Characterization of the stem-branch leading to the ForaminiferaA more detailed study of the complete SSU rRNA sequences used in Figure 6.1, taking
into account the secondary structure model proposed by Neefs et al. (1993), reveals that
170 particular nucleotides in specific positions define the Foraminifera, when compared
with ‘crown’ Eukaryotes. These particular nucleotides are not randomly scattered
throughout the molecule: 57 per cent of them are the result of substitutions in more
variable, external regions of the SSU rRNA secondary structure (49 per cent of the
alignment), whereas 24 per cent of them are the result of substitutions in internal single-
strand regions (21 per cent of the alignment), and only 19 per cent of them are the result
of substitutions in the most conserved, internal stems (30 per cent of the alignment).
Besides, several of the substitutions that occurred in stem regions, especially at the
periphery of the molecule, apparently led to local modifications of the secondary
structure of the SSU rRNA, which can concern only a few base pairings or extend to a
whole helix. Furthermore, expansions specific to the Foraminifera can be observed in
several terminal and internal loops (unpublished data).
The 170 substitutions that occurred in the stem-lineage leading to Foraminifera affected
the base composition of the molecule. The mean G/C content of the 40 non-
foraminiferan sequences that we used in our analyses is 44.7 per cent, ranging from 43
per cent in some ciliates to 47 per cent in the haptophytes, whereas it is 48.2 per cent for
the foraminiferan sequences, ranging from 47.2 per cent in Globorotalia inflata to 48.9 per
cent in Reticulomyxa filosa. However, a spatial zonation can be observed in the G/C content
of foraminiferan SSU rRNA. In the most conserved internal helices and loops, it ranges
from 48 to 50 per cent, whereas in the external hairpins, it ranges from 45.5 to 47.8 per
cent. Finally, in the universal variable regions V1 to V9, which were excluded from our
phylogenetic analyses, the mean G/C content is very low (24.5 per cent, ranging from 20
per cent in some allogromiids to 39 per cent in some rotaliids), and in the regions of
expansion specific to Foraminifera, it can even get close to zero in some groups.
These observations suggest that in the stem-lineage leading to Foraminifera, a base
composition bias towards G/C led to a reinforcement of the internal secondary structure
of the SSU rRNA. Then, an inverse, secondary base composition bias towards A/T
occurred in the external, more variable regions of the molecule. However, this secondary
bias did not, apparently, affect all lineages in equal proportion, and did not affect the
more conserved regions of the SSU rRNA, where the specific substitutions that
accumulated during the stem-lineage acceleration became fixed in most species. This
would mean that different bias and constraints affect different parts of the molecule.
Implications for phylogenetic inference and estimation of
divergence timesThe episodic evolution of ribosomal genes in the foraminiferan stem-lineage has a strong
impact on the inference of the phylogenetic position of the group. In previous
phylogenetic analyses of SSU and LSU rRNA genes, Foraminifera are placed in the lower
part of the eukaryotic tree, among the earliest mitochondrial protists (Pawlowski et al.
1994, 1996). This position was interpreted as the result of rapid rates in LSU and SSU
rRNA (Sogin 1997). Indeed, the analysis of foraminiferan actincoding gene sequences
116 JAN PAWLOWSKI AND CÉDRIC BERNEY