AOBPreview originally published online on August 7, 2008
Annals of Botany 2008 102(4):491-507; doi:10.1093/aob/mcn128
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Molecular Phylogeny of the Neotropical Genus Christensonella (Orchidaceae, Maxillariinae): Species Delimitation and Insights into Chromosome Evolution
1 Department of Botany, Biology Institute, P.O. Box 6109, University of Campinas, UNICAMP, 13083-970, Brazil
2 Departamento de Botânica, Centro de Ciências Biológicas, Universidade Federal de Pernambuco, Recife, PE, 50670-420, Brazil
3 Florida Museum of Natural History, University of Florida, Dickinson Hall, Gainesville, FL, 32611-7800, USA
4 Departamento de Botânica, IB, Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, 91501-970, Brazil
5 Department of Genetics and Evolution, Biology Institute, P.O. Box 6109, University of Campinas, UNICAMP, 13083-970, Brazil
* For correspondence. Present address: Departamento de Genética, Universidade de São Paulo/ESALQ, Caixa Postal 83, Piracicaba, SP, 13400-970, Brasil. E-mail skoehler{at}esalq.usp.br
Received: 5 October 2007 Returned for revision: 18 March 2008 Accepted: 23 June 2008 Published electronically: 7 August 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
Background and Aims: Species' boundaries applied within Christensonella have varied due to the continuous pattern of variation and mosaic distribution of diagnostic characters. The main goals of this study were to revise the species' delimitation and propose a more stable classification for this genus. In order to achieve these aims phylogenetic relationships were inferred using DNA sequence data and cytological diversity within Christensonella was examined based on chromosome counts and heterochromatin patterns. The results presented describe sets of diagnostic morphological characters that can be used for species' identification.
Methods: Phylogenetic studies were based on sequence data of nuclear and plastid regions, analysed using maximum parsimony and maximum likelihood criteria. Cytogenetic observations of mitotic cells were conducted using CMA and DAPI fluorochromes.
Key Results: Six of 21 currently accepted species were recovered. The results also support recognition of the ‘C. pumila’ clade as a single species. Molecular phylogenetic relationships within the ‘C. acicularis–C. madida’ and ‘C. ferdinandiana–C. neowiedii’ species' complexes were not resolved and require further study. Deeper relationships were incongruent between plastid and nuclear trees, but with no strong bootstrap support for either, except for the position of C. vernicosa. Cytogenetic data indicated chromosome numbers of 2n = 36, 38 and 76, and with substantial variation in the presence and location of CMA/DAPI heterochromatin bands.
Conclusions: The recognition of ten species of Christensonella is proposed according to the molecular and cytogenetic patterns observed. In addition, diagnostic morphological characters are presented for each recognized species. Banding patterns and chromosome counts suggest the occurrence of centric fusion/fission events, especially for C. ferdinandiana. The results suggest that 2n = 36 karyotypes evolved from 2n = 38 through descendent dysploidy. Patterns of heterochromatin distribution and other karyotypic data proved to be a valuable source of information to understand evolutionary patterns within Maxillariinae orchids.
Key words: Chromosome number, Christensonella, Cymbidieae, cytotaxonomy, fluorochrome staining, Maxillaria, Maxillariinae, molecular phylogenetics, species delimitation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
The subtribe Maxillariinae (sensu Whitten et al., 2000) comprises a monophyletic group of neotropical orchids with approximately 600–700 species, which are characterized by the presence of a distinct column foot and mentum, four rounded or ovoid pollinia, and a broad, open stigma (Whitten et al., 2000). Although Maxillariinae orchids form a major component of the epiphytic vegetation in the Neotropics, particularly the large and diverse genus Maxillaria, they are still taxonomically poorly known. For most species complexes within this genus there is no consensus on how many species should be recognized. In addition, available keys for species' identification are incomplete and subgeneric classifications are highly artificial.
With the goals of re-evaluating generic boundaries and defining well-supported clades within the subtribe Maxillariinae, Whitten et al. (2007) obtained sequence data of multiple regions for 604 specimens belonging to this group. One of the strongly supported clades supported in this study (represented by 61 specimens; Whitten et al., 2007) corresponds to the genus Christensonella, also known as the ‘Maxillaria madida’ group. This clade was partially recognized by Christenson (2002) as Maxillaria section Urceolatae (Table 1). Plants within section Urceolatae were characterized as densely caespitose small orchids bearing sulcate and cylindrical pseudobulbs with 2–4 needle-like leaves each, subsessile single-flowered inflorescences, flowers with an entire lip, ligulate callus and a long column with unadorned clinadrium and anther, elongated basally in a short foot (Christenson, 2002). Later, Szlachetko et al. (2006) transferred all eight species of this section Urceolatae to Christensonella. Szlachetko et al. (2006) also broadened the circumscription of Christenson (2002) to include plants with fleshy-to-coriaceous leaves varying from linear to oblong, and conduplicate to cylindrical. Nine species were transfered by Szlachetko et al. (2006), resulting in 17 species assigned to Christensonella (Table 1). Based on the molecular phylogenetic study of Whitten et al. (2007) and according to clear morphological characters, Blanco et al. (2007) transferred four additional species to Christensonella (Table 1), increasing the total number of species in this genus to 21. Species belonging to Christensonella can also be easily recognized by their whitish-to-yellow or red-to-maroon coloured flowers, generally with a shiny spot in the midlobe lip and, commonly, by the roots with annular expansions of the velamen (Blanco et al., 2007). Most species are restricted to south-eastern South America, ranging from the south of Bahia State, in Brazil, to Misiones, in Argentina. A few species, however, are restricted to Central America and western South America, occurring from Bolivia up to southern Mexico.
|
Although Christensonella is easily distinguishable from other Maxillariinae, species' identification within it has been very challenging. Species' delimitation within this genus, especially the south-eastern Brazilian species, has varied widely among taxonomic treatments, in part due to the mosaic distribution of diagnostic characters among taxa, but also because of the continuously variable nature of morphological traits (Cogniaux, 1904; Hoehne, 1953; Pabst and Dungs, 1977; Butzin and Senghas, 1996; Table 1).
In the first taxonomic treatment of the genus Maxillaria in Brazil, Cogniaux (1904) recognized 19 taxa (17 species and two new varieties), which agree with the morphological circumscription of Christensonella. Later, in several publications concerning the taxonomy of Brazilian species of Maxillaria, Hoehne (1947, 1952, 1953) described several new taxa, recognizing a total of 31 that are morphologically consistent with Christensonella (Hoehne, 1953). The diagnostic characters regarded as important by Hoehne (1953) to identify species within this group were number and shape of leaves, shape of perianth segments, and lip morphology (Table 1). Species' delimitation in Hoehne's work was clearly influenced by the typological species' concept, with many taxa being recognized as a function of high polymorphism of vegetative and flower characters among populations (Hoehne, 1953).
In their classification of Brazilian orchids, Pabst and Dungs (1977) divided the species of Maxillaria into several alliances, primarily according to vegetative traits. They grouped all the Brazilian species currently recognized as Christensonella into four different alliances based upon the number and shape of leaves (Pabst and Dungs 1977). The only exception was Maxillaria uncata [= Christensonella uncata], which was assigned to a distinct alliance based on the presence of a long rhizome and an undivided lip (Pabst and Dungs, 1977).
In the last edition of Schlechter's Die Orchideen (Butzin and Senghas, 1996), nine species of Maxillaria, currently recognized as belonging to Christensonella (Szlachetko et al., 2006), were divided into three main groups, according to leaf shape and plant architecture. The placement of the species M. madida [= C. madida] in two different groups per se reflects the high infra-specific polymorphism of vegetative characters used for assigning group membership and the difficulty of classifying taxa within this clade based solely on morphological data. Interestingly, Schlechter was the only author, apart from Szlachetko et al. (2006), to consider C. uncata (which occurs from north-western South America to Mexico) and C. nardoides (from Peru) to belong to this clade (mostly restricted to south-east South America) based on the presence of fleshy leaves, erect pseudobulbs with brownish papery bracts, and medium-size, partially closed flowers, varying from yellowish to brownish.
The available taxonomic treatments fail to provide clear boundaries for species within Christensonella, possibly because of the continuous variable nature of morphological characters among current recognized species, especially plant architecture, leaf shape and lip morphology (Cogniaux, 1904; Hoehne, 1953; Pabst and Dungs, 1977; Butzin and Senghas, 1996). Some species belonging to this group also present extremely high phenotypic plasticity of vegetative characters, as observed for some cultivated specimens, probably related to different light conditions and humidity levels (S. Koehler, pers. obs.). Such extreme morphological variation within and among putative species and infra-specific taxa resulted in morphologically undiagnosable species.
The main goals of this study were to understand patterns of diversification within the genus Christensonella in order to revise species' delimitation and provide a more stable classification for this group. To achieve such aims, we (1) inferred phylogenetic relationships within Christensonella based on sequence data from the plastid trnL intron and trnL–F intergenic spacer, the plastid matK gene, the atpB-rbcL spacer, and nuclear ribosomal internal transcribed spacers [ITS1-2] DNA regions; and (2) described sets of diagnostic morphological characters that can be used for species' identification according to the molecular patterns here obtained. Additionally, cytological diversity within Christensonella was also examined based on chromosome counts and heterochromatin patterns. Despite being one of the largest subtribes of the Orchidaceae, cytogenetic data on Maxillariinae species such as Christensonella are scarce and chromosome evolution within this genus remains poorly understood (Cabral et al., 2006; Whitten et al., 2007). Cytogenetic data are an important source of information that might help phylogenetic studies (Dobigny et al., 2004). Both the variation in chromosome numbers and the study of patterns of heterochromatin distribution have proved to be valuable tools in species' and generic delimitation (Brandham, 1999; Guerra, 2000; Brasileiro-Vidal et al., 2007), including within the subtribe Maxillariinae (Cabral et al., 2006).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
Since the monophyly of Christensonella was confirmed by Whitten et al. (2007), only two outgroup species, Maxillaria crocea and M. ochroleuca, were included in this study for tree rooting purposes (although the sister group to Christensonella is not resolved; Whitten et al., 2007). Sampling included 48 specimens of the ingroup with at least two individuals of all species except for C. cogniauxiana and Christensonella sp. Whitten 2310 (an undescribed species). The study was based on a total of 182 sequences, 127 of which were already published as part of a study of generic delimitation of the subtribe Maxillariinae (Whitten et al., 2007; Table 2). Sequence data was added from the plastid trnL intron and trnL–F intergenic spacer and complementary data of the other three regions was obtained for additional specimens (Table 2). All samples were vouchered as herbarium specimens, including the ones used for cytogenetic studies (Tables 2 and 3). Samples Koehler 73, Koehler 79, Koehler 91 and Koehler 240 were excluded from the ITS matrix, as were samples Koehler 173, Koehler 243 and Whitten 951 from the plastid matrix due to the poor quality of the data.
|
|
We followed the species' criteria of Hoehne (1953) for preliminary identification and discussion of species limits within Christensonella, since it comprises the most recent taxonomic treatment available for the majority of species belonging to this clade. Three species recognized by Hoehne (1953), Maxillaria heterophylla, M. mosenii and M. plebeja, have not been transferred to Christensonella yet and therefore are treated here as Maxillaria. We based our re-evaluation of species' limits on well-supported monophyletic groups of specimens that could be identified by sets of morphological and/or cytological characters.
Laboratory protocols
DNA was extracted from fresh plant tissues (leaves and flowers) from plants available in cultivation according to Doyle and Doyle (1987) and scaled down to 1-mL extraction volumes following the protocol described by Whitten et al. (2000), except that all total DNA extracts were purified with QIAquick columns (Qiagen Inc.) prior to amplification. Amplification was performed in 25–50-mL reactions, with 2·5 mM MgCl2, and Sigma buffers (Sigma Inc.), 0·2–0·4 mM of each primer, 1 U of Taq polymerase and 50–300 ng of template. In all ITS amplifications, betaine (Sigma Inc.) was added (1·0 mM final concentration) to the PCR mix to relax secondary structure. Amplification and sequencing primers used for ITS and trnL–F regions are those of Sun et al. (1994) and Taberlet et al. (1991), respectively. Some amplifications for trnL–F using primers C and F produced multiple bands; for these samples, the region was amplified in two separate reactions using primer pairs C + D and E + F. For the amplification of matK and atpB–rbcL spacer regions, primers specifically designed for Maxillariinae orchids were used (Whitten et al., 2007). The matK + trnK intron region was usually amplified as a single piece, using the primers -19F (Goldman et al., 2001) and trnK2R (Johnson and Soltis, 1994); primers 308F and 1100F (Whitten et al., 2007) were used as additional internal sequencing primers. Some taxa were amplified using the primers 56F and 1520R that yielded a shorter but nearly complete portion of matK (Whitten et al., 2007). The atpB–rbcL intergenic spacer was amplified with the primers Max F and Max R (Whitten et al., 2007). Protocols for the amplification reactions were as follows. ITS: 10 min initial denaturation at 99 °C, 30 cycles of 94 °C denaturation for 45 s, 60 °C annealing for 45 s, 72 °C extension for 1 min; trnL-F: 32 cycles of 94 °C denaturation for 30 s, 61 °C annealing for 30 s, 72 °C extension for 75 s; matK and atpB–rbcL spacer: 33 cycles of 94 °C denaturation for 45 s, 60 °C annealing for 45 s, 72 °C extension for 2 min. Amplified products were purified with QIAquick PCR cleaning column and filtration kit (Qiagen Inc.) and directly sequenced on Applied Biosystems, Inc (ABI) 373/377 or 3100/3500 automated sequencers using standard dye-terminator according to the manufacturer's protocols, except that cycle sequencing reactions were scaled down to 5 mL. Both strands were sequenced to assure accuracy in base calling.
Data analysis
Alignment
The software packages ‘Sequence NavigatorTM’, ‘AutoassemblerTM’ (ABI) and ‘SequencherTM’ (Gene Codes Corporation) were used to edit and assemble complementary and overlapping sequences. Each individual base position was examined for agreement of the two strands. DNA sequences were aligned manually using Se-Al (Rambaut, 2002), and gaps were coded as missing values. Terminal priming regions were excluded, as were regions where alignment was ambiguous or where extensive length variation occurred. All the aligned matrices are available upon request from S. Koehler.
Data exploration
The null hypothesis of base frequency stationary among sequences was evaluated using the chi-square heterogeneity test as implemented in PAUP* 4·0b10 (Swofford, 2000). The g1 statistic was used to determine if the phylogenetic signal was significantly non-random (Hillis and Huelsenbeck, 1992). The left skew of tree distributions was obtained in PAUP* based on 10 000 randomly generated trees. Possible incongruence between nuclear and chloroplast genomes was assessed with the incongruence length difference (ILD) test (Farris et al., 1994), implemented in PAUP* as the partition homogeneity test using 1000 replicates and excluding uninformative characters to avoid over-estimation of the amount of incongruence (Lee, 2001).
Phylogenetic analyses
Phylogenetic analyses were initially conducted with a heuristic search under the maximum parsimony (MP) criterion of Fitch (unordered characters, equal weights to all changes; Fitch, 1971), excluding uninformative characters, and with ACCTRAN optimization. The search strategy for all data sets used 10 000 addition sequence replicates by stepwise addition holding ten trees per replicate, TBR branch swapping on best trees, MULTREES on, saving no more than ten optimal trees per replicate. To assess support for internal clades we performed 1000 bootstrap pseudo-replicates (Felsenstein, 1985) of ten addition sequence replicates by stepwise addition holding one tree per replicate. The categories of bootstrap support considered in this study were: unsupported (<50 %); weak (50–74 %); moderate (75–84 %); strong (85–100 %) (Whitten et al., 2000). Since simulation experiments have shown that high levels of homoplasy can decrease the accuracy of phylogenetic inference under the parsimony criterion (Huelsenbeck and Hillis, 1993), we also employed the successive weighting strategy (SW) for maximum parsimony analyses (Farris, 1969; Carpenter, 1994). Optimization of successive weighting analyses was carried out considering 1000 addition sequence replicates and SPR branch swapping with characters being reweighed according to the rescaled consistency index until tree scores were not improved. Then, a final analysis considering the same search strategy applied to the unweighted data was conducted (10 000 addition sequence replicates, TBR branch swapping).
For maximum likelihood analyses (ML), alternative nested models of DNA sequence evolution were first evaluated with likelihood ratio tests as implemented in MODELTEST 3·7 (Posada and Crandall, 1998; 
= 0·01). The best-fit model of DNA sequence evolution with its estimated parameters was then input into detailed maximum-likelihood tree searches performed in PAUP*. Starting trees were obtained using ten addition sequence replicates by stepwise addition holding 1 tree per replicate, with further SPR branch swapping. Starting branch lengths were obtained using the Rogers–Swofford approximation method with the branch-length optimization of Newton–Raphson. Confidence of the ML trees obtained was assessed by bootstrap analyses based on 100 pseudo-replicates using the fast reduced search option in PAUP*.
Cytogenetic studies
Root tips were collected and pretreated in 0·002 M 8-hydroxyquinoline for 20 h at 8 °C. Samples were then fixed in Carnoy's solution (ethanol/glacial acetic acid, 3:1, v/v) for 2 h at room temperature and stored at –20 °C. Root tips were digested with 2 % cellulase and 20 % pectinase (both w/v) for 90 min at 37 °C. The meristem was subsequently isolated and squashed in 45 % acetic acid. After removing the cover-slip the slides were air-dried and aged for 3 d at room temperature.
The aged slides were double stained according to Schweizer and Ambros (1994) with CMA (0·5 mg mL–1, 1 h) and DAPI (2 µg mL–1, 30 min), and mounted in McIlvaine's (pH 7·0) buffer–glycerol (1 : 1, v/v) containing 2·5 mM MgCl2. After being aged at room temperature for at least 3 d for fluorochrome stabilization, the best slides were analysed using a Leica DMLB microscope and the cell images were captured with a COHU digital camera using the QFISH software (Leica).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
Phylogenetic analyses
The test for phylogenetic signal based on random-tree distributions showed that all data sets contain significant phylogenetic information (g1 values: ITS = –0·40, plastid = –0·55). We were unable to detect any significant heterogeneity in base frequencies among taxa using the chi-square heterogeneity test for the ITS data set (P = 0·99). However, the chi-square test rejected the hypothesis of base frequency homogeneity for the plastid data set (P = 0·004). The null hypothesis of congruence between the nuclear and plastid data sets was strongly rejected by the ILD test (P = 0·001). As visual inspection of nuclear and plastid topologies also indicated they are discordant, we chose not to discuss the results based on a single combined analysis (but see Discussion, below). Tree statistics of MP and SW analyses are summarized in Table 4. Bootstrap analyses were conducted only for unweighted data.
|
ITS
The MP and ML statistics for the ITS analyses are given in Tables 4 and 5, respectively. Most traditionally recognized species (Hoehne, 1953; Pabst and Dungs, 1977) are not monophyletic according to the ITS tree (Fig. 1). Five morphologically distinct species are strongly supported as monophyletic: Christensonella echinophyta, C. nardoides, C. pachyphylla, C. uncata and C. vernicosa. Two species' complexes, composed of species with extremely similar morphology, also emerged as monophyletic groups from this analysis: the ‘Christensonella acicularis–C. madida–Maxillaria mosenii’ clade and the ‘Christensonella minuta–C. pumila–Maxillaria plebeja’ clade: these are designated here as ‘C. acicularis–C. madida’ (Fig. 1A) and ‘C. pumila’ (Fig. 1D), respectively. The clade Christensonella cogniauxiana–C. ferdinandiana–C. juergensii–Maxillaria heterophylla–C. neowiedii (= ‘C. ferdinandiana–C. neowiedii’ clade, Fig. 1C), comprises two morphologically distinct groups, the species C. ferdinandiana and the ‘C. cogniauxiana–C. juergensii–M. heterophylla–C. neowiedii’ group. The currently accepted species C. pacholskii also appears as a monophyletic group in the ITS analysis, but with weak bootstrap support (Fig. 1I). Other strongly supported clades obtained from both MP and ML analyses of ITS data were (1) the ‘C. acicularis–C. madida’ clade + C. nardoides + Christensonella sp; (2) C. uncata + C. vernicosa + C. pacholskii; and (3) C. echinophyta + C. pachyphylla + the ‘C. pumila’ clade + the ‘C. ferdinandiana–C. neowiedii’ clade (Fig. 1). Two clades were only recovered in the ML and SW analyses: ‘C. acicularis–C. madida’ clade + C. nardoides and (C. pachyphylla, ‘C. ferdinandiana–C. neowiedii’ clade, C. echinophyta, ‘C. pumila’ clade) + (C. uncata, C. vernicosa, C. pacholskii)) (Fig. 1).
|
|
Plastid
The MP and ML statistics for the plastid analyses are given in Tables 4 and 5, respectively. Several clades with bootstrap support greater than 80 % in common with the ITS data set were recovered by MP, SW and ML analyses of plastid data (Fig. 2): the ‘C. acicularis–C. madida’ clade (Fig. 2A), the ‘C. pumila’ (Fig. 2D) clade, C. echinophyta (Fig. 2E), C. vernicosa (Fig. 2H), C. pacholskii (Fig. 2I) and C. uncata (Fig. 2G). Although the clade ‘C. ferdinandiana–C. neowiedii’ was also recovered by the MP, SW and ML strict consensus trees, it was not supported in the bootstrap consensus tree (Fig. 2C).
|
Despite the many terminal clades in common with the ITS data set, deeper nodes in the ML, MP and SW plastid trees were incongruent with those based on ITS data (Figs 1 and 2). The plastid trees did not support the clade ‘C. uncata + C. vernicosa + C. pacholskii’ as sister to the clade ‘C. pumila’ + ‘C. juergensii–C. ferdinandiana’ + C. echinophyta + C. pachyphylla. Instead, the latter is indicated as sister to the ‘C. acicularis–C. madida’ clade, with C. vernicosa embedded in it, although none of these alternatives received bootstrap support greater than 50 % (nodes collapsed). The plastid tree also supports C. nardoides as sister to Christensonella sp., whereas ITS data supports the former as sister to the ‘C. acicularis–C. madida’ clade. However, none of these clades received bootstrap support higher than 50 %: MP analyses of plastid data did not even support Christensonella sp. and C. nardoides as sister to the ‘C. acicularis–C. madida’ clade (Fig. 2).
Cytogenetic studies
Cytogenetic data for the species analysed in the present study are summarized in Table 3. Three different chromosome numbers were found: 2n = 36 (C. ferdinandiana, C. heterophylla, C. pachyphylla, C. pumila); 2n = 38 (C. acicularis, C. madida, M. mosenii), and 2n = 76 (M. mosenii var. echinochila) (Figs 3, 4). All species studied revealed symmetric karyotypes with predominance of acrocentric chromosomes, except for the species with 2n = 36, which also showed at least two metacentric or submetacentric chromosomes (Fig. 3).
|
CMA+ bands varied in number and were mainly terminal or sub-terminal on the long arm of acrocentric chromosome pairs, while DAPI+ bands varied in number but were always proximal to the centromere (Table 3). Heteromorphism of CMA+ bands was observed in all species with 2n = 38 (‘C. acicularis–C. madida’ clade) and in C. pumila, belonging to the ‘C. pumila’ clade. Species with 2n = 36 had fewer CMA+ and DAPI+ bands than species with 2n = 38. There were 2–3 CMA+ blocks and 4–16 DAPI+ bands in species with 2n = 36 (Fig. 3), whereas the 2n = 38 species (Fig. 4) exhibited 3–8 CMA+ bands and 14–20 DAPI+ bands (up to 32 bands in the tetraploid individuals; Fig. 4B). Among 2n = 36 species there were always DAPI+ bands present in at least one meta- or sub-metacentric pair. On the other hand, species with 2n = 38 displayed no meta- or sub-metacentric chromosome with bands and showed both heterochromatin types, CMA+ and DAPI+, in some chromosome pairs (arrowheads in Fig. 4C). The species C. ferdinandiana, with 2n = 36, was remarkable for bearing only one chromosome pair with duplicated DAPI+ bands, presumably each one at one side of the centromere (Fig. 3A).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
Nuclear vs. plastid incongruence
Incongruence among different data partitions comprises a rather complex subject in phylogenetic systematics that has received increasing attention over the last decades (Bull et al., 1993; Huelsenbeck et al., 1996; Cunningham, 1997; Reed and Sperling, 1999). Several studies have demonstrated that incongruence may be caused by distinct categories of bias, namely random and systematic errors and independent evolutionary histories of partitions (Swofford et al., 1996; Reed and Sperling, 1999). In this study, we attempted to reduce random error by broad sampling of taxa as well as different genome regions. Sampling efforts were mainly centred on highly polymorphic species groups (the ‘C. pumila’, ‘C. ferdinandiana–C. neowiedii’ and ‘C. acicularis–C. madida’ clades) and bootstrap analyses were performed to assess confidence in the results obtained. Despite the sampling efforts, absence of sufficient phylogenetic signal is likely to be the problem in deep levels of the recovered trees. While the ITS trees (MP and ML) suggest the (C. pacholskii + C. vernicosa + C. uncata) clade to be sister to the (C. pachyphylla + C. echinophyta + ‘C. pumila’ + ‘C. juergensii–C. ferdinandiana’) clade, the ML plastid topology indicates the latter (including C. vernicosa) as sister to the (‘C. acicularis–C. madida’ + C. nardoides + Christensonella sp.) clade. None of these alternatives, however, had bootstrap values greater than 50 % (Figs 1 and 2), suggesting absence of sufficient phylogenetic signal. Combined analyses (results not shown), considering both MP and ML criteria, support phylogenetic patterns indicated by the nuclear topology, but also with no bootstrap support. Soft incongruence also seems to be the reason for low resolution within and between terminal clades.
The results also suggest the occurrence of strong genealogical discordance concerning the position of C. vernicosa, since alternative placements of this species in nuclear and plastid trees under all search criteria received high bootstrap values (Figs 1 and 2). Four additional MP analyses were performed excluding, one at a time, outgroups as well as all samples of C. vernicosa, C. echinophyta and C. uncata plus C. pacholskii to check for possible occurrence of long-branch attraction artefacts (Siddall and Whiting, 1999), but the trees obtained did not result in any distinct topologies (results not shown). Both combined analyses (MP and ML criteria) supported the position of C. vernicosa as sister to (C. pachyphylla + C. echinophyta + ‘C. pumila’ clade + ‘C. ferdinandiana–C. neowiedii’ clade) with bootstrap values of 90 % and 51 %, respectively (tree not shown).
Hybridization and introgression represent potential causes of incongruence among phylogenetic trees (Mansion et al., 2005; Buckley et al., 2006; but see Wolfe and Randle, 2004 for additional causes of incongruence). The occurrence of recent and rapid divergence of species, sympatric populations and generalist pollinators in this group (S. Koehler, unpubl. res.) certainly reinforces the likelihood of reticulation events. Presumed hybrid individuals from natural populations bearing intermediate phenotypes have been reported for Christensonella (Hoehne, 1953; Onishi, 1974), although such a scenario remains to be demonstrated for C. vernicosa. Both lineage-sorting and reticulation processes can result in similar phylogenetic patterns (Holder et al., 2001). Further studies, considering comparative analyses of a large number of uni- and biparental inherited markers and detailed geographical sampling within and between species (e.g. Comes and Abbott, 2001) are necessary to assess the role of each process in the diversification of Christensonella.
Species' delimitation
Despite of the incongruence among topologies based on chloroplast and nuclear data sets, there was consensus concerning delimitation of monophyletic species defined by molecular data and morphological/cytological characters within Christensonella. Table 6 gives the current species' delimitation used to identify species for this study and the species' concepts proposed here (also indicated in Fig. 1), with diagnostic morphological characters indicated. Chromosome numbers are indicated in Figs 1 and 2. All further nomenclatural rearrangements suggested will be presented in a forthcoming taxonomic revision of the genus (S. Koehler, currently in preparation).
|
The molecular data presented here support the current delimitations of the species (Figs 1 and 2, Table 6): C. nardoides (clade B), C. echinophyta (clade E), C. pachyphylla (clade F), C. uncata (clade G), C. vernicosa (clade H) and C. pacholskii (clade I), which are also easily characterized by sets of diagnostic morphological characters (Table 6). Within species from south-eastern Brazil, C. echinophyta (clade E, Table 6) and C. vernicosa (clade H, Table 6) comprise plants up to 5 cm tall with bifoliate pseudobulbs bearing needle-like leaves and flowers with pedicels always longer than the adjacent pseudobulb. Christensonella echinophyta has white-pinkish flowers with elongate segments, while C. vernicosa has yellow flowers with ovate-oblanceolate segments (Barbosa Rodrigues, 1996). Another currently recognized species from south-eastern Brazil corroborated by our results is C. pachyphylla (clade F, Table 6), a species growing up to 25 cm tall. It is distinguished by its unifoliate pseudobulbs bearing a thick, lanceolate leaf and pale-yellow flowers with pedicels shorter than the adjacent pseudobulbs.
The results indicate that the Amazonian species included in the genus Christensonella belong to two distinct subclades (Figs 1 and 2). One subclade consists of C. nardoides (clade B) and Christensonella sp. (Whitten 2310, an undescribed species from Peru; photograph in Butzin and Senghas, 1996, as Maxillaria paulistana Hoehne). Christensonella nardoides is easily recognized by its 3–4 needle-like leaves on each pseudobulb and brownish to red-purplish flowers with a short viscidium (Bennett and Christenson, 1993). Christensonella sp. is morphologically very similar to some populations of C. madida (= M. madida var. monophylla Cogn.), being a large plant with big, reddish flowers and unifoliate pseudobulbs.
The other north-western South American clade indicated by our results includes C. uncata (clade G, Table 6) and C. pacholskii (clade I, Table 6), which are morphologically very similar, bearing unifoliate pseudobulbs and elongated flowers with an extremely long stipe (up to 25 mm). They are easily distinguished by the colour of flowers and leaf morphology. Christensonella pacholskii can be recognized by its membranaceous, linear leaves and red-brownish flowers (Christenson, 2003), whereas the widespread, vegetatively highly polymorphic C. uncata is easily identified by its fleshy-to-coriaceous leaves and white-to-lavender flowers (Atwood and Mora de Retana, 1999). Variation within C. uncata deserves further attention, and possibly two or more taxa are currently embedded in the current C. uncata species concept, as was suggested by Atwood and Mora de Retana (1999).
Three clades also emerged as well-supported groups of species. These are the ‘C. acicularis–C. madida’ clade (clade A, Table 6), the ‘C. ferdinandiana–C. neowiedii’ clade (clade C) and the ‘C. pumila’ clade (clade D) (Figs 1 and 2). They all include species from south-eastern Brazil that have been shown to be very difficult to distinguish, since they are defined by continuously variable morphological characters. The ‘C. pumila’ clade (clade D, Table 6) currently comprises at least three species, as demonstrated by our analyses (C. pumila, M. plebeja and C. minuta; Figs 1 and 2). Hoehne (1953) distinguished them based on the size and shape of leaves and on the shape and colour of flowers, but overlapping of morphological diagnostic characters is evident as soon as one attempts to identify specimens within this group. In addition, such taxa are neither geographically nor ecologically isolated, being restricted to the humid and seasonally dry forests of south-eastern Brazil. Thus, morphological variation is not reflected in geographic distribution or habitat variation, reinforcing the recognition of a single, polymorphic species for this group, namely C. pumila. Despite being highly polymorphic, C. pumila can be easily distinguished from other species of Christensonella by its pseudobulbs bearing one coriaceous-to-fleshy leaf and small yellowish-red flowers with pedicels always shorter than the adjacent pseudobulb.
Current species' limits within the ‘C. acicularis–C. madida’ clade (clade A, Figs 1 and 2, Table 6) are also blurred by continuous variation of leaf shape and flower size. Plants within this clade can be distinguished from others within Christensonella as larger plants up to 30 cm tall, generally with reddish-brown flowers with pedicels always shorter than the adjacent pseudobulb, chromosome numbers of 2n = 38, 76 with generally more CMA+ and DAPI+ (Table 3, Fig. 3). Most flowers produce a strong, fruity, watermelon-like fragrance that is very distinctive. However, in contrast to the ‘C. pumila’ clade, morphological variation within ‘C. acicularis–C. madida’ is restricted geographically and ecologically. Christensonella acicularis has traditionally been described as a more delicate species with bifoliate pseudobulbs, needle-like leaves and smaller flowers, and is restricted to forested habitats of south-eastern Brazil, while C. madida and M. mosenii correspond to more robust plants with larger pseudobulbs and flowers, with 1–2 leaves varying from linear-lanceolate to cylindrical. Tetraploidy was observed in two accessions of M. mosenii var. echinochila (2n = 76), an ecomorphotype restricted to rock outcrop formations of south-eastern Brazil. Habitat differentiation could have contributed to the initial establishment of polyploids, as autotetraploidization may be caused by high rates of formation of unreduced gametes, induced by harsh environments (Ramsey and Schemske, 1998; Soltis et al., 2003). This result suggests polyploidization may have had an important role in the diversification of this group, as already demonstrated for other orchid species (Del Prete et al., 1991; D'Emerico et al., 2002).
Species' boundaries among C. madida, M. mosenii and C. acicularis have never been questioned, despite the fact there are at least six names currently available for this complex. The sequence data presented here are not informative enough to allow elucidation of the patterns of diversity within this clade. CMA/DAPI banding patterns, in general, support ecological and geographical subdivisions within this group (Table 3, Fig. 3). Further studies, utilizing more informative molecular markers at population levels and more detailed morphological and cytogenetic data, are necessary before any taxonomic change is proposed.
Another highly polymorphic well-supported clade in our analyses is the ‘C. ferdinandiana–C. neowiedii’ clade (clade C, Figs 1 and 2, Table 6). Contrary to the other two species' complexes discussed above, there are no diagnostic morphological characters for this one. This clade comprises two morphologically very distinct groups, the currently accepted species C. ferdinandiana and the C. cogniauxiana–C. juergensii–M. heterophylla–C. neowiedii species' complex (or the ‘C. neowiedii’ complex). The DNA sequence data do not support these two groups as distinct clades. However, C. ferdinandiana is morphologically very distinct from other taxa in this clade, being easily characterized by its unifoliate and flat pseudobulbs, unique in the ‘C. madida’ complex, and by its yellow flowers with pedicels always shorter than the adjacent pseudobulb. Another distinctive feature of this species is the presence of a single large metacentric chromosome pair with duplicated proximal DAPI+ bands (Fig. 3A). Species of the ‘C. neowiedii’ complex can be distinguished from C. ferdinandiana by its bifoliate, cylindrical pseudobulbs and generally reddish-to-dark-purple flowers, with pedicles always longer than the adjacent pseudobulb. In contrast to C. ferdinandiana, species' delimitation within the ‘C. neowiedii’ complex is extremely unclear and vague, since diagnostic characters once again vary continuously among taxa. Such intricate morphological variation is well illustrated by the fact that it is possible to observe in the same locality, such as Campos de Jordão (São Paulo State, Brazil), individuals with flat, lanceolate leaves and reddish, smaller flowers blooming together with needle-leaved specimens with larger, dark-purple flowers, with intermediate morphotypes growing between them.
One of the problems with the phylogenetic species' concept is dealing with recent diverged lineages, since they usually are not reciprocally monophyletic (Coyne and Orr, 2004). Clearly the molecular markers used in this study did not present enough variation to distinguish between C. ferdinandiana and the ‘C. neowiedii’ group, despite the fact that morphology and cytogenetic data do separate them. Considering our present lack of knowledge of phylogenetic patterns within this group, we suggest a conservative approach for species' delimitation for the ‘C. ferdinandiana–C. neowiedii’ clade based on morphological characters. Although our results do not support any of these entities as monophyletic, morphological characters can be used, at least in a first instance, to distinguish C. ferdinandiana and a broad ‘C. neowiedii’, as described above.
Cytogenetic patterns of diversification
Cytogenetic data gathered for Christensonella show two general patterns. The species C. ferdinandiana, C. pachyphylla, C. pumila and M. heterophylla are characterized by 2n = 36, few CMA+ bands and occurrence of sub-meta- and metacentric chromosomes, while C. acicularis, C. madida and M. mosenii have karyotypes with 2n = 38, more CMA+ bands and an apparent lack of sub-meta- and metacentric chromosomes. While 2n = 38, 40 has been shown to occur in most of the taxa sampled for the core subtribe Maxillariinae (sensu Whitten et al., 2007), the occurrence of chromosome numbers of 2n = 36 is much more restricted in this group. Only 12 out of the 68 taxa sampled for chromosome numbers in the core Maxillariinae have 2n = 36 or less: 2n = 28 for Cryptocentrum standleyi; 2n = 30 for C. lehmanii; 2n = 32 for M. arachnitiflora, 2n = 34 for M. fulgens, M. hedwigae, M. notylioglossa and M. rufescens; and 2n = 36 for M. barbosae, M. bicallosa, M. desvauxiana, M. microdendron and M. cf. luteoalba (Blumenschein and Paker, 1963; Carnevali 1991; Carnevali Fernandez-Concha, 1996; Brandham, 1999; Felix and Guerra, 2000; Whitten et al., 2007).
Disploidy has already been indicated as a mechanism of karyotype evolution in other groups of orchids, as in the section Fimbriatae of Lycaste (Ryan et al., 2000) and in the genera Cephalanthera (Schwarzacher and Schweizer, 1982) and Paphiopedilum (Cox et al., 1998). The probable mechanism involved in the dysploid differentiation of chromosome numbers in Christensonella is centric fusion or fission, as suggested by the presence of a single, large metacentric chromosome pair with duplicated proximal DAPI+ bands in C. ferdinandiana. The fact that acro- or telocentric chromosomes with proximal DAPI+ bands have been observed in the species studied with 2n = 38 and that all the sampled species with 2n = 36 had at least one sub-meta- or metacentric chromosome pair with a proximal DAPI+ band, not found in any sampled species with 2n = 38, suggests the occurrence of fusion/fission changes.
Felix and Guerra (2000) had earlier suggested that the probable chromosome base number for Maxillaria is x2 = 20, which places the numbers 2n = 36, 38 as evolving from a sequence of descendent disploidy events from 2n = 40. In this scenario, the ancestral condition for chromosome numbers in Christensonella would be 2n = 38 and the ‘C. acicularis–C. madida’ clade would have conserved the plesiomorphic state of chromosome number in the genus. Moreover, the common ancestor of species with 2n = 36 would have experienced further descendent disploidy due to centric fusion resulting in at least one sub-meta- or metacentric chromosome pair. Preliminary data on chromosome counts of C. uncata confirmed the occurrence of 2n = 36 for this species (J. S. Cabral, unpubl. res.). This information supports the phylogenetic tree based on nuclear (and combined) sequence data, as it places M. uncata as the sister clade of the other species with 2n = 36 (Fig. 1).
Nevertheless, little is still known about patterns of karyotypic evolution in the Maxillariinae subtribe. Only 68 taxa of the 354 species recognized for the core subtribe Maxillariinae have been sampled for chromosome numbers (see review in Whitten et al., 2007). The evolution of chromosome numbers in Christensonella could not be fully assessed by the phylogenetic and karyotypic data available. The sister group of Christensonella is still unclear (Whitten et al., 2007) and plastid and nuclear topologies presented here indicate conflicting patterns of diversification within this group. Further cytogenetic studies, considering additional samples within Maxillariinae as well as complementary phylogenetic data, are necessary for an accurate inference of the evolution of cytogenetic patterns in Christensonella.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
DNA sequence data and cytological analyses were used to investigate species' boundaries within the neotropical Christensonella. Six currently accepted species were recovered by the phylogenetic analyses presented here (C. echinophyta, C. nardoides, C. pachyphylla, C. pacholskii, C. vernicosa and C. uncata). Our results also support the recognition of the ‘C. pumila’ clade as a single, polymorphic species based on diagnostic morphological characters as well as by an overlapping geographic distribution of current species recognized within this clade. Two additional clades, the ‘C. acicularis–C. madida’ clade and the ‘C. ferdinandiana–C. neowiedii’ clade, demand further investigation since patterns of diversification remained obscured within both groups. For now, we propose the recognition of a broadly defined C. acicularis for the ‘C. acicularis–C. madida’ clade and two morphological complexes within the ‘C. ferdinandiana–C. neowiedii’ clade (namely C. ferdinandiana and C. neowiedii s.l.) based on diagnostic morphological features described above. Complementary studies considering more populations as well as data from different molecular markers are already being developed in order to better understand diversification patterns within these clades. Patterns of heterochromatin distribution as well as karyotypic data certainly deserve greater attention as a valuable complementary source of information to understand evolutionary patterns within Maxillariinae, as well as to assist in the revision of species/generic boundaries within this subtribe. It is clear that speciation within Orchidaceae is a complex issue that must be explored in greater detail. The question of how such a vast range of morphological diversity was shaped and what is the role of ecology behind the formation of new species remains a fascinating subject to be further explored in this group of plants.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
We sincerely thank the institutions that provided specimens for study, as well for all the help of their staff: Instituto de Botânica de São Paulo, Escola Superior de Agricultura ‘Luiz de Queiroz’ and Ecuagenera Ltd.; A. P. de Moraes, J. Aranha, L. Kollmann, R. Custódio, R. Farias-Singer, R. Gonçalves for all the lab assistance and/or help on fieldwork; A. dos Santos, A. O. Simões, C. van den Berg, F. Barros, G. Machado, J. Joseph, J. Whitton, P. Moraes, V. Bittrich and two anonymous referees for valuable discussions and insights on phylogenetic analyses and reading of this manuscript. Plants were collected under the permit No. 016–IC–FLO–DBAPMA from the Ministerio del Ambiente of Ecuador. Funding for this work was provided by Fundação de Amparo à Pesquisa do Estado de São Paulo (02/02161-7 and 06/58983-6 to S.K.; 01/08958-1 to R.B.S.; 02/02162-3 to M.C.E.A.); Coordenação de Aperfeiçoamento de Pessoal de Nível Superior to J.S.C.; Conselho Nacional de Desenvolvimento Científico e Tecnológico to M.C.E.A. and M.G.; Prance Fellowship in Neotropical Botany from the Kew Latin American Reasearch Fellowship Programme to S.K.; US National Science Foundation Project ‘Systematics of Maxillariinae (Orchidaceae): generic delimitation, pollinator rewards, and pollination’ (DEB-234064 to N.H.W. and W.M.W.).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSLITERATURE CITED |
|---|
-
Atwood JT, Mora de Retana DE. Flora Costaricencis: Orchidaceae subtribes Maxillarinae and Oncidiinae. Fieldiana, Botany n.s. (1999) 40:1–182.
Barbosa Rodrigues J. Maxillaria echinophyta Barb. Rodr. In: Iconographie des Orchidées du Brésil. Vol. 1: The Illustrations.—Sprunger S, ed. (1996) Basle: Friedrich Reinhardt Verlag.
Bennett DE, Christenson EA. Plate 107: Maxillaria nardoides Kraenzl. In: Icones Orchidacearum Peruvianum—Christenson EA, ed. (1993) Lima, Peru: Privately published.
Blanco MA, Carnevali G, Whitten WM, Singer RB, Koehler S, Williams NH, et al. Generic realignments in Maxillariinae (Orchidaceae). Lankesteriana (2007) 7:515–537.
Blumenschein A, Paker IU. Número de cromossomas nas Maxillariinae (Orchidaceae). Ciência e Cultura (1963) 15:255.
Brandham P. Cytogenetics. In: Genera Orchidacearum. Vol I.—Pridgeon AM, Cribb PJ, Chase MW, Rasmussen FN, eds. (1999) New York: Oxford University Press. 67–80.
Brasileiro-Vidal AC, dos Santos-Serejo JA, Soares Filho W dos S, Guerra M. A simple chromosomal marker can reliably distinguish Poncirus from Citrus species. Genetica (2007) 129:273–279.[CrossRef][Web of Science][Medline]
Buckley TR, Cordeiro M, Marshall DC, Simon C. Differentiating between hypotheses of lineage sorting and introgression in New Zealand alpine cicadas (Maoricicada Dugdale). Systematic Biology (2006) 55:411–425.[CrossRef][Web of Science][Medline]
Bull JJ, Huelsenbeck JP, Cunningham CW, Swofford DL, Waddell PJ. Partitioning and combining data in phylogenetic analysis. Systematic Biology (1993) 42:384–397.
Butzin F, Senghas K. Subtribus: Maxillariinae. In: Rudolf Schlechter, Die Orchideen.—Brieger FG, Maatsch R, Senghas K, eds. (1996) Berlin, Wien: Blackwell Wissenschafts-Verlag. 1727–1792.
Cabral JS, Felix LP, Guerra M. Heterochromatin diversity and its co-localization with 5S and 45S rDNA sites in chromosomes of four Maxillaria species (Orchidaceae). Genetics and Molecular Biology (2006) 29:659–664.[Web of Science]
Carnevali G. The cytology and the pollinaria in the Maxillariinae (Orchidaceae) (1991) Missouri, USA: University of Missouri, St. Louis. M.S. thesis.
Carnevali Fernandez-Concha G. Systematics, phylogeny and twig epiphytism in Cryptocentrum (Orchidaceae). (1996) Missouri, USA: University of Missouri, St. Louis. Ph.D. thesis.
Carpenter JM. Successive weighting, reliability and evidence. Cladistics (1994) 10:215–220.[CrossRef][Web of Science]
Christenson EA. Vue d'ensemble du genere Maxillaria. Richardiana (2002) 2:41–65.
Christenson EA. Two Maxillarias, one old and one new. Orchid Review (2003) 111:287–289.
Cogniaux A. Orchidaceae III. Maxillarinae. In: Flora Brasiliensis—von Martius CFP, Eichler AG, Urban I, eds. (1904) Vol. 3(6). Lipsiae: Frid. Fleischer. 1–78.
Comes HP, Abbott RJ. Molecular phylogeography, reticulation, and lineage sorting in Mediterranean Senecio sect. Senecio (Asteraceae). Evolution (2001) 55:1943–1962.[Web of Science][Medline]
Cox AV, Abdelnour GJ, Bennett MD, Leitch IJ. Genome size and karyotype evolution in the slipper orchids (Cypripedioideae: Orchidaceae). American Journal of Botany (1998) 85:681–687.[Abstract]
Coyne JA, Orr HA. Speciation (2004) Sunderland, MA: Sinauer Associates. 459–470.
Cunningham CW. Is congruence between data partitions a reliable predictor of phylogenetic accuracy? Systematic Biology (1997) 46:464–478.[CrossRef][Web of Science][Medline]
D'Emerico S, Cozzolino S, Pellegrino G, Pignone D, Scrugli A. Karyotype structure, supernumerary chromosomes and heterochromatin distribution suggest a pathway of karyotype evolution in Dactylorhiza (Orchidaceae). Botanical Journal of the Linnean Society (2002) 138:85–91.[CrossRef][Web of Science]
Del Prete C, Mazzola P, Miceli P. Karyological differentiation and speciation in C. Mediterranean Anacamptis (Orchidaceae). Plant Systematics and Evolution (1991) 174:115–123.[CrossRef][Web of Science]
Dobigny G, Ducroz J-F, Robinson TJ, Volobouev V. Cytogenetics and cladistics. Systematic Biology (2004) 53:470–484.
Doyle JJ, Doyle JL. A rapid DNA isolation procedure for small quantaties of fresh leaf tissue. Phytochemical Bulletin of the Botanical Society of America (1987) 19:11–15.
Farris JS. A successive approximation approach to character weighting. Systematic Zoology (1969) 18:374–385.
Farris JS, Källersjö M, Kluge AG, Bult C. Testing the significance of incongruence. Cladistics (1994) 10:315–319.[CrossRef][Web of Science]
Felix LP, Guerra M. Cytogenetics and cytotaxonomy of some Brazilian species of Cymbidioid orchids. Genetics and Molecular Biology (2000) 23:957–978.[Web of Science]
Felsenstein J. Confidence limits on phylogenies: an approach using the bootstrap. Evolution (1985) 39:783–791.[CrossRef][Web of Science]
Fitch WM. Toward defining the course of evolution: minimum change for a specific tree topology. Systematic Zoology (1971) 20:406–416.[Abstract]
Goldman DH, Freudenstein JV, Kores PJ, Molvray M, Jarrell DC, Whitten WM, et al. Phylogenetics of Arethuseae (Orchidaceae) based on plastid matK and rbcL sequences. Systematic Botany (2001) 26:670–695.[Web of Science]
Guerra M. Patterns of heterochromatin distribution in plant chromosomes. Genetics and Molecular Biology (2000) 23:1029–1041.[Web of Science][Medline]
Hillis DH, Huelsenbeck JP. Signal, noise, and reliability in molecular phylogenetic analyses. Journal of Heredity (1992) 83:189–195.
Hoehne FC. Reajustamento de algumas espécies de Maxillarias do Brasil, com a criação de dois novos gêneros para elas. Arquivos de Botânica do Estado de São Paulo (1947) 2:65–73.
Hoehne FC. Espécies e variedades novas de Orchidaceas do Brasil. Arquivos de Botânica do Estado de São Paulo (1952) 2:121–136.
Hoehne FC. Orchidaceas. In: Flora Brasilica—Hoehne FC, ed. (1953) Vol 10(12,7). São Paulo, Brazil: Secretaria da Agricultura. 223–341.
Holder MT, Anderason JA, Holloway AK. Difficulties in detecting hybridization. Systematic Biology (2001) 50:978–982.
Huelsenbeck JP, Hillis DM. Success of phylogenetic methods in the four-taxon case. Systematic Biology (1993) 42:247–264.
Huelsenbeck JP, Bull JJ, Cunningham CW. Combining data in phylogenetic analysis. Trends in Ecology and Evolution (1996) 11:152–158.[CrossRef]
Johnson LA, Soltis DE. MatK DNA sequences and phylogenetic reconstruction in Saxifragaceae s.s. Systematic Botany (1994) 19:143–156.[CrossRef][Web of Science]
Lee MSY. Uninformative characters and apparent conflict between molecules and morphology. Molecular Biology and Evolution (2001) 18:676–680.
Mansion G, Zeltner L, Bretagnolle F. Phylogenetic patterns and polyploid evolution within the Mediterranean genus Centaurium (Gentianaceae–Chironieae). Taxon (2005) 54:931–950.[Web of Science]
Onishi E. Sobre o polimorfismo do grupo Maxillaria madida (Orchidaceae, Maxillarinae). (1974) Brazil: Escola Superior de Agronomia Luiz de Queiroz, Universidade de São Paulo. Ph.D. Thesis.
Pabst GFJ, Dungs F. Orchidaceae Brasiliensis II (1977) Hildesheim: Brücke-Verlag Kurt Schmersow.
Posada D, Crandall KA. Modeltest: testing the model of DNA substitution. Bioinformatics (1998) 14:817–818.
Rambaut A. Se-Al Sequence Alignment Program, version 2·0a11 (2002) Oxford, UK: Department of Zoology, University of Oxford. Available at http://evolve.zoo.ox.ac.uk/.
Ramsey J, Schemske DW. Pathways, mechanisms, and rates of polyploid formation in flowering plants. Annual Review in Ecology and Systematics (1998) 29:467–501.[CrossRef]
Reed RD, Sperling FAH. Interaction of process partitions in phylogenetic analysis: an example from the swallowtail butterfly genus Papilio. Molecular Biology and Evolution (1999) 16:286–297.[Abstract]
Ryan A, Whitten WM, Johnson MAT, Chase MW. A phylogenetic assessment of Lycaste and Anguloa (Orchidaceae: Maxillarieae). Lindleyana (2000) 15:33–45.
Schwarzacher T, Schweizer D. Karyotype analysis and heterochromatin differentiation with Giemsa C-banding and fluorescent counterstaining in Cephalanthera (Orchidaceae). Plant Systematics and Evolution (1982) 141:91–113.[CrossRef][Web of Science]
Schweizer D, Ambros PF. Chromosome banding. In: Methods in molecular biology—Gosden JR, ed. (1994) vol. 29. Totowa, NJ: Humana Press. 97–112.[Medline]
Siddall ME, Whiting MF. Long-branch abstractions. Cladistics (1999) 15:9–24.[Medline]
Soltis DE, Soltis PS, Tate JA. Advances in the study of polyploidy since plant speciation. New Phytologist (2003) 161:173–191.[CrossRef][Web of Science]
Sun Y, Skinner DZ, Liang GH, Hulbert SH. Phylogenetic analysis of Sorghum and related taxa using internal transcribed spacers of nuclear ribossomal DNA. Theoretical and Applied Genetics (1994) 89:26–32.[Web of Science]
Swofford DL. PAUP*: Phylogenetics analysis using parsimony (and other methods), version 4·0b10 (2000) Sunderland, MA: Sinauer Associates.
Swofford DL, Olsen GJ, Waddell PJ, Hillis DM. Phylogenetic inference. In: Molecular systematics—Hillis DM, Moritz C, Mable BK, eds. (1996) 2nd edn. Sunderland, MA: Sinauer. 407–514.
Szlachetko DL, Mytnik-Ejsmont J, Górniak M, 
miszek M. Genera et species orchidalium. 15. Maxillarieae. Polish Journal of Botany (2006) 51:57–59.
Taberlet PL, Geilly G, Bouvet J. Universal primers for amplification on three non-coding regions of chloroplast DNA. Plant Molecular Biology (1991) 17:1105–1109.[CrossRef][Web of Science][Medline]
Whitten WM, Williams NH, Chase MW. Subtribal and generic relationship of Maxillarieae (Orchidaceae) with emphasis on Stanhopeinae: combined molecular evidence. American Journal of Botany (2000) 87:1842–1856.
Whitten WM, Williams NH, Blanco MA, Endara L, Neubig K, Koehler S, Carnevali G, Singer R. Molecular phylogenetics of Maxillaria and related genera (Orchidaceae: Cymbidieae) based upon combined molecular data sets. America Journal of Botany (2007) 94:1860–1889.[CrossRef]
Wolfe AD, Randle CP. Recombination, heteroplasmy, haplotype polymorphism, and paralogy in plastid genes: implications for plant molecular systematics. Systematic Botany (2004) 29:1011–1020.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
Related articles in Ann Bot:
- ContentSnapshots
Ann Bot 2008 102: NP.[Extract] [Full Text]
This article has been cited by other articles:
|
|
 |
|
  R. Begum, S. S. Alam, G. Menzel, and T. Schmidt Comparative molecular cytogenetics of major repetitive sequence families of three Dendrobium species (Orchidaceae) from Bangladesh Ann. Bot., October 1, 2009; 104(5): 863 - 872. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||