The endosymbiotic origin of plastids
Cyanobacteria
Plants
Plastids in Coleochaete
Primary Endosymbiosis
Zygnema
Glaucocystis
Gigartina, Fucus
Secondary Endosymbiosis
Conflict between rRNA and rbcL on plastid phylogeny
16S rRNA phylogeny from Douglas and Turner, 1991
rbcL phylogeny from Morden et al., 1992
Hypotheses concerning conflict
tufA as a third informative molecule
Insertion in tufA of cyanobacteria and plastids
Maximum likelihood analysis of tufA data
Bootstrap analysis
Maximum parsimony analysis of tufA data
Distance analysis (Dayhoff distances, neighbor joining tree)
Comparison of bootstrap values among the three methods
Bootstrap analysis
A statistical resampling method that provides a measure of how consistently the data support given phylogenetic conclusions
Each bootstrap value represents a "taxon bipartition", i.e., a division of all of the OTUs into two groups.
Bootstrapping does *not* tell you whether or not the group is correct
Rather it provides a measure of the strength of signal within the data, given the assumptions of the analytical method
rbcL has a different evolutionary history than other plastid genes
A reasonable concern would be that the rbcL analysis are misleading because of problems with the analyses
The effect of base composition
Base compositional bias is one form of pattern in the data that can affect phylogenetic conclusions
The tufA data do show compositional bias
However, methods that are thought to perform better under such conditions (e.g., LogDet) produce similar results
In addition to tufA, several other molecules have taxonomic sampling adequate to address the origin of the primary plastid lineages
These consistently agree with rRNA, with only the rubisco genes showing a different pattern
This congruence provides strong support for the conclusion that all plastids are derived from cyanobacteria
Congruence among five molecules, with rbcL as the exception
Because rbcL seems to have a different evolutionary history than the other plastid genes, the only direct source of information concerning its evolutionary history is study of the gene itself.
Maximum parsimony tree of rbcL (using amino-acid data)
Critical features of the rubisco tree
Form I rubiscos are divided into two types
Within-types comparisons show 70% or higher amino acid identity.
Between-type comparisons show at most 60% amino acid identity.
Several groups of organisms are divided between the two major types of form I rubisco
This pattern is not consistent with other phylogenetic information
Comparing hypotheses of horzontal gene transfer and gene duplication
The phylogenetic distribution of rubisco genes could be explained either by gene transfer or paralogy.
Documented transfers of rbcL
Summary of endosymbioses
Supplementary Reading
Delwiche, C.F., and J.D. Palmer. 1996. Rampant horizontal gene transfer and duplication of rubisco genes in eubacteria and plastids. Mol. Biol. Evol. 13:873-882.
Delwiche, C.F. 1999. Tracing the thread of plastid diversity through the tapestry of life. American Naturalist 154:S164-S177. Available for University of Maryland community as an electronic journal: http://www.journals.uchicago.edu/AN/journal/contents/v154nS4.html