170
PH. BULET and S. UTTENWEILER-JOSEPH
methodology has the disadvantage of imposing a relative selectivity resulting in
a possible loss of certain DIMs compared to the MALDI-MS mapping achieved
on the crude hemolymph.
In a pilot experiment, the hemolymph of 20 unchallenged and 20 bacteriachallenged flies, collected in acidified water, were subjected separately to microbore
(Imm of internal diameter) reversed-phase HPLC separation with a linear gradient
of acetonitrile (for details, see Uttenweiler-Joseph et al. 1998). All the HPLC fractions were screened by UV spectroscopy and MALDI -TO F MS in order to detect the
DIMs initially characterized by their molecular mass in the fingerprint performed
on the crude hemolymph. This strategy allowed to detect DIMs 1 to 19 in the fractions collected from the HPLC analysis of hemolymph from infected Drosophila
(data not shown). The peptide nature of these DIMs was authenticated by subjecting
the hemolymph sample to pronase, a treatment that totally abolished all UV and
mass signals corresponding to DIMs 1-19 (data not shown). Following this first
step of purification we were unable to detect DIMs 20 to 24.
Different methodologies can be applied to identify purified peptides/proteins
depending on the sequence strategy which can be used: in silico or de novo
sequencing. If the genome of the biological model is already known, in silico
characterization (using mostly nucleotide sequences as source of information)
can be performed even if only partial information on the amino acid sequence
(obtained by ESI-MS/MS or Edman degradation) or tryptic fragment masses
(MALDI peptide mapping) are available (for review, see Roepstorff 1997). If
genome information is not available, de novo sequencing must be performed in
order to get accurate peptide sequence information necessary to unambiguously
identify the protein by complete sequencing orland after cDNA cloning. De novo
sequencing by MS is not yet straightforward but the new technology of
quadrupole-TOF is promising (Shevchenko et al. 1997).
As the Drosophila genome is only expected to be fully sequenced by the turn
of the century, we choose to sequence the DIMs by a more classical technique:
Edman chemistry. However such an approach has different limitations to unambiguously define a primary amino acid sequence: (i) the peptide must have a non
blocked N-terminus and (ii) the quantity of purified peptide must be sufficient
(approximately 1-10 picomoles). After purification to homogeneity from an
hemolymph sample collected from less than 140 immune-challenged flies, we
fully sequenced DIM 4 and partially DIMs 1 and 2 that appeared to be isoforms.
The full primary structure of DIM 2 was obtained by molecular cloning of cDNA
prepared from RNA of immune-challenged adult Drosophila. The information
obtained after cDNA sequencing established that DIM 2 does not result from the
degradation of a larger protein (for more details on the molecular biology analysis see Uttenweiler-Joseph et al. 1998). Unfortunately, database searching using
the amino acid sequences obtained on DIMs 1,2 and 4 did not reveal any significant similarity with already known pep tides that could be used to understand the
precise role of these compounds in Drosophila immunity. Three additional DIMs
(DIMs 10, 13 and 16) were purified to homogeneity but found to be resistant to
Edman degradation.
The structural characterization of peptides detected by MALDI- TOF MS directly
from a biological complex mixture is challenging without any purification step.
PH. BULET and S. UTTENWEILER-JOSEPH
methodology has the disadvantage of imposing a relative selectivity resulting in
a possible loss of certain DIMs compared to the MALDI-MS mapping achieved
on the crude hemolymph.
In a pilot experiment, the hemolymph of 20 unchallenged and 20 bacteriachallenged flies, collected in acidified water, were subjected separately to microbore
(Imm of internal diameter) reversed-phase HPLC separation with a linear gradient
of acetonitrile (for details, see Uttenweiler-Joseph et al. 1998). All the HPLC fractions were screened by UV spectroscopy and MALDI -TO F MS in order to detect the
DIMs initially characterized by their molecular mass in the fingerprint performed
on the crude hemolymph. This strategy allowed to detect DIMs 1 to 19 in the fractions collected from the HPLC analysis of hemolymph from infected Drosophila
(data not shown). The peptide nature of these DIMs was authenticated by subjecting
the hemolymph sample to pronase, a treatment that totally abolished all UV and
mass signals corresponding to DIMs 1-19 (data not shown). Following this first
step of purification we were unable to detect DIMs 20 to 24.
Different methodologies can be applied to identify purified peptides/proteins
depending on the sequence strategy which can be used: in silico or de novo
sequencing. If the genome of the biological model is already known, in silico
characterization (using mostly nucleotide sequences as source of information)
can be performed even if only partial information on the amino acid sequence
(obtained by ESI-MS/MS or Edman degradation) or tryptic fragment masses
(MALDI peptide mapping) are available (for review, see Roepstorff 1997). If
genome information is not available, de novo sequencing must be performed in
order to get accurate peptide sequence information necessary to unambiguously
identify the protein by complete sequencing orland after cDNA cloning. De novo
sequencing by MS is not yet straightforward but the new technology of
quadrupole-TOF is promising (Shevchenko et al. 1997).
As the Drosophila genome is only expected to be fully sequenced by the turn
of the century, we choose to sequence the DIMs by a more classical technique:
Edman chemistry. However such an approach has different limitations to unambiguously define a primary amino acid sequence: (i) the peptide must have a non
blocked N-terminus and (ii) the quantity of purified peptide must be sufficient
(approximately 1-10 picomoles). After purification to homogeneity from an
hemolymph sample collected from less than 140 immune-challenged flies, we
fully sequenced DIM 4 and partially DIMs 1 and 2 that appeared to be isoforms.
The full primary structure of DIM 2 was obtained by molecular cloning of cDNA
prepared from RNA of immune-challenged adult Drosophila. The information
obtained after cDNA sequencing established that DIM 2 does not result from the
degradation of a larger protein (for more details on the molecular biology analysis see Uttenweiler-Joseph et al. 1998). Unfortunately, database searching using
the amino acid sequences obtained on DIMs 1,2 and 4 did not reveal any significant similarity with already known pep tides that could be used to understand the
precise role of these compounds in Drosophila immunity. Three additional DIMs
(DIMs 10, 13 and 16) were purified to homogeneity but found to be resistant to
Edman degradation.
The structural characterization of peptides detected by MALDI- TOF MS directly
from a biological complex mixture is challenging without any purification step.