356
w.T. Starn et al.
Tree topologies showed four sequence clusters of paralogous teleost
hemoglobin sequences, which may originate from three gene
duplications in ancestral teleost species. The two most recently derived
clusters consisted of the sequences of the "major" and "minor"
hemoglobin chains of the Notothenioids, respectively, together with
those of tuna (Thunnus thynnus) and red gumard (Chelidonichthys
kumu), which is a scorpaeniform. The Notothenioids belong to the
teleost order Perciformes [3], together with tuna and snapper
(Chrysophrys auratus), but not with red gurnard. The position of
sequenced hemoglobin chains of snapper in another cluster in the trees
indicated a paralogous relationship with the hemoglobins of the other
Perciformes. The a- and p-chains of the major components of the
Notothenioids formed a monophyletic group in all investigated trees,
with the nonAntarctic Pseudaphritis as their sister taxon. The minor
chains also were a monophyletic group. The families Nototheniidae and
Bathydraconidae appeared to be paraphyletic.
Here we present new analyses of teleost hemoglobins in which we
have included deduced amino acid sequences of two sets of adjacent aand f3-globin genes from Notothenia coriiceps (unpublished) and of the
almost completed sequences of one of the hemoglobins of Cottoperca
gobio (10 and 6 residues still are unknown in the a- and p-chains,
respectively; unpublished).
Methods
Sequences were analyzed using maximum parsimony under the
heuristic search option (tree bisection-reconnection, local and global
swapping, simple and random stepwise addition) using the PAUP
computer program, version 3.1.1 [4]. Gaps (deletions) are considered as
"missing data". Searches were done under the "protpars" assumption, in
which the number of codon differences between replaced amino acids is
taken into account [5]. Bootstrap analysis (100 replicates) using a 50%
majority rule was used to provide an estimate of tree stability. In all
analyses the hemoglobins of human and coelecanth were used as
outgroups.
Results and Discussion
Two most parsimonious trees (MPT) requiring 685 substitutions were
found for the a-chains and three, each requiring 826 substitutions, for
the j3-chains. As the trees were again highly congruent, we decided to
analyze trees of tandemly arranged a- and l3-chains as in our previous
w.T. Starn et al.
Tree topologies showed four sequence clusters of paralogous teleost
hemoglobin sequences, which may originate from three gene
duplications in ancestral teleost species. The two most recently derived
clusters consisted of the sequences of the "major" and "minor"
hemoglobin chains of the Notothenioids, respectively, together with
those of tuna (Thunnus thynnus) and red gumard (Chelidonichthys
kumu), which is a scorpaeniform. The Notothenioids belong to the
teleost order Perciformes [3], together with tuna and snapper
(Chrysophrys auratus), but not with red gurnard. The position of
sequenced hemoglobin chains of snapper in another cluster in the trees
indicated a paralogous relationship with the hemoglobins of the other
Perciformes. The a- and p-chains of the major components of the
Notothenioids formed a monophyletic group in all investigated trees,
with the nonAntarctic Pseudaphritis as their sister taxon. The minor
chains also were a monophyletic group. The families Nototheniidae and
Bathydraconidae appeared to be paraphyletic.
Here we present new analyses of teleost hemoglobins in which we
have included deduced amino acid sequences of two sets of adjacent aand f3-globin genes from Notothenia coriiceps (unpublished) and of the
almost completed sequences of one of the hemoglobins of Cottoperca
gobio (10 and 6 residues still are unknown in the a- and p-chains,
respectively; unpublished).
Methods
Sequences were analyzed using maximum parsimony under the
heuristic search option (tree bisection-reconnection, local and global
swapping, simple and random stepwise addition) using the PAUP
computer program, version 3.1.1 [4]. Gaps (deletions) are considered as
"missing data". Searches were done under the "protpars" assumption, in
which the number of codon differences between replaced amino acids is
taken into account [5]. Bootstrap analysis (100 replicates) using a 50%
majority rule was used to provide an estimate of tree stability. In all
analyses the hemoglobins of human and coelecanth were used as
outgroups.
Results and Discussion
Two most parsimonious trees (MPT) requiring 685 substitutions were
found for the a-chains and three, each requiring 826 substitutions, for
the j3-chains. As the trees were again highly congruent, we decided to
analyze trees of tandemly arranged a- and l3-chains as in our previous