8 Molecular Phylogenetic and Evolutionary Studies of Parasitic Plants Daniel L. Nickrent, R. Joel Duff, Alison E. Colwell, Andrea D. Wolfe, Nelson D. Young, Kim E. Steiner, and Claude W. dePamphilis The parasitic nutritional mode is a frequently 1995). For the vast majority of parasitic plants, evolved adaptation in animals (Price, 1980), as negative effects upon the host are difficult to de­ well as in flowering plants (Kuijt, 1969). Het­ tect, yet others (e.g., Striga, Orobanche) are se­ erotrophic angiosperms can be classified as ei­ rious weeds of economically important crops ther mycotrophs or as haustorial parasites. The (Kuijt, 1969; Musselman, 1980; Eplee, 1981; former derive nutrients via a symbiotic relation­ Stewart and Press, 1990; Press and Graves, ship with mycorrhizal fungi. Haustorial para­ 1995). sites, in contrast, directly penetrate host tissues The degree of nutritional dependence on the via a modified root called a haustorium and host varies among haustorial parasites. Hemi­ thereby obtain water and nutrients. Although parasites are photosynthetic during at least one such categories are often a matter of semantics, phase of their life cycle and derive mainly water we use the term parasite in a strict sense to refer and dissolved minerals from their hosts. Oblig­ to haustorial parasites. Angiosperm parasites are ate hemiparasites require a host plant to com­ restricted to the dicot subclasses Magnoliidae, plete their life cycles whereas facultative hemi­ Rosidae, and Asteridae; have evolved approxi­ parasites do not. Hemiparasites can be found in mately 11 times; and represent approximately Laurales (Cassytha), Polygalales (Krameria), 22 families, 265 genera, and 4,000 species, that and all families of Santalales. In Solanales is, about 1% of all angiosperms (Fig. 8.1). Ow­ (Cuscuta) and Scrophulariales, some species ing to their unique adaptations, parasitic plants are chlorophyllous hemiparasites whereas other have long been the focus of anatomical, mor­ species are achlorophyllous holoparasites. Holo­ phological, biochemical, systematic, and eco­ parasites represent the most extreme manifesta­ logical research (Kuijt, 1969; Press and Graves, tion of the parasitic mode because they lack This work was supported by grants from the National Science Foundation (DEB 94-07984 to DLN, DEB 91-20258 to CWO, and BIR 93-03630 to ADW), the Special Research Program of the Office of Research Development Administration, SIUC and the University Research Council of Vanderbilt University. Thanks go to C. Augspurger, W. Barthlott, I. Beaman, D. E. Bran, S. Carlquist, W. Forstreuter, I. Leebens-Mack, A. Markey, C. Marticorena, D. McCauley, S. Medbury, M. Melampy, Willem Meijer, B. Molloy, L. Musselman, R. N(lfayana, M. Nees, I. Paxton, S. Sargent, B. Swalla, W. Takeuchi, and M. Wetherwax for helpful discussions and/or for contributing plant material. The manuscript was improved by the criti­ cal comments of M. Bowe and an anonymous reviewer. 212 MOLECULAR SYSTEMATICS OF PLANTS /""/"""""~, ~ Sanlalales ~ Solanales , Ol"""c,,, (271190) , , IoIls.,.,.,ndr",,". (118) ~ ~ lor.nlh.ce•• (741cI. 700), , Opillacuc (9129) , , Sanlilleo•• (371480) ~ Boraginales ~ (Incl. Elcmoloptcl.ee.e) , l,nnoocial (315) ~ Vlsca.,.l. (6-7/350) /' ';J;;""""", Cynomoriales CynomOlloull (111·2) Scrophulariates Scrophul.... I""... s. I. >-oI;-.,........ I-~ (Incl. Orobancllae ..,) Q) ---- (ell. 78/1938) ~ Balanophorales Qj Silanophoflcen (3117) ~ Dac'Yllnlhle••• (314) 1-+-T3~ III D­ Sircophylace." (213) O lIlrlCOphlllee .. (6/9) '0 lophophylilceae (318) J: "C ;: !!! Cy1lnales :G Cy1In.ce.e (117) c:: o Z Raffieslales Rlllllllae.la .... (3119) Apodanlhac.ae (2119) IoIlIr.51.monlc ••" (112) Bdallophyron (III"') Hydnorales Hydnor.ceae (2114-18) Figure 8.1. The distribution of haustorial parasitism among angiosperms. This generalized diagram incorporates information from global molecular phylogenetic studies using rbcL (Chase, Soltis, Olmstead et al. 1993) and nuclear 18S rONA (Soltis, Soltis, Nickrent et aI., 1997). No attempt was made to show all taxa, only to indicate groups that were supported by both studies. Hemiparasitic angiosperms are enclosed within dashed borders and holoparasites by black borders. Both trophic modes occur in Scrophulariaceae s. I. and Cuscuta. Arrows that touch a group indicate that strong evidence exists for the placement of that parasitic taxon within the group. Uncertain affinities are indicated by arrows with question marks. The familial classification of the nonasterid holoparasites is modified from Tahktajan (1987); however the placement of these orders is not concordant with his superordinal classification. The number of genera and species is indicated in parentheses following each family name. For Scrophulariaceae s. I., only the parasitic members are tabulated. photosynthesis and must rely upon the host for (1988) by excluding Balanophoraceae, Medu­ both water and inorganic and organic nutrients. sandraceae, and Dipentodontaceae. Morpholog­ Six groups (orders or families-Fig. 8.1) are ical, cytological, and molecular evidence all represented entirely by holoparasites: Bal­ point toward the separation of Cynomoriaceae anophorales, Cynomoriaceae, Cytinaceae, Hyd­ from Balanophoraceae and of Cytinaceae from noraceae, Lennoaceae, and Rafflesiales. Rafflesiaceae (Takhtajan, 1987; Nickrent and The relationships shown in Figure 8.1 are Duff, 1996; Pazy et aI., 1996). Traditional clas­ based upon those of Takhtajan (1987) as well as sifications have often allied these holoparasites results of recent molecular analyses. For this pa­ with Santalales; however, considerable variation per, Santalales are considered, in a strict sense, can be seen in alternate classifications and such to include Olacaceae, Misodendraceae, Loran­ an affinity is not apparent from molecular inves­ thaceae, Opiliaceae, Santalaceae, and Viscaceae. tigations (see below). Given this, the term This composition differs from that of Cronquist nonasterid holoparasites will be used to distin- PARASITIC PLANTS 213 guish these plants from holoparasites in Asteri­ sites. Our goals are to demonstrate the utility of dae such as those found in Scrophulariales, Bor­ these molecular markers in documenting phylo­ aginales, and Solanales. The nonasterid holo­ genetic relationships (at the genus level and parasites are Balanophorales, Cynomoriales, above) and to show how parasitic plants repre­ Cytinales, Hydnorales, and Rafflesiales (Fig. sent unique models that can be used to study 8.1). Results of molecular phylogenetic studies molecular evolutionary and genetic processes indicate that the nonasterid holoparasites are not such as the structure, function, and evolution of closely related to each other (Nickrent and Duff, plant genomes. For example, the continuum of 1996), hence the term is applied for reference trophic modes in Scrophulariaceae s. 1. from purposes only. nonparasitic to hemiparasitic to holoparasitic Additional parasitic plants can be found in makes this group ideal for investigating ques­ Scrophulariaceae, a large family (over 250 gen­ tions concerning the evolution of parasitism and era) that is broadly and at present inexactly de­ the molecular changes that accompany adapta­ fined within Scrophulariales/Lamiales (Burtt, tion to a heterotrophic existence. Holoparasitic 1965; Cronquist, 1981; Thorne, 1992; Olmstead plants that show increased rates of molecular and Reeves, 1995). Although most species are evolution pose particular problems for phyloge­ completely autotrophic, the members of two netic analysis but at the same time provide tribes (Buchnereae, Pediculareae; [Pennell, intriguing subjects for studying genome reorga­ 1935]) display a wide range of parasitic modes nizations that accompany the loss of photosyn­ from fully photosynthetic, facultative hemipara­ thesis. sites, to nonphotosynthetic holoparasites. Orobanchaceae are a group of nonphotosyn­ Problems with the Classification thetic holoparasites closely related to the of Parasitic Plants holoparasitic Scrophulariaceae. A continuum of morphological and physiological traits unites The placement of many parasitic plants within the two, as do several "transitional genera" (Har­ the global angiosperm phylogeny is not dis­ veya, Hyobanche, and Lathraea) that have been puted. For example, despite questions about fa­ classified alternatively in one family or the other milial boundaries and interfamilial relationships (Boeshore, 1920; Kuijt, 1969; Minkin and Esh­ within Scrophulariales, it is clear that Scrophu­ baugh, 1989). Orobanchaceae are alternatively lariaceae are allied with other sympetalous di­ included within Scrophu1ariaceae (Takhtajan, cots of Asteridae s. 1. Such is not the case, how­ 1987; Thorne, 1992) or recognized, by tradition, ever, for the nonasterid holoparasites whose at the family level (Cronquist, 1981). Regardless higher-level classification still remains prob­ of rank, most workers are in agreement that lematic. Two processes that occur during the Orobanchaceae are derived from within the par­ evolution of advanced parasitism are reduction asitic Scrophulariaceae. Numerous lines of evi­ (and/or extreme modification) of morphological dence support this conclusion, including the features and convergence. The first may involve shared presence of several morphological char­ loss of leaves, chlorophyll, perianth parts, or acters (Boeshore, 1920; Weber, 1980), pollen even ovular integuments. Loss of features con­ features, and a derived chloroplast DNA restric­ founds