Deep divergences in the coffee family and the systematic position of Acranthera

Birgitta Bremer

Plant Syst Evol (2009) 278:101–123 DOI 10.1007/s00606-008-0138-4 ORIGINAL ARTICLE Deep divergences in the coffee family and the systematic position of Acranthera Catarina Rydin Æ Kent Kainulainen Æ Sylvain G Razafimandimbison Æ Jenny E E Smedmark Æ Birgitta Bremer Received: 31 March 2008 / Accepted: 28 November 2008 / Published online: 12 February 2009 Ó Springer-Verlag 2009 Abstract Despite extensive efforts, there are unresolved questions on evolutionary relationships in the angiosperm family Rubiaceae. Here, information from six loci and 149 Rubiaceae taxa provide new insights. Acranthera and Coptosapelta are strongly supported as sisters. Pollen grains of Acranthera possess several features common in Rubiaceae, but amongst potential similarities with the unusual grains of Coptosapelta are the nature of the apertures and the structure of the sexine. Luculia, Acranthera and Coptosapelta are excluded from the three subfamilies Ixoroideae, Cinchonoideae and Rubioideae. Sipaneeae and Condamineeae form a clade, sister to remaining Ixoroideae. Rondeletieae and Guettardeae are sisters to remaining Cinchonoideae. Colletoecema is sister to remaining Rubioideae, followed by the Urophylleae–Ophiorrhizeae clade. Nuclear ITS provided structured information at all phylogenetic levels, but the main gain from adding nrITS was the increased resolution. Average support values also increased but were generally high also without nrITS and the increase was not statistically significant. Keywords Anther-stigma complex Cinchonoideae Coptosapelta Ixoroideae Luculia Rubioideae Introduction Rubiaceae is one of the largest families of flowering plants, comprising more than 13,000 species (Govaerts et al. 2006). Distribution is worldwide, with a particularly high diversity in the tropics and subtropics. The family is a welldefined monophyletic group that can be easily recognised by (generally) opposite branching and phyllotaxis, interpetiolar stipules and sympetalous and epigynous flowers (Schumann 1891; Hutchinson 1973). Phylogenetic studies generally recognise three major lineages within Rubiaceae (Bremer et al. 1995; Bremer 1996b, 1999; Rova et al. 2002; Robbrecht and Manen 2006), often referred to as subfamilies Rubioideae, Ixoroideae and Cinchonoideae sensu Bremer et al. (1999). Subsequent studies have further investigated relationships within these subfamilies, for example Bremer and Manen (2000, Rubioideae), Andreasen and Bremer (2000, Ixoroideae) and Andersson and Antonelli (2005, Cinchonoideae). However, despite these extensive efforts, several questions on evolutionary relationships within Rubiaceae, including deep divergences in the family, have remained unanswered. We introduce some of the unresolved questions here. Acranthera C. Rydin Institute of Systematic Botany, University of Zürich, Zollikerstrasse 107, 8008 Zürich, Switzerland C. Rydin (&) K. Kainulainen S. G. Razafimandimbison J. E. E. Smedmark B. Bremer Department of Botany, Bergius Foundation, Royal Swedish Academy of Sciences, Stockholm University, 106 91 Stockholm, Sweden e-mail: Catarina.Rydin@botan.su.se Acranthera (Arnott 1838) is distributed in India, South to Central China and Central Malesia and consists of about 40 species of sparsely branched subshrubs (Bremekamp 1947; Govaerts et al. 2006). The flowers of Acranthera are unique within Rubiaceae and are characterised by the presence of united connective appendages, which in turn are united with the stigma by means of a columnar tissue (Puff et al. 1995). In the original description, Arnott (1838) 123 102 made a remark on a possible affinity to Mussaenda [now placed in the tribe Mussaendeae sensu Bremer and Thulin (1998) of Ixoroideae]. Bremekamp (1947) questioned this affinity in his monograph of the genus and considered the position of Acranthera unknown. He later classified Acranthera as a monogeneric tribe within Ixoroideae (Bremekamp 1966). Since then, only a few studies have investigated this genus. Puff et al. (1995) described the pollination ecology, morphology and anatomy of the stamens in selected Acranthera species and Kiehn (1995) included Acranthera in a survey of chromosome data. Furthermore, Acranthera was assigned to the tribe Sabiceeae (Cinchonoideae sensu Bremekamp 1966) based on the results of a morphologicalbased phylogenetic study by Andersson (1996). Bremer and Thulin (1998) did not include Acranthera in their molecular study but argued that its testa structure is different from that of Sabicea, being instead similar to that of Amphidasya. They postulated on a possible placement of Acranthera in Rubioideae. One paper has addressed the phylogenetic position of Acranthera based on molecular data; Alejandro et al. (2005) analysed trnT–L–F chloroplast data and the genus was resolved as sister to the rest of subfamily Rubioideae, with a relatively high statistical support. This study focused, however, on Mussaenda and allied genera. Luculia (Rubiaceae) was used as outgroup, and the sampling within Rubioideae and Cinchonoideae was limited. Coptosapelta and Luculia Coptosapelta consists of 16 species from South East Asia (Valeton 1923; Govaerts et al. 2006). They are woody vines with axillary, pentamerous flowers (Chao 1978). The genus was originally described by Korthals (1851) and placed in the tribe Cinchoneae (subfamily Cinchonoideae), but the morphology and phylogeny of the genus were later reinvestigated and debated by many authors (e.g. Verdcourt 1958; Bremekamp 1966; Robbrecht 1988; Andersson and Persson 1991). Bremekamp (1952, 1966) recognised the tribe Coptosapelteae in subfamily Ixoroideae. Luculia comprises four species of trees or shrubs with showy flowers, distributed in Himalaya, northern Thailand and southern China (Polunin and Stainton 1984; Govaerts et al. 2006). Luculia was also placed in Cinchoneae by Schumann (1891), and subsequent authors (e.g. Verdcourt 1958; Bremekamp 1966; Robbrecht 1988) did not disagree. Based on the phylogenetic analysis of morphological data, Andersson and Persson (1991) included Luculia and several other genera in a much wider circumscription of Coptosapelteae, which was later shown to be highly polyphyletic (Razafimandimbison and Bremer 2001). 123 C. Rydin et al. In phylogenetic studies based on molecular data, Luculia and Coptosapelta have typically had isolated, unresolved or poorly supported positions, often amongst the basal nodes within the family (Bremer and Jansen 1991; Bremer et al. 1995; Bremer 1996b; Bremer et al. 1999; Bremer and Manen 2000; Rova et al. 2002; Robbrecht and Manen 2006). The diversity of results that has been presented indicates that the position(s) of Luculia and Coptosapelta is not confidently resolved. The two genera form a clade in some studies (e.g. Andersson and Persson 1991; Robbrecht and Manen 2006), in others they appear more distantly related to each other (Bremer et al. 1999). In some studies, Luculia is sister to remaining Rubiaceae (Bremer and Jansen 1991, Coptosapelta not included), in others they are sister to Cinchonoideae–Ixoroideae (Robbrecht and Manen 2006). Urophylleae and Ophiorrhizeae The subfamily Rubioideae was proposed by Bremekamp (1952), based e.g. on the presence of raphide idioblasts, and was formally described by Verdcourt (1958). Andersson and Rova (1999) and Bremer and Manen (2000) addressed the phylogeny of the subfamily but some results were poorly supported and/or differed between the studies. For example, Andersson and Rova (1999) found a sister relationship between Urophylleae and Ophiorrhiza, this clade being the sister of remaining Rubioideae. Bremer and Manen (2000), who used a larger sample of species and more characters, found a basal grade within Rubioideae, with Ophiorrhizeae as the earliest diverging clade, followed by Urophylleae and Lasiantheae. Aims of this study After more than 60 phylogenetic studies during the last 18 years (adjusted from Bremer in press) many aspects of Rubiaceae evolution are now relatively well understood. There are, however, phylogenetic questions that remain unanswered, which hampers further studies addressing for example biogeography and geographical origin, molecular dating of divergences, ancestral state reconstruction and character evolution within the family. We address deep divergences in Rubiaceae with special emphasis on Acranthera, and we investigate the usefulness of nrITS for analysing deep divergences in Rubiaceae. Materials and methods Selection of species and laboratory procedures We selected 149 taxa for the present study (Table 1), representing the major clades within Rubiaceae. We included Deep divergences in the coffee family and the systematic position of Acranthera 85 terminals (representing 16 tribes) from Rubioideae, 26 terminals (representing 13 tribes) of Ixoroideae, 11 terminals (representing 7 tribes) of Cinchonoideae, and in addition seven terminals of Acranthera, eight of Coptosapelta and four of Luculia. Eight outgroup taxa from the sister group of Rubiaceae (the other families within Gentianales, Backlund et al. 2000) were selected and sampled at the generic level. Ingroup sequences were sampled at the species level. We utilised information from six loci: five chloroplast regions (rbcL, rps16 intron, ndhF, atpB–rbcL spacer, trnT–L–F region) and the internal transcribed spacer of the nuclear ribosomal DNA (nrITS1, 5.8S, nrITS2). We used sequences from GenBank whenever available and we also produced 249 new sequences for this study. GenBank accession numbers are given in Table 1. DNA was extracted, amplified and sequenced using standard procedures outlined in Kårehed and Bremer (2007). References to primers are given in Table 2. Sequence fragments were assembled using the Staden package (Staden 1996). Alignment Alignments of rbcL, rps16, ndhF, atpB–rbcL spacer and trnT–L–F could easily be performed by eye using the software Se–Al v.2.0 (Rambaut 1996). Insertion/deletion events were visually inferred, following the alignment criteria outlined in Oxelman et al. (1997). Gaps were treated as missing data in the alignment and added as binominal characters (absent or present) at the end of the matrix. In order to investigate if nrITS could be utilised for investigating deep divergences in Rubiaceae, we performed an initial alignment using Clustalx/Clustalw (Chenna et al. 2003). From the resulting alignment, it was obvious that most of the region could very easily be aligned over the entire family. Two short regions, one located in nrITS1, the other in nrITS2, were not properly aligned in Clustal and we edited the output from Clustal by eye. We made a simple parsimony analysis to evaluate the amount of information in nrITS. The resulting tree was partly collapsed in basal parts, but added valuable information on higher-level relationships. We continued by adding nrITS to the combined data set and compared results from bootstrap analyses including and excluding nrITS. We further conducted a bootstrap analysis on the combined six-gene data set where we removed the two regions (mentioned above), which were more difficult to align. Parts removed correspond to positions 173-236 and 537-541 in the nrITS sequence of Luculia gratissima (GenBank accession: EU145344). 103 Phylogenetic reconstruction We analysed each gene separately, including and excluding information from indels. In order to evaluate the usefulness of nrITS, we performed combined analyses including and excluding nrITS (5-cp data set; six-gene data set). We further analysed the combined six-gene data set, including and excluding information from indels. All matrices were analysed with two approaches: Bayesian inference and parsimony. Bayesian analyses were performed in MrBayes 3.1 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). For each single gene data set, the best performing evolutionary model was identified under three different model selection criteria: Akaike information criterion (AIC) (Akaike 1973), AICc (a second order AIC, necessary for small samples) and the Bayesian information criterion (BIC) (Schwartz 1978). We performed these calculations in software MrAIC ver. 1.4.3 (Nylander 2004). Indels were treated as a morphological partition. For single gene analyses, one million generations were run, with a sample frequency of 1,000 and four parallel chains. Prior probabilities were specified as follows (according to output from MrAIC): a flat Dirichlet prior probability (all values set to 1.0) was selected for the substitution rates (revmatpr) and the nucleotide frequencies (statefreqpr). The prior probability for the shape parameter of the gamma distribution of rate variation (shapepr) was uniformly distributed in the interval (0.1, 50.0). For analyses using a gamma distribution with a proportion of invariable sites, we specified a prior probability for this proportion (pinvarpr), uniformly distributed on the interval (0.0, 1.0). For combined analyses, five million generations were run. We partitioned the combined data set in several ways. First, we included all sequence data into a single partition and analysed it together with the morphological partition. Second, we included all chloroplast regions in one partition, and specified a separate partition for the nuclear ribosomal ITS. Indels constituted a separate morphological partition as before. We further excluded gap coding information and partitioned the molecular data into two partitions: chloroplast data and nrITS data. Finally, we specified seven partitions, one for each gene region, and one for indels. In all analyses, partitions were unlinked so that each partition was allowed to have its own set of parameters. Convergence of runs was assessed from the average standard deviation of split frequencies, chain swap information and potential scale reduction factors. To investigate the usefulness of nrITS in the present study, we performed further analyses on the combined data set, (1) excluding nrITS, and (2) excluding potentially 123 104 123 Table 1 The data matrix Taxon Voucher (of previously unpublished sequences) Acranthera Arn. ex Meisn. (sp. 1) Ridsdale 2470 (L) Classification rbcL –Coptosapelteae AM11719835 EU145477* rps16 35 ndhF atpB-rbcL spacer trnT/F nrITS EU145400* – AJ84740842 – Acranthera Arn. ex Meisn. (sp. 2) Bremer 1731 (UPS) –Coptosapelteae AM117199 EU145478* – EU145312* EU145524* – Acranthera atropella Stapf KH Kjeldsen 54 (AAU) –Coptosapelteae – EU145480* – – EU145525* – Acranthera frutescens Valeton AD Poulsen 52 (AAU) –Coptosapelteae EU145449* EU145475* EU145398* EU145310* EU145522* EU145345* Acranthera grandiflora Bedd. Klackenberg & Lundin 541 (S) –Coptosapelteae EU145448* EU145474* EU145397* EU145309* EU145521* – Acranthera siamensis (Kerr) Bremek. Larsen 45665 (AAU) –Coptosapelteae EU145450* EU145476* EU145399* EU145311* EU145523* EU145346* Acranthera siamensis(?) (Kerr) Bremek. Puff 990826-1/1 (WU) –Coptosapelteae AM11720035 EU145479* EU145401* EU145313* – – IXO-Gardenieae Z6884419 AF20097429 – – AF20102829 AJ22483517 Tonkin 200 (UPS)36 IXO-Alberteae Y1870817 EU145491* AJ23628216 – AJ62011847 AJ22484217 Clark & Watt 736 (UPS) APOCYNACEAE RUB-Urophylleae X917607 Y1184414 AJ4310324 AF12927124 AJ0119825 – DQ3591616 EU145337* AJ4309074 EU145576* DQ3588806 EU145383* GENTIANACEAE L143898 – AJ2358299 DQ1316956 AJ49019044 AJ48986444 Aidia micrantha (K.Schum.) Bullock ex F. White Alberta magna E. Mey. Alstonia scholaris (L.) R.Br. Amphidasya ambigua (Standl.) Standl. Anthocleista Afzel. ex R.Br. Anthospermum herbaceum L.f. Bremer, 3093 (UPS) Arcytophyllum aristatum Standl. Argostemma hookeri King Malaysia, Wanntorp s.n. (S) 1 16 RUB-Anthospermeae X83623 EU145496* AJ236284 RUB-Spermacoceae AJ2885952 AF33334820 – – RUB-Argostemmateae Z6878821 EU145497* EU145419* AJ2340322 EU145545* EU145356* – – AM26690237 AM26698937 2 AM266813 37 AJ234028 2 EU145544* EU145355* AF33334920 AM18206157 Batopedina pulvinellata Robbr. RUB-Knoxieae AJ288596 Bertiera guianensis Aubl. IXO-Bertiereae AJ22484517 AF20098329 – – AF15267012 AJ22484117 Bouvardia ternifolia (Cav.) Schltdl.) (syn. Bouvardia glaberrima) RUB-Spermacoceae X836261 AF00275811 – X7647841 DQ3591656 DQ3588846 Calycophyllum candidissimum (Vahl) DC. IXO-Condamineeae X836271 AF00403011 AJ23628516 DQ1317086 AF15264612 DQ3588866 Dirichletia glaucescens Hiern (syn. Carphalea glaucescens) Catesbaea spinosa L. RUB-Knoxieae Z6878921 AM26681737 AJ23628716 – AM26690637 AM26699337 CINCH-Chiococceae X836281 AF00403211 AF15270612 AY76388013 1 11 Cephalanthus occidentalis L. CINCH-Naucleeae X83629 Chiococca alba (L.) Hitchc. CINCH-Chiococceae L143948 AF00403411 CINCH-Chiococceae 1 11 Cinchona pubescens Vahl Coccocypselum condalia Pers. Pirani & Bremer 4891 (SPF) X83630 AF004033 AF004035 35 AM11734335 – AJ236288 16 AJ13083516 AJ235843 9 DQ131710 6 DQ1317116 AJ233990 2 AF152692 12 AY76381313 AJ346963 3 AJ3468833 DQ3588876 AY53835615 AM117217 EU145499* EU145420* EU145324* EU145547* EU145358* RUB-Coussareeae X8714522 EU145500* EU145421* EU145325* EU145548* EU145359* Coffea arabica L. IXO-Coffeeae X836311 AF00403811 AJ23629016 X7036440 DQ15384546 DQ15360946 C. Rydin et al. RUB-Coussareeae Coccocypselum hirsutum Bartl. ex CT 908, Bremer 2700 (S) DC. Taxon Voucher (of previously unpublished sequences) Classification rbcL trnT/F nrITS Colletoecema dewevrei (De Wild.) S Lisowski 47195 (K) E.M.A. Petit RUB-Colletoecemateae EU14545761 AF12927224 EU14540961 DQ1317136 EU14553261 EU145353* Condaminea corymbosa (Ruiz & Pav.) DC. IXO-Condamineeae Y1871316 AF00403911 AJ23629116 – AF10240643 – rps16 ndhF atpB-rbcL spacer Coptosapelta diffusa (Champ.) Steenis (specimen 1) Bartholomew et al. 847 (AAU) –Coptosapelteae EU145452* EU145482* EU145403* EU145315* EU145527* EU145347* Coptosapelta diffusa (Champ.) Steenis (specimen 2) Steward et al. 594 (S) –Coptosapelteae EU145453* EU145483* EU145404* AJ2339872 DQ3591666 DQ3588826 Coptosapelta flavescens Korth. (specimen 1) Puff 950720-1/2 (WU) –Coptosapelteae Y1871416 EU145484* AJ23629216 EU145316* AM11735435 EU145348* Coptosapelta flavescens Korth. (specimen 2) Gardette et al. EG1716 (K) –Coptosapelteae EU145454* EU145485* EU145405* EU145317* EU145528* EU145349* Coptosapelta flavescens Korth. (specimen 3) Larsen et al. 31147 (AAU) –Coptosapelteae – EU145488* EU145408* – EU145531* EU145352* Coptosapelta montana Korth. ex Valeton Clemens & Clemens 40864 (K) –Coptosapelteae EU145451* EU145481* EU145402* EU145314* EU145526* – Coptosapelta tomentosa Valeton ex K.Heyne (specimen 1) Beusekom & Charoenpol 1741 –Coptosapelteae (AAU) EU145455* EU145486* EU145406* EU145318* EU145529* EU145350* Coptosapelta tomentosa Valeton ex K.Heyne (specimen 2) Beusekom & Charoenpol 1933 –Coptosapelteae (AAU) EU145456* EU145487* EU145407* EU145319* EU145530* EU145351* Coussarea hydrangeifolia (Benth.) Fuentes 5504 (GB) Benth. & Hook.f. ex Müll. Arg. RUB-Coussareeae EU145460* EU145501* EU145422* EU145326* EU145549* EU145360* Coussarea macrophylla (Mart.) Müll. Arg. RUB-Coussareeae Y1184714 AF00404011 – – AF15261212 (C. sp) – Cremaspora triflora (Thonn.) K.Schum. IXO-Cremasporeae Z6885619 AF20099029 – DQ1317186 AF20104029 AJ22482417 – AM11735635 – EU145550* – AF15270112 AY76389113 AJ2340152 – AY51406153 AJ2340192 DQ66213832 EU145364* Cremocarpon lantzii Bremek. Razafimandimbison 517 (UPS) RUB-Psychotrieae AM11722235 AM11729635 – 2 – X836321 AF00404411 AM11734535 DQ1317206 RUB-Morindeae Z6879321 AF33164720 – Danais xanthorrhoea (K. Schum.) Bremer 3079 (UPS) Bremek. RUB-Danaideae Z6879421 AM11729735 AJ23629316 Declieuxia cordigera Mart. & Zucc. ex Schult. & Schult.f. Pirani & Bremer 4893 (SPF) RUB-Coussareeae AM11722435 AM11729835 EU145423* EU145327* EU145551* EU145361* Declieuxia fruticosa (Willd. ex Roem. & Schult.) Kuntze B. Hammel 20875 (MO, CR) RUB-Coussareeae AJ00217723 DQ1317216 EU145552* EU145362* Rodriguez 10 (K) RUB-Coussareeae AJ288599 Cubanola domingensis (Britton) Aiello CINCH-Chiococceae Damnacanthus indicus C.F. Gaertn. 123 EU145503* – 105 EU145502* Cruckshanksia hymenodon Hook. & Arn. AJ234004 2 Deep divergences in the coffee family and the systematic position of Acranthera Table 1 continued 106 123 Table 1 continued Taxon Voucher (of previously unpublished sequences) Dentella repens (L.) J.R.Forst. & G.Forst. Classification rbcL rps16 ndhF atpB-rbcL spacer trnT/F nrITS RUB-Spermacoceae – AF33337020 – – AF38154049 – Dibrachionostylus kaessneri (S.Moore) Bremek. Strid 2564 (UPS) RUB-Spermacoceae AJ61621128 AF00276111 – – EU145574* – Didymaea alsinoides (Cham. & Schltdl.) Standl. Keller 1901 (CAS) RUB-Rubieae Z6879521 – – AJ2340362 EU145570* – Diplospora polysperma Valeton Ridsdale IV.E.130 (L) IXO-Coffeeae AJ28670318 – AM11730135 – EU145538* – Dunnia sinensis Tutcher (Specimen 1) Yangchun 10, Ge et al. 2002 RUB-Dunnieae EU145467 EU14551561 EU14544261 EU14533961 EU14558361 EU145390* Dunnia sinensis Tutcher (Specimen 2) Taishan 10, Ge et al. 2002 RUB-Dunnieae EU14546861 EU14551661 EU14544361 EU14534061 EU14558461 EU145391* Dunnia sinensis Tutcher (Specimen 3) Zhuhai 12, Ge et al. 2002 RUB-Dunnieae EU14546961 EU14551761 EU14544461 EU14534161 EU14558561 EU145392* Dunnia sinensis Tutcher (Specimen 4) Longmen 12, Ge et al. 2002 RUB-Dunnieae EU14547061 EU14551861 EU14544561 EU14534261 EU14558661 EU145393* Dunnia sinensis Tutcher (Specimen 5) Xinhui 16, Ge et al. 2002 RUB-Dunnieae EU14547161 EU14551961 EU14544661 EU14534361 EU14558761 EU145394* IXO-Condamineeae Y1871516 AF15263712 – Emmenopterys henryi Oliv. Unknown Rubiaceae (GenBank name: Ernodea littoralis Sw.) Faramea multiflora A Rich. Bremer et al. 3331 (UPS) 61 RUB-Spermacoceae AJ288601 RUB-Coussareeae Z6879621 AM11730235 AJ23629416 2 AF002763 11 AF00404811 35 35 DQ1317286 2 – AJ234025 – EU145424* EU145328* AF10242243 EU145412* 6 EU145534* – – EU145363* Ferdinandusa speciosa Pohl Malme 2442 (UPS) IXO-Condamineeae AM117226 Feretia aeruginescens Stapf Bremer 3137 (UPS) IXO-Octotropideae Z6885719 AM11730535 – – EU145539* Fernelia buxifolia Lam. de Block s.n. (BR) IXO-Octotropideae AJ28670418 AM11730635 – DQ1317366 EU145540* – – X7645941 – – AJ0119845 L3640038 AJ2339852 AF10242843 39 DQ398604 X7789345 DQ3588816 DQ39863939 – – EU145569* – EU145533* AY73029430 AJ84740742 – AF38153949 – 27 AM117304 11 Galium album Mill. RUB-Rubieae X81090 Gelsemium Juss. Gentiana L. GELSEMIACEAE GENTIANACEAE L143978 L143988 RUB-Psychotrieae AM11722835 AF36984526 Geophila obvallata Didr. Guettarda uruguensis Cham. & Schltdl. Gynochthodes coriacea Blume Hedyotis fruticosa L. Q Luke 9037 (FR) X5-127, Gillis 9575 (FTG) 1 CINCH-Guettardeae X83638 RUB-Morindeae? AJ2886032 RUB-Spermacoceae 21 Z68799 1 CINCH-Hillieae X83642 Houstonia caerulea L. RUB-Spermacoceae AJ2886042 AJ4310334 AJ4310344 EU145489* AJ236297 16 AM11731135 – – – – AM117315 35 AF33337920 AJ236298 – DQ131739 6 16 AJ234026 2 AJ233993 2 AM117362 35 AF38152449 – – DQ01270658 DQ01277458 Hydnophytum formicarum Jack RUB-Psychotrieae X83645 1 AF001339 11 – X76480 41 – AF03491223 C. Rydin et al. Hillia triflora (Oerst.) C.M. Taylor AF004050 DQ131735 Taxon Voucher (of previously unpublished sequences) Classification rbcL rps16 ndhF atpB-rbcL spacer trnT/F nrITS Hymenodictyon floribundum (Hochst. & Steud.) Rob. Puff 861109-3/1 (WU) CINCHHymenodictyoneae AJ3470153 AF00405811 EU145411* DQ1317426 AY53845415 AJ3469053 IXO-Ixoreae X836461 AM11732135 AJ23629916 – AJ62011747 AJ22482617 – EU145573* – Ixora coccinea L. Kohautia caespitosa Schnizl. Bremer et al. 42566B (UPS) Kopsia fruticosa (Roxb.) A.DC. Kraussia floribunda Harv. RUB-Spermacoceae 21 Z68800 AM117324 8 APOCYNACEAE IXO-Octotropideae X91763 Z6885819 35 – – AJ235824 AM11732535 – 9 10 – DQ1317466 AM295091 AM11736835 – – Lasianthus kilimandscharicus K.Schum. H. Lantz 119 (UPS) RUB-Lasiantheae AM11723735 AM11732735 EU145426* EU145330* DQ66214732 EU145366* Lasianthus lanceolatus (Griseb.) Urb. Taylor 11719 (MO) RUB-Lasiantheae AM11723835 AF00406211 – EU145331* EU145554* EU145367* Lasianthus pedunculatus E.A. Bruce Andreasen 71 (UPS) RUB-Lasiantheae Z6880221 EU145504* EU145427* AJ2340032 EU145555* EU145368* Lasianthus strigosus Wight Bremer & Bremer 3902 (UPS) RUB-Lasiantheae AM11723935 EU145505* EU145428* – EU145556* EU145369* Lerchea bracteata Valeton Axelius 343 (S) RUB-Ophiorrizeae AJ2886102 EU145508* EU145433* AJ2339972 EU145561* EU145374* Luculia grandifolia Ghose Bremer 2713 (S) –Luculieae X836481 AM90059360 AM11734635 AJ2339862 AJ3469293 AJ3468963 Luculia gratissima (Wall.) Sweet Cult in Univ. Conn. Storres 870064 (no voucher) –Luculieae AM11724335 AJ4310364 AJ4309114 EU145344* Luculia intermedia Hutch. Howick et al. HOMC1524 (K) –Luculieae – EU145473* EU145396* – EU145520* – Luculia pinceana Hook. NN Thin et al. 3061 (AAU) EU145447* EU145472* EU145395* DQ1317496 AM1173713 – Manostachya ternifolia E.S. Martins Bamps & Martins 4410 (UPS) RUB-Spermacoceae AJ61621328 AM11732835 – – EU145572* – – – EU145568* – Margaritopsis nudiflora (Griseb.) Ekman 10248 (UPS) K. Schum. (Syn. Margaritopsis acuifolia) Maschalocorymbus corymbosus (Blume) Bremek. Mitchella repens L. Ridsdale 2471 (L) Mitrasacmopsis quadrivalvis Jovet Kayombo et al. (UPS) –Luculieae AM11724735 AF00134011 RUB-Urophylleae AJ2886112 AM90061160 – – EU145577* EU145384* RUB-Morindeae Z6880521 AF00144111 – – AB10353554 AB103536 RUB-Spermacoceae AJ61621428 AM11732935 EU145439* EU145336* EU145575* EU145382* AJ32007825 AJ23630016 AJ2340132 AF15261612 AY76284355 – AJ2358289 DQ1316976 – – DQ1317546 EU145535* AJ84685842 RUB-Morindeae AJ318448 Mostuea brunonis Didr. GELSEMIACEAE L144048 Gillis 10838 (FTG) Mycetia malayana (G. Don) Craib Novotny et al. (2002) 1 IXO-Mussaendeae X83652 RUB-Argostemmateae Z6880621 CINCH-Naucleeae RUB-Anthospermeae 1 X83653 X836541 25 – 16 EU145493* AJ130836 AF00277111 – AJ2340332 AF15262212 – EU145410* – EU145320* – AJ3469583 AF15262312 AJ3468973 AF25792731 25 AJ320080 AF00274111 107 123 Nauclea orientalis (L.) L. Coprosma granadensis Mutis ex L.f. (syn. Nertera granadensis) EU145308* RUB-Psychotrieae Morinda citrifolia L. Mussaenda erythrophylla Schumach. & Thonn. AJ0119875 Deep divergences in the coffee family and the systematic position of Acranthera Table 1 continued 108 123 Table 1 continued Taxon Voucher (of previously unpublished sequences) Classification rbcL rps16 ndhF atpB-rbcL spacer trnT/F nrITS Neurocalyx championii Benth. ex Thwaites Thor 601 (S) RUB-Ophiorrizeae EU145463* EU145509* EU145435* – EU145563* EU145376* Neurocalyx zeylanicus Hook. B & K Bremer 937 (S) RUB-Ophiorrizeae Z6880721 AM90059460 EU145434* 35 31 AJ2339952 EU145562* EU145375* – – EU145543* AF25793031 Normandia neocaledonica Hook.f. Munzinger 532 (MO) RUB-Anthospermeae AM117250 Oldenlandia corymbosa L. Ophiorrhiza elmeri Merr. RUB-Spermacoceae RUB-Ophiorrizeae X836551 EU145464* AF33338120 EU145510* AJ13083716 EU145436* – – AF38153749 EU145564* AY85405359 EU145378* RUB-Ophiorrizeae X836561 AF00406411 Ophiorrhiza mungos L. Kjeldsen & Poulsen 233 (AAU) Bremer 3301 (UPS) 2 Oreopolus glacialis (Poepp.) Ricardi RUB-Coussareeae AJ288612 Paederia foetida L. RUB-Paederieae AF33237320 Palicourea crocea (Sw.) Schult RUB-Psychotrieae Palicourea guianensis Aubl. RUB-Psychotrieae Parapentas silvatica (K. Schum.) Bremek. AM117253 X83657 AJ13083816 – DQ6621516 EU145377* 11 – – – – AF00406511 – AJ2340062 AF15261912 – AF147510 33 – – – AF14932233 AF001345 11 AF004042 35 – RUB-Knoxieae AF257931 1 AM266849 37 – – AF152615 2 12 AM266937 37 AY63555456 AM26702337 – AJ234021 EU145440* EU145338* AF10246743 EU145386* Pauridiantha symplocoides (S. Moore) Bremek. Lantz 123 (UPS) RUB-Urophylleae AY53850215 AF00406811 Pauridiantha paucinervis (Hiern) Bremek. Bremer 3090 (UPS) RUB-Urophylleae Z6881121 AM90060060 AJ23630216 AJ2339982 EU145578* EU145385* Pentas lanceolata (Forssk.) Deflers RUB-Knoxieae X836591 AM26687537 AJ23630416 X7647941 AM26696337 AB24727552 Pentodon pentandrus (Schumach. & Thonn.) Vatke, Oesterr. RUB-Spermacoceae X836601 AF00361211 AJ2340242 – – – Pouchetia baumanniana Büttner (syn. Pouchetia gilletii) Kiehn HBV sub RR-81-31 (WU) IXO-Octotropideae Z6885919 AM11733635 – – EU145541* – Praravinia suberosa (Merr.) Bremek. Sabah: Ridsdale no voucher RUB-Urophylleae AJ2886162 EU145514* – EU145579* EU145387* Pravinaria leucocarpa Bremek. Beaman 7950 (S) RUB-Urophylleae AJ2886172 AM90061360 EU145441* AJ2340012 EU145580* EU145388* X7648141 AY53846915 AF07203823 AF15261412 AF07199823 – AF14940033 Psychotria kirkii Hiern RUB-Psychotrieae X83663 Psychotria pittieri Standl. RUB-Psychotrieae – Psychotria poeppigiana Müll. Arg. Pyrostria hystrix (Bremek.) Bridson RUB-Psychotrieae Bremer 3791 (UPS) 1 AF410728 34 AF00274611 21 Z68818 AF002748 35 AM117262 Retiniphyllum pilosum (Spruce ex Wurdack & Adderley 43270 Benth.) Müll.Arg. (S) IXO-Retiniphylleae AF33165420 Rhachicallis americana (Jacq.) Hitchc. CINCH-Rondeletieae X836641 AM117338 35 AJ236307 16 – – – AJ234018 2 47 AJ31511450 EU145418* – AJ620168 AF00407611 – – EU145536* – AF00407311 – – AF15274712 AY73030130 C. Rydin et al. IXO-Vanguerieae 11 – Taxon Voucher (of previously unpublished sequences) Classification rbcL rps16 ndhF atpB-rbcL spacer trnT/F nrITS Rondeletia odorata Jacq. Bremer & Andreasen 3504 (UPS) CINCH-Rondeletieae Y1185714 EU145490* AJ2358459 EU145321* AF15274112 AY73030730 RUB-Rubieae X836661 – DQ3591676 X7646541 – DQ3588856 Rubia tinctorum L. Sabicea aspera Aubl. Andersson et al. 1941 (NY) IXO-Sabiceeae AY538508 15 AF004079 11 EU145416* – AY538475 15 Bremer et al. 4018-B18 (UPS) IXO-Sabiceeae Bremer & al 4038-BB38 (UPS) RUB-Lasiantheae EU145459* EU145494* AM11726935 AF12927524 EU145415* EU145429* DQ131781 EU145332* AJ847396 EU145557* AJ84688342 EU145370* Saldinia A. Rich. ex DC. (specimen 2) Kårehed et al. 257 (UPS) RUB-Lasiantheae EU145461* EU145430* EU145333* EU145558* EU145371* Schismatoclada sp. Baker Razafimandimbison & Ravelonarivo 373 (MO) Adam 20116 (UPS) RUB-Danaideae AM11727135 AM11734135 EU145425* EU145329* EU145553* EU145365* RUB-Schizocoleeae AM11727235 EU14549861 – EU14532361 EU14554661 EU145357* RUB-Schradereae Y1185914 AF00361711 AJ2340142 AF15261312 – 27 11 Schizocolea linderi (Hutch. & Dalziel) Bremek. Schradera sp K. Krause – 42 AM409008 Sabicea diversifolia Pers. Saldinia A. Rich. ex DC. (specimen 1) EU145506* 6 41 Sherardia arvensis L. K. Andreasen 345 (SBT) RUB-Rubieae X81106 – X76458 EU145571* – Sipanea biflora (L.f.) Cham. & Schltdl. Rova et al. 2005 (S) IXO-Sipaneeae AY53850915 AF00408511 EU145413* DQ1317886 AF15267512 AY55511648 Sipanea hispida Benth. ex Wernham Irwin et al. 34756 (UPS) IXO-Sipaneeae EU145458* EU145492* EU145414* EU145322* AY55510748 AY55512248 IXO-Sipaneeae – AF24302230 – Sipanea pratensis Aubl. Spermacoce remota Lam. RUB-Spermacoceae Spigelia L. Spiradiclis bifida Kurz J. B. H. 55 (S) 21 Z68823 LOGANIACEAE Y11863 14 AF004082 – AF004093 11 – AF15267712 AY55511548 AJ236309 16 – – – AJ235840 9 – – AF17800451 RUB-Ophiorrizeae EU145465* EU145511* EU145437* – EU145565* EU145379* Strychnos L. LOGANIACEAE L144108 AF00409411 AJ2358419 DQ1316916 AF10248443 – Thecorchus wauensis (Schweinf. ex Hiern) Bremek. RUB-Spermacoceae AM11728235 AM26690137 – – AM26698737 AM26707037 Theligonum cynocrambe L. RUB-Theligoneae X836681 Tricalysia cryptocalyx Baker 19 AF00408711 – X8168040 AF15262112 – 11 – – AF15266912 AJ22482717 IXO-Coffeeae Z68854 AF004088 Andersson & Nilsson 2304 (GB) RUB-Lasiantheae EU145462* EU145507* EU145431* EU145334* EU145559* EU145372* Trichostachys sp. Hook.f. B. Sonké 1725 (UPS) RUB-Lasiantheae AJ2886262 AM90059560 EU145432* DQ1317926 EU145560* EU145373* DQ1317936 EU145582* – AJ2340022 EU145581* EU145389* – EU145542* AJ22483917 60 Urophyllum arboreum (Reinw. ex Boeea 7887 (S) Blume) Korth. RUB-Urophylleae – AM900617 Urophyllum ellipticum (Wight) Thwaites Vangueria madagascariensis J.F. Gmel. Lundqvist 11085 (UPS) RUB-Urophylleae AJ2886272 AM90061960 – Bremer 3077 (UPS) IXO-Vanguerieae X836701 – – AJ13084016 109 123 Trichostachys aurea Hiern Deep divergences in the coffee family and the systematic position of Acranthera Table 1 continued 110 123 Table 1 continued Taxon Voucher (of previously unpublished sequences) Classification rbcL rps16 ndhF atpB-rbcL spacer trnT/F nrITS Virectaria major (K. Schum.) Verdc. Reekmans 10916 (UPS) IXO-Sabiceeae Y1186114 EU145495* EU145417* AJ2339892 EU145537* EU145354* Xanthophytum borneense (Valeton) Axelius Axelius 316 (S) RUB-Ophiorrizeae EU145466* EU145513* EU145438* EU145335* EU145567* EU145381* Xanthophytum capitellatum Ridl. Ridsdale 2473 (L) RUB-Ophiorrizeae AJ2886282 EU145512* – AJ2339962 EU145566* EU145380* Total number of taxa in single gene data sets 141 141 91 97 135 105 Total number of characters in single gene data sets 1402 1602 ? 23 2243 ? 7 1098 ? 18 3219 ? 18 925 (677)** Number of variable characters 527 1029 1172 605 1837 608 (386)** Number of phylogenetically informative characters 404 648 856 395 1145 504 (309)** Evolutionary model employed (AICc weights) GTRIG GTRG GTRG GTRG GTRIG GTRIG Conflicts between Bayesian and parsimony analyses No No No No No No Conflicts between results including/excluding indels – No No No No – Notes. Classification: SUBFAMILY ABBREVIATION-Tribe. For outgroup taxa, only the FAMILY name is given. New classification in bold. Detailed information on methods and results is presented in the text * Previously unpublished sequence. ** Numbers within brackets represent values when parts of the nrITS alignment were removed. Published sequences: 1: Bremer et al. (1995). 2: Bremer and Manen (2000). 3: Razafimandimbison and Bremer (2002). 4: Bremer et al. (2002). 5: Oxelman et al. (1999). 6: J-F Manen (GenBank unpublished). 7: Sennblad and Bremer (1996). 8: Olmstead et al. (1993). 9: Backlund et al. (2000). 10: ME Endress et al. (GenBank unpublished). 11: Andersson and Rova (1999). 12: Rova et al. (2002). 13: Motley et al. (2005). 14: Bremer et al. (1998). 15: Andersson and Antonelli (2005). 16: Bremer et al. (1999). 17: Andreasen et al. (1999). 18: Andreasen and Bremer (2000). 19: Andreasen and Bremer (1996). 20: L Andersson (GenBank unpublished). 21: Bremer (1996b). 22: Bremer (1996a). 23: Nepokroeff et al. (1999). 24: Piesschaert et al. (2000a). 25: Novotny et al. (2002). 26: Andersson (2001). 27: Manen and Natali (1995). 28: Thulin and Bremer (2004). 29: Persson (2000). 30: JHE Rova (GenBank unpublished). 31: CL Anderson et al. (GenBank unpublished). 32: M Backlund (GenBank unpublished). 33: L Andersson and C Taylor (GenBank unpublished). 34: Andersson (2002). 35: B Bremer (in prep.). 36: A Mouly (unpublished). 37: Kårehed and Bremer (2007). 38: Olmstead and Reeves (1995). 39: XL Zhang et al. (GenBank unpublished). 40: Natali et al. (1995). 41: Manen et al. (1994). 42: Alejandro et al. (2005). 43: Struwe et al. (1998). 44: Yuan et al. (2003). 45: Gielly and Taberlet (1996). 46: O Maurin et al. (GenBank unpublished). 47: Lantz and Bremer (2004). 48: Delprete and Cortes-B (2004). 49: Church (2003). 50: Lantz et al. (2002). 51: Gould and Jansen (1999). 52: Nakamura et al. (2006). 53: P Ding et al. (GenBank unpublished). 54: J Yokoyama et al. (GenBank unpublished). 55: AD Proujansky and DL Stern (GenBank unpublished). 56: CW Dick and E Bermingham (GenBank unpublished). 57: D Wolff and S Liede-Schumann (GenBank unpublished). 58: Church and Taylor (2005). 59: CI Yuan (GenBank unpublished). 60: Smedmark et al. (2008). 61: Rydin et al. (2008) C. Rydin et al. Deep divergences in the coffee family and the systematic position of Acranthera Table 2 Primers Sequence 50 –30 /Reference DNA region Primer names rbcL 50 F, 30 R and Bremer et al. (2002) 427F rbcL Z895R rps16 nrITS F and 2R Oxelman et al. (1997) ITSForwRub CCTTATCATTTAGAGGAAGGAG Zurawski, DNAX Research institute nrITS ITSRevRub CCTCCGCTTATTGATATGC nrITS P17 and 26S-82R Popp and Oxelman (2001) nrITS P25 Oxelman (1996) ndhF 2F Rydin et al. (2008) ndhF 1000R Rydin et al. (2008) ndhF 720F Rydin et al. (2008) ndhF 1700R Rydin et al. (2008) ndhF 1320F Rydin et al. (2008) ndhF 2280R Rydin et al. (2008) atpB-rbcL spacer rbcL50 R Rydin et al. (2008) atpB-rbcL spacer atpB50 R Rydin et al. (2008) trnT-L-F A1 Bremer et al. (2002) trnT-L-F 940R Rydin et al. (2008) trnT-L-F 820F Rydin et al. (2008) trnT-L-F IR Bremer et al. (2002) trnT-L-F 1250F Rydin et al. (2008) trnT-L-F trnT-L-F D 1880F Taberlet et al. (1991) Rydin et al. (2008) trnT-L-F 2670R Rydin et al. (2008) ambiguous parts of nrITS (specified above). We used Wilcoxon-signed rank tests implemented in VassarStats (Lowry 2008) to test for significant changes in posterior probabilities and bootstrap estimates between analysis including or excluding nrITS. Parsimony analyses were performed in Paup* version 4.0b10 for Unix (Swofford 1998), for single gene data sets, as well as for combined data sets including and excluding Table 3 Results of selected combined analyses 111 nrITS. Most parsimonious trees were calculated using the heuristic search option, 500 random sequence additions, tree bisection reconnection branch swapping. Support values were obtained by using bootstrap in Paup*, performing 1,000 bootstrap replicates, each with 10 random sequence additions with settings as before. A majority rule consensus tree was produced from the resulting trees, in which nodes with a bootstrap support \50% were collapsed. Pollen morphology Anthers with in situ pollen of Acranthera tomentosa R.Br. ex Hook.f., voucher: Vidal 5001 (P), were mounted on cleaned aluminium stubs and initially investigated under a stereomicroscope. For scanning electron microscopy (SEM), the material was coated with gold for 90 s in a sputter coater, and examined with a Hitachi Field Emission scanning electron microscope at 5 kV. Results Data The aligned six-gene data set included 149 terminals and 10,555 characters, from which 1,402 derived from rbcL, 1,602 from rps16, 2,243 from ndhF, 1,098 from atpB–rbcL spacer, 3,219 from trnT–L–F: 3,219, 925 from nrITS and 66 from indels (see also Table 1). The nrITS alignment with potentially ambiguous parts removed contained 677 characters. The number of variable and informative characters, number of supported nodes and average support values are given for single gene analyses in Table 1 and for combined analyses in Table 3. Model choice For each single gene analysis, the best performing model according to the corrected Akaike information criterion 5 regions ? 6 regions ? 6 regions ? indels 6 regions indels indels (parts of nrITS removed) (no indels) Number of characters in matrix 9,630 10,555 10,307 10,489 Number of variable characters 5,228 5,778 5,614 5,712 Number of informative characters 3,449 3,952 3,757 3,886 Number of supported nodes (bootstrap) 120 129 128 128 Number of supported nodes (Bayesian) 133 136 – 133 Average support (bootstrap) 90.84 92.15 89.56 90.60 Average support (Bayesian) 96.60 97.15 – 96.66 123 112 C. Rydin et al. 100 (100) 95 (82) 100 (100) 87 (71) 100 (100) 53 (58) 100 (100) 57 (70) 100 (100) 100 (100) 100 (100) 100 (100) 98(100) 91 (91) 100 (100) 100 (100) 100 (100) 88 (90) 93 (---) 63 (---) 78(85) 100 (100) --- (---) 100 (100) 100 (100) 100 (100) 72(67) 79 (78) 100 (100) 100 (100) 100 (100) 91 (90) 100 (---) 96 (---) 100 (100) 100 (100) 100 (---) --- (---) 100 (100) 100 (84) 100 (---) 81 (55) 100 (99) 100 (---) 85 (---) 99 (89) 100 (100) 100 (100) 100 (100) 100 (---) 100 (100) 57 (---) ( --- )(---) 66 (---) 96(---) 100 (100) 85 (---) 100 (100) 92 (86) 100 (100) Cinchonoideae 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 87 (60) 100 (100) 67 (82) 76 (81) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (98) 100 (100) 98 (96) 95(58) 74 (---) 100 (100) 100 (100) 99 (83) 92 (---) 59 (---) 100 (100) ---(---) 98 (97) 100 (100) 100 (100) 100 (100) 94 (72) 100 (100) 100 (100) Ixoroideae 100 (100) 82 (---) 100 (100) 100 (100) 96 (92) 90 (56) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 82 (82) 100 (100) 100 (100) 72 (88) 55 (54) 100 (100) 100 (100) Alstonia Kopsia Anthocleista Gentiana Gelsemium Mostuea Strychnos Spigelia Luculia grandifolia Luculia gratissima Luculia pinceana Luculia intermedia Acranthera grandiflora Acranthera frutescens Acranthera sp. 1 Acranthera atropella Acranthera sp. 2 Acranthera siamensis Acranthera siamensis? Coptosapelta montana Coptosapelta flavescens 2 Coptosapelta flavescens 3 Coptosapelta diffusa 1 Coptosapelta diffusa 2 Coptosapelta flavescens 1 Coptosapelta tomentosa 1 Coptosapelta tomentosa 2 Catesbaea spinosa Cubanola domingensis Chiococca alba Cinchona pubescens Hillia triflora Cephalanthus occidentalis Nauclea orientalis Hymenodictyon floribundum Guettarda uruguensis Rhachicallis americana Rondeletia odorata Alberta magna Bertiera guianensis Coffea arabica Tricalysia cryptocalyx Diplospora polysperma Cremaspora triflora Feretia aeruginescens Fernelia buxifolia Kraussia floribunda Pouchetia baumanniana Aidia micrantha Ixora coccinea Pyrostria hystrix Vangueria madagascariensis Retiniphyllum pilosum Mussaenda erythrophylla Sabicea diversifolia Sabicea aspera Virectaria major Emmenopterys henryi Condaminea corymbosa Calycophyllum candidissimum Ferdinandusa speciosa Sipanea biflora Sipanea hispida Sipanea pratensis Outgroup Luculieae Coptosapelteae Chiococceae Cinchoneae Hillieae Naucleeae Hymenodictyeae Guettardeae Rondeletieae Alberteae Bertiereae Coffeeae Cremasporeae Octotropideae Gardenieae Ixoreae Vanguerieae Retiniphylleae Mussaendeae Sabiceeae Condamineeae Sipaneeae Rubioideae Fig. 1 Relationships within the tribes Luculieae and Coptosapelteae; and the subfamilies Cinchonoideae and Ixoroideae, estimated using Bayesian inference of phylogeny based on molecular data from chloroplast regions rbcL, rps16 intron, ndhF, atpB–rbcL spacer, trnT–L–F, the nuclear ribosomal ITS and indels. Posterior probabilities are given above branches, bootstrap values (under parsimony) below. Support values from the analyses of chloroplast data (excluding nrITS) are given (in brackets) (AICc, Akaike 1973) was selected. AICc is appropriate when the ratio between sample size and number of parameters is small (n/K \ 40, Burnham and Anderson 2003, p. 66), but also for higher ratios because AICc will then converge to AIC (Posada and Buckley 2004). Empirically, the three criteria indicated the same best performing model for all matrices. For the rbcL, trnT–L–F and nrITS data, the general time reversible model (Tavare 123 Deep divergences in the coffee family and the systematic position of Acranthera 100 (100) 100 (99) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 52 (97) 96(96) --- (---) 87 (98) 100 (100) 100 (100) (57) 69 (78) 100 (100)100 (100) 10063(100) --- (76) 98 (97) 100 (100) --- (---) 100 (100) 100 (100) 80(---) 100 (100) 100 (100) 94 (---) 100 (100) 100 (98) 100 (100) 100 (100) 100 (100) 100 (100) --- (95) 100 (100) 98 (99) 100 (100) Spermacoceae alliance 100 (100) 100 (100) 100 (98) --- (---) 75 (54) 63 (51) 100 (100) 97 (100) 98 (100) 52 (75) 100 (100) 100 (100) 60(---) --- (---) 69(50) --- (---) 100(100) 100 (100) 100 (100) 100 (100) 100 (100) 99(99) 82 (80) 100 (100) 66 (69) 100 (100) 98 (97) 100 (100) 79 (84) 100 (100) 96 (87) 54 (57) 88 (64) 100 (100) --- (---) 82 (84) 100 (100) 99 (96) 100 (100) 100 (99) 100 (100) 100 (100) 97 (98) 93 (98) 97 (96) 85 (86) 100 (---) 100 (100) 100 (100) 100 (---) 90 (92) 97 (99) 100 (100) 100 (100) --- (80) 94 (95) 100 (100) 100 (100) 97 (98) 100 (100) 85 (59) 100 (100) --- (---) 100 (100) 100 (100) 100 (100) 100 (100) (100) 89 (96) 100 (100) 100 100 (100) 88 (84) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 98 (100) 99 (98) 100 (100) 56 (---) 100 (100) 70 (66) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (79) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) Psychotrieae alliance 100 (100) 99 (99) 100 (100) 100 (100) 100 (100) 100 (100) 100 (100) 95 (99) 100 (100) 100 (100) 100 (---) 82 (66) 100 (100) 100 (100) 100 (100) 100 (100) 98 (90) 100 (100) 100 (100) 90 (93) 71 (96) 53 (78) 100 (100) 100 (100) 100 (100) 100 (100) Rubioideae Coprpsma granadensis Normandia neocaledonica Anthospermum herbaceum Mycetia malayana Argostemma hookeri Paederia foetida Didymaea alsinoides Rubia tinctorum Sherardia arvensis Galium album Theligonum cynocrambe Dunnia sinensis 1 Dunnia sinensis 2 Dunnia sinensis 3 Dunnia sinensis 4 Dunnia sinensis 5 Danais xanthorrhoea Schismatoclada sp Batopedina pulvinellata Pentas lanceolata Dirichletia glaucescens Parapentas silvatica Hedyotis fruticosa Manostachya ternifolia Kohautia caespitosa Bouvardia ternifolia Spermacoce remota Oldenlandia corymbosa Thecorchus wauensis Unknown Rubiaceae Arcytophyllum aristatum Houstonia caerulea Dibrachionostylus kaessneri Mitrasacmopsis quadrivalvis Dentella repens Pentodon pentandrus Schizocolea linderi Morinda citrifolia Gynochthodes coriacea Damnacanthus indicus Mitchella repens Schradera subandina Palicourea guianensi s Palicourea crocea Psychotria poeppigiana Psychotria pittieri Margaritopsis nudiflorav Geophila obvallata Psychotria kirkii Cremocarpon lantzii Hydnophytum formicarum Coccocypselum condalia Coccocypselum hirsutum Declieuxia cordigera Declieuxia fruticosa Cruckshanksia hymenodon Oreopolus glacialis Coussarea macrophylla Coussarea hydrangeifolia Faramea multiflora Lasianthus kilimandscharicus Lasianthus pedunculatus Lasianthus lanceolatus Lasianthus strigosus Saldinia sp. 1 Saldinia sp. 2 Trichostachys aurea Trichostachys sp. Lerchea bracteata Ophiorrhiza mungos Ophiorrhiza elmeri Spiradiclis bifida Xanthophytum capitellatum Xanthophytum borneense Neurocalyx zeylanicus Neurocalyx championii Amphidasya ambigua Maschalocorymbus corymbosus Praravinia suberosa Pravinaria leucocarpa Urophyllum arboreum Urophyllum ellipticum Pauridiantha paucinervis Pauridiantha symplocoides Colletoecema dewevrei 113 Anthospermeae Argostemmateae Paederieae Rubieae Theligoneae Dunnieae Danaideae Knoxieae Spermacoceae Schizocoleeae Morindeae Mitchelleae Schradereae Psychotrieae s.l. Coussareeae Lasiantheae Ophiorrhizeae Urophylleae Colletoecemateae Fig. 2 Relationships within subfamily Rubioideae, estimated using Bayesian inference of phylogeny based on molecular data from chloroplast regions rbcL, rps16 intron, ndhF, atpB–rbcL spacer, trnT–L–F, the nuclear ribosomal ITS and indels. Posterior probabilities are given above branches, bootstrap values (under parsimony) below. Support values from the analyses of chloroplast data (excluding nrITS) are given (in brackets) 1986) with gamma distributed rates (Yang 1993) and a proportion of invariable sites was selected (GTR ? I ? C). For the rps16, ndhF and the atpB-rbcL spacer, GTR ? C was selected (Table 1). For combined analyses with less than seven partitions, GTR ? C was selected for the chloroplast partition. 123 114 C. Rydin et al. Analyses Table 4 Test for significance of differences in support values, when including/excluding nrITS Combined data set Including/excluding nrITS z P (twotailed) Bayesian posterior probabilities (Rubiaceae) z = 0.98 P = 0.3271 As described in ‘‘Materials and methods’’, the combined data set was analysed several times, partitioning the data set in different ways. These analyses resulted in nearly identical topologies, but with slight differences in resolution and support values. We observed no supported (i.e. C50% posterior probability and/or bootstrap support) conflicts between results obtained from the different combined analyses. Figures 1, 2 show the results from the Bayesian analysis including information from indels (two data partitions: nucleotide data and indels). Bootstrap values of 50% or more are plotted on the Bayesian tree. We have further indicated (within brackets) support values from the 5-cp gene analysis (excluding nrITS). Usefulness of nrITS for addressing deep divergences in Rubiaceae Including nrITS generally increased resolution and support values over the entire phylogeny (Figs. 1, 2; Table 3). Some nodes received a lower support when nrITS was added, but overall resolution and average support (arithmetic mean) increased. The phylogeny based on rbcL, rps16, ndhF, the atpB–rbcL spacer and trnT–L–F (excluding nrITS) had 120 supported nodes with an average bootstrap value of 90.84%. The tree also based on nrITS data had 129 supported nodes with an average bootstrap support of 92.15%. For Bayesian analyses, the analysis excluding nrITS had 133 supported nodes with an average posterior probability of 96.60%. Including nrITS yielded 136 supported nodes with an average posterior probability of 97.15%. However, the increase in mean support values was not statistically significant (Table 4), neither in Bayesian analyses (z = 0.98, P = 0.327), nor in bootstrap analyses (z = 0.92, P = 0.358). For subfamily Rubioideae, mean bootstrap support was slightly lowered when including nrITS, but the difference was not significant (z = -0.46, P = 0.6455). In Bayesian analyses, support values increased also in Rubioideae when including nrITS, but again not significantly (z = 0.16, P = 0.8729). The topology from the analysis of six genes, excluding potentially ambiguous sites in nrITS, was basically the same as for the complete six-gene topology but support values generally decreased and some resolution was lost (Table 3). The sister relationship between Luculia and the Acranthera–Coptosapelta clade was for example collapsed in this tree, as was the case in the 5-cp analysis, excluding nrITS altogether (Fig. 1). 123 Bayesian posterior probabilities (clade Ab) –a b a Bayesian posterior probabilities (clade B ) – b –a –a Bayesian posterior probabilities (clade C ) z = 0.16 P = 0.8729 Bootstrap values (Rubiaceae) z = 0.92 P = 0.3576 Bootstrap values (clade A) –a –a Bootstrap values (clade B) z = 1.55 P = 0.1211 Bootstrap values (clade C) z = -0.46 P = 0.6455 Wilcoxon signed-rank test a ns/r too small Clade A: Luculia–Coptosapelta–Acranthera; clade B: Cinchonoideae-Ixoroideae; clade C: Rubioideae b Single gene analyses We found no major conflicts between single gene data sets and no conflicts within each region (between parsimony and Bayesian analyses, or when including or excluding gap information, see also Table 1). The position of a few taxa varied between single gene data sets and supported deviations are presented below. Phylogeny—the combined data set Deep divergences and the Luculia–Acranthera– Coptosapelta clade All ingroup taxa were resolved in three (or four) major clades (Figs. 1, 2). 1: The Luculia–Acranthera–Coptosapelta clade (which collapsed in the 5-cp analysis into one Luculia clade and one Acranthera–Coptosapelta clade); 2: the Cinchonoideae–Ixoroideae clade; 3: the Rubioideae clade. Support was very high for the latter two groups (support values are presented as follows [Bayesian posterior probability including nrITS (posterior probability excluding nrITS)/bootstrap support including nrITS (bootstrap support excluding nrITS)]: Cinchonoideae– Ixoroideae [100 (100)/100 (98)] and Rubioideae [100 (100)/100 (100)]. Luculia, Acranthera and Coptosapelta fell outside these groups. Acranthera and Coptosapelta were sister groups in all analyses [100 (100)/100 (100)], a result which to our knowledge has not been presented before. Luculia was sister to the Acranthera–Coptosapelta clade with relatively high Bayesian posterior probability, but low bootstrap support and only recovered when information from the entire nrITS was included [93 (–)/63 (–)]. Deep divergences in the coffee family and the systematic position of Acranthera All currently recognised species of Luculia were included in this study and we show that the genus is monophyletic [100 (100)/100 (100)]. Our results also support the monophyly of Acranthera [100 (100)/100 (100)] and Coptosapelta [100 (100)/100 (100)]. Results within the Cinchonoideae–Ixoroideae clade Support values for Cinchonoideae and Ixoroideae (Fig. 1) were high [100 (100)/100 (100)]. Within Cinchonoideae, Rondeletieae–Guettardeae [100 (100)/100 (100)] was sister to a large clade comprising Hymenodictyeae, Naucleeae, Hillieae, Cinchoneae and Chiococceae [96 (–)/85 (–)]. Hymenodictyeae and Naucleeae formed a clade [100 (100)/ 100 (100)]. Hillieae was sister to Cinchoneae and Chiococceae [100 (–)/– (–)]. Within Ixoroideae, Sipaneeae and Condamineeae were sister groups [99 (83)/92 (–)], and this clade was sister to remaining Ixoroideae [96 (92)/90 (56)]. Sabiceeae and Mussaendeae [100 (100)/94 (72)] comprise the next diverging clade, followed by Retiniphyllum. The position of Retiniphyllum was strongly supported. Within remaining Ixoroideae, Vanguerieae and Ixoreae [100 (100)/100 (100)] were sister to a clade comprising Alberta, Coffeeae, Bertiereae, Cremasporeae, Octotropideae and Aidia [100 (100)/100 (100)], within which Alberta was sister to the two sister clades [100 (100)/100 (100)]: Coffeeae– Bertiereae [100 (100)/87 (60)] and Cremasporeae– Octotropideae–Aidia [95 (58)/74 (–)]. Within the latter, Aidia was sister to a Cremasporeae–Octotropideae clade [100 (100)/82 (–)]. Results within Rubioideae Subfamily Rubioideae (Fig. 2) was well supported [100 (100)/100 (100)]. Colletoecema dewevrei was sister to remaining Rubioideae with high support [100 (100)/95 (99)]. The next diverging clade consisted of Urophylleae [100 (100)/100 (100)] and Ophiorrhizeae [100 (100)/100 (100)], which grouped together with relatively high support [100 (–)/82 (66)]. Lasiantheae [100 (100)/100 (100)] was the next diverging group, followed by Coussareeae [100 (100)/100 (100)], which was sister group to the Psychotrieae and Spermacoceae alliances [98 (100)/52 (75)]. The Psychotrieae alliance [100 (100)/100 (100)] was here represented by 15 species. Schizocolea linderi was highly supported as sister to a clade comprising the remaining sampled taxa [100 (100/90 (92)]. The remaining species comprised two sister groups: 1) Mitchelleae–Schradereae [93 (98)/85 (86)], sister to Morindeae [100 (100)/100 (99)], and 2) Psychotrieae s.l. [100 (100)/100 (100)]. The Spermacoceae alliance [100 (100)/100 (100)] comprised two major clades: The first was here represented 115 by Anthospermeae, Argostemmateae, Paederieae, Rubieae, Theligoneae and Dunnieae [100 (100)/52 (97)]. Amongst these, Anthospermeae [100 (100)/100 (100)] was the earliest diverging group, followed by Argostemmateae [100 (100)/100 (100)]. The next diverging clade [– (76)/ –(–)] comprised Dunnieae [100 (100)/100 (100)] and its sister clade [100 (100)/69 (78)], which consisted of Paederieae and Rubieae–Theligoneae [100 (100)/100 (100)]. Within the second major clade of the Spermacoceae alliance [100 (98)/ –(–)], Danaideae [100 (100)/100 (100)] was sister to Knoxieae–Spermacoceae [100 (100)/97(100)]. Phylogeny—single gene data sets Generally, single gene analyses produced the same topologies as those obtained from the combined data set. There are some minor deviations and we arbitrarily decided that differences with a Bayesian posterior probability higher than 85%, and/or a bootstrap support higher than 50% can be considered ‘‘supported’’. Such differences are presented here (posterior probability/bootstrap index). rbcL The results from the Bayesian analysis of the rbcL data resolved Ophiorrhizeae and Urophylleae as a basal grade (instead of a clade) within Rubioideae (95/– for Rubioideae except Ophiorrhizeae). The result was not supported in the bootstrap analysis of the rbcL data. rps16 The analyses of the rps16 data resolved Luculia, Acranthera and Coptosapelta as sister to the Cinchonoideae– Ixoroideae clade (93/76). ndhF In the ndhF tree, the Acranthera–Coptosapelta clade was sister to remaining Rubiaceae including Luculia (98/85). Colletoecema was sister to Lasiantheae (90/–). This relationship was not supported in bootstrap analyses. Sipaneeae and Condamineeae formed a basal grade (not a clade) within Ixoroideae. Support for Condamineeae and remaining Ixoroideae was low (55/80). trnT–L–F In analyses based on the trnT–L–F data, the Acranthera– Coptosapelta clade was sister to Rubioideae (88/80). There were some differences amongst major clades in the Spermacoceae alliance, regarding the positions of Danaideae, Anthospermeae and Argostemmateae. The differences had 123 116 C. Rydin et al. The nrITS data resolved Colletoecema as sister to the Urophylleae–Ophiorrhizeae clade (82/76). Coussareeae grouped together with a collapsed Anthospermeae (95/–). This relationship was not supported in the bootstrap analysis. protrusion. The sexine is (micro)reticulate-perforate but differs probably between mesocolpial and apocolpial areas. Structures tentatively interpreted as aborted grains (ovoid, about 3 lm long, roughly undulating-palliate surface and apertures, not shown), were numerously present amongst the grains. Note: this SEM study represents a preliminary overview of characters found in grains (not acetolysed) from one specimen. Further studies are needed to provide more details and detect potential inter and intraspecific variation in Acranthera pollen. Pollen morphology Taxonomic implications Because our results strongly support Acranthera as sister to Coptosapelta, which has unique pollen morphology (Verellen et al. 2004), we made a preliminary SEM study of Acranthera pollen. Acranthera pollen (Fig. 3) is triangular (rarely quadrangular) in shape and spheroidal to subspheroidal, with a polar axis of about 17 lm and equatorial diameter of 18–22 lm. They have three (rarely four) apertures positioned at the angles. The apertures are of a compound, colporate type. The ectoaperture is a short colpus (6–8 lm long), with acute to obtuse endings. The mesoaperture is a pore with a diameter of about 3–4 lm. Each mesoaperture is covered by an apertural Based on the results, we describe four new tribes and one new tribal circumscription. Our decisions are based on the principles of classification outlined in Backlund and Bremer (1998). Acranthera is strongly supported as sister to Coptosapelta and we have included Acranthera in the tribe Coptosapelteae. Considering the persisting difficulties to find support for a close relationship between Luculia and other species of Rubiaceae, we have chosen to describe the new monogeneric tribe Luculieae. Luculieae and Coptosapelteae are clearly excluded from the three subfamilies Ixoroideae, a Bayesian posterior probability of 80–90% but were not present in the bootstrap tree. These results are further investigated elsewhere. nrITS Fig. 3 Pollen grains of Acranthera tomentosa (SEM): a Polar view. Acranthera pollen is generally triangular in shape, spheroidal to subspheroidal and about 18–22 lm in equatorial diameter. The sexine is (micro)reticulate-perforate. The grains have three apertures positioned at the angles. b Equatorial view. c The apertures are of a compound, colporate type; the ectoaperture is a short colpus and the mesoaperture is a pore. Each mesoaperture is covered by an apertural protrusion. d Polar view. Acranthera grains are rarely quadrangular with four apertures positioned at the angles. Scale bars 5 lm 123 Deep divergences in the coffee family and the systematic position of Acranthera Cinchonoideae and Rubioideae, but we do not propose a new subfamily for the Luculia–Coptosapelta–Acranthera clade at this point. The clade is relatively well supported (93%) in the Bayesian analysis of the six-gene data set, but poorly supported in bootstrap analysis (63%), and collapsed in five-gene data sets. Further studies are needed to confirm the monophyly of the Luculia–Acranthera– Coptosapelta clade. Three genera, Colletoecema, Schizocolea and Dunnia, are lone sister lineages of large clades comprising several well-defined tribes. They cannot be implemented in any of the existing tribes and we have therefore described the new monogeneric tribes Colletoecemateae, Schizocoleeae and Dunnieae (see below). Discussion In order to address deep divergences in Rubiaceae, we sampled a large data set comprising 149 terminals and nearly 11,000 characters. The project has thus had potential to address a number of previously unresolved relationships and conflicting results throughout the family. Morphology and character evolution are discussed but obvious morphological support for major groups defined by molecular data may be difficult to find. The usefulness of nrITS Nuclear ribosomal ITS has previously been used for resolving higher-level relationships within Rubiaceae (e.g. Andreasen et al. 1999) but not for addressing the phylogeny of the entire family. A comparison of the topologies from analyses including and excluding nrITS shows that when nrITS is included, resolution and/or support increase for relationships within several groups of interest here, for example, the sister relationships between Urophylleae and Ophiorrhizeae, between Sipaneeae and Condamineeae and between Luculia and the Coptosapelta–Acranthera clade (Figs. 1, 2). There are also nodes (for example in the Spermacoceae alliance), for which support values decrease when nrITS is included and we conducted a bootstrap analysis on the combined six-gene data set, excluding two short regions of nrITS where homology assessments were difficult and potentially ambiguous. The resulting topology was nearly identical to that obtained from the complete six-gene data set, but slightly less well resolved and with a distinctly lower average support value (Table 3). In the present study, nrITS thus provided structured information, which resulted in increased resolution. Nuclear ITS also contributed to an increase in average support, however, many nodes were well-supported also without information from 117 nrITS and the increase in support values was not statistically significant. New insights into evolutionary relationships— Acranthera The sister relationship between Acranthera and Coptosapelta is very well supported in all combined and single gene analyses except in the analysis of nrITS, where the node is present but less well supported (94/–). Our results further support the monophyly of the two genera. To our knowledge, these results have not been presented before. Although the Acranthera–Coptosapelta clade is well supported by molecular data, we find no unambiguous morphological support for the relationship. Bremekamp (1947, p. 273) discussed a potential synapomorphy for Acranthera and Coptosapelta: the style functioning as a temporary depository for pollen, a ‘‘receptaculum pollinis’’. However, Puff et al. (1995) considered such a structure in Acranthera a misconception and they consequently refuted this synapomorphy for Coptosapelta and Acranthera. Further, even though Bremekamp (1947) suggested secondary pollen presentation as a potential synapomorphy for Acranthera and Coptosapelta, he argued that the united apical connective appendage in Acranthera is a feature unique within Rubiaceae and similar to the morphology of stamens in Apocynaceae. Puff et al. (1995) also considered the ‘‘anther–style and stigma complex’’ of Acranthera unique within Rubiaceae, in structure as well as function. Pollen grains of Coptosapelta possess several features unique within Rubiaceae (Verellen et al. 2004). They are pororate and may have up to 10 apertures (even if 3–4 apertures are most common), they lack columellae and they have ‘‘droplets’’ on the inner nexine (Verellen et al. 2004). Acranthera pollen has so far not been thoroughly documented (but see Mathew and Philip 1983) and in order to get an indication on whether Acranthera pollen shares some of the features of Coptosapelta grains, we performed a preliminary SEM study of the outer surface of the grains and the nature of the apertures (Fig. 3). Several characters of Acranthera pollen are common in Rubiaceae and probably plesiomorphic. Acranthera grains are not pororate (like Coptosapelta grains) but colporate, which is considered the plesiomorphic condition in the family (Dessein et al. 2005). The size of Acranthera grains (18–22 lm in equatorial diameter) fits within the 20– 40 lm, which is most common in Rubiaceae (Dessein et al. 2005). The triangular (rarely quadrangular) shape is more unusual but occurs according to Dessein et al. (2005) for example in Tapiphyllum Robyns (i.e. Vangueria Juss.) and Psydrax Gaertn. (Vanguerieae, Ixoroideae). Apertural protrusions (papillae-forming onci), pollen buds and 123 118 structures that cover the aperture (opercula) have been reported for several genera of Rubiaceae, but to our knowledge, not for Coptosapelta. There are some potential similarities between Acranthera and Coptosapelta pollen. The short ectocolpi of Acranthera could perhaps be compared with the ectopores of Coptosapelta and the microreticulate to perforate sexine in Acranthera is similar to that described for some species of Coptosapelta (Verellen 2002). However, pollen characters in Acranthera need to be further studied (e.g. the presence or absence of columellae, ‘‘droplets’’ on the inner nexine, the nature of the apertural protrusions) before any hypotheses on synapomorphies can be put forward. The enigmatic Luculia Our study included all four currently recognised species of Luculia (Govaerts et al. 2006) and we show that the genus is monophyletic, but its relationship to other species of Rubiaceae remains uncertain. The clade comprising Luculia, Acranthera and Coptosapelta is here only supported in some of the single gene analyses (atpB-rbcL spacer and nrITS) and in combined analyses including nrITS. However, no analysis resulted in a well-supported alternative position for Luculia. Further, there is biogeographical support for a relationship between these three South East Asian genera and a relationship between Luculia and Coptosapelta has been indicated in other recent studies (Robbrecht and Manen 2006). The Luculia–Acranthera–Coptosapelta clade is equally puzzling from a morphological perspective as is the Acranthera–Coptosapelta clade. Korthals (1851) very briefly mentioned some similarities between Luculia and Coptosapelta regarding the form of the seed, but he did not specify this further. Bremekamp (1947, p. 261) considered corolla aestivation, insertion of the stamens in the corolla tube and many-seeded fruits important regarding the systematic position of Acranthera, but these characters provide no support for the Luculia–Acranthera–Coptosapelta clade. Corolla aestivation is imbricate in Luculia (Bremer and Struwe 1992), valvate in Acranthera (Bremekamp 1947) and contorted in Coptosapelta (Andersson and Persson 1991). Filaments are inserted at the base of the corolla tube in Acranthera (Bremekamp 1947), but at about one-third from the mouth of the corolla tube in Coptosapelta and Luculia (Andersson and Persson 1991). All three genera have many-seeded fruits (Sweet 1826; Korthals 1851; Bremekamp 1947), but this character is common in Rubiaceae and probably plesiomorphic. Pollen characters also show little resemblance between the three genera. Luculia grains are small to medium-sized, 22–24 lm in polar axis (Murray 1990), spheroidal, 3(–4)colporate and with a reticulate tectum (Dessein et al. 2005). 123 C. Rydin et al. These character states probably represent primitive states within the family (Dessein et al. 2005) so even though grains of Coptosapelta are (oblate)spheoidal (Verellen et al. 2004), and Acranthera grains are (tri)colporate (the present study), these respective similarities with Luculia grains are likely plesiomorphic. The more specialised respective features of Coptosapelta and Acranthera pollen, e.g. the pororate pollen of Coptosapelta and the triangular shape of Acranthera grains, are not present in Luculia. Early divergences within the family Despite that we have used a relatively extensive sampling of taxa and characters in this study, the major clades of the family form a basal trichotomy: (1) the Luculia–Acranthera–Coptosapelta clade, (2) a clade consisting of the subfamilies Cinchonoideae and Ixoroideae, (3) subfamily Rubioideae (Figs. 1, 2). Robbrecht and Manen (2006) argued, based on parsimony analyses of 15 selected species and eight gene regions, that Luculia and Coptosapelta (Acranthera was not investigated) are ‘‘basal to the rest of Cinchonoideae’’ (i.e. sister to the Cinchonoideae–Ixoroideae clade). However, this conclusion is not supported by their results. Their combined analysis had no support for the position of these genera (Robbrecht and Manen 2006, Fig. 2) and the super tree analysis placed Luculia and Coptosapelta as sister to the rest of the family, not sister to the Cinchonoideae– Ixoroideae clade (Robbrecht and Manen 2006, Fig. 4a). Results from super tree analyses are difficult to evaluate; trees from the literature often contain some poorly supported nodes, which consequently may decrease accuracy of the super tree. Further, some information in the original data sets is lost, because the character information is simplified into a phylogeny (de Queiroz and Gatesy 2007). When analysing a combined data set, it is possible to get increased support for relationships that are not supported, perhaps not even present, in the single gene analyses (see e.g. Kluge 1989; Olmstead and Sweere 1994). This has, however, not been the case regarding basal relationships in Rubiaceae. Different gene regions produce contradicting (poorly supported) results and the combined analyses are unresolved (the present study and Robbrecht and Manen 2006). Ixoroideae Sipaneeae and Condamineeae form a strongly supported clade, which is sister to the remaining Ixoroideae. Sabiceeae and Mussaendeae are sisters (see also Alejandro et al. 2005) and comprise the next diverging clade, followed by Retiniphylleae. Two additional well-supported relationships within Ixoroideae have not been presented Deep divergences in the coffee family and the systematic position of Acranthera 119 In our study, Rondeletieae and Guettardeae form a clade, sister to the remaining Cinchonoideae. The result is well supported but differs from that reported in Andersson and Antonelli (2005), where Naucleeae and Hymenodictyeae constituted the sister clade to the remaining Cinchonoideae. The sister-group relationship between Naucleeae and Hymenodictyeae, previously shown by Razafimandimbison and Bremer (2001) and later endorsed by Andersson and Antonelli (2005), is further supported by our analyses, as well as by pollen morphology (Verellen et al. 2007). However, an extended sampling in Cinchonoideae is needed to further address the relationships and evolution of the group (see Manns and Bremer 2008). data and did not include representatives from all genera. We show that Coussarea–Faramea constitutes the sister clade to remaining genera. Oreopolus and Cruckshanksia have long been considered related based on morphology (Taylor 1996), but few phylogenetic studies have included Cruckshanksia. We confirm, with high support, the close relationship between Oreopolus and Cruckshanksia. Heterophyllaea Hook.f. also belongs to this group (Andersson and Rova, 1999), sister to the Oreopolus–Cruckshanksia clade (Rydin et al. 2006). These three genera are all restricted to the western parts of South America. The Neotropical genera Coccocypselum and Declieuxia are sisters and results from Rydin et al. (2006) highly support the inclusion of Hindsia Benth. ex Lindl. in this clade, as sister to Declieuxia. Piesschaert et al. (2000b) discussed morphological as well as biogeographical support for the Coccocypselum ? Declieuxia–Hindsia clade. The tribe Danaideae is here sister to the Knoxieae– Spermacoceae clade. The posterior probability for this relationship is high, but the clade is collapsed in bootstrap consensus trees. In Bremer and Manen (2000) Danaideae was sister to the remaining Spermacoceae alliance (with very low bootstrap support). More research is needed to further assess the position of Danaideae. Rubioideae Conclusions The sister relationship between Colletoecema dewevrei and remaining Rubioideae is here confirmed with high support (see also Robbrecht and Manen 2006; Rydin et al. 2008). The next diverging clade comprises the East Asian Ophiorrhizeae and the pantropical Urophylleae. This is consistent with Andersson and Rova (1999), but the tribes have otherwise often had an unresolved position at the base of Rubioideae or they have formed a basal grade, being subsequent sister groups to the rest of the subfamily (Bremer and Manen 2000; Robbrecht and Manen 2006; Razafimandimbison et al. 2008). The sister-group relationship between Ophiorrhizeae and Urophylleae is well supported, but as often is the case for major groups in Rubiaceae, obvious morphological support is difficult to find. Spiradiclis bifida, is here sister to Ophiorrhiza (Fig. 2), but a rps16 sequence (Rydin et al. 2006) nested Spiradiclis caespitosa Blume within Ophiorrhiza. The monophyly of the two genera needs to be investigated further. Coussareeae is a morphologically variable group, restricted to the New World. Most species occur in lowland rainforests, but the monotypic genus Oreopolus inhabits the Andean regions. Several studies have contributed to our understanding of relationships between the genera in Coussareeae (Andersson and Rova 1999; Bremer and Manen 2000), but they were based on the less amounts of The systematic position of Acranthera, a long-debated question, is resolved; Acranthera and Coptosapelta are sisters. Acranthera is considered unique within Rubiaceae in reproductive characters and obvious morphological synapomorphies for the Acranthera–Coptosapelta clade are currently not known, but the well-supported result in all our analyses leaves little doubt about their close relationship. We performed a preliminary study of the pollen grains of Acranthera in an attempt to find synapomorphies with the unique pollen of Coptosapelta, but most characters of Acranthera grains (for example size, the colporate grains with three apertures positioned at angles and the reticulate sexine) are common in Rubiaceae and probably plesiomorphic. There are some potential (derived) similarities though; future studies may reveal new insights on morphological features of the clade. Luculia is sister to Acranthera–Coptosapelta but the clade is only well-supported in Bayesian analyses including nrITS. Nuclear ITS has traditionally been utilised mainly for studying higher-level relationships, e.g. within a genus, but it cannot be a priori assumed that homology assessments are impossible for certain loci at certain taxonomic levels. Here, nrITS provided structured information on deep divergences, as well as on higher-level relationships in Rubiaceae, and appear particularly useful in Cinchonoideae and Ixoroideae. before: Retiniphylleae sister to the (Vanguerieae–Ixoreae) ? (Alberteae–remaining Ixoroideae) clade (Fig. 1). It should be noted, however, that no representatives of Posoquerieae and Henriquezieae are included in the present study. Further, Sipaneeae and Condamineeae are not sisters but form a grade to remaining Ixoroideae in our ndhF analyses and this is consistent with results found in Kainulainen et al. (in press). Cinchonoideae 123 120 Basal relationships within the three subfamilies Rubioideae, Cinchonoideae and Ixoroideae are indicated in the present study, but deep divergences in the family were not resolved. Single gene regions produced contradicting (poorly supported) results and combined analyses resulted in a basal polytomy consisting of (1) Luculia–Acranthera– Coptosapelta, (2) an Ixoroideae-Cinchonoideae clade, (3) Rubioideae. Like for example major relationships amongst seed plants (Burleigh and Mathews 2007a, 2007b); mosses and worts (Qiu et al. 2006); the position of Equisetum (Schuettpelz et al. 2006) and relationships within the angiosperm clades Ericales (Schoenenberger et al. 2005), Lamiales (Wortley et al. 2005) and Malpighiales (APGII 2003), early radiation patterns within Rubiaceae have not been unambiguously resolved despite that large amounts of data have been analysed. In cases when molecular markers produce conflicting results, other kinds of data, for example structural rearrangements in the genomes, developmental biology and comparative morphology, may be useful when discriminating between alternative hypotheses. Acknowledgments We thank the curators of the herbaria A.A.U., B.R., G.B., K., S., P. and U.P.S. for loan of herbarium material, biomedical technicians Anbar Khodabandeh (Bergius Foundation, Royal Academy of Sciences) and Keyvan Mirbakhsh (Stockholm University, Sweden) for assistance, Jürg Schönenberger (Stockholm University) for suggestions for improvement of SEM investigations, Charlotte Taylor (Missouri Botanical Garden) for sharing unpublished information on Dunnia sinensis, Peter Endress (University of Zürich), Jan-Thomas Johansson, Per-Ola Karis (Stockholm University), Elmar Robbrecht (National Botanic Garden, Belgium) and an anonymous reviewer for valuable suggestions and comments on the text. The study was supported by grants from the Swedish Research Council to C.R. and B.B., and from the Knut and Alice Wallenberg Foundation to B.B. Appendix: FAMILY—RUBIACEAE JUSS. Tribe Luculieae Rydin and B. Bremer, tribus nov. Type: Luculia Sweet Diagnosis: Arbuscula. Calyx 5-merous, corolla 5-mera, tubo longo vix supra dilatato. Flores heterostyli. Antherae intra tubum subsessiles semiexsertae. Stigmata 2, ovarium 2-loculare, loculis polyspermis. Fructus baccatus, semina minuta. Description: Small trees or shrubs, opposite phyllotaxis. Stipules deciduous, lanceolate to linear. Flowers large, showy, pentamerous, heterostylous. Stamens inserted in narrow corolla tube, filaments short. Ovary bilocular, fruit baccate, seeds small, numerous. Genus included: Luculia Sweet Useful publications: Murray (1990); Bremer et al. (1999). Tribe Coptosapelteae Bremek. ex S. Darwin, Taxon 25: p. 600, (Darwin 1976), emend. Rydin and B.Bremer 123 C. Rydin et al. Type: Coptosapelta Kort. Description: Sparsely branched subshrubs or vines. Flowers usually pentamerous (rarely 4 or 6 parted). Ovary bilocular, fruit a capsule, seeds numerous. Chromosome basic number 10–11, Acranthera x10 (Kiehn 1995), Coptosapelta x11 (Verdcourt 1958; Puangsomlee and Puff 2001). Note: The new circumscription is based on molecular evidence presented in this paper. Morphological synapomorphies are not known at this point. Genera included: Coptosapelta Kort., Acranthera Arn. ex Meisn. Useful publications: Alejandro et al. (2005); Verellen et al. (2004); Puangsomlee and Puff (2001); Bremer et al. (1999); Kiehn (1995); Puff et al. (1995); Bremekamp (1947); Valeton (1923); Rydin et al. (this study). SUBFAMILY—RUBIOIDEAE VERDC. Bull. Jard. Bot. État Brux. 28: 280 (1958) Tribe Colletoecemateae Rydin and B. Bremer, tribus nov. Type: Colletoecema E.M.A. Petit Diagnosis: Arbores vel fructices, stipulis integris. Inflorescentiae axillares floribus multis conglomeratis. Flores heterostyli, 5-meri. Calyx cupuliformis, corolla tubiformis, stamina filamentis longis sub sinibus corollae adfixis. Ovarium 2-loculare, ovulo 1. Fructus drupaceus, pyrena 2loculare, semina albumine satis molli et oleoso, embryo teres. Description: Small trees or shrubs. Inflorescences axillary, flowers pentamerous, heterostylous. Stamens inserted in corolla tube. Ovary bilocular, one ovule per locule. Embryo long and narrow. Fruit a drupe, pyrenes bilocular. Genus included: Colletoecema E.M.A. Petit Useful publications: Petit (1963); Piesschaert et al. (2000a); Robbrecht and Manen (2006); Rydin et al. (2008). Schizocoleeae Rydin and B. Bremer, tribus nov. Type: Schizocolea Bremek. Diagnosis: Arbuscula. Stipulae in vaginam longam et angustam in fimbrias plerumque 8 fissam connatae. Flores in axillis foliorum dispositi. Calyx 5-merous, lobis e basi triangulari-setiformibus, hirsutis. Corolla hypocrateriformis, tubo calycem longitudine multo excedente. Stamina parte dilatata tubi inserta. Ovarium biloculare, loculis septo tenui separatis. Fructus baccatus, monospermus. Description: Small trees, stipules bordered with fine hairs. Flowers pentamerous, calyx triangular at base. Corolla extends beyond calyx, stamens inserted in corolla tube. Ovary bilocular with thin dissepiments separating the locules. Fruit a berry, one-seeded, surmounted by persistent calyx. Genus included: Schizocolea Bremek. Useful publications: Bremekamp (1950); Razafimandimbison et al. (2008); Rydin et al. (2008). Deep divergences in the coffee family and the systematic position of Acranthera Dunnieae Rydin and B. Bremer, tribus nov. Type: Dunnia Tutcher Diagnosis: Frutex. Inflorescentiae terminales, cymosae, floribus multis conglomeratis, bracteis magnis albis circumdatae. Flores 5-meri, calycis lobi minuti, persistentes. Corolla tubiformis, tubo calycem longitudine multo excedente. Fructus capsularis, 2-valvis, valvis 2-partitis. Semina numerosa. Description: Woody shrubs, stipules pubescent. Inflorescences terminal cymes, surrounded by enlarged, petaloid bracts. Flowers pentamerous, corolla tube extends out of calyx. Stamens inserted in corolla lobe. Pistil distylous. Fruit a capsule, seeds numerous. Diagnosis and description are based on the original publication of Dunnia (Tutcher 1905) and on observations made by C. Taylor (Missouri Botanical Garden, pers. com.). Genus included: Dunnia Tutcher Useful publications: Tutcher (1905); Ge et al. (2002); Chiang et al. (2002); Rydin et al. (2008). References Akaike H (1973) Information theory as an extension of the maximum likelihood principle. In: Petrov BN, Csaki F (eds) Second international symposium of information theory. Akademiai Kiado, Budapest, pp 267–281 Alejandro GD, Razafimandimbison SG, Liede-Schumann S (2005) Polyphyly of Mussaenda inferred from ITS and trnT-F data and its implication for generic limits in Mussaendeae (Rubiaceae). Am J Bot 92:544–557 Andersson L (1996) Circumscription of the tribe Isertieae (Rubiaceae). Opera Bot Belg 7:139–164 Andersson L (2001) Margaritopsis (Rubiaceae, Psychotrieae) is a pantropical genus. Syst Geogr Plants 71:73–85 Andersson L (2002) Relationships and generic circumscription in the Psychotria complex (Rubiaceae, Psychotrieae). Syst Geogr Plants 72:167–202 Andersson L, Antonelli A (2005) Phylogeny of the tribe Cinchoneae (Rubiaceae), its position in Cinchonoideae, and description of a new genus, Ciliosemina. Taxon 54:17–28 Andersson L, Persson C (1991) Circumscription of the tribe Cinchoneae (Rubiaceae)—a cladistic approach. Plant Syst Evol 178:65–94 Andersson L, Rova JHE (1999) The rps16 intron and the phylogeny of Rubioideae (Rubiaceae). Plant Syst Evol 214:161–186 Andreasen K, Bremer B (1996) Phylogeny of the subfamily Ixoroideae (Rubiaceae). In: Robbrecht E, Puff C, Smets E (eds) Second international Rubiaceae conference, proceedings, pp 119–138 Andreasen K, Bremer B (2000) Combined phylogenetic analysis in the Rubiaceae–Ixoroideae: morphology, nuclear and chloroplast DNA data. Am J Bot 87:1731–1748 Andreasen K, Baldwin BG, Bremer B (1999) Phylogenetic utility of the nuclear rDNA ITS region in subfamily Ixoroideae (Rubiaceae): comparisons with cpDNA rbcL sequence data. Plant Syst Evol 217:119–135 APGII (2003) An update of the angiosperm phylogeny group classification for the orders and families of flowering plants: APG II. Bot J Linn Soc 141:399–436 121 Arnott GAW (1838) Acranthera Arn. ex Meisn. In: Meisner CDF (ed) Plantarum vascularum Genera, p 115 Backlund A, Bremer K (1998) To be or not to be—principles of classification and monotypic plant families. Taxon 47:391–400 Backlund M, Oxelman B, Bremer B (2000) Phylogenetic relationships within the Gentianales based on ndhF and rbcL sequences, with particular reference to the Loganiaceae. Am J Bot 87:1029– 1043 Bremekamp CEB (1947) A monograph of the genus Acranthera Arn. ex Meisn. (Rubiaceae). J Arn Arb 28:261–307 Bremekamp CEB (1950) Schizocolea linderi (Hutch. et Dalz.) Brem. Hooker’s Icon Pl 35:tab. 3482 Bremekamp CEB (1952) The African species of Oldenlandia L. sensu Hiern et K. Schumann. Verh Kon Ned Akad Wetensch, Afd Natuurk, Tweede Sect 48:1–297 Bremekamp CEB (1966) Remarks on the position, the delimitation and the subdivision of the Rubiaceae. Acta Bot Neerl 15:1–33 Bremer B (1996a) Combined and separate analyses of morphological and molecular data in the plant family Rubiaceae. Cladistics 12:21–40 Bremer B (1996b) Phylogenetic studies within Rubiaceae and relationships to other families based on molecular data. Opera Bot Belg 7:33–50 Bremer B (in press) A historical perspective on molecular phylogenetics of Rubiaceae. Ann Mo Bot Gard Bremer B, Jansen RK (1991) Comparative restriction site mapping of chloroplast DNA implies new phylogenetic relationships within Rubiaceae. Am J Bot 78:198–213 Bremer B, Manen JF (2000) Phylogeny and classification of the subfamily Rubioideae (Rubiaceae). Plant Syst Evol 225:43–72 Bremer B, Struwe L (1992) Phylogeny of the Rubiaceae and the Loganiaceae: congruence or conflict between morphological and molecular data? Am J Bot 79:1171–1184 Bremer B, Thulin M (1998) Collapse of Isertieae, re-establishment of Mussaendeae, and a new genus of Sabiceeae (Rubiaceae); phylogenetic relationships based on rbcL data. Plant Syst Evol 211:71–92 Bremer B, Andreasen K, Olsson D (1995) Subfamilial and tribal relationships in the Rubiaceae based on rbcL sequence data. Ann Mo Bot Gard 82:383–397 Bremer B, Jansen RK, Oxelman B, Backlund M, Lantz H, Kim K (1999) More characters or more taxa for a robust phylogeny— case study from the coffee family (Rubiaceae). Syst Biol 48:413–435 Bremer B, Bremer K, Heidari N, Erixon P, Olmstead RG, Anderberg AA, Källersjö M, Barkhordarian E (2002) Phylogenetics of asterids based on 3 coding and 3 non-coding chloroplast DNA markers and the utility of non-coding DNA at higher taxonomic levels. Mol Phylogenet Evol 24:274–301 Burleigh JG, Mathews S (2007a) Assessing among-locus variation in the inference of seed plant phylogeny. Int J Plant Sci 168:111– 124 Burleigh JG, Mathews S (2007b) Assessing systematic error in the inference of seed plant phylogeny. Int J Plant Sci 168:125–135 Burnham KP, Anderson DR (2003) Model selection and multimodel inference: a practical information-theoretic approach. Springer, New York Chao J-M (1978) Rubiaceae. In: Li HEA (ed) Flora of Taiwan. Epoch Publishing Co., Ltd., Taipei Chenna R, Sugawara H, Koike T, Lopez R, Gibson TJ, Higgins DG, Thompson JD (2003) Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res 31:3497–3500 Chiang YC, Ge XJ, Chou CH, Wu WL, Chiang TY (2002) Nucleotide sequence diversity at the methionine synthase locus in endangered Dunnia sinensis (Rubiaceae): an evaluation of the positive selection hypothesis. Mol Biol Evol 19:1367–1375 123 122 Church SA (2003) Molecular phylogenetics of Houstonia (Rubiaceae): descending aneuploidy and breeding system evolution in the radiation of the lineage across North America. Mol Phylogenet Evol 27:223–238 Church SA, Taylor DR (2005) Speciation and hybridization among Houstonia (Rubiaceae) species: the influence of polyploidy on reticulate evolution. Am J Bot 92:1372–1380 Darwin SP (1976) The subfamilial, tribal and subtribal nomenclature of the Rubiaceae. Taxon 25:595–610 de Queiroz A, Gatesy J (2007) The supermatrix approach to systematics. Trends Ecol Evol 22:34–41 Delprete PG, Cortes-B R (2004) A phylogenetic study of the tribe Sipaneeae (Rubiaceae, Ixoroideae), using trnL-F and ITS sequence data. Taxon 53:347–356 Dessein S, Ochoterena H, de Block P, Lens F, Robbrecht E, Schols P, Smets E, Vinckier S, Huysmans S (2005) Palynological characters and their phylogenetic signal in Rubiaceae. Bot Rev 71:354–414 Ge XJ, Chiang YC, Chou CH, Chiang TY (2002) Nested clade analysis of Dunnia sinensis (Rubiaceae), a monotypic genus from China based on organelle DNA sequences. Conserv Genet 3:351–362 Gielly L, Taberlet P (1996) A phylogeny of the European gentians inferred from chloroplast trnL (UAA) intron sequences. Bot J Linn Soc 120:57–75 Gould KR, Jansen RK (1999) Taxonomy and phylogeny of a Gulf Coast disjunct group of Spigelia (Loganiaceae sensu lato). Lundellia (Austin, Tex.) 2:1–13 Govaerts R, Andersson L, Robbrecht E et al (2006) World Checklist of Rubiaceae. The Board of Trustees of the Royal Botanic Gardens, Kew. http://apps.kew.org/wcsp/home.do Huelsenbeck JP, Ronquist FR (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754–755 Hutchinson J (1973) Rubiaceae. The families of flowering plants. Clarendon Press, Oxford, pp 476–478 Kainulainen K, Mouly A, Khodabandeh A, Bremer B (in press) Molecular phylogenetic analysis of the tribe Alberteae (Rubiaceae), with description of a new genus, Razafimandimbisonia. Taxon Kårehed J, Bremer B (2007) The systematics of Knoxieae (Rubiaceae)-molecular data and their taxonomic consequences. Taxon 56:1051–1076 Kiehn M (1995) Chromosome survey of the Rubiaceae. Ann Mo Bot Gard 82:398–408 Kluge AG (1989) A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Syst Zool 38:7–25 Korthals PW (1851) Overzigt der Rubiaceën van de Nederlandschoostindische kolonien. Ned Kruidk Arch 2:98–114 Lantz H, Bremer B (2004) Phylogeny inferred from morphology and DNA data: characterizing well-supported groups in Vanguerieae (Rubiaceae). Bot J Linn Soc 146:257–283 Lantz H, Andreasen K, Bremer B (2002) Nuclear rDNA ITS sequence data used to construct the first phylogeny of Vanguerieae (Rubiaceae). Plant Syst Evol 230:173–187 Lowry R (2008) VassarStats: web site for statistical computation. Vassar College, Poughkeepsie. http://faculty.vassar.edu/lowry/ VassarStats.html Manen JF, Natali A (1995) Comparison of the evolution of ribulose– 1, 5-biphosphate carboxylase (rbcL) and atpB-rbcL noncoding spacer sequences in a recent plant group, the tribe Rubieae (Rubiaceae). J Mol Evol 41:920–927 Manen JF, Natali A, Ehrendorfer F (1994) Phylogeny of RubiaceaeRubieae inferred from the sequence of a cpDNA intergene region. Plant Syst Evol 190:195–211 123 C. Rydin et al. Manns U, Bremer B (2008) Intertribal relationships within subfamily Cinchonoideae s.str. (Rubiaceae). IV International Rubiaceae (Gentianales) Conference 44:45 Mathew PM, Philip O (1983) Studies in the pollen morphology of South Indian Rubiaceae. In: Nair PKK (ed) Advances in pollenspore research. Today and Tomorrow’s Printers, New Delhi Motley TJ, Wurdack KJ, Delprete PG (2005) Molecular systematics of the Catesbaeeae-Chiococceae complex (Rubiaceae): flower and fruit evolution and biogeographic implications. Am J Bot 92:316–329 Murray BG (1990) Heterostyly and pollen-tube interactions in Luculia gratissima. Ann Bot 65:691–698 Nakamura K, Chung SW, Kokubugata G, Denda T, Yokota M (2006) Phylogenetic systematics of the monotypic genus Hayataella (Rubiaceae) endemic to Taiwan. J Plant Res 119:657–661 Natali A, Manen JF, Ehrendorfer F (1995) Phylogeny of the Rubiaceae Rubioideae, in particular the tribe Rubieae - evidence from a noncoding chloroplast DNA-sequence. Ann Mo Bot Gard 82:428–439 Nepokroeff M, Bremer B, Sytsma KJ (1999) Reorganization of the genus Psychotria and tribe Psychotrieae (Rubiaceae) inferred from ITS and rbcL sequence data. Syst Bot 24:5–27 Novotny V, Basset Y, Miller SE, Weiblen GD, Bremer B, Cizek L, Drozd P (2002) Low host specificity of herbivorous insects in a tropical forest. Nature 416:841–844 Nylander JAA (2004) MrAIC.pl. Program distributed by the author. Evolutionary Biology Centre, Uppsala University, Uppsala Olmstead RG, Reeves PA (1995) Evidence for the polyphyly of the Scrophulariaceae based on chloroplast rbcL and ndhF sequences. Ann Mo Bot Gard 82:176–193 Olmstead RG, Sweere JA (1994) Combining data in phylogenetic systematics: an empirical approach using three molecular data sets in the Solanaceae. Syst Biol 43:467–481 Olmstead RG, Bremer B, Scott KM, Palmer JD (1993) A parsimony analysis of the Asteridae-sensu-lato based on rbcL sequences. Ann Mo Bot Gard 80:700–722 Oxelman B (1996) RAPD patterns, nrDNA ITS sequences and morphological patterns in Silene section Sedoineae (Caryophyllaceae). Plant Syst Evol 201:93–116 Oxelman B, Liden M, Berglund D (1997) Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Plant Syst Evol 206:393–410 Oxelman B, Backlund M, Bremer B (1999) Relationships of the Buddlejaceae s. 1. Investigated using parsimony jackknife and branch support analysis of chloroplast ndhF and rbcL sequence data. Syst Bot 24:164–182 Persson C (2000) Phylogeny of Gardenieae (Rubiaceae) based on chloroplast DNA sequences from the rps16 intron and trnL(UAA)F(GAA) intergenic spacer. N J Bot 20:257–269 Petit E (1963) Rubiaceae Africaneae X. Colletoecema, genre nouveau de Rubiaceae d’Afrique. Bull Jard Bot État Bruxelles 33:375– 380 Piesschaert F, Andersson L, Jansen S, Dessein S, Robbrecht E, Smets E (2000a) Searching for the taxonomic position of the African genus Colletoecema (Rubiaceae): morphology and anatomy compared to an rps16-intron analysis of the Rubioideae. Can J Bot 78:288–304 Piesschaert F, Huysmans S, Jaimes I, Robbrecht E, Smets E (2000b) Morphological evidence for an extended tribe—Coccocypseleae (Rubiaceae-Rubioideae). Plant Biol 2:536–546 Polunin O, Stainton A (1984) Flowers of the Himalaya. Oxford University Press, Oxford Popp M, Oxelman B (2001) Inferring the history of the polyploid Silene aegaea (Caryophyllaceae) using plastid and homoeologous nuclear DNA sequences. Mol Phylogenet Evol 20:474–481 Deep divergences in the coffee family and the systematic position of Acranthera Posada D, Buckley TR (2004) Model selection and model averaging in phylogenetics: advantages of Akaike information criterion and Bayesian approaches over likelihood ratio tests. Syst Biol 53:793–808 Puangsomlee P, Puff C (2001) Chromosome numbers of Thai Rubiaceae. N J Bot 21:165–175 Puff C, Igersheim A, Buchner R, Rohrhofer U (1995) United stamens of Rubiaceae. Morphology, anatomy; their role in pollination ecology. Ann Mo Bot Gard 82:357–382 Qiu YL, Li LB, Wang B, Chen ZD, Knoop V, Groth-Malonek M, Dombrovska O, Lee J, Kent L, Rest J, Estabrook GF, Hendry TA, Taylor DW, Testa CM, Ambros M, Crandall-Stotler B, Duff RJ, Stech M, Frey W, Quandt D, Davis CC (2006) The deepest divergences in land plants inferred from phylogenetic evidence. Proc Natl Acad Sci USA 103:15511–15516 Rambaut A (1996) Se-Al: Sequence Alignment Editor. Available at http://evolve.zoo.ox.ac.uk/ Razafimandimbison SG, Bremer B (2001) Tribal delimitation of Naucleeae (Cinchonoideae, Rubiaceae): inference from molecular and morphological data. Syst Geogr Pl 71:515–538 (publ. 2002) Razafimandimbison SG, Bremer B (2002) Phylogeny and classification of Naucleeae s.l. (Rubiaceae) inferred from molecular (ITS, rbcL, and trnT-F) and morphological data. Am J Bot 89:1027– 1041 Razafimandimbison SG, Rydin C, Bremer B (2008) Evolution and trends in the Psychotrieae alliance (Rubiaceae)—A rarely reported evolutionary change of many-seeded carpels from one-seeded carpels. Mol Phylogenet Evol 48:207–223 Robbrecht E (1988) Tropical woody Rubiaceae. Opera Bot Belg 1:1– 271 Robbrecht E, Manen J-F (2006) The major evolutionary lineages of the coffee family (Rubiaceae, angiosperms). Combined analysis (nDNA and cpDNA) to infer the position of Coptosapelta and Luculia, and supertree construction based on rbcL, rps16, trnLtrnF and atpB-rbcL data. A new classification in two subfamilies, Cinchonoideae and Rubioideae. Syst Geogr Plants 76:85–146 Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572– 1574 Rova JHE, Delprete PG, Andersson L, Albert VA (2002) A trnL-F cpDNA sequence study of the Condamineeae-RondeletieaeSipaneeae complex with implications on the phylogeny of the Rubiaceae. Am J Bot 89:145–159 Rydin C, Smedmark JEE, Bremer B (2006) Phylogeny, diversity and biogeography of four tribes in Rubioideae. In: Abstract of the third international Rubiaceae conference, p 65 Rydin C, Razafimandimbison SG, Bremer B (2008) Rare and enigmatic genera (Dunnia, Schizocolea, Colletoecema), sisters to species-rich clades: phylogeny and aspects of conservation biology in the coffee family. Mol Phylogenet Evol 48:74–83 Schoenenberger J, Anderberg AA, Sytsma KJ (2005) Molecular phylogenetics and patterns of floral evolution in the Ericales. Int J Plant Sci 166:265–288 Schuettpelz E, Korall P, Pryer KM (2006) Plastid atpA data provide improved support for deep relationships among ferns. Taxon 55:897–906 Schumann K (1891) Rubiaceae. In: Engler A, Prantl K (eds) Die natürlichen Pflanzenfamilien 4 (4). Wilhelm Engelmann, Leipzig 123 Schwartz G (1978) Estimating the dimensions of a model. Annu Stat 6:461–464 Sennblad B, Bremer B (1996) The familial and subfamilial relationships of Apocynaceae and Asclepiadaceae evaluated with rbcL data. Plant Syst Evol 202:153–175 Smedmark JEE, Rydin C, Razafimandimbison SG, Khan SA, LiedeSchumann S, Bremer B (2008) A phylogeny of Urophylleae (Rubiaceae) based on rps16 intron data. Taxon 57:24–32 Staden R (1996) The Staden sequence analysis package. Mol Biotechnol 5:233–241 Struwe L, Thiv M, Kadereit JW, Pepper AS-R, Motley TJ, White PJ, Rova JHE, Potgieter K, Albert VA (1998) Saccifolium (Saccifoliaceae), an endemic of Sierra de la Neblina on the BrazilianVenezuelan border, is related to temperate-alpine lineages of Gentianaceae. Harv Pap Bot 3:199–214 Sweet R (1826) Luculia gratissima. Br Fl Gard 2: t. 145 Swofford DL (1998) PAUP*. Phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland Taberlet P, Gielly L, Pautou G, Bouvet J (1991) Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol Biol 17:1105–1109 Tavare S (1986) Some probabilistic and statistical problems on the analysis of DNA sequences. In: Miura RM (ed) Some mathematical questions in biology—DNA sequence analysis. American Mathematical Society, Providence, pp 57–86 Taylor CM (1996) Taxonomic revision of Cruckshanksia and Oreopolus (Rubiaceae: Hedyotideae). Ann Mo Bot Gard 83:461–479 Thulin M, Bremer B (2004) Studies in the tribe Spermacoceae (Rubiaceae-Rubioideae): the circumscriptions of Amphiasma and Pentanopsis and the affinities of Phylohydrax. Plant Syst Evol 247:233–239 Tutcher WJ (1905) Description of some new species, and notes on other Chinese plants. J Linn Soc Bot 37:58–70 Valeton T (1923) The genus Coptosapelta Korth. Proc K Akad Wet Amsterdam 26:361–377 Verdcourt B (1958) Remarks on the classification of the Rubiaceae. Bull Jard Bot État Bruxelles 28:209–281 Verellen J (2002) Palynologische studie en revisie van Coptosapelta (Rubiaceae). Laboratorium voor Systematiek. Katholieke Universiteit, Leuven Verellen J, Smets E, Huysmans S (2004) The remarkable genus Coptosapelta (Rubiaceae): pollen and orbicule morphology and systematic implications. J Plant Res 117:57–68 Verellen J, Dessein S, Razafimandimbison SG, Smets E, Huysmans S (2007) Pollen morphology of the tribe Naucleeae and Hymenodictyeae (Rubiaceae-Cinchonoideae) and its phylogenetic significant. Bot J Linn Soc 153:329–341 Wortley AH, Rudall PJ, Harris DJ, Scotland RW (2005) How much data are needed to resolve a difficult phylogeny? Case study in Lamiales. Syst Biol 54:697–709 Yang Z (1993) Maximum likelihood estimation of phylogeny from DNA sequences when substitution rates differ over sites. Mol Biol Evol 10:1396–1401 Yuan YM, Wohlhauser S, Möller M, Chassot P, Mansion G, Grant J, Kupfer P, Klackenberg J (2003) Monophyly and relationships of the tribe Exaceae (Gentianaceae) inferred from nuclear ribosomal and chloroplast DNA sequences. Mol Phylogenet Evol 28:500–517 123

RELATED PAPERS

RELATED TOPICS

Log In

Deep divergences in the coffee family and the systematic position of Acranthera

Deep divergences in the coffee family and the systematic position of Acranthera

Related Papers

RELATED PAPERS

RELATED TOPICS