A method to correlate uninterpreted tandem mass spectra of modified peptides, produced under low-energy (10-50 eV) collision conditions, with amino acid sequences in a protein database has been developed. The fragmen-tation patterns observed in the tandem mass spectra of peptides containing covalent modifications is used to directly search and fit linear amino acid sequences in the database. Specific information relevant to sites of modification is not contained in the character-based sequence information of the databases. The search method consid-ers each putative modification site as both modified and unmodified in one pass through the database andsimul-taneously considers up to three different sites of modifica-tion. The search method will identify the correct sequence if the tandem mass spectrum did not represent a modified peptide. This approach is demonstrated with peptides containing modifications such as S-carboxymethylated cysteine, oxidized methionine, phosphoserine, phosphothreonine, or phosphotyrosine. In addition, a scanning approach is used in which neutral loss scans are used to initiate the acquisition of product ion MS/MS spectra of doubly charged phosphorylated peptides during a single chromatographic run for data analysis with the database-searching algorithm. The approach described in this paper provides a convenient method to match the nascent tandem mass spectra of modifiedpeptides to sequences in a protein database and thereby identify previously unknown sites of modification.

The synthesis of proteins in biological organisms proceeds by transcription of deoxyribonucleotide sequences to messenger RNA (mRNA). In eukaryotic organisms, mRNA is processed to remove introns and is then translated on the ribosomal complex to synthesize the protein. Almost all protein sequences are post-translationally modified in processes that may range from simple proteolytic cleavage to covalent modification of specific amino acid residues. 1 As many as 200 typesof covalent modifications may exist, and their biological functions are poorly understood. 1 Organisms can increase chemical diversity, control enzymatic activity, or transmit signals through posttranslational modification of protein structure. The initiatives to completely sequence the genomes of the human and numerous model organisms will provide complete sequence information for all geneproducts.* 12· 3