MBE Advance Access originally published online on January 24, 2008
Molecular Biology and Evolution 2008 25(4):748-761; doi:10.1093/molbev/msn022
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
Genes of Cyanobacterial Origin in Plant Nuclear Genomes Point to a Heterocyst-Forming Plastid Ancestor
* Institut für Botanik III, Heinrich-Heine Universität Düsseldorf, Universitätsstrasse 1, Düsseldorf, Germany
Department of Biology and Biochemistry, University of Houston
School of Biological and Chemical Sciences, Queen Mary, University of London, London, United Kingdom
E-mail: tal.dagan{at}uni-duesseldorf.de.
Accepted for publication January 20, 2008.
Plastids are descended from a cyanobacterial symbiosis which occurred over 1.2 billion years ago. During the course of endosymbiosis, most genes were lost from the cyanobacterium's genome and many were relocated to the host nucleus through endosymbiotic gene transfer (EGT). The issue of how many genes were acquired through EGT in different plant lineages is unresolved. Here, we report the genome-wide frequency of gene acquisitions from cyanobacteria in 4 photosynthetic eukaryotes—Arabidopsis, rice, Chlamydomonas, and the red alga Cyanidioschyzon—by comparision of the 83,138 proteins encoded in their genomes with 851,607 proteins encoded in 9 sequenced cyanobacterial genomes, 215 other reference prokaryotic genomes, and 13 reference eukaryotic genomes. The analyses entail 11,569 phylogenies inferred with both maximum likelihood and Neighbor-Joining approaches. Because each phylogenetic result is dependent not only upon the reconstruction method but also upon the site patterns in the underlying alignment, we investigated how the reliability of site pattern generation via alignment affects our results: if the site patterns in an alignment differ depending upon the order in which amino acids are introduced into multiple sequence alignment—N- to C-terminal versus C- to N-terminal—then the phylogenetic result is likely to be artifactual. Excluding unreliable alignments by this means, we obtain a conservative estimate, wherein about 14% of the proteins examined in each plant genome indicate a cyanobacterial origin for the corresponding nuclear gene, with higher proportions (17–25%) observed among the more reliable alignments. The identification of cyanobacterial genes in plant genomes affords access to an important question: From which type of cyanobacterium did the ancestor of plastids arise? Among the 9 cyanobacterial genomes sampled, Nostoc sp. PCC7120 and Anabaena variabilis ATCC29143 were found to harbor collections of genes which are—in terms of presence/absence and sequence similarity—more like those possessed by the plastid ancestor than those of the other 7 cyanobacterial genomes sampled here. This suggests that the ancestor of plastids might have been an organism more similar to filamentous, heterocyst-forming (nitrogen-fixing) representatives of section IV recognized in Stanier's cyanobacterial classification. Members of section IV are very common partners in contemporary symbiotic associations involving endosymbiotic cyanobacteria, which generally provide nitrogen to their host, consistent with suggestions that fixed nitrogen supplied by the endosymbiont might have played an important role during the origin of plastids.
Key Words: endosymbiosis • phylogenomics • multiple sequence alignment • plastid origin • nitrogen
Takashi Gojobori, Associate Editor