ABSTRACT From a phylogenetic perspective, the genus Manihot can be considered as an orphan group of plants, and the scientific knowledge acquired has been mainly related to cassava, one of the most important crops in poor tropical countries. The goal of the majority of evolutionary studies in the genus has been to decipher the domestication process and identify the closest relatives of cassava. Few investigations have focused on wild Manihot species, and the phylogeny of the genus is still unclear. In this study the DNA sequence variation from two chloroplast regions, the nuclear DNA gene G3pdh and two nuclear sequences derived from the 30-end of two cassava ESTs, were used in order to infer the phylogenetic relationships among a subset of wild Manihot species, including two species from Cnidoscolus as out-groups. Maximum parsimony and Bayesian analyses were conducted for each data set and for a combined matrix due to the low variation of each region when analyzed independently. A penalized likelihood analysis of the chloroplast region trnL–trnF , calibrated with various age estimates for genera in the Euphorbiaceae extracted from the literature was used to determine the ages of origin and diversification of the genus. The two Mesoamerican species sampled form a well-defined clade. The South American species can be grouped into clades of varying size, but the relationships amongst them cannot be established with the data avail- able. The age of the crown node of Manihot was estimated at 6.6 million years ago. Manihot esculenta vari- eties do not form a monophyletic group that is consistent with the possibility of multiple introgressions of genes from other wild species. The low levels of variation observed in the DNA regions sampled suggest a recent and explosive diversification of the genus, which is confirmed by our age estimates. Materials & Methods The species, gene regions analyzed and GenBank numbers are summarized in Table 1. The majority of Manihot accessions were In vitro propagated in order to get fresh leaves suitable for the DNA extraction (Fig. 2). Maximum parsimony heuristic searches were performed in PAUP* v. 4.0b10 (4) for each data set separately, and for the combined matrix. Cnidoscolus spp. were included as outgroups . 50% Bootstrap support with 100 replicates was estimated in PAUP*. The posterior probability of each node was estimated in MrBayes v. 3.1.2 (5, 6) . The Bayesian analysis was carried up for each data set, and for the combined matrix, specifying the parameters of the best-fit model for each data partition. The models were estimated with Modeltest v. 3.7 (6). Bayesian analyses were carried up using four simultaneous chains for 1 million generations, sampling every 100 generations. These results were visualized by using TreeView v. 1.6.6. (7). Age estimation of the Manihot crown node: Given the lack of well-suited Manihot fossils, in order to estimate the age of the Manihot crown node, we followed the method used by van Ee et al. (8) to date Croton diversification. Therefore, in addition to the Manihot species sampled, we incorporated additional sequence data that permitted use of known Euphorbiaceae fossils as calibration points. By including Glycydendron and Hevea , a minimum date of 23 my can be assigned to their divergence as shown by the presence of Glycydendron in palynological samples from the early Miocene. Hippomane mancinella and its sister taxon Bonania cubana form a node that can be constrained to a minimum age of 40 my based on the oldest reliable age estimate of the fossil Crepetocarpon perkinsii . By including Acalypha californica and a representative of its sister genus ( Mareya micranta ), another node is created that can be constrained to a minimum of 61 my based on the estimated age of the Acalypha type fossil pollen from the early paleocene. Davis et al . (9) and Wikström et al. (10) estimated a Cretaceous origin for Euphorbiaceae, with an origin of Croton as early as 65 mya, at the Cretaceous/Tertiary boundary. The two fossil-based dates are used as minimum age constraints for the applicable nodes in a trnL–trnF phylogeny and the maximum age constraint, the estimated age of 114 my of Euphorbiaceae, is taken from a thorough sampling of the Malpighiales that used four macrofossils and 11 palynofossils. We are aware of the potential for problems with clade-age calibrations due to the few fossils available ( 11 ), but currently these fossils represent the best available minimum age constraints within the family. Divergence times were estimated on a Bayesian phylogram obtained from the trnL–trnF dataset using penalized likelihood (PL) with logarithmic penalizing function in the program r8s v.. 1.71 ( 12 ). RESULTS Figure 3. Phylogenetic tree of Manihot based on the Bayesian analysis of the combined data set. Posterior probability values and bootstrap percentages are given above and below each node, respectively. The scale of the number of changes per site is also shown. The numbers of the corresponding taxonomic sections sensu Rogers and Appan (1) are given in front of each species. The clades with additional support from other nuclear DNA sequences are indicated with an asterisk. The trees obtained with maximum parsimony methods for the same data set, had a consistency index, CI = 0.74 and a retention index, RI = 0.73. Figure 1. Variation of leaf morphology in six Manihot species (1. M. caerulescens 2. M. fruticulosa 3. M. alutacea 4. M. violacea 5. M. anomala 6. M. epruinosa ). (http://sciweb.nybg.org/science2/VirtualHerbarium.asp) Figure 2. In vitro propagation of wild Manihot species obtained from the Genetic Resources Unit collection. DISCUSION Genetic variability in Manihot - A great difference in the number of variable sites between the chloroplast, and the nuclear data sets of Manihot was noticeable in this study (Table 2). The plastid regions presented lower percentages of variable, and parsimony informative sites, as obtained in previous phylogenetic studies of the genus (13). This result together with the low number of substitutions per site (Fig. 3), may be reflecting the recent origin of species as proposed by Rogers & Appan (1). The higher levels of genetic variation of the nuclear data sets are indicative of a very heterogeneous genome, which is still changing (2, 14, 15). The G3pdh gene region was characterized by a high percentage of phylogenetic informative sites, and low levels of homoplasy (Table 2). These results, and the high mutation rates found in this gene region (16) make it a good option to conduct evolutionary studies in the genus Manihot . The 3’-end sequences of the cassava ESTs (clsi5l6 and clsi3i7) presented both high, and low percentages of variation, due to its location near the 3’-UTR region of the gene (3). Due to the stochastic nature of mutation in this region, it is better to have a higher sampling of these nuclear regions in order to evaluate their utility for phylogenetic inference, since we cannot confirm the orthologous nature of all the sequences generated. For this reason they were not included in the phylogenetic analyses of Manihot . Phylogenetic analysis of Manihot : The South American species form a monophyletic group whose topology is better resolved when the data are combined ( Fig. 3), and when the models of sequence evolution are taken into account. In this case the non-monophyly of Manihot esculenta subspecies is evident. This might be a result of interspecific hybridization that is common between the wild species and the crop, because the reproductive barriers have not been completely established ( 1, 14 ). Some branches are congruent with the phenetic relationships proposed by Allem (18) , like the close relationship between M. esculenta subsp. peruviana and M. pilosa ( Fig. 1 ), which are part of the ‘‘secondary gene pool of cassava”, or the species that he considers to be the closest relatives of M. esculenta , based on their phenetic and ecogeographic similarities. On the other hand, the position of M. pruinosa contradicts the results of Olsen & Schaal (19), who placed it as a closely related potentially hybridizing species of cassava. The position of some species in the phylogenies differed when data were combined as well as when parsimony or Bayesian methods were used and may be a consequence of reticulation and introgression events that cannot be inferred from phylogeny alone. The clade composed by Manihot alutacea , M. cecropiaefolia , M. longipetiolata , M. orbicularis , and M. sparsifolia is a well-supported relationship, common to all the phylogenies obtained. All these species are found in the Cerrado ecosystem of Brazil (Goiás) and have a shrubby habit. They belong to the Quinquelobae section ( 1 ), where species such as M. jacobinensis and M. violacea are also included. However, the last species form a separate branch in the tree ( Fig. 3 ), making the section paraphyletic. This might be an effect of the characters chosen by Rogers and Appan (1) to describe this section, which are related with the leaves. It is possible that these characteristics are shared by different species for reasons other than inheritance from a common ancestor. Another clade common to all the phylogenies was formed by M. glaziovii , M. carthaginensis , M. epruinosa and M. guaranitica (Fig. 1). All the species form a part of different taxonomic sections, except for M. glaziovii and M. epruinosa , which belong to the Glaziovianae section, characterized by the presence of trees and tall shrubs (1 ); however these growth habits are also present in M. carthaginensis , and M. guaranitica , which belong to the sections Carthaginenses and Anysophyllae , respectively. The close relationship among M. glaziovii , M. epruinosa and M. guaranitica is congruent with the results obtained by Fregene et al. (14), based on RFLP data. The results obtained demonstrate that a taxonomic revision of the genus Manihot is necessary due to the inconsistency of the sections proposed by Rogers & Appan (1). Biogeography of the genus Manihot: The genetic variability and divergence dates reported in this study indicate a recent origin of the genus Manihot during the late Miocene. The place of origin of the genus is equivocal. In addition, the limited sampling of Mesoamerican species and the topology of the trees obtained do not provide strong evidence regarding the chronological ancestry of those species. In South America species diversification probably began in the Cerrado ecosystem of eastern-central Brazil, taking into account the great number of species distributed there (1, 2). The Cerrado has been thought to be an old ecosystem, existing since the Cretaceous before the final separation of Africa and South America (19, 20). However, recent studies based on molecular dating of phylogenies of species endemic to the Cerrado propose a very recent origin of 4–8 mya for these elements (21, 22). The diversification of Manihot , with ca. 80% of its species belonging to the Cerrado, approximately 6.6 mya coincides with the recent origin of these Cerrado lineages. (1) Rogers, D. J. & Appan S. G. 1973. Manihot and Manihotoides (Euphorbiaceae): A Computer Assisted Study. Flora Neotropica (Monograph No. 13). Hafner Press, New York. (2) Bertram, R. B. 1993. Application of molecular techniques to genetic resources of cassava (Manihot esculenta Crantz, Euphorbiaceae): interspecific evolutionary relationships and intraspecific characterization. PhD thesis, University of Maryland. (3) Lopez, C., Piégu, B., Cooke, R., Delseny, M., Tohme, J. & Verdier, V. 2005. Using cDNA and genomic sequences as tools to develop SNP strategies in cassava (Manihot esculenta Crantz). Theor. Appl. Genet. 110: 425-431. (4) Swofford, D. L. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4.0b10 for MacintoshTM. Sinauer Associates, Sunderland, Massachusetts. (5) Ronquist, F., & Huelsenbeck, J. P. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572-1574. (6) Posada, D. & Crandall, K. A. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817-818. (7) Page, R. D. M. 1996. TREEVIEW: An application to display phylogenetic trees on personal computers. Computer Applications in the Biosciences 12: 357-358. (8) van Ee, B.W., Berry, P.E., Riina, R., Gutiérrez-Amaro, J.E., 2008. Molecular phylogenetics and biogeography of the Caribbean-centered Croton subgenus Moacroton (Euphorbiaceae s.s.). Bot. Rev. 74, 132–165. (9) Davis, C.C., Webb, C.O., Wurdack, K.J., Jaramillo, C.A., Donoghue, M.J., 2005. Explosive radiation of Malpighiales supports a mid-Cretaceous origin of modern tropical rain forests. A. Nat. 165, E36–E65. (10) Wikström, N., Savolainen, V., Chase, M.W., 2001. Evolution of the angiosperms: calibrating the family tree. Proc. R. Soc. Lond. B-Bio. 268, 2211–2220. (11) Heads, M., 2005. Dating nodes on molecular phylogenies: a critique of molecular biogeography. Cladistics 21, 62–78. (12) Sanderson, M.J., 2003. r8s: inferring absolute rates of molecular evolution and divergence times in the absence of a molecular clock. Bioinform Appl Note 19, 301–302. (13) Schaal, B. A., Olson, P. D., Prinzie, T. P., Carvalho, L. J. C. B., Tonukari, J. & Hayworth, D.A. 1994. Cassava Biotechnology Network Proceedings (Cent. Int. Agric. Trop., Cali, Colombia), Pp. 62-70. (14) Fregene, M. A., Vargas, J., Angel, F., Tohme, J., Asiedu, R. A., Akorada, M. O. y Roca, W. M. 1994. Chloroplast DNA and nuclear ribosomal DNA variability in cassava (Manihot esculenta Crantz) and its wild relatives. Theor. Appl. Genet. 89: 719-727. (15) Roa, A. C., Maya, M. M., Duque, M. C., Tohme, J., Allem, A. C. & Bonierbale, M. W. 1997. AFLP analysis of relationships among cassava and other Manihot species. Theor Appl Genet. 95: 741-750. (16) Olsen, K. M. & Schaal, B. A. 1999. Evidence on the origin of cassava: Phylogeography of Manihot esculenta. Proc. Natl. Acad. Sci. USA 96: 5586-5591. (17) Olsen, K.M., 2002. Population history of Manihot esculenta (Euphorbiaceae) inferred from nuclear DNA sequences. Mol. Ecol. 11, 901–911. (18) Allem, A.C., 2002. The origins and taxonomy of cassava. In: Hillocks, R.J., Thersh, J.M., Belloti, A.C. (Eds.), Cassava: Biology, Production and Utilization. CABI Publishing, New York, pp. 1–16. (19) Cole, M.M., 1960. Cerrado, caatinga, and pantanal. The distribution and origin of the savanna vegetation of Brazil. Geogr. J. 126, 168–179. (20) Ratter, J.A., Ribeiro, J.F., Bridgewater, S.B., 1999. The Brazilian Cerrado vegetation and threats to its biodiversity. Ann. Bot. 80, 223–230. (21) Pennington, R.T., Lewis, G.P., Ratter, J.A., 2006. An overview of the plant diversity, biogeography and conservation of neotropical savannas and seasonally dry forests. In: Pennington, R.T., Lewis, G.P., Ratter, J.A. (Eds.), Neotropical Savannas and Seasonally Dry Forests: Plant Diversity, Biogeography and Conservation. CRC Press, Boca Raton, Florida, USA, pp. 1–29. (22) Simon, M., Grether, R., de Queiroz, L.P., Skema, C., Pennington, T., Hughes, C.E., 2008. Origins and diversification of the Cerrado flora in Brazil: evidence from multiple time-calibrated molecular phylogenies. Abstracts Botany 2008. http:// 2008.botanyconference.org/engine/search/index.php?func=detail&aid=813.
Juliana Chacón a, b , Santiago Madriñán a, * , Daniel Debouck b , Fausto Rodriguez b & Joe Tohme b