54
J. R. YATES et al.
second mass analyzer separates the fragment ions and passes them to a detector.
Other types of tandem mass spectrometers have been developed such as ion
traps and quadrupole-time-of-flight instruments that are also capable of peptide
sequencing. Peptide ions fragment primarily at amide bonds creating a pattern
that can be interpreted to reveal the amino acid sequence. A drawback to peptide
sequencing using tandem mass spectrometry, however, is that interpretation of
the data can be time-intensive. Ionization techniques such as electrospray ionization (ESI) provided better and more robust integration of peptide sequencing
methods with liquid chromatography improving the sensitivity of analysis
(Covey et al. 1991; Griffin et al. 1991). Sample manipulation is streamlined as
protein digests can be directly loaded onto a reversed-phase column, separated,
and introduced directly into the tandem mass spectrometer for sequencing.
Introduction of chromatographically separated peptides into tandem mass spectrometers is greatly augmented by an under-appreciated technology employing
data-dependent instrument control algorithms to control operation of the tandem mass spectrometer (Yates et al. 1995). By using data dependent acquisition,
tandem mass spectra can be acquired with higher efficiency than through manual control of conditions. As a result many more tandem mass spectra are
acquired, taxing data interpretation.
In 1994 Eng et al discovered the use of protein sequence databases to interpret
tandem mass spectra of pep tides and identify the protein of origin (Eng et al.
1994). In this process tandem mass spectra are used to search databases using
both the molecular weight of the peptide and the fragmentation pattern from the
tandem mass spectrum. Each tandem mass spectrum is independently searched
through the database, and matched to the amino acid sequence with the best fit.
When tandem mass spectra of pep tides from the same protein are present and
matched to the protein sequence, considerable confidence is added to the identification. Furthermore, Eng et al showed this process capable of identifying proteins present in mixtures (Eng et al. 1994). This protein identification strategy
using tandem mass spectra presented a new approach for the analysis of protein
mixtures. Subsequently, McCormack et al showed direct identification of proteins
in mixtures obtained from several different types of molecular biology experiments (McCormack et al. 1997). Link et al extended this approach to the identification of proteins enriched from subcellular compartments of cells (Link et al.
1997). A large collection of proteins was identified from the periplasmic space of
E. coli. The strengths of this approach are rapid identification of proteins in mixtures, increased sensitivity, and the potential for comprehensive identification of
the proteins present. We have extended this approach to the analysis of protein
complexes and developed new technologies to advance the process.
2
Protein Identification in Mixtures
Identification of proteins in mixtures is highly dependent on computer software
to process and analyze the data. Our efforts in this area have encompassed the
development of computer algorithms to perform subtractive analysis, search
databases with increased speeds, and simplification of the data review process.
J. R. YATES et al.
second mass analyzer separates the fragment ions and passes them to a detector.
Other types of tandem mass spectrometers have been developed such as ion
traps and quadrupole-time-of-flight instruments that are also capable of peptide
sequencing. Peptide ions fragment primarily at amide bonds creating a pattern
that can be interpreted to reveal the amino acid sequence. A drawback to peptide
sequencing using tandem mass spectrometry, however, is that interpretation of
the data can be time-intensive. Ionization techniques such as electrospray ionization (ESI) provided better and more robust integration of peptide sequencing
methods with liquid chromatography improving the sensitivity of analysis
(Covey et al. 1991; Griffin et al. 1991). Sample manipulation is streamlined as
protein digests can be directly loaded onto a reversed-phase column, separated,
and introduced directly into the tandem mass spectrometer for sequencing.
Introduction of chromatographically separated peptides into tandem mass spectrometers is greatly augmented by an under-appreciated technology employing
data-dependent instrument control algorithms to control operation of the tandem mass spectrometer (Yates et al. 1995). By using data dependent acquisition,
tandem mass spectra can be acquired with higher efficiency than through manual control of conditions. As a result many more tandem mass spectra are
acquired, taxing data interpretation.
In 1994 Eng et al discovered the use of protein sequence databases to interpret
tandem mass spectra of pep tides and identify the protein of origin (Eng et al.
1994). In this process tandem mass spectra are used to search databases using
both the molecular weight of the peptide and the fragmentation pattern from the
tandem mass spectrum. Each tandem mass spectrum is independently searched
through the database, and matched to the amino acid sequence with the best fit.
When tandem mass spectra of pep tides from the same protein are present and
matched to the protein sequence, considerable confidence is added to the identification. Furthermore, Eng et al showed this process capable of identifying proteins present in mixtures (Eng et al. 1994). This protein identification strategy
using tandem mass spectra presented a new approach for the analysis of protein
mixtures. Subsequently, McCormack et al showed direct identification of proteins
in mixtures obtained from several different types of molecular biology experiments (McCormack et al. 1997). Link et al extended this approach to the identification of proteins enriched from subcellular compartments of cells (Link et al.
1997). A large collection of proteins was identified from the periplasmic space of
E. coli. The strengths of this approach are rapid identification of proteins in mixtures, increased sensitivity, and the potential for comprehensive identification of
the proteins present. We have extended this approach to the analysis of protein
complexes and developed new technologies to advance the process.
2
Protein Identification in Mixtures
Identification of proteins in mixtures is highly dependent on computer software
to process and analyze the data. Our efforts in this area have encompassed the
development of computer algorithms to perform subtractive analysis, search
databases with increased speeds, and simplification of the data review process.