang (75) argued also that the number of n 36 reported for O. uniflora (Jensen, 51) is rather the result of misinterpretation than a true hexaploid based on x 2x h. They do not propose any scenario for the relationships between the two basic chromosome numbers, x and x. Cubero (96) considered x to be too high for an ancestral basic chromosome number and again suggested x h 6 as the original basic number. He argued that x arose via polyploidization (Fig. 6IIb), thus becoming the lowest basic chromosome number currently known in Orobanche and related genera. Further polyploidizations of x 2x h led either to 2x 4x h in O. sects. Gymnocaulis and Myzorrhiza (Fig. 6III) or resulted in a series of dysploid basic chromosome numbers deviating from 2x, such as x 4x h 5 in O. sect. Orobanche, x 4x h 4 20 in Cistanche and Conopholis, x 4x h 3 21 in some Cistanche populations, and 3x 6x h 36 in some populations of O. uniflora (Fig. 6V). We present a new model, which minimizes the number of dysploidization steps at the expense of polyploidization steps, which seems to be more likely in Orobanche and related genera for three reasons: (1) the basic chromosome numbers (x) found today are very stable within groups; (2) no linking dysploid numbers are known, indicating that dysploidy is a rather rare evolutionary event in this group; and (3) polyploidy occurs independently in several lineages within Orobanche and thus seems to be a relatively frequent evolutionary event.

9 March 2004] SCHNEEWEISS ET AL. KARYOLOGY OF OROBANCHE AND RELATED GENERA 447 In the common ancestors of the clade comprising Cistanche, Conopholis, Diphelypaea, and Orobanche (Young et al., 99), two basic chromosome numbers were present, x h 5 and x h 6. They could be connected via dysploidy (Fig. 6I), but it is impossible to decide which number was ancestral and which was derived. None of these putative ancestral basic chromosome numbers is known from any extant member of this group. Polyploidization of x h 6 led to x 2x h (sects. Myzorrhiza and Trionychon; Fig. 6IIb) and further to x 4x h (sect. Gymnocaulis; Fig. 6III), which numbers are synapomorphic characters of the Phelipanche group (Fig. 5). Similarly, polyploidization of x h 5, led to x h 2x h 10 (not found in any present species; Fig. 6IIa) and subsequently to x 4x h 20 (Fig. 6IV) as found today in Cistanche and Conopholis. From this number, a descending dysploid change resulted in x 4x h 1 (Fig. 6VI), a synapomorphic character of the Orobanche group (Fig. 5). Because there is no evidence that Cistanche is the direct ancestor of the Orobanche group (Young et al., 99; Fig. 5), it is possible that from x h 5orx h 10 the basic chromosome number x 20 evolved independently in Cistanche and the ancestor of the Orobanche group. This latter hypothesis is supported by chromosome morphology, which differs substantially between Cistanche (large mostly metacentric chromosomes) and the Orobanche group (small to medium sized submeta- and acrocentric chromosomes). Alternative scenarios, like the derivation of x from x h 10 or vice versa (Fig. 6VII), would require more dysploidization events and are therefore less likely under the assumptions explained earlier. After polyploidization and following speciation, karyotypes of different groups may evolve in different modes. Polyploidization can be accompanied by the loss of DNA and thus a decrease of chromosome and genome size (Raina et al., 94), although this was shown mostly for allopolyploids (Eckhardt, 2001; Ozkan et al., 2001). This mode might be present in the Phelipanche group, where the diploid species of sect. Trionychon have considerably bigger chromosomes than the polyploid ones of sect. Myzorrhiza, and in the Orobanche group, where the polyploid O. austrohispanica and gracilis are among those species with the smallest chromosomes. On the contrary, polyploidization and speciation can also be associated with an increase of genome and chromosome size due to, for instance, differential amplification of species-specific repetitive sequences or retrotransposons (Kumar and Bennetzen, 99; Staginnus et al., 99). The large chromosomes of Cistanche might be an example of that mode of karyotype evolution. Fig. 6. Scheme of possible pathways of basic chromosome number evolution in Orobanche and related genera. Black numbers are chromosome numbers actually found in present members of this group, gray numbers are hypothesized ancestral numbers; bold arrows indicate polyploidization, thin arrows dysploidization events. For details see Discussion. Intraspecific ploidy level polymorphism The presence of more than one ploidy level is known only from three species of Orobanche, but only in O. transcaucasica a correlation of cytotype distribution with the presence of biogeographically and ecologically defined races might be present. The diploid cytotype was found in northeast Turkey growing on Rhus coriaria (Anacardiaceae) in xerophytic shrublands, while the tetraploid cytotype occurs in Georgia in mesophytic forests growing probably on Carpinus betulus (syn. C. caucasica, Betulaceae; G. M. Schneeweiss, personal observation). More data are needed for a sounder evaluation of cytotype distribution and possible correlations with the presence of different races. Meiotic irregularities and aneuploidy For two more intensively studied Orobanche species (O. crenata, O. gracilis), a relatively high frequency (up to 90%) of meiotic irregularities and subsequent aneuploidy has been reported (Greilhuber and Weber, 75; Moreno et al., 79; Palomeque, 79). These include uneven distribution of chromosomes to the microspores, laggards, or the formation of uni- and multivalents. Pazy and Plitman (96) and Pazy (98) report a frequency of 70% of univalent formation from asynapsis during first meiotic division in a population of Cistanche tubulosa. Most of the species investigated in our survey have regular meiosis. Irregularities, if present, are confined to single flowers or even anthers. This is also the case in the polyploid taxa investigated (e.g., Fig. 3E), where a higher frequency of meiotic irregularities might be expected due to their likely autopolyploid origin. Meiotic irregularities seem to depend strongly on the condition of the individual plant. We found in one individual of Orobanche arenaria, from which flower buds from an inflorescence in a late stage of flowering were used, a high number of meiotic irregularities (Fig. 2F, G) leading to malformation of pollen (data not shown), while a second individual from the same population had normal meiosis (Fig. 2E) and pollen development. Frequently, the distal flower buds in an inflorescence do not develop fully (G. M. Schneeweiss, personal observation), which might be due to internal factors, such as the age of flower buds, or may be stimulated by external factors, such as environmental stress or host condition. Further studies are currently being carried out to investigate if the DNA fragmentation during prophase I in PMCs in species from all sections of Orobanche analyzed is due to programmed apoptosis. Conclusions Investigations of karyological features of Orobanche and related genera contribute significantly to a better understanding of systematics and evolution of this fascinating group of holoparasitic plants. Distribution of basic chromosome numbers and chromosome morphology confirm the presence of three phylogenetically independent lineages, the

10 448 AMERICAN JOURNAL OF BOTANY [Vol. 91 Phelipanche group (x, ; O. sects. Gymnocaulis, Myzorrhiza, and Trionychon), the Orobanche group (x ; Diphelypaea, O. sect. Orobanche), and the genus Cistanche (x 20). This corroborates results from molecular phylogenetic analyses that Orobanche in its most frequently used circumscription is not monophyletic and eventually will have to be disintegrated. Polyploidization played an important role in the evolution of all three groups, because all current basic chromosome numbers are likely of palaeopolyploid origin. The pronounced difference in chromosome size in Cistanche and the Orobanche group indicates that polyploidization, even if based on the same hypothetical basic chromosome number, might occur independently. LITERATURE CITED AIGBOKHAN, E. I., D. K. BERNER, AND L. J. MUSSELMAN. 98. Reproductive ability of hybrids of Striga aspera and Striga hermonthica. Phytopathology 88: BECK-MANNAGETTA, G Monographie der Gattung Orobanche. Theodor Fischer, Cassel, Germany. BECK-MANNAGETTA, G. 30. IV Orobanchaceae. In A. Engler [ed.], Das Pflanzenreich. Regni Vegetabilis Conspectus. Wilhelm Engelmann, Leipzig, Germany. CUBERO, J. I. 96. Cytogenetics in Orobanchaceae: a review. In M. T. Moreno, J. I. Cubero, D. Berner, D. Joel, L. J. Musselman, and C. Parker [eds.], Advances in parasitic plant research, Junta de Andalucía, Consejerı a de Agricultura y Pesca, Sevilla, Spain. ECKHARDT, N A sense of self: the role of DNA-sequence elimination in allopolyploidization. The Plant Cell 13: GARDÉ, A. 52. Nota cariologica sobre tres Orobancáceas portuguesas. Genetica Iberica 3: GREILHUBER, J., AND A. WEBER. 75. Aneusomaty in Orobanche gracilis. Plant Systematics and Evolution 1: HECKARD, L. R., AND T. I. CHUANG. 75. Chromosome numbers and polyploidy in Orobanche (Orobanchaceae). Brittonia 27: HOLUB, J. 77. New names in Phanerogamae 6. Folia Geobotanica et Phytotaxonomica : HOLUB, J. 90. Some taxonomic and nomenclatural changes within Orobanche s. l. (Orobanchaceae). Preslia 62: 3 8. IWO, G. A., S. W. H. HUSAINI, AND G. O. OLANIYAN. 93. Cytological observations and distribution of Striga species in central part of Nigeria. Feddes Repertorium 104: JENSEN, H. W. 51. The normal and parthenogenetic forms of Orobanche uniflora L. La Cellule 54: KADRY, A. E. R. 52. The development of microsporangium and pollen grains in Cistanche tinctoria (Forssk.) G. Beck. Botaniska Notiser 52: KUMAR, A., AND J. L. BENNETZEN. 99. Plant retrotransposons. Annual Reviews in Genetics 33: LÖVE, Á. 54. Cytotaxonomical evaluation of corresponding taxa. Vegetatio 5 6: 2 2. LÖVE, Á. 83. IOPB chromosome number reports LXXX. Taxon 32: MORENO, M. T., J. I. CUBERO, AND A. MARTÍN. 79. Meiotic behavior in Orobanche crenata. In L. J. Musselman, A. D. Worsham, and R. D. Eplee [eds.], Proceedings on the second international symposium on parastic weeds, North Carolina State University, Raleigh, North Carolina, USA. OZKAN, H., A. A. LEVY, AND M. FELDMAN Allopolyploidy-induced rapid genome evolution in the wheat (Aegilops Triticum) group. The Plant Cell 13: PALOMEQUE, T. 79. Estudios cariologicos en especies españolas de los generos Cistanche Hoffmanns. & Link. y Orobanche L. (Orobanchaceae). Ph.D. thesis, Universidad de Granada, Granada, Spain. PAZY, B. 98. Diploidization failure and apomixis in Orobanchaceae. Botanical Journal of the Linnean Society 8: PAZY, B., AND U. PLITMAN. 96. Asynapsis in Cistanche tubulosa (Orobanchaceae). Plant Systematics and Evolution 201: PILLAI, R. V. R. 77. Chromosome number of Orobanche indica Hain-Buch. Science and Culture 43: RAINA, S. N., A. PARIDA, K. K. KOUL, S. S. SALIMATH, M. S. BISHT, V. RAJA, AND T. N. KHOSHOO. 94. Associated chromosomal DNA changes in polyploids. Genome 37: REVEAL, J. L., AND R. MORAN. 77. Miscellaneous chromosome counts of western American plants IV. Madroño : SCHNEEWEISS, G. M., A. COLWELL, J.-M. PARK, C.-G. JANG, AND T. F. STUES- SY. In press. Phylogeny of holoparasitic Orobanche (Orobanchaceae) inferred from nuclear ITS-sequences. Molecular Phylogenetics and Evolution. SOJÁK, J. 72. Nomenklatorické poznámky (Phanerogamae). Časopis Národniho Muzea, odd. přírodovědny 140: SRIVASTAVA, G. D. 39. Contribution to the morphology of Orobanche aegyptiaca Pers. Proceedings of the National Academy of Sciences, India 9: STAGINNUS, C., P. WINTER, C.DESEL, T.SCHMIDT, AND G. KAHL. 99. Molecular structure and chromosomal localization of major repetitive DNA families in the chickpea (Cicer arietinum L.) genome. Plant Molecular Biology 39: TERYOKHIN, E. S., G. V. SHIBAKINA, N.B.SERAFIMOVICH, AND T. I. KRAV- TSOVA. 93. Opredelitelj Sarasichovich Florii SSSR (Determinator of broomrapes of the USSR flora). Nauka, Leningrad, Russia. WOLFE, A. D., AND C. W. DEPAMPHILIS. 98. The effect of relaxed functional constraints on the photosynthetic gene rbcl in photosynthetic and nonphotosynthetic parasitic plants. Molecular Biology and Evolution 15: YOUNG, N. D., K. E. STEINER, AND C. W. DEPAMPHILIS. 99. The evolution of parasitism in Scrophulariaceae/Orobanchaceae: plastid gene sequences refute an evolutionary transition series. Annals of the Missouri Botanical Garden 86: ŽUKOV, N. I. 39. Citologičeskoe issledovanie redukcionnogo delenija nekotoryh vidov zarazih. Sb. rabot po selekcii, genetike i semenovedeniju tabaka i mahorki, Vsesojuzn. N.-I. Inst. Tabačnoj i Mahoroc ˇnoj Promys ˇ- lennosti, Krasnodar 139: