H3 replacement variants arose at least twice in evolution
Although hv2 is clearly an H3 replacement variant by its
expression, its sequence is distinct from those of other H3
replacement variants. To analyze the evolution of the H3 variants,
a phylogenetic tree was constructed using the neighbor-joining
(NJ) method of Saitou and Nei (36) on 73 H3 sequences compiled
from the GenBank and PIR databases (see Materials and
Methods). The total evolutionary divergence between two
sequences is represented by the sum of the horizontal branch
lengths between the two sequences. Vertical distances are for
illustration purposes only.
The NJ tree for H3 (Fig. 4A) indicates that the three types
of H3 replacement variants represented (animal H3.3, plant
H3.ITJ and Tetrahymena hv2 and H3(2)) are not monophyletic
that is, they do not have a single common ancestor. The H3 to
which the Tetrahymena variants hv2 and H3(2) are most closely
related is the Tetrahymena major H3. Animal H3.3s branch off
the tree prior to the divergence of plants and animals, while plant
H3.HI is found within the monophyletic plant branch. Although
this suggests that plant H3.rU may have evolved from an earlier
H3.3-type gene, other NJ trees which do not contain the
A.nidulans sequence have a different topology, in which animal
H3.3s and fungal H3s are monophyletic and branch after the
divergence of plant H3s (data not shown). To investigate this
discrepancy, a maximum parsimony analysis was carried out on
a subset of H3 proteins from major phylogenetic lineages with
the program PAUP (37) (Fig. 4B). The use of a subset containing
fewer taxa permits a more rigorous analysis using parsimony
methods. The parsimony tree places the divergence of animals
H3.3s and fungal H3s two branches deeper than plant H3.m with
a high degree of confidence. A NJ tree of the same subset was
identical to one of the minimum length parsimony trees (data not
shown). This confirms our NJ analyses that suggested that animal
H3.3s evolved after the plant-animal divergence and were closely
related to animal and fungal H3s. We carried out additional
parsimony and NJ analyses on several different subsets of H3
proteins and found that most NJ and all parsimony phylogenies
placed animal H3.3s and fungal H3s on the same branch, after
the divergence of plants (data not shown). Thus, the parsimony
method appears to be more robust with respect to this data set,
and suggests that animal, plant and Tetrahymena H3 replacement
variants arose independendy. Both NJ and parsimony analyses
agree unequivocally that Tetrahymena replacement variant arose
independently of other H3 variants.