Intron Dynamics in Ribosomal Protein Genes
Citation: Yoshihama M, Nguyen HD, Kenmochi N (
Intron Dynamics in Ribosomal Protein Genes
Maki Yoshihama 0 1
Hung D. Nguyen 0 1
Naoya Kenmochi kenmochi@med 0 1
0 Academic Editor: Oliver Hofmann, South African National Bioinformatics Institute , South Africa
1 Frontier Science Research Center, University of Miyazaki , Kiyotake, Miyazaki , Japan
The role of spliceosomal introns in eukaryotic genomes remains obscure. A large scale analysis of intron presence/absence patterns in many gene families and species is a necessary step to clarify the role of these introns. In this analysis, we used a maximum likelihood method to reconstruct the evolution of 2,961 introns in a dataset of 76 ribosomal protein genes from 22 eukaryotes and validated the results by a maximum parsimony method. Our results show that the trends of intron gain and loss differed across species in a given kingdom but appeared to be consistent within subphyla. Most subphyla in the dataset diverged around 1 billion years ago, when the ''Big Bang'' radiation occurred. We speculate that spliceosomal introns may play a role in the explosion of many eukaryotes at the Big Bang radiation.
-
INTRODUCTION
Many spliceosomal introns, which are non-coding DNA
sequences, exist in eukaryotic nuclear genes. Their role in the
genome, however, remains poorly understood. From the view of
eukaryotic evolution, it is very important to know why exon/
intron structures of genes differ across species and what the effects
of intron gain and loss are. In order to clarify these issues, we must
first reconstruct the process of intron gain and loss during
eukaryotic evolution. This task became possible recently with the
availability of many completely sequenced genomes. In a
representative study, Rogozin et al. [1] compiled a dataset of 684 gene
orthologs from eight eukaryotes and used a maximum parsimony
method to infer the evolution of introns in this dataset. The results
of applying maximum likelihood methods to the same dataset were
reported later [24]. Although the number of species in the dataset
is not very large and the different methods inferred different
patterns of intron gain and loss, it became clear that: (i) from 15%
to 25% of present-day introns were already present in the last
common ancestor of plantae, metazoa, and fungi, and (ii) many
introns were gained after this divergence [14].
We have recently compiled a dataset of ribosomal protein (RP)
genes [5]. RP genes offer several advantages for studying intron
evolution [68]. First, they exist in all species and, as they are
involved in the vital process of translation, they are well conserved
throughout evolution [9,10]. Thus, it is fairly easy to compare
intron positions in RP genes across a wide range of distantly
diverged species. Second, there are a large number of conserved
RP gene families. For instance, 79 distinct RPs are found in
humans and of these 79, 78 are also found in yeast. Third, introns
also exist in RP genes of very deep-branching eukaryotes that
harbor very few introns, such as Giardia lamblia [11,12]. With these
advantages, we expect that RP genes will become a powerful tool
for discovering the roles of spliceosomal introns.
RESULTS
Compilation of the dataset and phylogenetic
analysis
We compiled a dataset of 76 RP gene orthologs from 22
eukaryotes. The phylogenetic tree of these 22 species is depicted
in Figure 1. These 22 species belong to four kingdoms, metazoa,
fungi, protozoa, and plantae, and cover 14 different subphyla. The
conserved regions of this dataset included 2,961 introns located at
1,182 different positions. To the best of our knowledge, this is the
first time a dataset with this many gene families and species has
been used for studying intron evolution.
Patterns of intron gain and loss in 22 species
We first used our recently developed maximum likelihood (ML)
method [4] to infer the process of intron gain and loss (Figure 2A).
We also used a maximum parsimony method to validate the result
of the ML method, because the ML method may produce
unreliable results when the data sample is small (Figure 2B). Since
the two results show similar patterns of intron gain and loss in most
subphyla of the dataset (the largest differences are in the two plant
subphyla), the results from the ML method were used for
subsequent analyses.
The most significant feature in Figure 2 is that species belonging
to a given subphylum show similar trends of intron gain and loss.
For example, all three species in insecta (subphylum 3 in Figure 2)
show trends toward decreasing introns, whereas all three species in
pezizomycotina (subphylum 5) show trends toward increasing
introns. There is, however, no consensus trend of intron gain and
loss among species of a given kingdom. This fact is most notable in
the fungus kingdom. The subphyla pezizomycotina (subphylum 5)
and hymenomycetes (subphylum 8) trended toward increasing
introns, whereas the three other subphyla [saccharomycotina
(subphylum (...truncated)