316
Chi-Hing C. Cheng
somehow evolutionarily related. It was not obvious however how an
AFGP gene could evolve from a trypsinogen gene as the amino acid
sequences of the two proteins share no similarity whatsoever.
To decipher this relationship, we isolated and sequenced trypsinogen
gene and cDNA from the same notothenioid, D. mawsoni, and compared
them to the Dm AFGP gene and cDNA sequences. The results and
comparative analyses that lead to our discovery of the molecular
mechanism involved in the transformation of a trypsinogen gene to an
AFGP gene have been published in detail recently [29]. The essential
aspects of the evolutionary process are shown in Fig. 2, and described
below.
The front and tail segments of Dm trypsinogen and AFGP genes where
high degree of sequence identity (93-96%) resides clearly were recruited
and incorporated to form the new gene. The front segment of trypsinogen
gene, namely El and Il inclusive of a run of gt dinucleotides, was
recruited and used as is in the AFGP gene. Exon 1 which contains the
coding sequence for the signal peptide would provide the secretory signal
for the emerging new antifreeze protein. A large insertion (~1. 7 kbp) was
subsequently incorporated into II. The tail end of trypsinogen gene,
namely the last exon (exon 6) which encodes the C-terminus 50 residues of
trypsin and the 3'UTR, was recruited and converted into the 3' sequence of
AFGP gene starting from the penultimate codon in the AFGP coding
region through the downstream 3' flanking sequence (same as 3'UTR).
This region corresponds to the segment of plaice trypsinogen cDNA that
shares >70% identity with AFGP 3' flanking sequence in the initial
database search.
The putative ancestral sequence that would give rise to the AFGP coding
sequence comprising E2 in the AFGP gene required more intensive
searches. At the very end of the front segment of sequence similarity, a 9nucleotide ThrAlaAla coding element, acagcggca (splice sequence in
italics) was found straddling the junction ofIl and E2 of trypsinogen gene
(Fig. 2). This 9-nt sequence corresponds very closely to the coding
sequences of the ThrAlaAla repeats in the AFGP polyproteins (Fig. 2),
indicating strongly that it is the extant version of the ancestral sequence
that gave rise to the AFGP tripeptide repeats through iterative duplications.
It is very likely that the ancestral 9-nt element was acagc1:gca (g for 1:
substitution in extant version), reflected in gct being the majority (61 %,
versus 36% of gcg) of the codons for the first Ala in the AFGP tripeptide
repeats (Fig. 2). A single nucleotide substitution of this codon will provide
the same amino acid Ala (gct to gcg), or the periodic Pro (gct to cct) seen
in this position in the tripeptide repeats. The almost exclusive bias for cct
as the Pro codon (96%) (Fig. 2) in the AFGP polyprotein genes lends
strong support to this reasoning.
Chi-Hing C. Cheng
somehow evolutionarily related. It was not obvious however how an
AFGP gene could evolve from a trypsinogen gene as the amino acid
sequences of the two proteins share no similarity whatsoever.
To decipher this relationship, we isolated and sequenced trypsinogen
gene and cDNA from the same notothenioid, D. mawsoni, and compared
them to the Dm AFGP gene and cDNA sequences. The results and
comparative analyses that lead to our discovery of the molecular
mechanism involved in the transformation of a trypsinogen gene to an
AFGP gene have been published in detail recently [29]. The essential
aspects of the evolutionary process are shown in Fig. 2, and described
below.
The front and tail segments of Dm trypsinogen and AFGP genes where
high degree of sequence identity (93-96%) resides clearly were recruited
and incorporated to form the new gene. The front segment of trypsinogen
gene, namely El and Il inclusive of a run of gt dinucleotides, was
recruited and used as is in the AFGP gene. Exon 1 which contains the
coding sequence for the signal peptide would provide the secretory signal
for the emerging new antifreeze protein. A large insertion (~1. 7 kbp) was
subsequently incorporated into II. The tail end of trypsinogen gene,
namely the last exon (exon 6) which encodes the C-terminus 50 residues of
trypsin and the 3'UTR, was recruited and converted into the 3' sequence of
AFGP gene starting from the penultimate codon in the AFGP coding
region through the downstream 3' flanking sequence (same as 3'UTR).
This region corresponds to the segment of plaice trypsinogen cDNA that
shares >70% identity with AFGP 3' flanking sequence in the initial
database search.
The putative ancestral sequence that would give rise to the AFGP coding
sequence comprising E2 in the AFGP gene required more intensive
searches. At the very end of the front segment of sequence similarity, a 9nucleotide ThrAlaAla coding element, acagcggca (splice sequence in
italics) was found straddling the junction ofIl and E2 of trypsinogen gene
(Fig. 2). This 9-nt sequence corresponds very closely to the coding
sequences of the ThrAlaAla repeats in the AFGP polyproteins (Fig. 2),
indicating strongly that it is the extant version of the ancestral sequence
that gave rise to the AFGP tripeptide repeats through iterative duplications.
It is very likely that the ancestral 9-nt element was acagc1:gca (g for 1:
substitution in extant version), reflected in gct being the majority (61 %,
versus 36% of gcg) of the codons for the first Ala in the AFGP tripeptide
repeats (Fig. 2). A single nucleotide substitution of this codon will provide
the same amino acid Ala (gct to gcg), or the periodic Pro (gct to cct) seen
in this position in the tripeptide repeats. The almost exclusive bias for cct
as the Pro codon (96%) (Fig. 2) in the AFGP polyprotein genes lends
strong support to this reasoning.