Bse Chemtax

Biochemical Systematics and Ecology 33 (2005) 1087e1106 www.elsevier.com/locate/biochemsyseco

Veronica: Chemical characters for the support of phylogenetic relationships based on nuclear ribosomal and plastid DNA sequence data Dirk C. Albach a,1,*, Søren R. Jensen b, Fevzi O¨zgo¨kce c, Rene´e J. Grayer d a

Department of Higher Plant Systematics, Institute of Botany, University of Vienna, Rennweg 14, A-1030 Vienna, Austria b Department of Chemistry, The Technical University of Denmark, DK-2800 Lyngby, Denmark c Department of Biology, Faculty of Science and Arts, Yu¨zu¨ncu¨ Yil University, 65080 Van, Turkey d Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AB, UK Received 15 December 2004; accepted 6 June 2005

Abstract Molecular phylogenetic analyses have revealed many relationships in Veronica (Plantaginaceae) never anticipated before. However, phytochemical characters show good congruence with DNA-based analyses. We have analysed a combined data set of 49 species and subspecies derived from the nuclear ribosomal ITS-region (18 new sequences) and the plastid trnL-F region (12 new sequences) of Veronica emphasizing subgenera Chamaedrys and Pocilla and separate analyses of subgenera Pentasepalae (ITS only) and Pseudolysimachium. Results for subgenus Chamaedrys show that European and Asian perennial species are monophyletic sister groups with the annual species consecutive sisters to them. All species of Veronica that contain cornoside are found in this subgenus, although some species seem to have secondarily lost the ability to produce this compound. Subgenera Pocilla and Pentasepalae are well supported sister groups characterized by the occurrence of 8-hydroxyflavones. The traditional subsection Biloba of subgenus Pocilla is biphyletic with Veronica intercedens being clearly * Corresponding author. Tel.: C49 6131 39 23169; fax: C49 6131 39 23524. E-mail address: [email protected] (D.C. Albach). 1 Present address: Institute of Special Botany, Johannes Gutenberg-Universita¨t Mainz, Bentzelweg 9b, 55099 Mainz, Germany. 0305-1978/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.bse.2005.06.002

1088

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

separate from the rest of the group. This result is mirrored by the unusual phytochemical arsenal of V. intercedens, which is the only species in the genus analysed to date to contain melittoside and globularifolin. Subgenus Pentasepalae appears to be a clade of diverse lineages from southwestern Asia and a single European clade. Species shown to have 6hydroxyflavones do not form a monophyletic group. Subgenus Pseudolysimachium seems to have originated in Eastern Asia. 6-Hydroxyflavones acylated with phenolic acids are common in this subgenus but may have originated only later in the evolution of the group. Possible chemotaxonomic markers for other groups are discussed. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Veronica; Plantaginaceae; Chemosystematics; Phylogenetic analysis; ITS; trnL-F region

1. Introduction Veronica is a genus of the Plantaginaceae sensu APG (1998) formerly placed in Scrophulariaceae with about 500 species (Albach et al., 2004a). It is distributed over most of the Northern Hemisphere and in many parts of the Southern Hemisphere, and is ecologically highly diverse with species growing in aquatic to dry steppe habitats from sea level to high alpine regions. This diversity and the fact that many species have beautiful blue flowers may explain the interest Veronica has drawn for a long time. Extensive breeding programs and morphological studies between 1900 and 1955, i.e. by Watzl (1910) and Lehmann and his pupils (e.g., Lehmann, 1908, 1937; Ro¨mpp, 1928; Ha¨rle, 1932; Riek, 1935; Schlenker, 1936; Riek-Ha¨ußermann, 1943) have been followed by cytological and morphological analyses by M.A. Fischer (1967, 1972, 1975, 1987), Martı´ nez-Ortega (Martı´ nez-Ortega et al., 2000; Martı´ nez-Ortega and Rico, 2001) and floristic surveys (e.g., Borissova, 1955; Hartl, 1966e1968; M.A. Fischer, 1978, 1981). Cladistic analyses incorporating much of the morphological data have been published by Hong (1984) and Kampny and Dengler (1997). Phytochemistry of Veronica and related species of the Northern Hemisphere has been studied by several authors with an emphasis on iridoid compounds (e.g., Grayer-Barkmeijer, 1973; Lahloub et al., 1993; Taskova et al., 2002, 2004) and flavonoid compounds (e.g., Grayer-Barkmeijer, 1978, 1979; Peev, 1982; Albach et al., 2003). Recently, molecular techniques and phylogenetic analyses have been applied to Veronica and related genera (Wagstaff and Garnock-Jones, 1998; Albach and Chase, 2001; Albach et al., 2004b,c). These studies have helped revolutionizing our ideas about the evolution of the genus and led to a new, phylogenetic infrageneric classification of Veronica (Albach et al., 2004a). Combined with the vast amount of information from other aspects of their biology, we now have a much better understanding of how major groups in Veronica are delimited and related, and how they have evolved. This information combined in a phylogenetic framework opens a new perspective for character evolution in the genus and helps focus on those clades that have not been studied intensively in the past. One area of the study of plant evolution that has shown strong correlations with the results of DNA sequence analyses is chemotaxonomy (Grayer et al., 1999). The

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1089

objective of the present paper is therefore to investigate the distribution of phytochemical compounds, which may be useful as an independent test of relationships suggested in DNA sequence analyses. Taskova et al. (2004) have compared the distribution of iridoid compounds in the genus Veronica with dendrograms based on DNA sequences. However, interesting chemotaxonomic characters can also be found among flavonoids and verbascoside-like compounds of Veronica. In two accompanying studies, we investigate the phytochemistry of additional species focusing on unusual flavonoids (acylated 6- and 8-hydroxylated flavone glycosides, Albach et al., 2005) and the iridoid glucoside, ajugol, and on the non-iridoid compound, cornoside, in the genus (Jensen et al., 2005). The distribution of these characters is discussed here in the light of a phylogenetic hypothesis derived from DNA sequences from the nuclear ribosomal internal spacer region (ITS 1, 5.8 S rDNA, ITS 2) and the plastid trnL-F region (trnL intron, trnL 3# exon, trnL-F spacer).

2. Materials and methods 2.1. Plant sampling For a larger analysis, 49 accessions were sequenced for the ITS- and trnL-F region with an emphasis on subgenera Chamaedrys and Pocilla, which appeared as most interesting in the phytochemical analyses (Albach et al., 2005). Taxa from all other subgenera sensu Albach et al. (2004a) were added to the data matrix (Table 1). Veronica montana and Veronica beccabunga were designated as outgroups based on previous analyses (Albach and Chase, 2001; Albach et al., 2004b,c). Due to limited variation in the trnL-F region of subgenus Pentasepalae, a smaller matrix of 15 ITSsequences including eight new sequences (Table 1) for species of subgenus Pentasepalae was analysed separately and rooted with Veronica czerniakowskiana based on previous analyses (Albach et al., 2004b,c). Finally, nine ITS- and five trnLF-sequences available from subgenus Pseudolysimachium (Table 1) were analysed in separate analyses with Veronica glandulosa, V. beccabunga and Veronica gentianoides as outgroups based on previous analyses (Albach et al., 2004b,c). Eleven of the trnL-F and 19 of the ITS-sequences are reported here for the first time. Voucher specimens were made for all plants used in this study (Table 1). 2.2. DNA extraction, amplification and sequencing The protocol followed previous studies (Albach and Chase, 2001; Albach et al., 2004b). Total genomic DNA was extracted from herbarium material and silica geldried leaf samples according to the 2x CTAB procedure of Doyle and Doyle (1987). The trnL intron, 3# exon, and trnL-F spacer (hereafter trnL-F ) were amplified with primers c and f of Taberlet et al. (1991). ITS-sequences were amplified and sequenced using the primers 17SE (Sun et al., 1994) and ITS4 (White et al., 1991). PCR products were run on a 1.0% TBE-agarose gel, cut out of the gel, and cleaned using QIAquickÔ PCR purification and gel extraction kit (Qiagen GmbH, Hilden,

1090

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

Table 1 Voucher specimens and GenBank accession numbers for sequences used in the phylogenetic analysis Species

Subgenus

Voucher

ITS GenBank acc. no.

trnL-F GenBank acc. no.

V. agrestis V. amoena

Pocilla Pocilla

AF509784 AF509786

AF513335 AF513334

V. argute-serrata e Afghanistan V. argute-serrata e Turkey V. armena* V. arvensis V. beccabunga V. biloba V. bombycina V. campylopoda e Jordan V. campylopoda e Turkey V. capillipes V. catarractae

Pocilla

Albach 386, WU Schelkovnikov and Schmidt 5.4.1907, TBS Rechinger 35177, WU

AF509787

AF513337

Pocilla

Albach 640, WU

AY673605

AY673623

Pentasepalae Chamaedrys Beccabunga Pocilla Pentasepalae Pocilla

AF313040 AF313002 AF313015 AY673606 AF486353 AF486364

e AF486340 AF486403 AY673625 AF486376 AF486370

Pocilla

Struwe 1404, WU Albach 147, WU Albach 122, K Albach 636, WU Struwe 1403, WU Scho¨nswetter and Tribsch 4152, WU Albach 656, WU

AY673608

AY673624

Pocilla Hebe

Rechinger 51188, WU Garnock-Jones 2403, CHR

AY673607 AY034859

V. V. V. V. V. V. V. V.

caucasica* ceratocarpa chamaedryoides chamaedrys chamaepithyoides cinerea* crista-galli cuneifolia subsp. isaurica V. czerniakowskiana V. dahurica V. dichrus* V. filiformis V. fruticulosa V. gentianoides V. glandulosa V. glauca

Pentasepalae Pocilla Chamaedrys Chamaedrys Triangulicapsula Pentasepalae Cochlidiosperma Pentasepalae

Albach 326, WU O¨ztu¨rk 429, WU Albach 393, WU Albach 121, K UA 174, SALA Albach and Chase 113, K Dolmkanov 17.4.1983, TBS Struwe 1409, WU

AF486357 AY741514 AY673611 AF313003 AF509796 AY144458 AF509799 AF486354

AY673627 AY540887/ AY540901 e AY673626 AY673631 AF486377 AF511477/8 e AF486367 AF486372

Pentasepalae Pseudolysimachium Pentasepalae Pocilla Stenocarpon Beccabunga Veronica Pellidosperma

AF486362 AF313023 AF312998 AF486363 AF313004 AF313018 AF313008 AF313006

AF486371 AF411479/80 e AF486368 AF486383 AF486401 AF486394 AF486395

V. V. V. V. V. V. V. V. V. V. V.

Pseudolysimachium Pseudolysimachium Pocilla Pocilla Pentasepalae unplaced Chamaedrys Stenocarpon Chamaedrys Pentasepalae Pseudolysimachium

Terme 39177E, EVIN BG Jena, WU Struwe 1407, WU Albach 298, WU Albach 71, WU Albach 72, WU Fischer 713/98, WU Chase s.n., K/M. A. Fischer 9, 7.4.1999, WU BG Bonn, WU Sun and Kim 12022, WU Albach 667, WU Albach 666, WU Albach 70, WU Fries et al. 2016, BR Albach 484, WU Schickhoff 1377, GOET Dickoree 13042, GOET Struwe 1411, WU Albach 66, WU

e AF515210 AY673610 AY673609 AF313000 AY540867 AY673612 AY540868 AY673613 AF312997 AF313021

AY486449 AF486406 AY486439 AY673628 AF513341 AY540872 AY673633 AY486442 AF486378 e AF486407

incana nakaiana intercedens e 1 intercedens e 2 jacquinii javanica krumovii lanosa laxa liwanensis* longifolia

1091

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106 Table 1 (continued) Species

Subgenus

Voucher

ITS GenBank acc. no.

trnL-F GenBank acc. no. e AY673634 AY673632 AF486397 AF486388 AY540890/ AY540903 AF513339 AY673630 e AF513340 e AF513336 AF486369 e AF513338 AY540881/ AY540895 e AF486405 e e e AF513343 AF513333 AF486396 AF486374 AF486379 AF510426

V. V. V. V. V. V.

longifolia e UK magna micans missurica montana nivea

Pseudolysimachium Chamaedrys Chamaedrys Synthyris Veronica Derwentia

McSheahan 48, K Albach 360, WU Scho¨nswetter 2567, WU Chase s.n., K Albach 151, WU CHR512486

AY673619 AY673615 AY673616 AF313019 AF313014 AF037382

V. V. V. V. V. V. V. V. V. V.

oltensis opaca ovata paederotae peduncularis* persica polita propinqua* rubrifolia salicifolia

Pentasepalae Pocilla Pseudolysimachium Pentasepalae Pentasepalae Pocilla Pocilla Pentasepalae Pocilla Hebe

Struwe 1405, WU Polatschek 8.7.90, W Matsumoto s.n., WU Klein 7901, WU Albach 325, WU Fay 175, K Albach 146, WU Albach 327, WU n.n., WU CHR512466

AF312995 AY673617 AY673618 AF509783 AF486356 AF509785 AF509818 AF486361 AF509788 AF037385

V. V. V. V. V. V. V. V. V. V. V.

schmidtiana spicata e cult. spicata e UK spicata e UK teucrium* thessalica triloba triphyllos turrilliana verna vindobonensis

Pseudolysimachium Pseudolysimachium Pseudolysimachium Pseudolysimachium Pentasepalae Stenocarpon Cochlidiosperma Pellidosperma Pentasepalae Chamaedrys Chamaedrys

S. Umezawa 20130, WU BG Bonn 09333, WU Fay e Avon Gorge, K Fay e Avon Gorge, K Albach 119, K Raus and Rogl 5072, SALA Albach 242, WU Albach 244, WU Albach 278, WU Albach 149, WU Fischer, cult. Wien, WU

AY673620 AF313022 AY673621 AY673622 AF312999 AF509792 AF509804 AF509795 AF486360 AF509789 AY673614

Species with an * were only used in the analysis of ITS-sequences for V. subgenus Pentasepalae.

Germany) following the manufacturer’s protocols. Sequencing reactions (10 ml) were carried out using the Taq DyeDeoxy Terminator Cycle Sequencing mix (Applied Biosystems Inc.). Reactions were run out on a Prism 377 automated sequencer (Applied Biosystems Inc.), and both strands were sequenced. Sequences were assembled and edited using Sequence NavigatorÔ (Applied Biosystems Inc.). Assembled sequences were manually aligned prior to analysis. Aligned sequence matrices are available from DCA by request. 2.3. Sequence analysis Insertions and deletions are frequent in both DNA regions. We have scored those gaps that are neither found in mononucleotide repeats nor are only 1 bp long as present/absent to preserve the information included in them. All matrices were analysed with PAUP* 4.0b10 (Swofford, 1998) using heuristic parsimony methods.

1092

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

Ten runs of random taxon addition (10 replicates each) using tree bisection reconnection (TBR) were conducted with MulTrees (keeping multiple shortest trees) in effect and no tree limit. Bootstrap percentages were assessed using 1000 replicates and TBR-branch swapping with a maximum of 100 trees per replicate.

3. Results and discussion 3.1. ITS Aligned sequences of the ITS-region included 759 characters including 4 gaps scored as present/absent (one having a CI ! 1). One hundred and ninety-seven characters were potentially parsimony informative. We found 121 most parsimonious trees, which required 943 steps (Fig. 1; CI Z 0.46, RI Z 0.71). Relationships among the subgenera are not well supported except for the clade consisting of subgenera Chamaedrys, Pocilla and Pentasepalae (72 BP). However, monophyly of subgenera is well supported. Within subgenus Chamaedrys (99 BP), the annual species Veronica arvensis and Veronica verna are consecutive sisters to the perennial species. Among the latter, the Asian species, Veronica laxa and Veronica magna, are sister to the European species (75 BP). Veronica chamaedrys is most closely related to Veronica micans (90 BP). Within subgenus Pocilla (100 BP), two clades are strongly supported that resemble subsection Agrestes in the circumscription of Lehmann (1908; 96 BP) and subsection Biloba sensu Ro¨mpp (1928; 100 BP). Within subsection Agrestes, Veronica polita, Veronica persica and Veronica opaca form one moderately supported clade (87 BP), as does the group of Veronica ceratocarpa, Veronica filiformis, and Veronica agrestis. A position of Veronica amoena close to subsect. Agrestes has hitherto not been suggested. Within subsection Biloba, Veronica arguteserrata is not monophyletic, one sequence being more closely related than the other to Veronica campylopoda. Veronica biloba and Veronica capillipes are strongly supported sisters (99 BP). Veronica intercedens, usually considered part of subsection Biloba, does not seem to belong to the group in a strict sense. 3.2. trnL-F Aligned sequences of the trnL-F region included 1067 characters including 16 gaps scored as present/absent with 136 characters potentially parsimony informative. We stopped saving trees when 1000 most parsimonious trees that required 475 steps were found (Fig. 2; CI Z 0.79, RI Z 0.83). Repeating the search five times did not result in searches finding shorter trees. Relationships among subgenera do find some bootstrap support above 50%. Within subgenus Chamaedrys (83 BP), the branching pattern differs from that in the ITS-analysis by V. verna switching its place with V. magna and V. laxa, which renders the perennial species of the subgenus nonmonophyletic. Veronica vindobonensis is sister to V. micans (59 BP) rather than found in a clade together with Veronica krumovii and Veronica chamaedryoides as in the ITS-analysis. Within subgenus Pocilla (90 BP), subsection Biloba is paraphyletic with

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1093

Fig. 1. One of 121 most parsimonious trees from the analysis of the ITS-region for 49 taxa of Veronica. Numbers above branch and before the slash are branch lengths, numbers behind the slash are bootstrap percentages. Branches not present in strict consensus tree are dotted with branch lengths in italics.

1094

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

Fig. 2. One of more than 1000 most parsimonious trees from the analysis of the trnL-F region for 49 taxa of Veronica. Numbers above branch and before the slash are branch lengths, numbers behind the slash are bootstrap percentages. Branches not present in strict consensus tree are dotted with branch lengths in italics.

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1095

respect to subsection Agrestes and Veronica rubrifolia. Within subsection Agrestes, V. amoena is sister to the rest of the subsection. Subgenus Pentasepalae (97 BP) is characterized by a lack of resolution due to low sequence divergence except for the sister-group relationship of V. czerniakowskiana to the rest of the subgenus (75 BP). 3.3. Combined data set The complete data set of both regions combined included 1826 characters including 20 gaps scored as present/absent with 333 characters potentially parsimony informative. We found 384 most parsimonious trees that required 1441 steps (Fig. 3; CI Z 0.56, RI Z 0.73). The analysis differs from that of the single data sets in an overall increase in bootstrap support in all relationships. Branching pattern among subgenera Pocilla, Pentasepalae, and Chamaedrys resembles that in the analysis of ITS. In subgenus Chamaedrys (100 BP), the annual species are paraphyletic with respect to the perennials (71 BP). Branching pattern in subgenus Pocilla (100 BP) is similar to that of the trnL-F analysis with subsection Biloba paraphyletic to subsection Agrestes and V. rubrifolia and V. amoena sister to the rest of subsection Agrestes. Due to the lack of variation in the trnL-F data set, relationships within subgenus Pentasepalae (95 BP) are the same as in the analysis of ITS. 3.4. ITS-data set for subgenus Pentasepalae The 15 sequences for subgenus Pentasepalae included 36 potentially parsimonyinformative characters but no potentially parsimony-informative indel. We found three most parsimonious trees that required 133 steps (Fig. 4; CI Z 0.78, RI Z 0.66). Bootstrap support is generally low with the exception of the European species from subsection Pentasepalae and the association of Veronica peduncularis and Veronica caucasica. 3.5. Analyses of subgenus Pseudolysimachium The analysis of nine ITS-sequences from six species from subgenus Pseudolysimachium included 76 potentially parsimony-informative characters and resulted in two most parsimonious trees (Fig. 5A) with 175 steps (CI Z 0.89, RI Z 0.86). The analysis of five trnL-F-sequences from five species included 24 potentially parsimony-informative characters and resulted in a single most parsimonious tree (Fig. 5B) with 71 steps (CI Z 0.99, RI Z 0.97). The most parsimonious trees from the analysis of ITS are reasonably well resolved, but the one from the analysis of trnL-F is not, because of limited sequence variability. Both analyses agree in Veronica nakaiana from Eastern Asia being sister to the rest of the subgenus. In the analysis of ITS-sequences, V. nakaiana is consecutively followed by two other East Asian species, Veronica dahurica and Veronica schmidtiana. However, bootstrap support is relatively low except for the monophyly of the subgenus, of those two species where more than one sequence was available, and of the clade comprising the European Veronica spicata and Veronica longifolia together with

1096

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

Fig. 3. One of 384 most parsimonious trees from the combined analysis of the ITS-region and the trnL-F region for 49 taxa of Veronica. Numbers above branch and before the slash are branch lengths, numbers behind the slash are bootstrap percentages. Branches not present in strict consensus tree are dotted with branch lengths in italics. Branches with bootstrap support above 70% appear in double thickness, those above 95% in triple thickness.

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1097

Fig. 4. One of three most parsimonious trees. The other two show Veronica cuneifolia either as sister to Veronica cinerea or as sister to V. subsect. Pentasepalae. Numbers above the branches indicate branch lengths, those below the branch bootstrap support.

1098 D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

Fig. 5. Phylograms for sequence analyses of subgenus Pseudolysimachium. Branch lengths (before the dash) and bootstrap percentages above 50% (after the dash) are indicated. (A) Strict consensus and one of two most parsimonious trees of the analysis of the ITS-region (the other most arsimonious tree combines V. spicata and V. longifolia with V. ovata as sister to both). (B) Most parsimonious tree of the analysis of the trnL-F region.

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1099

Veronica ovata. Sequence variability within V. spicata is surprisingly high. Even two individuals within one English population differ by three nucleotide substitutions, which is also supported by strong differentiation in AFLP-fingerprint patterns (Albach, unpubl.). 3.6. Discussion of the DNA sequence analyses The present analysis of DNA sequence data focuses on two subgenera (subg. Pocilla and Chamaedrys), which appeared especially interesting regarding their phytochemical arsenal. Relationships between subgenera have been the focus of previous publications (Albach and Chase, 2004; Albach et al., 2004b,c). Subgenus Pocilla is sister to subgenus Pentasepalae in analyses of both nuclear ITS and plastid trnL-F sequences with weak support in the separate analyses (ITS: 57 BP; trnL-F: 68 BP) but moderate support in the combined analysis (84 BP). This sister-group relationship has also been found in all previous publications (Albach and Chase, 2001, 2004; Albach et al., 2004b,c). The present analysis extends taxon sampling in subgenus Pocilla, allowing for the first time a discussion of relationships in this subgenus. As mentioned before, subgenus Pocilla includes species belonging to the Agrestis-group sensu Lehmann (1908) and ‘‘Verwandtschaftsgruppe’’ Biloba sensu Ro¨mpp (1928). The analysis supports subsection Agrestes in the circumscription of Lehmann (1908) including V. agrestis, V. filiformis, V. ceratocarpa, V. polita, V. persica, and V. opaca. Apart from the two strictly European tetraploid species V. agrestis and V. opaca, all other species originate from the hyrcanian-caucasian region with three of them (V. ceratocarpa, Veronica francispetae e both diploid; Veronica siaretensis e unknown ploidy) restricted to this region. V. polita, V. filiformis and V. persica (the first two diploid, the latter tetraploid) have become cosmopolitan weeds at different times. The occurrence of V. filiformis in this group is peculiar because it differs from all other species of the subgenus in being perennial and self-incompatible (Lehmann, 1944), whereas all other species are annuals and at least those of subsection Agrestes self-compatible (Lehmann and Schmitz-Lohner, 1954). Self-incompatibility in subsection Biloba is unlikely. Morphological differences between the species of subsection Agrestes have been discussed by Thaler (1951) and Fischer (1987). Several publications have dealt with the origin of the polyploid species in this subsection. Based on karyological and biogeographical evidence, Beatus (1936) hypothesized that all three tetraploid species (V. agrestis, V. opaca, V. persica) are autopolyploids, V. persica derived from V. filiformis and the other two derived from V. polita. Lehmann (1940) contradicted Beatus with respect to V. persica, which he considered to be also derived from V. polita. Fischer (1987) however, argued based on morphological intermediacy that V. persica is an allopolyploid species derived from a cross of V. polita and V. ceratocarpa. Evidence presented here from both maternally- and biparently-inherited DNA sequences supports a contribution of V. polita to the genomes of V. persica and V. opaca, although it cannot be confirmed whether they are autopolyploid or allopolyploids, because ITS-sequence could be converging towards the maternal parent. The position of V. agrestis as sister to V. filiformis is rather surprising as a close relationship has not been proposed before

1100

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

based on morphological evidence, although their flavonoid profiles are very similar (Grayer-Barkmeijer, 1979). Further investigations are necessary to elucidate the ancestry of the polyploid species in this group. Subsection Biloba is characterized by pairwise fusion of the calyces, more or less recurved pedicels when fruiting and in most cases strongly bilobed capsules. Our analyses (Figs. 1e3) do not support this subsection to be monophyletic in its original circumscription by Lehmann (1910; see also Ro¨mpp, 1928). All analyses indicate V. intercedens to be more closely related to V. rubrifolia than to the rest of the subsection. V. intercedens was split from subsection Biloba by Elenevsky (1977) together with Veronica cardiocarpa to subsection Cardiocarpae based on the verticillate cauline leaves in these species. This is also supported by the unusual iridoid phytochemistry of V. intercedens, which is the only species in the genus to contain melittoside and globularifolin (Albach et al., 2003). V. cardiocarpa has not yet been investigated phytochemically. V. rubrifolia does not share any of the characters of subsection Biloba or subsection Cardiocarpae mentioned above and its position in the phylogeny remains doubtful until further annual species from southwest Asia have been included in the analyses. Analysis of the subgenus Pentasepalae supports the sister-group relationship of V. czerniakowskiana to the rest of the subgenus (Figs. 1e4) found in previous analyses (Albach et al., 2004b,c). However, using cpDNA, subgenus Chamaedrys is sister to a larger clade of subgenera Pentasepalae and Pocilla, subgenus Stenocarpon and the Hebe complex in analyses of the plastid trnL-F region (Fig. 2, Albach et al., 2004b,c) and the plastid rps16 intron (Albach and Chase, 2004). The association of the perennial V. chamaedrys with the annual V. arvensis was first noted by Albach and Chase (2001). Both species have never been considered closely related because the distinction between annuals with a terminal inflorescence and those perennial species that only have lateral inflorescences was considered the major division in the genus (Bentham, 1846; Ro¨mpp, 1928). Apart from this, V. arvensis and V. verna have seeds that differ strongly from V. chamaedrys and relatives (Martı´ nez-Ortega and Albach, in preparation). Phytochemical similarity of these species (absence of catalpol esters in both species, Taskova et al., 2004; Jensen et al., 2005) is therefore valuable support for the results of DNA sequence analyses. The results for subgenus Chamaedrys presented here (Figs. 1e3) show that the perennial species are further divided into an Asian group (V. magna, V. laxa) and a European group. The European group can be termed V. chamaedrys sensu lato because all species have been included in V. chamaedrys at one time or another. Apart from the tetraploid V. chamaedrys all other species sampled are diploid. No definite answer can be given regarding the ancestors of V. chamaedrys but it is noteworthy that in all analyses V. chamaedrys seems to be closer to the Central European species V. vindobonensis and V. micans than to the strictly Balkanian species V. krumovii and V. chamaedryoides. An origin of V. chamaedrys in Central Europe seems therefore more likely. However further diploid populations of V. chamaedrys in Central and Eastern Europe are known (Fischer, 1973; M.A. Fischer, pers. comm.). Those populations so closely resemble the tetraploid V. chamaedrys that they have not been separated from it at any taxonomic rank.

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1101

Due to this fact they should be considered in future analyses investigating the ancestors of the tetraploid form of V. chamaedrys. Subgenus Pentasepalae is a species-rich group with its 70 species mostly found in Turkey and Iran. Its geographical outliers, subsect. Pentasepalae in Europe and subsect. Petraeae in the Greater Caucasus, are the only well supported groups (97 BP and 83 BP, respectively; Fig. 4) in this subgenus. A larger analysis with more informative characters is needed to resolve the relationships of this group. Subgenus Pseudolysimachium is a highly diverse group with approximately 25 species currently accepted despite approximately 400 names and combinations available. This taxonomic complexity demonstrates the high morphological variability within the subgenus and explains the difficulties for taxon sampling in molecular analyses. Another problem in systematic analyses of this subgenus is the probably widespread interfertility and hybridisation between taxa (e.g. Ha¨rle, 1932). The current analyses are therefore only intended as broad guidelines in analyses of AFLPfingerprints (Albach, unpubl.). Nevertheless, one important result is the paraphyly of East Asian species and the derived position of European species, which indicates an East Asian origin of the subgenus. This result is important because it points to those species that may resemble most the ancestral species of this subgenus. Among those species is V. schmidtiana, which is the only species that was not associated with this subgenus before 1968 (Yamazaki, 1968), because it has a flattened capsule, which is common among other species of Veronica but only found in a few Eastern Asian species of subgenus Pseudolysimachium. V. schmidtiana is also distinct from the rest of the subgenus in pollen morphology (Hong, 1984) and its rosette habit, which is frequently found among other species of Veronica (e.g. subgenus Synthyris) but only rarely among the more derived members of subgenus Pseudolysimachium. V. schmidtiana, however, shares with the rest of the subgenus the unusual chromosome number of x Z 17 (e.g. Sakai, 1935). Its inflorescence is long and terminal but not as dense and spike-like as those of most other species. The loosely flowered inflorescence is associated with shorttubed corollas and is found in several Eastern Asian species of this subgenus but not in its derived members. An emphasis of future studies should therefore be put on the East Asian species and not only study its relationship based on DNA-markers but also study its phytochemistry (see below), seed and pollen. 3.7. Review of chemotaxonomic markers in Veronica Molecular systematics has often given results that were hitherto unsuspected, which in turn has reinitiated research in other fields of systematics (e.g. Williams and Friedman, 2004). Phytochemistry is a prime example for this. Sometimes taxonomically useful characters were overlooked and only later appreciated (Grayer et al., 1999) but much too often phytochemistry is only incompletely known, which prevents a full assessment of its use in classifying plants. Veronica is a fortunate exception to this. A comparison of our analyses of ITS- and plastid trnL-F sequence data with published reports on iridoids, flavonoids and verbascoside-like compounds reveals several compounds that are good indicators of relationships, several of them previously not recognized. Due to the lack of information on iridoids in the Hebe

1102

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

complex and the complexity of information on flavonoids in the Hebe complex (Bayly et al., 2000, 2002), we focus here on species from the Northern Hemisphere. Comparisons reveal not only synapomorphies for clades, so far exclusively characterized by nucleotide differences but also reveal much about the evolution of these phytochemical compounds. Convergence in phytochemical arsenal seems to be widespread. The convergence in the production of acetylated 8-hydroxyflavone allosyl glycosides between Veronica (Grayer-Barkmeijer, 1979; Toma´s-Barbera´n et al., 1988; Albach et al., 2003) and Stachys (Lamiaceae, Lenherr et al., 1984; Toma´s-Barbera´n et al., 1988) is just one example. Reversals to ancestral conditions seem to be frequent despite long divergence from ancestors that also contained these compounds. This supports the idea of Grayer et al. (1999) that those genes responsible for the biosynthesis of such compounds may be retained in the genome despite lack of expression. Alternatively, these reversals may occur if additional steps in the synthesis of secondary compounds are secondarily lost again. Analyses of genes underlying the biosynthetic pathways such as those in the flavonoid biosynthetic pathway (Shirley, 1996; Quattrocchio et al., 1999; Zufall and Rausher, 2004) will be necessary to determine the fate of the underlying genes. Among those compounds that are chemotaxonomically useful in Veronica are iridoids, flavonoids and verbascoside-like compounds. Mussaenoside is one of the few iridoid compounds in Veronica not derived from aucubin and catalpol. It was first detected by Grayer-Barkmeijer (1973, 1979) and then identified by Afifi-Yazar and Sticher (1981). Subsequent studies showed that it is almost exclusively found in a clade comprising Veronica officinalis, Veronica aphylla and Veronica urticifolia (axillary inflorescences) and Veronica alpina and Veronica bellidioides (terminal inflorescences) (Grayer-Barkmeijer, 1979; Taskova et al., 2002, 2004), which used to be classified in two different sections. Grayer-Barkmeijer (1973, 1979) already noted that species producing mussaenoside also share the same basic chromosome number (x Z 9) and have similar seeds. However, the clade was not formally recognized until recently (subgenus Veronica, Albach et al., 2004a), because former intrageneric classifications of Veronica relied on the presence or absence of a terminal inflorescence, an unstable character in this group. Two other iridoids that are likely to be chemotaxonomically useful are melittoside and globularifolin, which are found in V. intercedens (Albach et al., 2003). Investigation of related species is necessary to evaluate their distribution, but occurrence in Veronica tenuissima and V. cardiocarpa seems likely because they also do not blacken while drying, an indication that aucubin is lacking. Finally, ajugol has been found in members of Veronica subsection Biloba (Jensen et al., 2005), and as a dominant compound is a chemotaxonomic marker for this subsection, although it is also found in trace amounts in V. arvensis (subgenus Chamaedrys) (Jensen et al., 2005). Other species of Veronica contain aucubin, catalpol and its esters, whereas the specific combination of esters may be significant for various groups at various taxonomic levels (Taskova et al., 2002, 2004), no ester is confined to a group at subgeneric level or below. Among flavonoids, 8-hydroxyflavones are the most prominent chemotaxonomic marker supporting the sister-group relationship of subgenus Pocilla and subgenus

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1103

Pentasepalae (Albach et al., 2005). Whereas subgenus Pocilla contains annual species, subgenus Pentasepalae is exclusively perennial. Life history (annuals versus perennials) has been considered an essential character for the classification (e.g. Ro¨mpp, 1928), but annuals have been shown to be derived in parallel multiple times within Veronica (Albach et al., 2004b). Within these two subgenera, a reversal to the production of 6-hydroxyflavones is also inferred several times (Albach et al., 2005). Less important chemotaxonomic markers among the flavonoids of Veronica are 6-hydroxyflavones acylated with phenolic acids (Albach et al., 2005). These compounds, which we called spicosides AeF (Albach et al., in press), occur in five out of eight investigated species of subgenus Pseudolysimachium. They are lacking in V. dahurica and V. nakaiana, inferred to be the earliest diverging species in the subgenus (Fig. 5), which contrasts with a possible presence of these compounds in the ancestor of the subgenus. Two of the spicosides have also been found in V. thymoides ssp. pseudocinerea (Saracoglu et al., 2004) belonging to subgenus Pentasepalae, where this character is thought to have originated independently (Albach et al., 2005). Other flavonoid compounds that may be chemotaxonomic markers still need closer investigation. Grayer-Barkmeijer (1979) noticed the restricted occurrence of luteolin 7,3#-O-glucuronide in Veronica fruticans, Veronica fruticulosa and Veronica saturejoides. Previously these species were thought to belong to two different subsections, but they are now considered more closely related (Albach et al., 2004b). Scutellarein-glycosides may be a fourth flavonoid chemotaxonomic marker in Veronica being only found in members of section Beccabunga with the exception of Veronica peregrina (Grayer-Barkmeijer, 1979). The position of V. peregrina in section Beccabunga was unrecognised until recently (Albach and Chase, 2001; Albach et al., 2004c). Although the absence of scutellarein-glycosides may argue against its inclusion, the overall similarity of the flavonoid arsenal (Grayer-Barkmeijer, 1979) is clearly in support of the DNAbased results. This highlights the trend already observed among iridoids that while specific chemotaxonomic markers may be lacking, the overall similar combination of compounds is generally supporting results from DNA-based molecular systematics. Finally, cornoside, a compound related to verbascoside, has been found in both annual and perennial species of subgenus Chamaedrys, but not outside this subgenus in Veronica (Jensen et al., 2005). It is therefore an important character supporting a clade that until now has only been distinguished by DNA sequence data (Figs. 1e3; Albach et al., 2004b,c). Phytochemical characters may also be important as chemosystematic markers at a lower taxonomic level in Veronica, e.g. the presence or absence of a diosmetin glycoside in the species-complex of Veronica hederifolia (Peev, 1982). Further examples have been given by Grayer-Barkmeijer (1973, 1978, 1979) and Taskova et al. (1997, 2002), but many more may be found after closer inspection. Much chemotaxonomic work is currently being carried out on the Hebe complex in New Zealand, where several examples of chemosystematic markers useful at the species level have been found (e.g. Bayly et al., 2002).

1104

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

Acknowledgments We thank the Studienstiftung des deutschen Volkes for a doctoral scholarship to D.C.A. and financial support by FWF (Fonds zur Fo¨rderung der wissenschaftlichen Forschung) project P15336.

References Afifi-Yazar, F.U¨., Sticher, O., 1981. Ladroside (Z6#-caffeoyl-mussaenoside), a new iridoid glucoside from Veronica officinalis L. (Scrophulariaceae) and the elucidation of the absolute configuration at C (8) of mussaenoside. Helv. Chim. Acta 64, 16e24. Albach, D.C., Chase, M.W., 2001. Paraphyly of Veronica (Veroniceae; Scrophulariaceae): evidence from the internal transcribed spacer (ITS) sequences of nuclear ribosomal DNA. J. Plant Res. 114, 9e18. Albach, D.C., Chase, M.W., 2004. Incongruence in Veroniceae (Plantaginaceae): evidence from two plastid and a nuclear region. Mol. Phylogenet. Evol. 32, 183e197. Albach, D.C., Grayer, R.J., Jensen, S.R., O¨zgo¨kce, F., Veitch, N.C., 2003. Acylated flavone glycosides from Veronica. Phytochemistry 64, 1295e1301. Albach, D.C., Martı´ nez-Ortega, M.M., Fischer, M.A., Chase, M.W., 2004a. A new classification of the tribe Veroniceae e problems and a possible solution. Taxon 53, 429e452. Albach, D.C., Martı´ nez-Ortega, M.M., Chase, M.W., 2004b. Veronica: parallel morphological evolution and phylogeography in the Mediterranean. Plant Syst. Evol. 246, 177e194. Albach, D.C., Martı´ nez-Ortega, M.M., Fischer, M.A., Chase, M.W., 2004c. Evolution of Veroniceae: a phylogenetic perspective. Ann. Missouri Bot. Gard. 91, 275e302. Albach, D.C., Grayer, R.J., Kite, G.C., Jensen, S.R., 2005. Veronica: acylated flavone glycosides as chemosystematic markers. Biochem. Syst. Ecol. 33 (11), 1167e1177. Angiosperm Phylogeny Group, 1998. An ordinal classification for the families of flowering plants. Ann. Missouri Bot. Gard. 85, 531e553. Bayly, M.J., Garnock-Jones, P.J., Mitchell, K.A., Markham, K.R., Brownsey, P.J., 2000. A taxonomic revision of the Hebe parviflora complex (Scrophulariaceae), based on morphology and flavonoid chemistry. N.Z. J. Bot. 38, 165e190. Bayly, M.J., Kellow, A.V., Mitchell, K.A., Markham, K.R., deLange, P.J., Harper, G.E., GarnockJones, P.J., Brownsey, P.J., 2002. Descriptions and flavonoid chemistry of new taxa in Hebe sect. Subdistichae (Scrophulariaceae). N.Z. J. Bot. 40, 571e602. Beatus, R., 1936. Die Veronica-Gruppe Agrestis der Sektion Alsinebe Griseb., ein Beitrag zum Problem der Artbildung. Z. indukt. Abstammungs- Vererbungsl. 71, 353e381. Bentham, G., 1846. Scrophulariaceae. In: de Candolle, A. (Ed.), Prodromus Systematis Naturalis Regni Vegetabilis. Victor Masson, Paris, pp. 448e491. Borissova, A.G., 1955. Veronica. In: Shishkin, B.K., Bobrov, E.G. (Eds.), Flora SSSR. Izdatel’stvo Akademii Nauk SSSR, Moskva, Leningrad, pp. 373e557. Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. Bot. Soc. Am. 19, 11e15. Elenevsky, A.G., 1977. Sistema roda Veronca L. Bjull. Moskovsk. Obsc. Isp. Prir., Otd. Biol. 82, 149e160. Fischer, M., 1967. Beitra¨ge zur Cytotaxonomie der Veronica hederifolia-Gruppe (Scrophulariaceae). O¨sterr. Bot. Z. 114, 189e233. Fischer, M., 1972. Neue Taxa, Chromosomenzahlen und Systematik von Veronica subsect. Acinifolia (Ro¨mpp). Stroh. O¨sterr. Bot. Z. 121, 413e437. Fischer, M., 1975. Untersuchungen u¨ber den Polyploidkomplex Veronica cymbalaria agg. (Scrophulariaceae). Plant Syst. Evol. 123, 97e105. Fischer, M.A., 1973. Notizen zur Systematik, Chromosomenzahl und Verbreitung einiger Veronica-Sippen in Ka¨rnten. Carinthia 83, 379e388.

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

1105

Fischer, M.A., 1978. Veronica L. In: Davis, P.H. (Ed.), Flora of Turkey and the East Aegean Islands University Press, Edinburgh, pp. 689e753. Fischer, M.A., 1981. Veronica. In: Rechinger, K.H. (Ed.), Flora des iranischen Hochlandes und der umrahmenden Gebirge. Akademische Druck- und Verlagsanstalt, Graz, pp. 52e165. Fischer, M.A., 1987. On the origin of Veronica persica (Scrophulariaceae) e a contribution to the history of a neophytic weed. Plant Syst. Evol. 155, 105e132. Grayer, R., Chase, M.W., Simmonds, M.S.J., 1999. A comparison between chemical and molecular characters for the determination of phylogenetic relationships among plant families: an appreciation of Hegnauer’s ‘‘Chemotaxonomie der Pflanzen’’. Biochem. Syst. Ecol. 27, 369e393. Grayer-Barkmeijer, R.J., 1979. Chemosystematic investigations in Veronica L. (Scrophulariaceae) and related genera. PhD thesis, University of Leiden. Grayer-Barkmeijer, R.J., 1973. A chemosystematic study of Veronica: iridoid glucosides. Biochem. Syst. Ecol. 1, 101e110. Grayer-Barkmeijer, R.J., 1978. Flavonoids in Parahebe and Veronica: a chemosystematic study. Biochem. Syst. Ecol. 6, 131e137. Ha¨rle, A., 1932. Die Arten und Formen der Veronica-Sektion Pseudolysimachia Koch auf Grund systematischer und experimenteller Untersuchungen. Bibl. Bot. 26, 1e86. Hartl, D., 1966e1968. Scrophulariaceae. In: Hartl, D.G.W. (Ed.), Gustav Hegi, Illustrierte Flora von Mitteleuropa. Verlag Paul Parey, Berlin Hamburg, pp. 16e469. Hong, D.-Y., 1984. Taxonomy and evolution of the Veroniceae (Scrophulariaceae) with special reference to palynology. Opera Bot. 75, 5e60. Jensen, S.R., Albach, D.C.., Ohno, T., Grayer, R.J., 2005. Veronica: iridoids and cornoside as chemotaxonomic markers. Biochem. Syst. Ecol. 33, 1031e1047. Kampny, C.M., Dengler, N.G., 1997. Evolution of flower shape in Veroniceae. Plant Syst. Evol. 205, 1e25. Lahloub, M.-F., Zaghloul, M.G., Afifi, M.S., Sticher, O., 1993. Iridoid glucosides from Veronica anagallisaquatica. Phytochemistry 33, 401e405. Lehmann, E., 1908. Geschichte und Geographie der Veronica-Gruppe Agrestis. Bull. Herb. Boissier 8, 229e660. Lehmann, E., 1910. U¨ber Merkmalseinheiten in der Veronica-Sektion Alsinebe. Z. Bot. 2, 577e602. Lehmann, E., 1937. Die Gattung Veronica in entwicklungsgeschichtlicher Betrachtung. Cytologia Fuji Jubilaei Volumen, 903e919. Lehmann, E., 1940. Polyploidie und geographische Verbreitung der Arten der Gattung Veronica. Jahrb. Wiss. Bot. 89, 461e542. Lehmann, E., 1944. Veronica filiformis Sm., eine selbststerile Pflanze. Jahrb. Wiss. Bot. 91, 395. Lehmann, E., Schmitz-Lohner, M., 1954. Entwicklung und Polyploidie in der Veronica-Gruppe Agrestis. Z. indukt. Abstammungs- Vererbungsl. 86, 1e34. Lenherr, A., Lahloub, M.F., Sticher, O., 1984. Three flavonoid glycosides containing acetylated allose from Stachys recta. Phytochemistry 23, 2343e2345. Martı´ nez-Ortega, M.M., Rico, E., 2001. Seed morphology and its systematic significance in some Veronica species (Scrophulariaceae) mainly from the western Mediterranean. Plant Syst. Evol. 228, 15e32. Martı´ nez-Ortega, M.M., Sanchez Sanchez, J., Rico, E., 2000. Palynological study of Veronica Sect. Veronica and Sect. Veronicastrum (Scrophulariaceae) and its taxonomic significance. Grana 39, 21e31. Peev, D.R., 1982. Different substitution tendencies of leaf flavones in the Veronica hederifolia group (Scrophulariaceae). Plant Syst. Evol. 140, 235e242. Quattrocchio, F., Wing, J., van der Woude, K., Souer, E., de Vetten, N., Mol, J., Koes, R., 1999. Molecular analysis of the anthocyanin2 gene of Petunia and its role in the evolution of flower color. Plant Cell 11, 1433e1444. Riek, R., 1935. Systematische und pflanzengeographische Untersuchungen in der Veronica-Sektion Chamaedrys Griseb. Repert. Spec. Nov. Regni Veg. Beih. 79, 1e68. Riek-Ha¨ußermann, C., 1943. Vergleichend-anatomische und entwicklungsgeschichtliche Untersuchungen u¨ber Samen in der Gattung Veronica. Beih. Bot. Centralbl. 62, 1e60. Ro¨mpp, H., 1928. Die Verwandtschaftsverha¨ltnisse in der Gattung Veronica. Repert. Spec. Nov. Regni Veg. Beih. 50, 1e171.

1106

D.C. Albach et al. / Biochemical Systematics and Ecology 33 (2005) 1087e1106

Sakai, K.-I., 1935. Studies on the chromomosome number in alpine-plants. II. Jap. J. Genet. 11, 68e73. Saracoglu, I., Varel, M., Harput, U.S., Nagatsu, A., 2004. Acylated flavonoids and phenol glycosides from Veronica thymoides subsp. pseudocinerea. Phytochemistry 65, 2379e2385. Schlenker, G., 1936. Systematische Untersuchungen u¨ber die Sektion Beccabunga der Gattung Veronica. Repert. Spec. Nov. Regni Veg. Beih. 90, 1e40. Shirley, B.W., 1996. Flavonoid biosynthesis: ‘new’ functions for an ‘old’ pathway. Trends Plant Sci. 1, 377e382. Sun, Y., Skinner, D.Z., Liang, G.H., Hulbert, S.H., 1994. Phylogenetic analysis of Sorghum and related taxa using internal transcribed spacers of nuclear ribosomal DNA. Theor. Appl. Genet. 89, 26e32. Swofford, D.L., 1998. PAUP) Phylogenetic Analysis using Parsimony ()and Other Methods). Sinauer Associates, Sunderland, Massachusetts. Taberlet, P., Gielly, L., Pautou, G., Bouvet, J., 1991. Universal primers for amplification of three noncoding regions of chloroplast DNA. Plant Mol. Biol. 17, 1105e1109. Taskova, R., Peev, D., Handjieva, N., Baranovska, I., 1997. A taxonomic study of some Veronica species. Phytol. Balcanica 3, 79e84. Taskova, R., Peev, D., Handjieva, N., 2002. Iridoid glucosides of genus Veronica s.l. and their systematic significance. Plant Syst. Evol. 231, 1e17. Taskova, R.M., Albach, D.C., Grayer, R.J., 2004. Phylogeny of Veronica e a combination of molecular and chemical evidence. Plant Biol. 6, 673e682. Thaler, I., 1951. Morphologisches u¨ber Veronica filiformis Smith und ihre Verwandten. Phyton 3, 216e226. Toma´s-Barbera´n, F.A., Grayer-Barkmeijer, R.J., Gil, M.I., Harborne, J.B., 1988. Distribution of 6-hydroxy-, 6-methoxy- and 8-hydroxyflavone glycosides in the Labiatae, the Scrophulariaceae and related families. Phytochemistry 27, 2631e2645. Wagstaff, S.J., Garnock-Jones, P.J., 1998. Evolution and biogeography of the Hebe complex (Scrophulariaceae) inferred from ITS sequences. N.Z. J. Bot. 36, 425e437. Watzl, B., 1910. Veronica prostrata L., V. teucrium L. und V. austriaca L. nebst einem Anhang u¨ber deren na¨chste Verwandte. Abh. zool.-botan. Ges. Wien. 5, 1e94. White, T.J., Bruns, T., Lee, S., Taylor, J., 1991. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis, M., Gelfand, D., Sninsky, J., White, T. (Eds.), PCR Protocols: A Guide to Methods and Applications. Academic Press, San Diego, pp. 315e322. Williams, J.H., Friedman, W.E., 2004. The four-celled female gametophyte of Illicium (Illiciaceae: Austrobaileyales): implication for understanding the origin and early evolution of monocots, eumagnoliids, and eudicots. Am. J. Bot. 91, 332e351. Yamazaki, T., 1968. On the genus Pseudolysimachion. J. Jap. Bot. 43, 405e412. Zufall, R.A., Rausher, M.D., 2004. Genetic changes associated with floral adaptation restrict future evolutionary potential. Nature 428, 847e850.