Technical Comments

Response to Comment on "A Green Algal Apicoplast Ancestor"

See allHide authors and affiliations

Science  04 Jul 2003:
Vol. 301, Issue 5629, pp. 49
DOI: 10.1126/science.1084684

We suggested lateral transfer of split cox2a and cox2b genes from a chlorophyte algal ancestor to an apicomplexan ancestor (1). Waller et al. (2) oppose this interpretation based on a phylogeny implying close affiliation of apicomplexan cox2a and cox2b genes with ciliate cox2 homologues; a 900-bp insertion that separates the corresponding domains in ciliate mitochondrial cox2 genes; and low hydrophobicity values for ciliate mitochondrial COXII and apicomplexan COXIIA and COXIIB subunits. They suggest that mitochondrial cox2 genes were split and transferred to the nuclei of the apicomplexan and chlorophyte ancestors during independent events.

We find their phylogenetic argument unconvincing because, as they admit, statistical support for the position of the ciliate sequences is lacking. Affiliation of ciliates and apicomplexans previously noted in mitochondrial sequence phylogenies should also be interpreted with caution because both lineages are rapidly evolving and may be artifactually grouped due to long-branch attraction (3). Furthermore, the addition of ciliate cox2 sequences to our previous phylogenetic analysis (1), does not show a relationship between ciliates and apicomplexans (Fig. 1, A and B). Nevertheless, we are pleased that in agreement with our findings in (1), Waller et al. (2) confirm the common branch for apicomplexan and chlorophyte COXIIA and COXIIB sequences—an otherwise unexpected affiliation.

Fig. 1.

Phylogenetic analysis of COXII. (A) Maximum likelihood (ML) tree. COXIIA and COXIIB (excluding MTS and extensions) were fused in silico as a single polypeptide and aligned with orthodox mitochondrial COXII sequences (14). (B) Phylogram showing branch lengths estimated with ML implementing the data, model, and parameters used to perform the ML search. COXII sequences for Tetrahymena pyriformis (NC_000862), T. thermophila (NC_003029), and Paramecium aurelia (NC_001324) were obtained from GenBank. (C) Multiple protein sequence alignment of the N-terminal regions of COXIIB sequences. Shown are the N-terminal extensions of nuclear cox2b genes from chlorophytes and apicomplexans (blue), the conserved residues in these domains (yellow), the conserved residues found only in chlorophytes and ampicomplexans (black background), the corresponding region of the 300-residue insertion in ciliate mitochondrial COXII (gray), and the conserved residues between ciliates and apicomplexans (bold). Droso (Drosophila melanogaster), Homo (Homo sapiens), Arabi (Arabidopsis thaliana), Mesos (Mesostigam viride), Theil (Theileria parva), Plasm (Plasmodium yoelli), Toxop (Toxoplasma gondii), Chlam (Chalmydomonas reinhardtii), Polyt (Polytomella sp.), Tther (Tetrahymena termophila), Tpyri (Tetrahymena pyriformis), and Param (Paramecium aurelia).

The presence of an insertion in the mitochondrial cox2 genes of ciliates does not necessarily imply kinship with apicomplexan fragmented cox2a and cox2b genes. The cox2 gene is susceptible to insertions between regions that encode the hydrophobic and hydrophilic domains of COXII. Insertions of variable lengths have been described in cox2 genes of unrelated organisms (4), including brown marine algae (NC_003055, NC_004024), microflagellates (NC_000946), and bacteria (NC_003888). Localization of insertions in cox2 is most likely constrained by the difficulty of inserting in the gene without disrupting the structure of the COXII protein (5). Ciliates also exhibit an insertion in cox1 and a fragmented mitochondrial nad1 (3). These rearrangements most likely occurred after the divergence of ciliates from apicomplexans and dinoflagellates (6), since no sequence remnants of insertions are present in cox1 in apicomplexans (7) or dinoflagellates (8). The cox1 and cox2 insertions were most likely absent in the alveolate ancestor and were acquired by ciliates independently. Therefore, the proposed vertical inheritance of cox2a and cox2b genes in apicomplexans is not directly supported by the presence of insertions in ciliate cox2.

The hydrophobicity analysis presented by Waller et al. (2) provides no evidence for a common origin of ciliate COXII and apicomplexan COXIIA and COXIIB sequences. We agree that hydrophobicity is one of the rate-limiting steps in functional gene transfer from the mitochondrion to the nucleus (5, 911). It is therefore not surprising that apicomplexan COXIIA and COXIIB subunits exhibit diminished hydrophobicity. Nevertheless, the low mean hydrophobicity values of ciliate mitochondrial COXII sequences merely indicate that, from the hydrophobicity point of view, the corresponding genes are ready to migrate—not that they migrated in the distant past. Ciliate mitochondrial DNA (mtDNA) contains more than the standard set of genes encoded by mtDNA in other eukaryotes (3) and therefore does not seem to show an increased rate of gene migration. The hydrophobicity of contemporary ciliate COXII is unlikely to bear on either the location of its gene or apicomplexan COXIIA and COXIIB phylogeny. In addition, the hydrophobicity analysis does not argue against the proposal that the original apicomplexan mitochondrial cox2 gene was eliminated after functional acquisition of chlorophyte cox2a and cox2b sequences (1).

COXIIB has a unique N-terminal extension that most likely arose since the splitting of the original mitochondrial cox2 gene of the chlorophyte ancestor (1). Protein sequencing has confirmed the presence of this extension in mature chlorophyte COXIIB (5). All chlorophyte and apicomplexan sequences share N-terminal extensions of COXIIB containing the conserved PxxxPxxY motif not present in the canonical COXII. These extensions seem homologous and suggest a common origin for chlamydomonad and apicomplexan COXIIA and COXIIB sequences. The proposal of Waller et al. (2) would imply that this domain conservation is due to convergent evolution, which we consider unlikely. The corresponding borders of the ciliate COXII insertions and the apicomplexan COXIIB extensions reveal no similarities (Fig. 1C).

We originally described rare characteristics that are shared solely between apicomplexans and certain chlorophyte algae (1), namely, the presence of nucleus-encoded split cox2a and cox2b genes and a conserved domain present in a region of COXIIB that is not derived from orthodox COXII. Functional fragmentation of a mitochondrial gene followed by functional integration in the nucleus is an extremely rare event, unlikely to happen several times. The current evidence suggests that this phenomenon was confined to the ancestor of chlorophyte algae. Our analyses support a close relationship between apicomplexan and chlorophyte cox2a and cox2b sequences specific to the mitochondrion, whereas analysis of apicoplast genome organization has suggested a red algal origin of the organelle (12). Clearly, there is still conflicting evidence for green versus red algal ancestry in the apicomplexans (13). Whatever the outcome of this debate, the presence of nucleus-encoded, fragmented cox2a and cox2b genes of green origin in apicomplexans must be considered whenever the evolutionary story of these parasites is reconstructed.

References and Notes

Stay Connected to Science

Navigate This Article