The Goal of Proteomics Is to Identify the Entire Complement of Proteins Elaborated by a Cell Under Diverse Conditions
While the sequence of the human genome is known, the picture it provides is both static and incomplete. As genes are switched on and off and mRNA molecules customized via alternative splicing, the spectrum of proteins synthesized varies by particular cell type, stages of growth or differentiation, and in response to external stimuli. Muscle cells express proteins not expressed by neural cells, and the type of subunits present in the hemoglobin tetramer undergo change pre- and postpartum. Many proteins undergo posttranslational modifications during maturation into functionally competent forms or as a means of regulating their properties. In order to obtain a more complete and dynamic molecular description of living organisms, scientists are working to determine the proteome, a term that refers to the identity, abundance, and state of modification of the entire suite of proteins expressed by an individual cell at a particular time. Since the proteome for each component cell of an organism is distinct and changes with time and circumstances, the ultimate, comprehensive human proteome constitutes a target of formidable size and complexity.
Simultaneous Determination of Hundreds of Proteins Is Technically Challenging
A key goal of proteomics is the identification of proteins whose levels of expression or modification correlate with medically significant events. In addition to their potential as diagnostic indicators, these protein biomarkers may provide important clues concerning the root causes and mechanisms of a specific physiologic condition or disease. First-generation proteomics employed SDS-PAGE or two-dimensional electrophoresis to resolve the proteins in a biologic sample one from another, followed by determination of the amino acid sequence of their amino terminus by the Edman method. Identities were determined by searching available polypeptide sequences for proteins that contained a matching N-terminal sequence as well as a similar Mr and, for 2D gels, pI.
These early efforts were constrained by the limited number of polypeptide sequences available and the difficulties in isolating polypeptides from the gels in sufficient quantities for Edman analysis. Attempts to increase resolving power and sample yield by increasing the size of the gels were only marginally successful. Eventually, the development of mass spectrometric techniques provided a means for protein sequence determination whose sensitivity was compatible with electro phoretic separation approaches.
Knowledge of the genome sequence of the organism in question greatly facilitated identification by providing a com prehensive set of DNA-encoded polypeptide sequences. It also provided the nucleotide sequence data from which to con struct gene arrays, sometimes called DNA chips, containing hundreds of distinct oligonucleotide probes. These chips could then be used to detect the presence of mRNAs containing complementary nucleotide sequences. While changes in the expression of the mRNA encoding a protein do not necessarily reflect comparable changes in the level of the corresponding protein, gene arrays were both less technically demanding and more sensitive than first-generation proteomic approaches, particularly with respect to low abundance proteins.
Second-generation proteomics coupled newly developed nanoscale chromatographic techniques with MS. The proteins in a biologic sample are first treated with a protease to hydrolyze them into smaller peptides that are then subject to reversed-phase, ion-exchange, or size-exclusion chromatography to apportion the vast number of peptides into smaller sub sets more amenable to analysis. These subsets are analyzed by injecting the column eluent directly into a double quadrupole or TOF mass spectrometer. Multidimensional protein identification technology (MudPIT) employs successive rounds of chromatography to resolve the peptides produced from the digestion of a complex biologic sample into several simpler fractions that can be analyzed separately by MS.
Today, advances in the capability and sensitivity of tandem mass spectrometers allow them to directly analyze complex samples. The elimination of prior proteolytic or chromatographic preparation will soon render it feasible to analyze the proteome of an individual cell.
Bioinformatics Assists Identification of Protein Functions
The functions of a large proportion of the proteins encoded by the human genome are presently unknown. Efforts continue to develop protein arrays or chips for directly testing the potential functions of proteins on a mass scale. However, while some protein functions are relatively easy to assay, such as protease or esterase activity, others are much less tractable. Data mining via bioinformatics permits researchers to compare amino acid sequences of unknown proteins with those whose functions have been determined. This provides a means to uncover clues to their potential properties, physiologic roles, and mechanisms of action. Algorithms exploit the tendency of nature to employ variations of a structural theme to perform similar functions in several proteins (eg, the Rossmann nucleotide binding fold to bind NAD(P)H, nuclear targeting sequences, and EF hands to bind Ca2+). These domains generally are detected in the primary structure by conservation of particular amino acids at key positions. Insights into the properties and physiologic role of a newly discovered protein thus may be inferred by comparing its primary structure with that of known proteins.
