sign in sign up
<<
Home , Amphidinium, Apicomplexa, Ciliate, Colpodella, Haemosporida, Oligohymenophorea

Proc. Natl. Acad. Sci. USA Vol. 92, pp. 5793-5797, June 1995

Evolutionary origin of and other based on rRNA (/ evolution/origin of //) ANANIAS A. ESCALANTE AND FRANCISCO J. AYALA* Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92717 Contributed by Francisco J. Ayala, March 7, 1995

ABSTRACT We have explored the evolutionary history of Table 1. Twenty Apicomplexa and their hosts the Apicomplexa and two related protistan phyla, Dinozoa and Definitive Ciliophora, by comparing the nucleotide sequences of small Intermediate or subunit ribosomal RNA genes. We conclude that the Plasmo- dium lineage, to which the malarial parasites belong, diverged Code Species host vector from other apicomplexan lineages (piroplasmids and coccid- Hematozoa, order ians) several hundred million years ago, perhaps even before Pfa Plasmodium Human the . The Plasmodium radiation, which gave rise to falciparum several species parasitic to humans, occurred 129 million Pma Plasmodium Human Mosquito years ago; Plasmodium of humans has indepen- malariae dently arisen several times. The origin of apicomplexans Pvi Human Mosquito (Plasmodium), dinoflagellates, and may be >1 billion Pbe Rodent Mosquito years old, perhaps older than the three multicellular king- Pga Plasmodium Bird Mosquito doms of , , and fungi. Digenetic parasitism gallinaceum independently evolved several times in the Apicomplexa. Class Hematozoa, order Bbo Babesiiq bovis Cattle, deer The Apicomplexa is a large and complex protist , Bca canis Canid Tick Bcb Babesia consisting of nearly 5000 described species, with as many as caballi Equid Tick 60,000 yet to be named (ref. 1, pp. 1-6; see also refs. 2 and 3). Beq Babesia equi Equid Tick The apicomplexans are all parasites, characterized by the Cfe felis Felid Unknown "apical complex," a structure that inspired the phylum's name. Thu buffeli Buffalo Tick Many species have economical and medical importance, Tpa Cattle Tick among them the malarial Plasmodium, one of the Tta Theileria taurotragi Cattle Tick greatest causes of mortality in the world. Tun Theileria sp. Antelope Tick? The and phylogeny of the Apicomplexa have been Class Coccidea the subject of controversy and frequent revision. At stake are Sgi gigantea Sheep Canid such matters as the evolutionary origin of Plasmodium- Nca caninum Unknown Canid whether it derives from gut parasites of or from Tgo Mammal Felid originally monogenetic (monoxenous) parasites of dipterans Cpa None Rodent (4-8). parvum We have analyzed the small subunit (SSU) rRNA genes Cwr Cryptosporidium None Rodent from 20 species that belong to three classes (we follow the wrairi taxonomy of ref. 2): Hematozoea (where Plasmodium is in- Class Perkinsidea cluded), Coccidea, and Perkinsidea. We use as outgroups 10 Per sp. None species from two related phyla-Dinozoa (dinoflagellates) and Accession nos. and references are as follows: Pfa, M19172 (9); Pma, Ciliophora (ciliates). M54897 (10); Pvi, X13926 (11); Pbe, M14599 (12); Pga, M61723 (13); We conclude that the Plasmodium lineage diverged from Bbo, L19078 (14); Bca, L19079 (14); Bcb, Z15104 (14); Beq, Z15105 other Apicomplexa several hundred million years ago, perhaps (14); Cfe, L19080 (14); Thu, Z15106 (14,15); Tpa L02366 (14, 15); Tta, earlier than the Cambrian, before the vertebrates () L19082 (14, 15); Tun, L19081 (14, 15); Sgi, L24384; Nca, L24380 (16); originated from their ancestral lineage. The Api- Tgo, M97703 (17); Cpa, L16996; Cwr, U11440; Per, L07375 (18). complexa phylum is very ancient, perhaps as old as the three are often considered the sister of thePlasmodium lineage multicellular kingdoms plants, fungi, and animals. (20). We also include 5 species of the class Coccidea, from which the Plasmodium lineage may have originated (1, 20, 21), MATERIALS AND METHODS and 1 species of the class Perkinsidea. The 10 species used as outgroups (Table 2) belong to the phyla Dinozoa and Cilio- The 20 Apicomplexa species (Table 1) represent three of the phora, akin to the Apicomplexa to the extent that Cavalier- four classes that make up the phylum (2). The Apicomplexa Smith (29) has included all three in the parvkingdom Alveo- phylum of Corliss (2) is essentially identical to the array lata. classified earlier as Sporozoa (19), except for the inclusion of We have aligned the 30 SSU rRNA sequences by means of Perkinsus, which remains controversial. We include 14 species the CLUSTAL-V program (30), with manual editing for maxi- of Hematozoea (5 species of Plasmodium, order Haemospo- mizing similarity between the sequences. The final alignment rida, and 9 species of the order Piroplasmida). The piroplasms consists of 1550 sites (of "1800 in each sequence).

The publication costs of this article were defrayed in part by page charge Abbreviations: ML, maximum likelihood; My, million years; NJ, payment. This article must therefore be hereby marked "advertisement" in neighbor-joining; SSU rRNA, small subunit rRNA. accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed. 5793 5794 Evolution: Escalante and Ayala Proc. Natl. Acad. Sci. USA 92 (1995)

Table 2. Ten species of dinoflagellates and ciliates, with accession comparing the SSU rRNA genes of several sets of increasingly nos. for SSU rRNA sequences divergent multicellular (mostly metazoan) organisms, for Accession which approximate times of divergence are known. Code Species no. Ref. Phylum Dinozoa, class Dinoflagellatea RESULTS Abe belauense L13719 22 The NJ trees are shown in Figs. 1 (based on all substitutions) Ata Alexandrium tamarense X56946 23 and 2 (transversions only). The five Plasmodium species form Pmi Prorocentrum micans M14649 24 a monophyletic clade in both trees, with full statistical reli- Ssp sp. M88509 25 ability (bootstrap value, 100%). The nine piroplasm species Smi Symbiodinium microadriaticum M88521 25 also form a monophyletic clade (99% and 98% bootstrap). Phylum Ciliophora, class The Plasmodium and piroplasm (two distinct orders Cca Colpidium campylum X56532 26 within the class Hematozoea) are sister clades in Fig. 2, but Eae aediculatus X03949 27 without statistical reliability (bootstrap, 66%; 70% is a com- Gch Glaucoma chattoni X56533 26 monly used boundary for statistical significance; ref. 40). In Ohe Opisthonecta henneguyi X56531 26 Fig. 1, the piroplasms appear as the sister clade to three Pte tetraurelia X03772 28 coccidian species (, Sarcocystis gigantea, and Toxoplasma gondii, family ) but also without Phylogenetic relationships are inferred by two methods: (i) statistical reliability (60%). neighbor joining (NJ) (31), with Tamura's three-parameter The three Sarcocystidae species (N. caninum, S. gigantea, distance content in these from and T. gondii) form a monophyletic clade in both figures, and (32). G+C genes ranges 35% so do the two other coccidians ( and to 50%; we use the averages for G+C content and transition/ Cryptosporidium wrairi, family Cryptosporidae), but the class transversion ratio. We calculate distances for all substitutions Coccidea is not a monophyletic clade. as well as for transversions only. Tree reliability is assessed by The phylum Apicomplexa forms a monophyletic clade, if the bootstrap method (33) with 1000 replications. All NJ Perkinsus is excluded, but the clade is statistically valid only analyses are performed using the program MEGA version 1.0 when transversions alone are used (82%, Fig. 2; bootstrap for (34). The genetic distances as well as the sequence alignment this clade in Fig. 1 is 51%). The anomalous genus Perkinsus is are available from the authors upon request. included within the clade (79% and 74% in Figs. (ii) Maximum likelihood (ML) (35) assumes specific tran- 1 and 2, respectively). The taxonomic and phylogenetic posi- sition/transversion ratios. We use seven transition/transver- tion of Perkinsus has been a matter of uncertainty and debate sion ratios (from 1 to 15), which yield similar results. Tree (1, 2, 18, 21, 29, 41). The dinoflagellate species (phylum topologies are compared as described (36). Analyses are Dinozoa), with Perkinsus included, form a monophyletic clade, performed with the algorithm DNAML of the PHYLIP package, and so do the ciliates (phylum Ciliophora). version 3.5c (ref. 37; program available from J. Felsenstein, We have attempted to resolve the ambiguous NJ phyloge- Department of Genetics, University of Washington, Seattle). netic relationships within the Apicomplexa phylum by ML We estimate time of divergence by using two rates of methods (data not shown). The tree that associates Plasmo- nucleotide substitutions: 2% (38, 39) and 0.85% per 100 dium and the piroplasms as sister clades (Fig. 2) is the best ML million years (My). We have obtained the 0.85% rate by tree under all transition/transversion values. But this tree is Pfa Pga 100 Pma Pvi Pbe

Eae 0 .01

FIG. 1. Phylogenetic relationships among 30 protistan species based on all substitutions. See Tables 1 and 2 for abbreviations of species names. Bootstrap values are shown on the branches; some values below 60 are omitted for clarity. Evolution: Escalante and Ayala Proc. Natl. Acad. Sci. USA 92 (1995) 5795

Pfa 100 Pga Pma Pvl 66 Pbe 95 Bca Bbo 98 Bcb Cfe 71 Beq Tpa 82 Tta Tbu Tun

6 .01 FIG. 2. Phylogenetic relationships among 30 protistan species based on transversions. See Tables 1 and 2 for abbreviations. not statistically different from trees that show the combination be calibrated by associating one or more points in a phylogeny of piroplasms and the coccidian Sasrcocystidae as the sister with precise times or geological events. clade ofPlasmodium (Fig. 1) or the association ofPlasmodium The substitution rate of SSU rRNA genes has been esti- and the coccidian Cryptosporidae as the sister clade to the mated to be 2% per 100 My (38, 39), although Moran et at (43) combination of piroplasms and Sarcocystidae. A tree that give a less precise rate of 2-4% per 100 My. One reservation associates coccidians with piroplasms, and this clade with against accepting at face value any one particular rate for the Plasmodium, is significantly worse than the best tree (only for SSU rRNA is that the set of nucleotides compared varies transition/transversion ratios of 10 or higher). from one study to another. Typically, as the array of lineages The phylogenetic affinity of Perkinsus with the dinoflagel- compared becomes more ancient, and hence more diverse, it lates is confirmed by the ML method. Any monophyletic tree becomes increasingly difficult to align the more variable that includes Perkinsus with the other Apicomplexa is statis- regions of the gene, so that only more conserved regions are tically worse than trees placing Perkinsus outside the Apicom- included in the comparisons. The estimated rate of evolution plexa. The exclusion of Perkinsus (and the class Perkinsidea, decreases accordingly. This effect has been noted by Escalante which comprises only another genus, ; ref. 2) from and Ayala (44): the genetic distances obtained for the rRNA the Apicomplexa conforms this phylum with the older phylum genes of Plasmodium species are 2-3 times larger when only Sporozoa (19), a name often treated as a synonym of Api- Plasmodium sequences are aligned than when other Apicom- complexa (ref. 2, p. 17). plexa genera are included. We have determined a rate applicable to our data as follows. We have selected sets of pairwise comparisons among SSU DISCUSSION rRNA gene sequences from increasingly divergent multicellu- The evolution of Plasmodium and other lar organisms for which reasonably accurate times of diver- Apicomplexa is of gence are known. The resulting alignment includes 1571 base particular interest because of the health and economic signif- pairs, close to the 1550 sites aligned for the 30 protistan icance of these parasites. Common views are that the Apicom- sequences. The regression of nucleotide substitutions on time plexa evolved from parasitic organisms of marine yields a rate of 0.85% substitutions per site per 100 My (Fig. (1, 20, 21); that Plasmodium evolved as a daughter clade from 3). the coccidians; and that the piroplasms are derived from, or at It is apparent in Figs. 1 and 2 that the rate of substitution of least are closely related to, Plasmodium (1, 20, 21). Some of the rRNA genes is not constant across all lineages. However, these views have been based on the postulate that "the we are primarily interested in the nodes shown in the simplified probable evolution of the Apicomplexa may be deduced from phylogeny of Fig. 4, but not even the branch averages yield that of the hosts," which leads to the conclusion, for example, equal substitution rates. The Plasmodium and lineages that the gregarines, which are exclusively parasitic on inver- evolve faster than the piroplasms, coccidians, or dinoflagel- tebrates, "surely evolved earlier than the other apicomplexan lates. The issue is which of the Apicomplexa substitution rates classes, which parasitize vertebrates" (ref. 3, p. 571). is most similar to the rate estimated for the multicellular There is no fossil record of apicomplexan organisms (42), organisms (Fig. 3). but time estimates of their evolution can be obtained by means We deal with the erratic behavior of the protistan molecular of a molecular clock. Two conditions need to be met: (i) the "clock" by making the following decisions (Table 3). The rate of evolution of the particular gene is constant through length of the branches from nod 1 (Fig. 4) is simply the average time for all lineages-i.e., there is a reliable clock; and (ii) the observed (Fig. 1) for the Plasmodium branches. The length of time rate of the clock is known, or alternatively the clock can the branches from node 2 is the average of the Plasmodium 5796 Evolution: Escalante and Ayala Proc. Natl. Acad. Sci. USA 92 (1995) lineage (0.094) and the average of the piroplasm and coccidian Vertebrates lineages (0.046). For node 3, we use the average of the Bony Fishes Plasmodium to dinoflagellate rate (0.103) and the coccidian Muiticellular Kingdoms | Reptiles and piroplasm to dinoflagellate (0.055). For node 4, we obtain / \ 4 4 jeBirds and Mammals the average of the Plasmodium to ciliate rate (0.133) and t 1 between ciliates and dinoflagellates, coccidians, or piroplasms (0.090). Times are obtained by using our estimated rate of 0.85% substitutions per site per 100 My; Table 3 also gives for Plasmodium comparison the times obtained with the 2% rate (38, 39). 2 The Coccidea include species parasitic to molluscs and r- Coccidians marine , but many are parasitic to mammals and other vertebrates (3). Many coccidians, such as Cryptosporidium, are 3 Piroplasms monogenetic parasites, able to complete their cycle within one single host (the "definitive" host; Table 1). But other 4 coccidians, including Neospora, Sarcocystis, and Toxoplasma, Dinoflagellates are digenetic (heteroxenous): their complete development requires two successive hosts. Table 3 gives 824 My for the Ciliates Apicomplexa radiation and thus for the age of the Coccidea. 1200 1000 800 600 400 200 Their old age implies that if they were parasites from the start, TIME My 0.10 FIG. 4. Simplified phylogeny of Apicomplexa, Dinozoa, and Cil- iophora. Scale above highlights the time of origin (first appearance of fossils) ofvertebrate groups. Uncertain origin of the , , and fungi kingdoms is indicated by parentheses. 0.08~~~~~~~~~~~~~ Vlill the coccideans' early hosts could not have been chordates or land organisms because these had not yet evolved. Even if we take 350 My (the lower estimate in Table 3) as the origin of the [] 0.06 *VI /< Coccidea, they would have existed >200 My before the origin of the modern mammal orders (such as Artiodactyla, Car- D00~~~~~~~~ *; nivora, and Primates) parasitized by coccidians. 0.04 The Haemosporida (the order to which Plasmodium be- longs) and the Piroplasmida are digenetic parasites: matura- tion of , fertilization, and sporogony occur in the I~~~~ hematophagous invertebrate vector; the rest of the life cycle is 0.02 completed in the blood of a . It is plausible that the monogenetic way of life was the primitive one, but the time 110R= 0.00 + 0.0085*10OMy z * ~~~~r=0.94 frame indicates that vertebrates most likely were not the original hosts. The current digenetic combination of an ar- 0.02 thropod (vector) and a vertebrate evolved relatively late in the 200 400 600 800 1000 1200 history of these lineages. My The issue remains as to which lineage, the vector's or the intermediate host's, represents the original monogenetic host. FIG. 3. Rate of evolution of the SSU rRNA genes estimated at 0.85 The adaptation ofPlasmodium and piroplasms to hemotrophy x 10-10 substitution per site per year (transitions and transversions). is often considered a derived character (3, 6, 8). However, Species included and GenBank accession nos. are as follows. Mam- some authors have proposed that the association of the mals: Homo sapiens, X03205; Mus musculus, X00686. Reptiles: Alli- Hematozoea with vertebrates is ancestral (1, 4, 21, 48). The gator mississippiensis, M59389; Heterodon platyrhinos, M59392. Birds: transmission by vectors and the origin of complex Gallus gallus, M59389; Turdus migratorius, M59402. Amphibians: Bufo life would have valliceps, M59386; Xenopus laevis, K01373. Bony fish: Fundulus het- cycles been, accordingly, a parasite adaptation eroclitus, M91180. Cartilaginous fish: Squalus acanthias, M91179. that evolved because it favors the parasites' rate of transmis- Agnatha: Lampetra aepyptera, M97573. Cephalochordate: Branchios- sion (49, 50). The phylogenetic relationships provide no de- toma floridae, M97571. , Insecta: Acyrthosiphon pisum, finitive clue for resolving this question. The Plasmodium X62623; Tenebrio molitor, X07801. Arthropods, Chelicerata: Eu- lineage predates the origin of the vertebrates. It is, therefore, rypelma californica, X13457;Amblyomma americanum, M60487. Mol- possible that the Plasmodium lineage evolved as a parasite in luscs: Placopecten magellanicus, X53899; Limicolaria kambeul, parallel with the lineage that eventually gave rise to the X66374. : Echinostoma caproni, L06567. Nematoda: Hae- vertebrates. The of monchus contortus, L04153. Jellyfish: Anemonia sulcata, X53498. divergence lizards, mammals, and birds : Microcionaprolifera, L10825. Plants: Zea mays, K02202;Alnus Table 3. Branch length and time of glutinosa, X54984. Fungi: Aspergillus fumigatus, M55626; Dipodascop- divergence sis uninucleata, U00969; Saccharomyces cerevisiae, J01353. Divergence Divergence times for these organisms are averages of the estimates (45-47). Ordinate values are half the pairwise genetic distances. Rate of Branch time, My substitution is the regression of genetic distance on time when setting Node Event length* 0.85t 2.Ot the origin at 0, which gives a fit of r = 0.94. Without the restriction that 1 Plasmodium radiation 0.011 129 55 it starts at 0, the + = regression is -0.697 0.009 x 100 My, with r 0.95. 2 Apicomplexa radiation 0.070 824 350 Roman numerals refer to the following I, vs. comparisons: mammals 3 Origin of Apicomplexa 0.079 reptiles vs. birds; II, amphibians vs. lineage I; III, agnatha (lamprey) 929 395 vs. vertebrates; IV, insects vs. chelicerata; V, protostomes (arthropods 4 Origin of the clade 0.112 1317 560 and molluscs) vs. lineage III; VI, acoelomates (flatworm) and Apicomplexa/Dinozoa pseudocoelomates () vs. lineage V; VII, porifera (sponge) vs. *Branch length is half the average genetic distance. radiata (jellyfish) and lineage VI; VIII, plants vs. fungi vs. animals. tRates in units of 10-10 substitution per site per year. Evolution: Escalante and Ayala Proc. Natl. Acad. Sci. USA 92 (1995) 5797 dates from the , 190-225 My ago (51), before the 15. Allsopp, B. A., Baylis, H. A., Allsopp, M. T., Cavalier-Smith, T., radiation of the Plasmodium genus. Fossil dipterans occur in Bishop, R. P., Carrington, D. M., Sohanpal, B. & Spooner, P. the Triassic and increase in diversity and abundance from the (1993) Parasitology 107, 157-165. 16. Ellis, J. T., Luton, K., Baverstock, P. R., Brindley, P., Nimmo, mid- through the Cenozoic (52, 53). The radiation of K. A. & Johnson, A. M. (1994) Mol. Biochem. Parasitol. 64, Plasmodium, which we have estimated at 129 My (Table 3), is 303-311. approximately coincident with the diversification of their 17. Johnson, A. M., Murray, P. J., Illana, S. & Baverstock, P. J. vectors' lineages. In any case, the molecular evidence indicates (1987) Mol. Biochem. Parasitol. 25, 239-246. that lateral transfers (i.e, new associations of parasite-vector 18. Goggin, C. L. & Barker, S. C. (1993) Mol. Biochem. Parasitol. 60, and parasite-vertebrate host) have occurred in the recent 65-70. evolution of malaria. The human Plasmodium falci- 19. Grasse, P. P. (1953) Traite de Zoologie (Masson, Paris), Vol. 1. parasites 20. Levine, N. D. (1985) inAn Illustrated Guide to the , eds. parum, , and Plasmodium vivax diverged Lee, J. J., Hunter, S. H. & Bovee, E. C. (Allen, Lawerence, KS), 129 My ago; yet they do not all parasitize the apes (or pp. 322-374. monkeys), from which the human lineage diverged much more 21. Levine, N. D. (1988) The Protozoan Phylum Apicomplexa (CRC, recently. Similarly, these three Plasmodium species are trans- Boca Raton, FL), Vol. 2. mitted by anopheline vectors of the same or very closely 22. McNally, K. L., Govind, N. S., Thome, P. E. & Trench, R. K. related species, which diverged from each other more recently (1994) J. Phycol. 30, 316-329. than the transmit. 23. Destombe, C., Cembella, A. D., Murphy, C. A. & Ragan, M. A. parasites they (1992) Phycologia 31, 121-124. Our results show that digenetic parasitism independently 24. Herzog, M. & Maroteaux, L. (1986) Proc. Natl. Acad. Sci. USA evolved several times in the Apicomplexa as suggested (8). The 83, 8644-8648. Apicomplexa radiation predates the origin of the land verte- 25. Rowan, R. & Powers, D. A. (1992) Proc. Natl. Acad. Sci. USA 89, brates (Table 3); yet digenetic parasitism affecting land ver- 3639-3643. tebrates has evolved in each of the three apicomplexan lineages 26. Greenwood, S. J., Sogin, M. L. & Lynn, D. H. (1991)1. Mol. Evol. shown in Fig. 4. 33, 163-174. The clade have at the 27. Sogin, M. L., Swanton, M. T., Gunderson, J. H. & Elwood, H. J. Apicomplexa/Dinozoa may originated (1986) J. Protozool. 33, 26-29. time of or earlier than the origin of the multicellular kingdoms. 28. Sogin, M. L. & Elwood, H. J. (1986) J. Mol. Evol. 23, 53-60. This is consistent with the conclusions of others (29, 54, 55). 29. Cavalier-Smith, T. (1993) Microbiol. Rev. 57, 953-994. Thus, the rate of 0.85 x 10-10 substitution per site per year is 30. Higgins, D. G., Bleasby, A. J. & Fuchs, R. (1992) Comput. Appl. approximately correct. Paleontological and molecular data Biosci. 8, 189-191. indicate that the radiation of red and green , ciliates, and 31. Saitou, N. & Nei, M. (1987) Mol. Biol. Evol. 4, 406-425. dinoflagellates occurred near the Mesoproterozoic-Neoprot- 32. Tamura, K. (1992) Mol. Biol. Evol. 9, 678-687. erozoic consistent with our estimate of 1317 33. Felsenstein, J. (1985) Evolution 39, 783-791. boundary (56), My 34. Kumar, S., Tamura, K. & Nei, M. (1993) MEGA, Molecular for the time of divergence of the ciliates from the ancestral Evolutionary Genetics Analysis (Pennsylvania State Univ. Press, lineage of dinoflagellates and Apicomplexa. University Park), Version 1.0. 35. Felsenstein, J. (1981) J. Mol. Evol. 17, 368-376. We are grateful to Walter Fitch, Richard Hudson, and Anthony 36. Kishino, H. & Hasegawa, M. (1989) J. Mol. Evol. 29, 170-179. James for comments on the manuscript. Work supported by a fellow- 37. Felsenstein, J. (1989) Cladistics 5, 164-166. ship from the Banco Interamericano de Desarrollo-Consejo Nacional 38. Ochman, H. & Wilson, A. C. (1987) J. Mol. Evol. 26, 74-86. de Ciencia y Tecnologia, Venezuela (A.A.E.) and National Institutes 39. Wilson, A. C., Ochman, H. & Prager, E. M. (1987) Trends Genet. of Health Grant GM42397 (F.J.A.). 3, 241-247. 40. Hillis, D. M. & Bull, J. J. (1993) Syst. Biol. 42, 182-192. 1. Levine, N. D. (1988) The Protozoan Phylum Apicomplexa (CRC, 41. Corliss, J. 0. (1984) BioSystems 17, 87-126. Boca Raton, FL), Vol. 1. 42. Margulis, L., McKhann, H. I. & Olendzenski, L. (1993) Illustrated 2. Corliss, J. 0. (1994) Acta Protozool. 33, 1-51. Guide of Protoctista (Jones & Bartlett, Boston). 3. Vivier, E. & Desportes, I. (1989) in Handbook ofProtoctista, eds. 43. Moran, N. A., Munson, M. A., Baumann, P. & Ishikawa, H. Margulis, L., Corliss, J. O., Melkonia, M. & Chapman, D. J. (1993) Proc. R. Soc. London B 253, 167-171. (Jones & Bartlett, Boston), pp. 549-573. 44. Escalante, A. A. & Ayala, F. J. (1994) Proc. Natl. Acad. Sci. USA 4. Manwell, R. D. (1955) Indian J. Malariol. 9, 247-253. 91, 11373-11377. 5. Garnham, P. C. C. (1966) Malaria Parasites and Other Haemo- 45. Clarkson, E. N. K. (1993) Invertebrate Paleontology and Evolution sporidia (Blackwell Scientific, Oxford), pp. 60-84. (Chapman & Hall, London), 3rd Ed., pp. 55-74. 6. Huff, C. A. (1938) Q. Rev. Biol. 13, 196-206. 46. Morris, S. C. (1993) Nature (London) 361, 219-225. 7. 47. Strickberger, M. W. (1990) Evolution (Jones & Bartlett, Boston). Mattingly, P. F. (1965) in Evolution of Parasites, ed. Taylor, 48. Baker, J. R. (1965) in Evolution ofParasites, ed. Taylor, A. E. R. A. E. R. (Blackwell Scientific, Oxford), pp. 29-45. (Blackwell Scientific, Oxford), pp. 1-27. 8. Barta, J. R. (1989) J. Parasitol. 75, 195-206. 49. Thompson, J. N. (1994) The Coevolutionary Process (Univ. Chi- 9. McCutchan, T. F., De la Cruz, V. F., Lal, A. A., Gunderson, cago Press, Chicago), pp. 102-111. J. H., Elwood, H. J. & Sogin, M. L. (1988) Mol. Biochem. Para- 50. Esch, G. W. & Fernandez, C. (1993) A Functional Biology of sitol. 28, 63-68. Parasitism (Chapman & Hall, New York), pp. 10-13. 10. Goman, M., Mons, B. & Scaife, J. G. (1991) Mol. Biochem. 51. Colbert, E. H. (1980) Evolution of the Vertebrates (Wiley, New Parasitol. 45, 281-288. York), 3rd Ed., pp. 1-13. 11. Waters, A. P., Higgins, D. G. & McCutchan, T. F. (1991) Proc. 52. Tasch, P. (1980) Paleobiology of the Invertebrates: Data Retrieval Natl. Acad. Sci. USA 88, 3140-3144. from the Fossil Record (Wiley, New York), pp. 607-617. 12. Gunderson, J. H., McCutchan, T. F. & Sogin, M. L. (1986) J. 53. Labandeira, C. C. & Sepkoski, J. (1993) Science 261, 310-314. Protozool. 33, 525-529. 54. Knoll, A. H. (1992) Science 256, 622-627. 13. Waters, A. P. & McCutchan, T. F. (1989) Nucleic Acids Res. 17, 55. Sogin, M. L., Edman, U. & Elwood, H. (1989) in The Hierarchy 2135. of Life, eds. Fernholm, B., Bremer, K. & J6rnvall, H. (Excerpta 14. Allsopp, M. T. E. P., Cavalier-Smith, T., De Waal, D. T. & Medica, New York), pp. 133-143. Allsopp, B. A. (1993) Parasitology 108, 147-152. 56. Knoll, A. H. (1994) Proc. Natl. Acad. Sci. USA 91, 6743-6750.