not retrieve yet the full sequence of caspartin, and more work has still to be done
before we elucidate its complete primary structure.
13.4.3.2 Molecular Data on the Nacre
The second strategy we chose was to construct a cDNA library from mantle tissues
of non-stressed juvenile specimens of Pinna nobilis, which were actively
calcifying. Briefly, the technique implies the extraction of total RNAs from a tissue,
the purification of the messenger RNAs by Oligo-dT resin, the conversion of the
transcripts into double-stranded cDNAs, the cloning of “adaptors” in the two
termini of the cDNAs, the selection of the longest cDNA by size-exclusion chromatography, the cloning of the cDNAs into a phage vector, and the packaging of the
construct into phage capsids. The resulting primary library, which yielded about
3.5Â10
5 clones, was amplified once.
At that point, because we did not have sequence information on shell proteins of
Pinna nobilis – let us recall that the first full sequence of a shell protein from the
pearl oyster Pinctada was published at the end of 1996 – we could not develop
degenerate probes (biotin or digoxigenin-labeled) for screening the constructed
library. As our constructed cDNA library was an expression library, we chose to
screen it with “our” homologous antibodies that we used before, those elicited
against the acid-soluble shell matrices of prisms and of nacre of Pinna nobilis.
This operation described elsewhere (Marin et al. 2000, 2003a) generated series of
clones, which were re-screened twice (secondary and tertiary screening) to purity.
The pure clones exhibited long cDNA inserts (above 2 kb), which encoded a putative
protein of 636 amino acids. The analysis of this ORF showed that the corresponding
acidic protein (theoretical pI ¼ 4.9) possesses three domains: a short N-terminus,
then a long central domain constituted of a succession of 13 almost identical repeats
put in tandem (403 residues in total), which represents two-third of the protein, then
a C-terminus (222 residues). Each tandem repeat, 31 amino acids long, is enriched in
proline and serine residues, and many of these serine residues are presumably
glycosylated. Although similarity search indicated only limited homologies with
known proteins, the most convincing hit was found with PGM, a pig gastric mucin.
Similarities with a mucin may not be fortuitous. Indeed, our protein possesses at
least two properties found in mucins, the main protein components found in mucus
secretions: the presence of tandem repeats, and the richness of these repeats in
proline and serine residues. In addition, like most of the mucins, our protein exhibits,
in the C-terminus, few cysteine residues. In mucins, these residues potentially allow
the formation of dimers first, and then, of multimeric insoluble gels. Because of these
features, we consequently chose to name this protein mucoperlin (Fig. 13.6).
Mucoperlin was overexpressed in a bacterial strain, and after a double purification, the single obtained band was used for generating a very specific polyclonal
antibody (Marin et al. 2000, 2003a) and for different in vitro experiments. Because
the overexpressed mucoperlin did not exhibit posttranslational modifications, it was
not active in “inhibition tests” e.g., neither in calcium-binding assay. The antibody
13 Acidic Shell Proteins of the Mediterranean Fan Mussel Pinna nobilis
377
before we elucidate its complete primary structure.
13.4.3.2 Molecular Data on the Nacre
The second strategy we chose was to construct a cDNA library from mantle tissues
of non-stressed juvenile specimens of Pinna nobilis, which were actively
calcifying. Briefly, the technique implies the extraction of total RNAs from a tissue,
the purification of the messenger RNAs by Oligo-dT resin, the conversion of the
transcripts into double-stranded cDNAs, the cloning of “adaptors” in the two
termini of the cDNAs, the selection of the longest cDNA by size-exclusion chromatography, the cloning of the cDNAs into a phage vector, and the packaging of the
construct into phage capsids. The resulting primary library, which yielded about
3.5Â10
5 clones, was amplified once.
At that point, because we did not have sequence information on shell proteins of
Pinna nobilis – let us recall that the first full sequence of a shell protein from the
pearl oyster Pinctada was published at the end of 1996 – we could not develop
degenerate probes (biotin or digoxigenin-labeled) for screening the constructed
library. As our constructed cDNA library was an expression library, we chose to
screen it with “our” homologous antibodies that we used before, those elicited
against the acid-soluble shell matrices of prisms and of nacre of Pinna nobilis.
This operation described elsewhere (Marin et al. 2000, 2003a) generated series of
clones, which were re-screened twice (secondary and tertiary screening) to purity.
The pure clones exhibited long cDNA inserts (above 2 kb), which encoded a putative
protein of 636 amino acids. The analysis of this ORF showed that the corresponding
acidic protein (theoretical pI ¼ 4.9) possesses three domains: a short N-terminus,
then a long central domain constituted of a succession of 13 almost identical repeats
put in tandem (403 residues in total), which represents two-third of the protein, then
a C-terminus (222 residues). Each tandem repeat, 31 amino acids long, is enriched in
proline and serine residues, and many of these serine residues are presumably
glycosylated. Although similarity search indicated only limited homologies with
known proteins, the most convincing hit was found with PGM, a pig gastric mucin.
Similarities with a mucin may not be fortuitous. Indeed, our protein possesses at
least two properties found in mucins, the main protein components found in mucus
secretions: the presence of tandem repeats, and the richness of these repeats in
proline and serine residues. In addition, like most of the mucins, our protein exhibits,
in the C-terminus, few cysteine residues. In mucins, these residues potentially allow
the formation of dimers first, and then, of multimeric insoluble gels. Because of these
features, we consequently chose to name this protein mucoperlin (Fig. 13.6).
Mucoperlin was overexpressed in a bacterial strain, and after a double purification, the single obtained band was used for generating a very specific polyclonal
antibody (Marin et al. 2000, 2003a) and for different in vitro experiments. Because
the overexpressed mucoperlin did not exhibit posttranslational modifications, it was
not active in “inhibition tests” e.g., neither in calcium-binding assay. The antibody
13 Acidic Shell Proteins of the Mediterranean Fan Mussel Pinna nobilis
377