Biology Reference
In-Depth Information
domains develop from a common origin but gradually acquire their own unique
identity in terms of structure and function. By contrast,
convergent evolution
refers
to the similarity between two protein structures, amino acid sequences or
nucleotide sequences due to their independent evolution from different origins
rather than through their possession of a common ancestor. Evolutionary conver-
gence therefore implies 'adaptive change in which lesser related entities come to
appear more related than they are' (Doolittle, 1994). The term convergence has
however been used in many different contexts with very different meanings.
Doolittle (1994) distinguished between
functional
convergence,
mechanistic
con-
vergence and
structural
convergence.
Functional convergence is exemplified by various pairs of enzymes that have
evolved independently to catalyse the same biochemical reactions, for example
superoxide dismutases, aldolases, alcohol dehydrogenases and topoisomerases
(Doolittle, 1994). Myoglobin from the abalone
Sulculus diversicolor
exhibits func-
tional convergence with vertebrate myoglobin although it is not in any way
homologous to it (Suzuki
et al
., 1996). Instead, the
Sulculus
myoglobin gene is evo-
lutionarily related to the human indoleamine 2,3-dioxygenase (
IDO
; 8p11-p12)
gene (Suzuki
et al
., 1996). The original
Sulculus
myoglobin gene may have been
lost and a modified
Ido
gene could have evolved as a substitute.
Perhaps the best example of mechanistic convergence is provided by the serine
proteases subtilisin and chymotrypsin which, although unrelated evolutionarily,
have independently evolved similar enzymatic mechanisms. Chymotrypsin has a
catalytic triad with a histidine at residue 57, an aspartate at residue 102 and a ser-
ine at residue 195. The bacterial protein subtilisin possesses a completely differ-
ent structure but also has a catalytic triad comprising an aspartate at residue 32, a
histidine at residue 64 and a serine at residue 221. The human genome contains
representatives of both families. Thus, the chymotrypsinogen B (
CTRB1
; 16q23),
chymotrypsin-like protease (
CTRL
; 16q22), trypsin 1 (
PRSS1
; 7q35), trypsin 2
(
PRSS2
; 7q35), elastase 1 (
ELA1
; 12q13), plasminogen (
PLG
; 6q26) and pro-
thrombin (
F2
; 11p11-q12) genes are members of the chymotrypsin family of ser-
ine proteases. Human genes encoding subtilisin-like serine proteases include the
proprotein convertases (
PCSK1
, 5q15-q21;
PCSK2
, 20p11;
PCSK4
, chromosome
19;
PCSK5
, 9q21.3), furin (
PACE
; 15q25-q26) and paired basic amino acid cleav-
ing enzyme 4 (
PACE4
; 15q26). Another example of mechanistic convergence is
provided by the cold shock domain protein family and the family of RNA-bind-
ing proteins that contain an RNA-binding domain. Both of these domains con-
tain conserved ribonucleoprotein motifs on similar single stranded nucleic
acid-binding surfaces (Graumann and Marahiel, 1996).
Structural convergence, on the other hand, may reflect the tendency of specific
amino acid sequences to fold into certain favored conformations. Thus, struc-
turally dissimilar families of transport proteins have been found to exhibit simi-
lar structural units consisting of six tightly packed
-helices which may comprise
all or part of a transmembrane channel (Saier, 1994).
In none of the above examples is there any evidence for
sequence convergence
. For
two sequences to be shown to display convergence in this strict sense, they would
not only have to be shown to be evolutionarily unrelated but chance would also
have to be excluded as a reason for their similarity. Indeed, sequence convergence
