A Plastid of Probable Green Algal Origin in Apicomplexan Parasites

See allHide authors and affiliations

Science  07 Mar 1997:
Vol. 275, Issue 5305, pp. 1485-1489
DOI: 10.1126/science.275.5305.1485


Protozoan parasites of the phylum Apicomplexa contain three genetic elements: the nuclear and mitochondrial genomes characteristic of virtually all eukaryotic cells and a 35-kilobase circular extrachromosomal DNA. In situ hybridization techniques were used to localize the 35-kilobase DNA of Toxoplasma gondii to a discrete organelle surrounded by four membranes. Phylogenetic analysis of the tufA gene encoded by the 35-kilobase genomes of coccidians T. gondii and Eimeria tenella and the malaria parasite Plasmodium falciparum grouped this organellar genome with cyanobacteria and plastids, showing consistent clustering with green algal plastids. Taken together, these observations indicate that the Apicomplexa acquired a plastid by secondary endosymbiosis, probably from a green alga.

Apicomplexan parasites contain two maternally inherited extrachromosomal DNA elements (1). The mitochondrial genome is a multicopy element of ∼6 to 7 kb encoding three proteins of the respiratory chain and extensively fragmented ribosomal RNAs (2). In addition, these parasites contain a 35-kb circular DNA molecule with no significant similarity to known mitochondrial genomes. The 35-kb element is similar to chloroplast genomes, containing an inverted repeat of ribosomal RNA genes and genes typically found in chloroplasts but not mitochondria (rpoB/C, tufA, and clpC) (3). The 35-kb DNA is also predicted to encode a complete set of tRNAs, numerous ribosomal proteins, and several unidentified open reading frames (3).

We used in situ hybridization to determine whether the 35-kb DNA is found within the parasite nucleus, mitochondrion, or cytoplasm or, alternatively, whether this molecule localizes to a previously unidentified DNA-containing organelle. We chose T. gondii for this project (rather than Plasmodium, in which the 35-kb element has been better characterized) for two reasons. First, there are approximately eight copies of the 35-kb circle per haploid genome in T. gondii tachyzoites, as opposed to approximately one copy in Plasmodium. Second, Toxoplasma offers much better ultrastructural resolution, because of its regular organization of intracellular organelles and well-defined apical region. To localize the 35-kb DNA, we hybridized extracellular tachyzoites with digoxigenin-labeled DNA probes that covered 10.5 kb of the 35-kb genomic sequence but excluded the ribosomal genes, to avoid cross-hybridization with the mitochondrial genome (4). We also targeted RNA transcripts derived from the 35-kb genome, using digoxigenin-labeled antisense RNA generated from putative rps4 sequences (5). The DNA:DNA or RNA:RNA hybrids were visualized by fluorescence in situ hybridization (FISH), and nuclear DNA was counterstained with the fluorescent dye YOYO-1. Examination by laser-scanning confocal microscopy revealed that the 35-kb DNA of T. gondii is localized to a specific region in the cell, adjacent to (but distinct from) the apical end of the parasite nucleus (Fig. 1A). Transcripts of rps4 were also concentrated in this region (Fig. 1B), suggesting that diffusion of 35-kb DNA-related transcripts is restricted by a physical (possibly membranous) barrier.

Fig. 1.

The 35-kb episomal genome and 35-kb derived RNA transcripts localize to a specific region adjacent to the nucleus in T. gondii tachyzoites. (A) Pseudocolor image of T. gondii tachyzoites hybridized with digoxigenin-labeled 35-kb genome-specific DNA (26). The DNA:DNA hybrids were visualized with rhodamine-conjugated anti-digoxigenin (red), and nuclear DNA was counterstained with YOYO-1 (green). Signals derived from the two fluorophores were collected independently by laser scanning confocal microscopy and merged with phase-contrast images simultaneously collected from the transmitted-light flow-through from the confocal microscope. (B and C) Localization of 35-kb DNA-encoded rps4 transcripts (27). Tachyzoites were hybridized with digoxigenin-labeled (B) antisense or (C) sense RNA generated in vitro from a cloned DNA fragment spanning the putative rps4 gene and visualized with rhodamine as above (red); nuclei were counterstained with YOYO-1 (green). (D and E) Extrachromosomal DNA in T. gondii tachyzoites. Fixed parasites were stained for 20 min at 25°C with ∼2 μg/ml of Hoechst 33258 in 1× SSC and examined by conventional epifluorescence microscopy with a Zeiss Axiovert 35 equipped with an ultraviolet filter set. A distinct extranuclear signal is seen in extracellular tachyzoites (D). Intracellular tachyzoites (E) orient in “rosettes,” with their apical ends pointed outward (28), permitting localization of the extranuclear DNA to the apical juxtanuclear region. (F through H) Co-localization of extranuclear DNA and 35-kb genome-specific sequences. Nuclei were labeled with YOYO-1 and extranuclear DNA with an antibody directed against DNA, followed by a fluorescein-conjugated secondary antibody (green). (Nuclear DNA was not labeled by the antibody to DNA except under extraction conditions, that destroyed parasite morphology, presumably because binding is blocked by chromatin-associated proteins.) The extranuclear DNA co-localizes with in situ hybridization probes derived from the 35-kb element (red). (F and G) Green and green + red images of the same field (containing two parasites); (H) green and red fluorescence signals from a different parasite, merged with the corresponding phase-contrast image. Scale bars, 5 μm.

Extranuclear DNA was not detected by YOYO-1 (or propidium iodide), presumably because of the low DNA concentrations typically found in non-nuclear organelles and the membrane-impermeable nature of these dyes. However, the extranuclear signal obtained by FISH resembled the pattern observed after staining with sensitive membrane-permeable DNA dyes such as Hoechst 33258 or 4′,6′-diamidino-2-phenylindole (DAPI) (Fig. 1, D and E). To compare the subcellular distribution of extranuclear DNA with the 35-kb DNA-derived FISH signal (Fig. 1, F though H), we used a monoclonal antibody to DNA because neither Hoechst nor DAPI stains are excited by the Kr-Ar laser that was available for confocal imaging and because in situ signals were difficult to detect on a conventional fluorescence microscope.

To examine the subcellular location of the 35-kb DNA more precisely, we hybridized frozen ultrathin sections with digoxigenin-labeled DNA probes (Fig. 2, A and B). Staining with antidigoxigenin followed by a secondary antibody and gold-conjugated protein A localized the 35-kb element to a membranous region adjacent to the nucleus but distinct from either the mitochondrion or the Golgi apparatus (large gold particles). Antibody to DNA also stained this area (small gold particles). In control experiments, probes prepared from plasmid vector DNA showed no hybridization, although the antibody to DNA still detected the membranous region just apical to the nucleus. The morphology of the membranous structure labeled by 35-kb DNA probes is difficult to resolve under the harsh conditions used for in situ hybridization, but conditions suitable for labeling with antibody against DNA alone revealed an organelle associated with multiple membranes (Fig. 2C). Thin sections through Epon-embedded parasites (which provide superior membrane preservation but do not permit antibody or in situ labeling) show that this organelle is invariably enclosed by four bilayer membranes (Fig. 2, D and E).

Fig. 2.

Ultrastructural localization of 35-kb genome-specific DNA to a unique organelle enclosed by four membranes in T. gondii tachyzoites. (A) Longitudinal ultrathin cryosection of T. gondii tachyzoites hybridized with digoxigenin-labeled probes derived from the 35-kb DNA (29). Digoxigenin was visualized with antibodies and protein A coupled to 10-nm gold particles. Samples were further incubated with monoclonal antibody directed against DNA (which does not stain intact chromatin in the parasite nucleus; see Fig. 1 legend), followed by a secondary antibody with protein A coupled to 5-nm gold (30). (B) Higher magnification of the region in (A) showing gold labeling. The 35-kb DNA probes hybridize specifically with a membranous region (*) just apical to the nucleus (Nu) but are distinct from the mitochondrion (m) and Golgi apparatus (g). (C) Immunogold labeling of extranuclear DNA (10-nm gold particles) in a T. gondii tachyzoite not subjected to in situ hybridization conditions. Membranes appear white in this negatively stained image. (D and E) Ultrathin sections through the apicoplast (*) of an Epon-embedded parasite (31). The organelle is surrounded by four membranes (stained black by uranyl acetate). The parasite in (E) is beginning to divide, as indicated by division of the Golgi and development of the two daughter “buds.” The apicoplast is flattened adjacent to the apical end of the nucleus and is divided between the two daughters early during endodyogeny.

Previous phylogenetic studies on the 35-kb genome suggested a plastid ancestry, but confidence in this assessment has been low because of the limited number of taxa and phylogenetic methods used (6). Genes identified on the 35-kb element include tufA, encoding the protein synthesis elongation factor Tu, a gene previously found useful for constructing molecular phylogenies (7). Phylogenetic analysis of tufA sequences from T. gondii, P. falciparum, and E. tenella (8) places the apicomplexan 35-kb element solidly within the plastids (Fig. 3). This placement is robust when either amino acid alignments or nucleotide alignments first and second codon positions are analyzed under a variety of phylogenetic methods, including maximum likelihood, parsimony, and distance methods (using either Kimura three-parameter or LogDet transformation) (9). The association of apicomplexan tufA genes with those of plastids does not appear to be caused by either the AT-rich or the divergent nature of the sequences (10). The similarity of apicomplexan and plastid tufA genes is also supported by the presence of two insertions characteristic of plastids and cyanobacteria, although the length of these insertions is variable among the Apicomplexa.

Fig. 3.

Molecular phylogenetic analyses of tufA genes from three apicomplexan 35-kb genomes and representative eubacteria, plastids, and mitochondria (32). Maximum likelihood finds the phylogeny that is statistically most likely to have given rise to the observed sequences. Neighbor joining is a cluster method, in this case using “LogDet” distances (−ln determinant). Parsimony finds the tree that requires the fewest inferred mutations to represent the data (9). Branch lengths are proportional to the number of inferred substitutions (or LogDet value); bootstrap values >40% are given above the corresponding branch (11). The column at the far right indicates the number of membranes surrounding the plastid for taxa in the parsimony tree. All three phylogenetic methods consistently group the apicomplexan 35-kb encoded tufA genes with green algal plastids.

All three phylogenetic methods used support monophyly of all plastids, including the apicomplexan 35-kb element. Resampling methods that test the internal consistency of phylogenetic patterns within the data gave bootstrap values (11) of 75, 39, and 88% for monophyly of plastids (for maximum likelihood, LogDet-neighbor-joining, and parsimony analyses, respectively) and 81, 69, and 94% for monophyly of plastids and cyanobacteria (Fig. 3). Parsimony analysis of nucleotide data scored only for transversion events also provides strong support for the clade composed of cyanobacteria and plastids (94%), and moderate support (78%) for plastid monophyly. The apicomplexan plastids were consistently placed among the green algae by all analytical methods used. Although support for green algal affinity was weak (bootstrap values of 41, 21, and 63%), these values are comparable to the level of support for green plastid monophyly when apicomplexans are excluded, yet the green plastids are known to be monophyletic on many other grounds (7). Trees constrained to place the Apicomplexa with nongreen plastids were consistently worse than those placing them with the green plastids, although the difference in likelihood was not significant by the Kishino-Hasegawa test (9).

Many investigators have assumed that the apicomplexan 35-kb genome is related to dinoflagellate plastids, on the basis of structural similarities between the Apicomplexa and dinoflagellates, and phylogenetic analyses of nuclear genes (12). Unfortunately, few dinoflagellate plastid genes have been examined, but there is considerable diversity of plastid form among dinoflagellates, and their plastids may have arisen from multiple distinct endosymbioses (13). Thus, it seems likely that the last common ancestor of all dinoflagellates was not photosynthetic and that the Apicomplexa and dinoflagellates acquired their plastids independently.

A structure consisting of multiple membranes has previously been described as the “Golgi adjunct” in Toxoplasma, and similar structures—variously termed the lamellärer körper, vacuoles plurimembranaires, spherical body, or Hohlzylinder—have been observed in other apicomplexan parasites (14). The cytological derivation of this structure has been unclear, but the demonstration that this organelle is associated with a plastid genome in Toxoplasma—combined with the monophyly of Toxoplasma, Plasmodium, and Eimeria tufAs in all analyses—argues for a single endosymbiotic organelle common to all apicomplexans. The apicomplexan plastid (abbreviated “apicoplast”) is an authentic plastid in all respects, albeit one that is probably incapable of photosynthesis.

Previous investigators have debated the number of membranes surrounding the apicoplast, suggesting that the appearance of multiple membranes may result from proximity to the endoplasmic reticulum or Golgi apparatus (14, 15). Although this organelle is closely associated with the Golgi, the fixation and staining conditions used for Fig. 2, D and E, commonly show four membranes. It is difficult to visualize distinct membranes all the way around the organelle (and serial sections necessarily lose definition at the top and bottom of the stack), but all of our micrographs are consistent with the four-membrane hypothesis, and many sections are clearly incompatible with ≤3 or≥5 membranes. The presence of four membranes enclosing the apicoplast suggests that it originated as a secondary endosymbiont (derived by ingestion of a eukaryote that itself harbored a plastid), analogous to the plastids of chlorarachniophytes and cryptomonads (16). This hypothesis is bolstered by the phylogenetic grouping of apicoplasts with green algal plastids, which presents a clear conflict with nuclear gene phylogenies (12, 17) and therefore provides prima facie support for a secondary endosymbiotic origin. The putative green algal origin of apicomplexan plastids should be testable through further phylogenetic analyses of plastid sequences and analysis of apicomplexan nuclear genes of potential green algal origin, such as phosphoglucose isomerase and enolase (18).

The function of the apicoplast remains unknown, but the parasite faithfully replicates this organelle, which divides by binary fission and is introduced into developing daughter parasites very early during replication (Fig. 2E). The apicoplast genome is certainly transcribed: Several transcripts have been identified by Northern (RNA) blot analysis (3, 19), rps4 transcripts localize to the same region as the 35-kb DNA (Fig. 1), and ribosomal RNA derived from the 35-kb circle has been localized to this organelle (15). Like other endosymbiotic genomes (20), the 35-kb element is presumed to be the remnant of a much larger precursor, most of whose original functions have been lost or transferred to the nuclear genome. Photosynthesis is the most familiar function of plastids, and evidence for a chlorophyll binding protein in Apicomplexa has been reported (21), although we have not been able to confirm these results in Toxoplasma. Plastids also play many other key metabolic roles—including biosynthesis of amino acids and fatty acids, assimilation of nitrate and sulfate, and starch storage (22)—and have been maintained in many nonphotosynthetic taxa over millions of years (23). The apicoplast has been suggested as a target for macrolide antibiotics in Toxoplasma (24) and may also be the target for rifampicin in Plasmodium (25). Further studies are likely to elucidate important aspects of plastid function and evolutionary history, in addition to identifying other parasite-specific targets for chemotherapy.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.

Stay Connected to Science

Navigate This Article