Vol. 60, Issue 3, 405-415, September 2001

MINIREVIEW

Mass Spectrometry of the Proteome

Ray Bakhtiar and Randall W. Nelson

Novartis Institute for Biomedical Research, East Hanover, New Jersey (R.B.); and Intrinsic Bioprobes, Inc., Tempe, Arizona (R.W.N.)

	Article

Top Article References

Biological polymers undergo numerous significant and fascinating interactions, such as post-translational modifications, noncovalent associations, and conformational changes. A valuable parameter for the characterization of a biopolymer is molecular mass (MM). The field of mass spectrometry (MS) has evolved rapidly in the last decade through the introduction of two ionization techniques, matrix-assisted laser desorption ionization (MALDI) and electrospray ionization (ESI). Modern methods of mass spectrometry, including ESI and MALDI, are ideally suited for accurate determination of the molecular mass of the biopolymers. Molecular weight measurements are now used routinely in the qualitative and quantitative analysis of macromolecules. In many cases, the small sample quantities (e.g., a few micrograms) limit the utility of nuclear magnetic resonance spectroscopy and X-ray crystallography for obtaining structural information. Thus, mass spectrometry, which may require only femtomolar or attomolar quantities offers an attractive alternative to the more traditional bioanalytical methods for rapid and sensitive measurements (Burlingame et al., 2000; Tomer, 2001). The ultimate goal of these experiments is to obtain sufficient information to map the complex molecular circuitry that operates within the cell (Aebersold and Goodlett, 2001; Godovac-Zimmermann and Brown, 2001). Herein, we present some of the aspects of the modern biological mass spectrometry for investigation of large molecules. For additional detailed technical descriptions, we have referred the reader to a number of recently published reports throughout the manuscript (see below).

MALDI and ESI Mass Spectrometry. The three main events during an MS analysis are ion production, ion transmission, and ion detection. In a typical MS experiment, the sample of interest is volatilized/ionized in the ionization source and guided via a series of electric and/or magnetic lenses to the detector. To control the motion of the ions during their transmission to the detector, it is necessary to control the influence of pressure and temperature on the ion mobility (kinetic energy). Therefore, a vacuum system with a pressure ranging from ~ 10⁵ to 10⁸ Torr is used in all mass spectrometers. The vacuum environment minimizes the interfering collision of analyte ions with the background neutral gaseous molecules and facilitates their transmission to the detector.

A majority of commercial mass spectrometers use an electron multiplier detector, which provides an internally amplified electrical current subsequent to exposure to charged ions. The ion current output corresponding to each specific analyte is then processed by instrument electronics and translated to MM. In the resulting mass spectrum, the ordinate indicates the relative intensity or abundance, whereas the abscissa shows the observed ratio of mass to the number of charges on the ions. The latter is referred to as mass-to-charge ratio or m/z. The basic tenet is that regardless of the ionization source, the mass spectrometer measures the m/z.

MALDI (Fig. 1a) uses pulses of laser light to desorb the analyte from a solid phase surface (analyte cocrystallized with a light-absorbing matrix) and yield gaseous ions. Pulsed laser radiation, typically from a nitrogen laser (output wavelength, 337 nm) is used to initiate the desorption/ionization event and to simultaneously generate a packet of ions of different m/z values. The matrix is typically a small organic molecule (e.g., nicotinic acid, 2,5-dihydroxybenzoic acid, -cyano-4-hydroxycinnamic acid) that has an absorption band that closely coincides with the energy of the laser radiation. The matrix is generally cocrystallized in large molar excess to the analyte; this facilitates the ionization of the analyte and minimizes the sample degradation from the laser radiation. For MALDI, the sample preparation procedure can be extremely crucial because the ion population depends upon the type of matrix and the presence of impurities. Depending on the specialist's experience and instrumentation, it is now possible to acquire MM information of a biopolymer using femtomolar to attomolar quantities of samples (Nelson and Krone, 1999; Nelson et al., 2000; Farmer and Caprioli, 1998). However, in some cases, additional sample quantities are required for detailed analyses such as peptide, polysaccharide, or oligonucleotide sequencing (Siuzdak, 1996; Dell and Morris, 2001).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   a, in a MALDI experiment, the sample is mixed or dissolved with an excess amount (e.g., 1 part sample to 10,000 parts matrix) of a matrix component (having an absorption wavelength, which matches closely with the laser wavelength). Upon laser irradiation, a plume of neutral molecules and ions are desorbed. The ions are then guided to the mass analyzer and the detector by electrostatic lenses. In contrast to ESI, MALDI generally does not yield multiple-charged ions, does not require mass spectral de-convolution, and is more suitable for analysis of complex mixtures. b, a simplified schematic diagram of an ESI source operating in the positive ion mode. A spray of fine droplets that contains the charged analyte and solvent molecules is generated upon the application of a high electrical tension on a needle. The voltage polarity of the metal capillary is positive or negative for positive or negative ion generation, respectively.

A complementary technique to MALDI is ESI (Fig. 1b), which produces single or multiply charged gaseous ions directly from a solution by generating a fine spray of highly charged droplets in the presence of a strong electric field. There are two widely proposed theories for ion formation in ESI. One theory suggests that ionized sample molecules are expelled from the droplets. Alternatively, it has been proposed that individual ionized sample molecules remain after continuous solvent evaporation and droplet fragmentation (Fernandez de la Mora et al., 2000).

As shown in Fig. 1b, a solution of analyte(s) and the solvent are introduced into a sampling metal capillary (~100 µm in internal diameter), which is charged with the application of an electrical voltage (4-5 kV). The voltage polarity of the metal capillary is positive or negative for positive or negative ion generation, respectively. At some point, the mutual repulsion between the ions at the surface becomes greater than the surface tension of the liquid. If the electrical field is sufficiently high, the spray commences and small charged droplets form. The ions generated by ESI carry multiple charges, provided the sample molecules have a MM of more than about 1000 Da. The characteristic feature of ESI that distinguishes it from other ionization techniques is that it generally imparts multiple charges to larger analyte molecules and the extent of multiple charging increases in near proportion with MM. The resulting highly charged molecular ions are thus within the m/z range in which most conventional mass spectrometers function quite well. It is the multiple-charging phenomenon that allows the assay of high-mass ions by mass analyzers with only a modest m/z range (Bakhtiar and Nelson, 2000; Bakhtiar and Tse, 2000; Thomas et al., 2000; Hofstadler and Griffey, 2001).

ESI does have limitations in that it is not very tolerant of the presence of salts, detergents, and inorganic buffers (MALDI has proven to be more amenable in such cases). Thus, to minimize signal suppression effects, ESI is often exploited as an interface between liquid chromatography (LC) or capillary electrophoresis (CE) and a mass spectrometer. Currently, LC-MS is an attractive tool in the analysis of complex mixtures in biochemical research and medical/diagnostic analysis. The up-front chromatographic separation aids in sample purification/enrichment from most common laboratory buffers and endogenous salts, and provides an additional useful parameter such as retention time.

Proteomics. The human body is estimated to contain ~35,000 protein-encoding genes (Lander et al., 2001). Furthermore, post-translational modification, mutation, degradation, and alternate RNA splicing routes dramatically increase the number of potential cellular proteins. The science of proteomics involves the detection, identification, and characterization of proteins produced by the genome, thus providing a functional annotation for the entire proteome (Anderson et al., 2001; Gabor Miklos and Maleszka, 2001). The proteome is a highly dynamic system that can be influenced by environmental variations (stress), such as quantitative changes in protein expression caused by exposure to xenobiotics. The information obtained from proteome analysis can aid in identifying therapeutic targets or surrogate markers in understanding the initiation and progression of disease states. Thus, proteomics research can be a valuable tool in drug discovery and offers scientists an integration of genomics, mRNA analysis, protein expression, and protein-protein interactions.

The extremely high degree of complexity of multicellular organisms warrants the need for a collection of sensitive and rapid analytical techniques to yield qualitative and quantitative information with high efficiency and accuracy. One of the most commonly used techniques for protein separation has been based on gel electrophoresis. Simple protein mixtures (<100 components) are normally separated using one-dimensional SDS-polyacrylamide gel electrophoresis (PAGE). On the other hand, for complex protein mixtures (e.g., cell or tissue extracts), the resolving power of a two-dimensional (2-D) SDS-PAGE may be required. In this approach, proteins are separated by an isoelectric point in the first dimension (preferably using immobilized pH gradient gels) and subsequently by their molecular masses in the second dimension. Visualization of the gel is achieved by staining, fluorescence tagging, or radioactive labeling, with some protocols having a detection limit of ~1 to 100 ng of protein (Gygi et al., 2000a; Rabilloud, 2000). However, visualization does not provide unambiguous protein identification; therefore, scientists typically use Western blotting or classical Edman sequencing for this purpose. Fortunately, the integration of the inherent benefits of MS (e.g., sensitivity, selectivity, and speed) with those conferred by protein and expressed sequence tag (EST) databases has led to a significant advancement in proteome characterization (Pandey and Mann, 2000; Yates, 2000; Mann and Pandey, 2001).

Figure 2 depicts a simplified strategy that is being used widely in high-throughput protein characterization. Typically, samples obtained from different cellular fractions are processed by 2-D SDS isoelectric-focusing gel electrophoresis. Each cell or tissue type may require a specific visualization approach such as Coomassie blue, silver staining, or fluorescence tagging for protein detection (Hancock et al., 1999; Williams, 1999; Lahm and Langen, 2000). Subsequently, gel images are electronically retrieved by high-resolution scanners and analyzed (spot-finding) using pattern recognition techniques against 2-D gel database queries (Miura, 2001). Sophisticated software packages can be employed to enhance contrast, subtract background, align images, remove artifacts, and perform gel comparison. Proteome maps are then compared against databases for identification of up- or down-regulation in a disease state. The resulting information can have the potential to identify biomarkers in clinical or toxicological studies (Alaiya et al., 2001; Bichsel et al., 2001).

View larger version (33K): [in this window] [in a new window]   Fig. 2.   A simplified proteomics scheme outlining steps involved in characterization of polypeptides (see text for details).

Conceptually, an approach that is similar and complementary but has higher accuracy (i.e., better than 10 ppm), sensitivity, and speed can be implemented with the aid of MS-based techniques. The gel "spots" can be excised, washed, subjected to proteolytic digestion, and characterized by MALDI- or ESI-MS in an automated fashion (Neubauer and Mann, 1999; Pandey et al., 2000; Westergren-Thorsson et al., 2001). Strategies such as the yeast two-hybrid system, affinity chromatography techniques, or prefractionation steps may be necessary to target a specific class of proteins prior to additional sample manipulations (Link et al., 1999; Gruninger-Leitch et al., 2000; Husi et al., 2000; Zuo and Speicher, 2000). Commonly, on-line chromatography coupled to MS analysis can be adapted to further separate complex protein or peptide mixtures (Chong et al., 2001). A number of software packages are currently available to perform query in large databases (e.g., bioinformatics approaches) and enhance the speed of the MS protein identification process (Clauser et al., 1999). Several of these software algorithms for protein or gene identification are available via the World Wide Web (Fenyo, 2000). At the present, nonredundant protein databases with ~350,000 entries and human EST databases with ~1,200,000 entries can yield most sophisticated and accurate identification output compared with the 2-D gel analysis (Jensen et al., 1999). In addition, 2-D gel sample components below MM of ~10,000 Da or above 100,000 Da that are not easily characterized can be scrutinized readily by MS. The following examples will clarify the above discussions.

Recently, a high-throughput protein identification [double parallel digestion (DPD)] method was reported by Sanchez and coworkers (Bienvenut et al., 1999). In the DPD approach, partially digested proteins were obtained using an immobilized trypsin membrane and subsequently transblotted. The resulting peptides were trapped on a polyvinylidene difluoride membrane and scrutinized by MS. The DPD approach was successfully applied to a mini-2D gel electrophoresis of Escherichia coli extract. A potential drawback of 2-DE-MS approach is its limited utility in the analysis of low-abundance (copy-number) or poor solubility proteins. This is particularly an issue in higher eukaryotic cells, which tend to be complex and contain a host of proteins with diverse isoelectric points and MM. Therefore, to achieve a meaningful visualization, the recommendations are: make improvements in the protein recovery steps (e.g., prefractionation, affinity LC), perform subcellular fractionation and purification of organelles, and/or start with larger amounts of proteins (Gygi et al., 2000a; Jung et al., 2000). Alternatively, elegant on-line multidimensional or tandem liquid chromatography-mass spectrometry strategies have been proposed as a substitute for the more widely used 2-D gel analysis (Davis et al., 2001; Washburn et al., 2001). These methods involve mixed bed microcapillary columns containing strong cation exchange, size exclusion, and/or reversed phase resins. The tandem LC approach offers the potential to avoid the need for time-consuming 2-DE and dramatically increases the number of peptides that can be identified in a complex mixture. In addition, it offers a wide dynamic range and alleviates the protein solubility problems encountered during 2-DE experiments, because the proteins undergo enzymatic digestion en masse (Haynes and Yates, 2000).

Post-translational modification of proteins plays a pivotal role in functional activity and signal transduction in all living organisms. The mass changes caused by post-translational modifications (i.e., acetylation, farnesylation, glycosylation, phosphorylation, methylation, sulfation) can be easily detected by ESI or MALDI (Knotts et al., 2001; Merrick et al., 2001; Nemeth et al., 2001). Identification of the protein fragment of increased mass after enzymatic digestion allows possible determination of the site of modification. For example, in the case of phosphorylation, radiolabeling with ³²P is not required and low levels of phosphopeptides (i.e., 200-300 femtomoles) can readily be identified with high efficiency and speed. Additional information is discerned by performing tandem mass spectrometry or collision-induced dissociation (CID) experiments. In a typical CID experiment, a beam of ions with a specific m/z (denoted as the precursor or parent ion) is selected and collided with a neutral and nonreactive gas phase target such as argon. These collisions result in subsequent fragmentation and product ions that are a direct consequence of dissociation of the precursor ion. Generally, the resulting fragmentation pattern is unique for all ions having a particular structure. CID experiments are particularly useful in partial or complete peptide sequencing (Baldwin et al., 2001; Oda et al., 2001). In this approach, low-energy CID experiment is used to generate spectra dominated by product ions resulting from cleavage at the amide bonds. Because the low-energy CID yields very little amino acid side-chain fragmentation, the interpretation of the CID spectra becomes relatively simple. In high-throughput proteome analysis, the observed fragment ions are often matched against the predicted fragment ions of all peptides in a database (Aebersold and Goodlett, 2001; Zhou et al., 2001). A series of scores are then assigned to each "hit" to rank the quality of the match. Clearly, this approach is useful for genetically well-characterized organisms in which substantial protein or cDNA sequence exists.

Figure 3 depicts the MS and tandem MS (MS/MS) of a synthetic cyclic phosphopeptide in positive and negative ion mode, respectively. The hexapeptide, cyclo[-Gly-Tyr(PO₃H₂)-Val-Pro-Met-Leu-], corresponds to the autophosphorylation segment around Tyr⁷⁵¹ of the platelet-derived growth factor receptor  subunit. This peptide serves as an inhibitor of the receptor tyrosine kinase/scr homology region 2 domains that mediate mitogenic signal transduction pathways (Barchi et al., 1996). After liquid chromatography or direct infusion of the sample, the molecular mass of the phosphopeptide is determined by ESI-MS in positive ion mode (Fig. 3a). To ascertain the presence of post-translational modification, the MS is switched to negative ion and the precursor or parent ion scanning mode. In this mode of operation, the phosphopeptide is fragmented using tandem mass spectrometry and yields a signature ion with high specificity at m/z 79 (i.e., loss of PO₃). Consequently, the resultant simplified mass spectrum exhibits only precursor ions that lost m/z 79 (Fig. 3b). This technique is particularly useful in detecting peptides containing phosphotyrosine residues, which are difficult to discern from the background chemical noise in a typical single stage ESI-MS experiment (Wilm et al., 1996; Vener at al., 2001). A similar strategy could be applied for the detection of glycopeptides containing N-hexosamines yielding characteristic ions at m/z 204 and m/z 168, corresponding to loss of the oxonium ion of N-hexosamine and the oxonium ion with a double water loss, respectively (Wilm et al., 1996). Recently, an alternative specific approach involving the detection of the immonium ion of phosphotyrosine at m/z 216.043 in positive precursor ion scanning mode combined with high-resolution quadrupole time-of-flight MS has been described (Steen et al., 2001).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   a, positive ESI mass spectrum of the hexapeptide, cyclo[-Gly-Tyr(PO3H2)-Val-Pro-Met-Leu-] using a triple quadrupole mass spectrometer yielding signals at m/z 741.4 and m/z 763.5 corresponding to [M+H]+ and [M+Na]+ ions, respectively; b, parent ion scan for m/z 79 (loss of PO3) in negative ion mode for the same peptide. The phosphopeptide is clearly evident in its singly charged form at m/z 739.5 corresponding to [M-H] ions.

Another related area in proteome analysis is protein-expression mapping, which is defined as the quantitative measurement of the protein dynamics in the specimen (e.g., cell, tissue, or body fluid) of interest. In this approach, the proteome analysis is typically performed in a subtractive fashion whereby the alterations in individual proteins for two or more states are compared. These so-called "cell-states" could refer to a cell before and after treatment/stress with xenobiotics or cells obtained from normal and pathological states. Protein-expression mapping is of interest in the early drug development (preclinical safety) stages and toxicological studies when exposure, effects, and susceptibility in laboratory animals are performed.

Although the de novo identification (qualitative analysis) of cellular proteins using MS technology is becoming widely recognized, the quantitative cellular stress response measurements provide an added benefit of the extent to which an organism is altered. In this regard, Smith and coworkers (Pasa-Tolic et al., 1999) demonstrated the utility of ultra-high resolution MS measurements in conjunction with the resolving power of capillary isoelectric focusing (CIEF) for characterization of cadmium stress response in Escherichia coli K-12 strain MG1655 cells. The cells were cultured in normal as well as rare-isotope (e.g., ¹³C, ¹⁵N, ²H) depleted media to provide internal standards for all detected proteins after cadmium-mediated stress. Abundance of several intact and isotopically distinctive proteins was monitored qualitatively for up to several hours using this method. Presumably, this type of stable-isotope approach eliminates the need for the 2-DE based analysis of crude mixtures by adapting on-line CIEF-MS. However, the above procedure is not easily amenable to the characterization of proteins in tissues. Furthermore, stable-isotope enriched media tend to be costly. Lastly, the protein identification must be performed before quantification to decipher the incorporation of stable isotopes within the specific sequence of protein.

Recently, an elegant quantitative microcapillary-LC-ESI-MS strategy for the analysis of protein mixtures in yeast Saccharomyces cerevisiae was reported by Aebersold and coworkers (Gygi et al., 1999, 2000b; Gygi and Aebersold, 2000; Griffin et al., 2001; Ideker et al., 2001). An isotope coded affinity tag (ICAT) that consisted of an affinity tag (biotin), a linker containing stable isotopes, and a reactive moiety with propensity to react with free sulfhydryl groups (e.g., cysteines) was used (Fig. 4). Two sets of cell states (or tissue extracts) were independently treated with the isotopically light and heavy (8 Da higher in MM by incorporation of ²H) ICAT reagents. The cells were combined and subjected to proteolytic cleavage. The ICAT-labeled peptides were isolated using the biotin tag and analyzed by microcapillary-LC-ESI-MS. Peptide sequence information was obtained by tandem mass spectrometry experiments and identified by computer searches against protein data banks. Quantification of proteins was performed by ratios of respective light and heavy ICAT-labeled peptides, which were generated using enzymatic digestion. The stable isotope labeling procedure was a clever approach to assist in the identification of two peptides with identical sequences and molecular mass values from two different cell-states. Because all the physical characteristics of two identical protein samples from two cell-states remain the same, the resulting peptide fragments obtained from enzymatic cleavages yielded identical mass spectra (Fig. 4). Thus, incorporation of specific stable isotopes in one cell-state resulted in mass shifts, which in turn served as an internal standard for all other cell-states within the same experiment. Because the light and heavy ICAT tagged peptides were chemically identical, one can safely assume that they would yield analogous MS detection (ionization) response and behave as mutual internal standards for quantification purposes. Other similar procedures such as biotinylation of cysteine containing peptides (Spahr et al., 2000), N-terminal labeling using light and 1-([H₄/D₄]nicotinoyloxy)succinimide esters (Munchbach et al., 2000), and derivatization of primary amine moieties with acetate and trideuteroacetate (Ji et al., 2000) have recently been reported. Currently, a drawback of the ICAT strategy is its limited scope to the cysteine containing proteins. This limitation should be alleviated with the introduction of ICAT reagents specific for other abundant amino acids.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Principle of quantitative proteomics by incorporation of a stable isotope-labeled tag. Cysteine amino acids of proteins in two different cell states (i.e., normal versus abnormal due to a stress cause by a drug or other environmental factors) are covalently modified by the tag. One tag contains 2H and the other incorporates 1H. The protein extracts from the two cell states are mixed, digested by trypsin, and separated by affinity chromatography. The resulting peptide mixtures are then subjected to LC-MS and tandem mass spectrometry. The ratios of labeled and unlabeled peptides (different by 8 Da) is the manifestation of the abundance of the gene in the two cell states. Subsequent, LC-MS/MS peptide sequencing experiments of these peptides can identify the gene product, which is being quantified. A computer-search algorithm that accesses large protein data banks is typically used during the identification process.

In general, the above protocols for qualitative and quantitative analysis of the key cellular proteins could aid in several areas of drug discovery and development such as toxicological and pharmacogenomics studies. For example, in the science of pharmacogenomics, genetic polymorphisms in transporters, drug-metabolizing enzymes (e.g., cytochromes P450, uridine 5'-triphosphate glucuronosyltransferases), receptors, and therapeutic target proteins have been postulated to be one of the underlying reasons in variable responses to drug treatments in patients (Broder and Venter, 2000; March, 2000; McLeod and Evans, 2001). Target validation that could predict a well-tolerated medicine with adequate efficacy in humans is a widely perceived challenge. Currently, these investigations are rather tedious and empirical. In most cases, the human genetic variations resulting in great heterogeneity in response to medications are realized in large studies at the postmarketing stages (i.e., population sizes exceeding 100,000). Some of these idiosyncratic responses are toxic and thus it is essential to identify them before phase III clinical studies (Gould Rothberg et al., 2000; Knowles et al., 2000). According to a meta-analysis published in the Journal of the American Medical Association (Lazarou et al., 1998), adverse or idiosyncratic drug reactions in hospitalized patients is one of the major causes of mortality. Consequently, it is of interest to elucidate the identity and pharmacogenomic traits (i.e., polymorphically expressed enzymes) of key cellular proteins and to design optimum medication for individual patients. MS technology offers a viable platform that can be utilized to assay differential protein expression after drug treatment.

Characterization of Clinical Biomarkers. The need and prospects for new and improved biomarkers in therapeutic area as clinical assessment tools are evident. It can be argued that once a biomarker is identified, it is more efficient to monitor its structural permutations and/or expression level by direct assay rather than the global 2-DE approach. Thus, there is still much impetus to develop technologies for detailing protein structure and accurately monitoring protein expression levels. Over the past 8 years, we have devoted much effort to developing devices and methods for the rapid MALDI-TOF analysis of proteins residing in complex biological media. Our approach, termed mass spectrometric immunoassay (MSIA) (Nelson et al., 1995) is shown in Fig. 5. In general, analytes are retrieved selectively from solution by repetitive flow through a receptor-derivatized porous frit constructed at the entrance of a wide-bore P-200 pipettor tip (MSIA-Tip). Once washed of the nonspecific compounds, the retained species are eluted onto a mass spectrometer target in preparation for mass spectrometry. MALDI-TOF then follows, with analytes detected at precise m/z values. Analyses are qualitative by nature but can be made quantitative by incorporating mass-shifted variants of the analyte into the procedure for use as internal standards (Nelson et al., 1994, 1995; Wang et al., 1996; Rudiger et al., 1999). The approach is ideally suited to high-throughput applications using robotics and automated MALDI-TOF instrumentation. Recently, we have explored the benefits of using the MSIA approach in structural characterization and quantification of -2-microglobulin (₂m), a low-molecular-mass protein, identified as the light chain of the class I major histocompatibility complex. Full details of the approach are given elsewhere (Tubbs et al., 2001). Briefly, anti-₂m polyclonal antibody MSIA-Tips were used in selective extraction of ₂m from various human biological fluids. Figure 6 shows MALDI-TOF and MSIA spectra obtained from a human urine sample. The MALDI-TOF spectrum shows a number of signals in the peptide region and an absence of signal for ₂m. The MSIA spectrum is dominated by signals from the ₂m, with few additional signals from nonspecified components.

View larger version (39K): [in this window] [in a new window]   Fig. 5.   a, illustration of the MSIA procedure. b, MSIA applied with robotics for high-throughput analyses.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   2-microglobulin MSIA screening of urine. Samples were prepared by dilution of urine in preparation buffer and repetitive flow incubation through the MSIA-Tip. Tips were washed using buffer and water before elution of retained compounds directly onto a mass spectrometer target using -cyano-4-hydroxycinnamic acid (saturated in 1:2 acetonitrile/water and 0.2% trifluoroacetic acid). 2m is observed as the predominant signal in the MSIA spectrum, whereas direct MALDI-TOF of urine yields little useful information.

Protein quantification using MALDI-TOF requires use of internal standards to compensate for varying laser intensities and spot-to-spot differences in sample composition that give rise to fluctuations in analyte ion signal (Nelson et al., 1994). Although proteins with characteristics unlike those of the analyte may be used as internal standards (as has been shown during protein quantification directly from mixtures (Nelson et al., 1994; Muddiman et al., 1995) or during MALDI-TOF quantification of affinity-retrieved species by addition of an internal reference standard to peptides eluted from beaded affinity reagent (Kuwata et al., 1998), internal reference standards that behave similarly to the analyte during laser desorption/ionization are generally preferred. This prerequisite is met during MSIA by choosing internal references that share sequence homology with the target protein: enzymatic/chemically modified versions of the targeted protein (Nelson et al., 1995; Krone et al., 1996; Wang et al., 1996), truncated/extended recombinant forms of the target proteins, the (same) target protein recombinantly expressed in isotopically-enriched media (e.g., ¹⁵N or ¹⁸O), or the same protein from a different biological species. Given that the receptor is able to capture both the target protein and the internal reference, MSIA can be designed around a single receptor system (Nelson et al., 1995; Wang et al., 1996). Alternatively, a two-receptor system can be considered where one receptor is used to retrieve the target protein and a separate receptor is used to retrieve the internal reference (Nelson et al., 1995; Krone at al., 1996; Rudiger et al., 1999).

The internal reference chosen for this study was equine ₂m (E₂m), which shares ~ 75% homology with its human counterpart (H₂m) and is ~300 Da less in mass than H₂m (thus, both species share similar characteristics and are easily resolved in the mass spectra). Although, no data could be found on the relative dissociation constants between the polyclonal anti-₂m IgG and H₂m or E₂m, preliminary studies showed that the antibody exhibited cross-reactivity sufficient to retain both species. Figure 7a shows spectra representing MSIA analyses of H₂m standards in a concentration range of 0.01-1.0 mg/l. Each spectrum, normalized to the E₂m signal, is one of 10 65-laser shots spectra taken for each calibration point. Plotting the average of the 10 normalized H₂m integrals for each standard versus the H₂m concentration results in the working curve shown in Fig. 7b. Linear regression fitting of the data yields I_H2m/I_E2m = 4.09 mg/l H₂m + 0.021 (R² = 0.983), with a working limit of detection at a S/N > 3 of 0.0025 mg/l (210 pM) and a limit of quantification of 0.01 mg/l (850 pM). The standard error of all points of the working curve is ~ 5%.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Quantitative 2m-MSIA. a, representative spectra of data used to generate working curve. Human 2m concentrations of 0.01 - 1.0 mg/l were investigated. Equine 2m (MM = 11,396.6 Da) was used as an internal standard. b, working curve generated using the data represented in a. The 2-decade range was spanned with good linearity (R2 = 0.983). Error bars reflect the S.D. of 10 repetitive 65-laser shot spectra taken from each sample. Inset, 2m- screening. Urine samples from five subjects were screened over a period of 2 days. The average value determined for healthy persons was 0.100 ± 0.021 mg/l. The level determined for an 86-year old female (F86) with a recent urinary tract infection indicated a significant increase in 2m concentration (3.23 ± 0.072 mg/l).

Urine samples were collected from four persons over a 2-day period: female, 31 years (F31); male, 30 years (M30); male, 36 years (M36); and, male 44 years (M44). All of the subjects were in a state of good health when the samples were collected. Results from MSIA of the urine samples are summarized in Fig. 7 (inset). The bars depict the ₂m concentration determined for each sample, and the inset spectra above each bar shows the respective H₂m signals normalized to E₂m. The data for the samples show remarkable consistency, with an average ₂m concentration of 0.100 ± 0.021 mg/l (high = 0.127 mg/l; low = 0.058 mg/l). An additional analysis was performed on a urine sample obtained from an 86-year old female (F86) who had recently suffered a renal infection, revealing a significantly higher ₂m concentration (3.23 ± 0.02 mg/l).

The mass-selective detection of MSIA makes possible the discovery and quantification of variants of ₂m that may be present in urine. During quantitative screening of the urine samples, a second, higher molecular mass species (m = +161 Da) was coextracted with the ₂m. The species is presumably a glycosylated (one hexose) form of ₂m and is observed most prominently in F86. Figure 8 shows an overlay of two MSIA spectra taken from the urine of F86 (diluted × 20) and M36 (no dilution; given for comparison). The level of glycosylated ₂m is much greater in F86 than in M36. The specific cause of the elevated level of the glyco-₂m, which is roughly the same concentration as wild-type ₂m in the urine samples from the healthy subjects, is at present uncertain. Note that because the two ₂m species are differentiated using mass spectrometry, the concentration of the wild type ₂m determined during MSIA accurately reflects the concentration of only the wild type ₂m and not the combination of both of the species. Thus, MSIA holds a particular advantage over other techniques that are unable to differentiate between similar forms of a target analyte. Elevated ₂m levels are used as a general indicator of immune system activity, whereas ₂m-glycosylation has been associated with more specific ailments [e.g., advanced glycosylated end-products associated with dialysis related amyloidosis (Niwa, 1997)]. MSIA is able to deconvolute these independent contributing factors and yield results that more accurately relate a specific biomarker with a specific disease state.

View larger version (33K): [in this window] [in a new window]   Fig. 8.   MSIA showing elevated level of 2m-glycan in 86-year-old female (F86, dark gray). During MSIA, a second signal is observed at m = +161 Da, indicating the presence of glycosylated 2m. MSIA is able to adequately resolve the two 2m forms resulting in a more accurate quantification of the nascent 2m, and possible quantification of the glycan. Such differentiation is important considering that the two 2m forms originate from (or are markers for) different ailments. MSIA of a healthy individual, showing little glycation, is given for comparison (light gray).

Conclusion and Future Outlook. During the past decade, dramatic progress in the field of mass spectrometry has resulted in a large increase in the number of commercially available MS instruments. Based on the overwhelming number of published manuscripts, it is clear that MS is becoming a pivotal bioanalytical tool in many biotechnology and biochemistry laboratories. MALDI- and ESI-MS allow the characterization of a myriad of small and large molecules with high sensitivity, speed, accuracy, and efficiency. MS-based techniques are becoming a permanent component of studies involving functional genomics, proteomics, early drug discovery, clinical diagnostics, and combinatorial chemistry.

Although the applications shown here have focused on protein quantification, there are numerous other applications in which the MSIA-Tips can be used in the qualitative (structural) characterization of proteins. In general, tips can be derivatized with affinity ligands and used for retrieval of proteins from mixtures. Screening of expression systems for production of the correct construct is a good example. Because most cloning vectors incorporate some general utility affinity tag into a recombinant protein, MSIA-Tips derivatized with antibodies toward these tags (e.g., anti-His-tag or anti-myc antibodies) can be used to selectively retrieve recombinant proteins from expression systems, after which MALDI-TOF is used to verify production of the correct protein via accurate mass determination. Alternatively, tagged-peptides containing nascent sequence of an expressed protein can be used to identify the expressed protein (Nelson et al., 1999). Such an application stands to find use in identifying components of DNA libraries that would otherwise be identified at the DNA level (through hybridization approaches or sequencing). Chelate surfaces can also be used to selectively retrieve metal-binding polypeptides for rigorous mass spectrometric characterization, as has been shown in the recent independent works of two laboratories (Posewitz and Tempst, 1999; Qian et al., 1999). Finally, in further-reaching applications, receptors may be immobilized for use in selective retrieval of receptor-interacting proteins from solution under native conditions. MSIA-Tips can be used in these applications to feed directly into MALDI-TOF or ESI instrumentation, or alternatively into one-dimensional gels, if further separation of retained compounds is necessary (Neubauer et al., 1998). These applications promise to complement protein-interaction reporter techniques such as yeast two-hybrid systems or instrumental techniques such as surface plasmon resonance-based biosensors with an emphasis, however, on the protein structural-analysis (e.g., identification) capabilities of mass spectrometry.

The future prospects of mass spectrometry are exciting. Advancements in miniaturization and chip technology using photolithography/etching fabrication (Nelson et al., 2000; Figeys and Pinto, 2001; Kodadek, 2001), identification of single-nucleotide polymorphisms (Griffin et al., 1999; Stoerker et al., 2000; Buetow et al., 2001), clinical diagnostic applications (Worrall et al., 2000; Chace, 2001), rapid enzyme kinetics analysis, (Bothner at al., 2000; Gross et al., 2000; Norris et al., 2001), and further adaptation of automation/robotics sample processing steps (Lopez et al., 2000; Kyranos et al., 2001) will continue to have a significant impact on modern medicine.

	Acknowledgments

R.B. thanks the Finnigan Corporation (San Jose, CA, USA) for providing Fig. 1b.

	Footnotes

Received December 19, 2001; Accepted June 11, 2001

This work was supported in part by National Institutes of Health Grant 1-R43-GM56603-01A2.

Dr. Ray Bakhtiar, Mail Stop: RY80L-109, Dept. of Drug Metabolism, Merck Research Laboratories, Rahway, NJ 07065. E-mail: ray_bakhtiar{at}merck.com or Dr. Randall Nelson, Intrinsic Bioprobes, 625 South Smith Road, Suite 22, Tempe, AZ 85281. E-mail: info{at}intrinsicbio.com

	Abbreviations

MM, molecular mass; MS, mass spectrometry; MALDI, matrix-assisted laser desorption ionization; ESI, electrospray ionization; LC, liquid chromatography; CE, capillary electrophoresis; PAGE, polyacrylamide gel electrophoresis; 2-D, two-dimensional; EST, expressed sequence tag; DPD, double parallel digestion; 2-DE, two-dimensional electrophoresis; CID, collision-induced dissociation; CIEF, capillary isoelectric focusing; ICAT, isotope coded affinity tag; TOF, time-of-flight; MSIA, mass spectrometric immunoassay; ₂m, -2-microglobulin; E₂m, equine -2-microglobulin; H₂m, human -2-microglobulin.

	References

Top Article References

• Aebersold R and Goodlett DR (2001) Mass spectrometry in proteomics. Chem Rev  101: 269-295[Medline].
• Alaiya AA, Oppermann M, Langridge J, Roblick U, Egevad L, Brandstedt S, Hellstrom M, Linder S, Bergman T, Jornvall H, et al. (2001) Identification of proteins in human prostate tumor material by two-dimensional gel electrophoresis and mass spectrometry. Cell Mol Life Sci  58: 307-311[Medline].
• Anderson NG, Matheson A and Anderson NL (2001) Back to the future: The human protein index (HPI) and the agenda for post-proteomic biology. Proteomics  1: 3-12[Medline].
• Bakhtiar R and Tse FLS (2000) Biological mass spectrometry: a primer. Mutagenesis  15: 415-430[Abstract/Free Full Text].
• Bakhtiar R and Nelson RW (2000) Electrospray ionization and matrix-assisted laser desorption ionization mass spectrometry. Biochem Pharmacol  59: 891-905[Medline].
• Baldwin MA, Medzihradszky KF, Lock CM, Fisher B, Settineri TA and Burlingame AL (2001) Matrix-assisted laser desorption/ionization coupled with quadrupole/orthogonal acceleration time-of-flight mass spectrometry for protein discovery, identification, and structural analysis. Anal Chem  73: 1707-1720[Medline].
• Barchi JR, Nomizu M, Otaka A, Roller PP and Burke TR, Jr (1996) Conformational analysis of cyclic hexapeptides designed as constrained ligands for the SH2 domain of the p85 subunit of phosphatidylinositol-2-OH kinase. Biopolymers  38: 191-208[Medline].
• Bichsel VE, Liotta LA and Petricoin EF, 3rd (2001) Cancer proteomics From biomarker discovery to signal pathway profiling. Cancer J  7: 69-78[Medline].
• Bienvenut WV, Sanchez J-C, Karmime A, Rouge V, Rose K, Binz P-A and Hochstrasser DF (1999) Toward a clinical molecular scanner for proteome research: Parallel protein chemical processing before and during Western blot. Anal Chem  71: 4800-4807[Medline].
• Bothner B, Chavez R, Wei J, Strupp C, Phung Q, Schneemann A and Siuzdak G (2000) Monitoring enzyme catalysis with mass spectrometry. J Biol Chem  275: 13455-13459[Abstract/Free Full Text].
• Broder S and Venter JC (2000) Sequencing the entire genomes of free-living organisms: The foundation of pharmacology in the new millennium. Annu Rev Pharmacol Toxicol  40: 97-132[Medline].
• Burlingame AL, Carr SA and Baldwin MA, (eds) (2000) Mass Spectrometry in Biology and Medicine. Humana Press, Totowa, NJ.
• Buetow KH, Edmonson M, MacDonald R, Clifford R, Yip P, Kelley J, Little DP, Strausberg R, Koester H, Cantor CR, et al. (2001) High-throughput development and characterization of a genome wide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Proc Natl Acad Sci USA  98: 581-584[Abstract/Free Full Text].
• Chace DH (2001) Mass spectrometry in the clinical laboratory. Chem Rev  101: 445-477[Medline].
• Chong BE, Hamler RK, Lubman DM, Ethier SP, Rosenspire AJ and Miller FR (2001) Differential screening and mass mapping of proteins from premalignant and cancer cell lines using nonporous reversed-phase HPLC coupled with mass spectrometric analysis. Anal Chem  73: 1219-1227[Medline].
• Clauser KR, Baker P and Burlingame AL (1999) Role of accurate mass measurement (+/ 10 ppm) in protein identification strategies employing MS or MS/MS and database searching. Anal Chem  71: 2871-2882[Medline].
• Davis MT, Beierle J, Bures ET, McGinley MD, Mort J, Robinson JH, Spahr CS, Yu W, Luethy R and Patterson SD (2001) Automated LC-LC-MS-MS platform using binary ion-exchange and gradient reversed-phase chromatography for improved proteomic analyses. J Chromatogr B  752: 281-291.
• Dell A and Morris HR (2001) Glycoprotein structure determination by mass spectrometry. Science (Wash DC)  291: 2351-2356[Abstract/Free Full Text].
• Farmer TB and Caprioli RM (1998) Determination of protein-protein interactions by matrix-assisted laser desorption/ionization mass spectrometry. J Mass Spectrom  33: 697-704[Medline].
• Fenyo D (2000) Identifying the proteome: software tools. Curr Opin Biotechnol  11: 391-395[Medline].
• Fernandez de la Mora J, Van Berkel GJ, Enke CG, Cole RB, Martinez-Sanchez M and Fenn JB. (2000) Electrochemical processes in electrospray ionization mass spectrometry. J Mass Spectrom  35: 939-952[Medline].
• Figeys D and Pinto D (2001) Proteomics on a chip: promising developments. Electrophoresis  22: 208-216[Medline].
• Gabor Miklos GL and Maleszka R (2001) Integrating molecular medicine with functional proteomics: realities and expectations. Proteomics  1: 30-41[Medline].
• Godovac-Zimmermann J and Brown LR (2001) Perspectives for mass spectrometry and functional proteomics. Mass Spectrom Rev  20: 1-57[Medline].
• Gould Rothberg BE, Ramesh TM and Burgess CE (2000) Integrating expression-based drug response and SNP-based pharmacogenetic strategies into a single comprehensive pharmacogenomics program. Drug Dev Res  49: 54-64.
• Griffin TJ, Hall JG, Prudent JR and Smith LM (1999) Direct genetic analysis by matrix-assisted laser desorption/ionization mass spectrometry. Proc Natl Acad Sci USA  96: 6301-6306[Abstract/Free Full Text].
• Griffin TJ, Gygi SP, Rist B, Aebersold R, Loboda A, Jilkine A, Ens W and Standing KG (2001) Quantitative proteomic analysis using a MALDI quadrupole time-of-flight mass spectrometer. Anal Chem  73: 978-986[Medline].
• Gruninger-Leitch F, Berndt P, Langen H, Nelboeck P and Dobeli H (2000) Identification of -secretase-like activity using a mass spectrometry-based assay system. Nature Biotechnol  18: 66-70[Medline].
• Gygi SP, Rist B, Gerber SA, Turecek F, Gelb MH and Aebersold R (1999) Quantitative analysis of complex protein mixtures using isotope-coded affinity tags. Nature Biotechnol  17: 994-999[Medline].
• Gygi SP, Corthals GL, Zhang Y, Rochon Y and Aebersold R (2000a) Evaluation of two-dimensional gel electrophoresis-based proteome analysis technology. Proc Natl Acad Sci USA  97: 9390-9395[Abstract/Free Full Text].
• Gygi SP, Rist B and Aebersold R (2000b) Measuring gene expression by quantitative proteome analysis. Curr Opin Biotechnol  11: 396-401[Medline].
• Gygi SP and Aebersold R (2000) Mass spectrometry and proteomics. Curr Opin Chem Biol  4: 489-494[Medline].
• Gross JW, Hegeman AD, Vestling MM and Frey PA (2000) Characterization of enzymatic processes by rapid mix-quench mass spectrometry: The case of dTDP-glucose 4,6-dehyratase. Biochemistry  39: 13633-13640[Medline].
• Hancock W, Apffel A, Chakel J, Hahnenberger K, Choudhary G, Traina JA and Pungor E (1999) Integrated genomic/proteomic analysis. Anal Chem  71: 742A-748A[Medline].
• Haynes PA and Yates JR, 3rd (2000) Proteomic profiling-pitfalls and progress. Yeast  17: 81-87[Medline].
• Hofstadler SA and Griffey RH (2001) Analysis of noncovalent complexes of DNA and RNA by mass spectrometry. Chem Rev  101: 377-390[Medline].
• Husi H, Ward MA, Choudhary JS, Blackstock WP and Grant SGN (2000) Proteomic analysis of NMDA receptor-adhesion protein signaling complexes. Nat Neurosci  3: 661-669[Medline].
• Ideker T, Thorsson V, Ranish JA, Christmas R, Buhler J, Eng JK, Bumgarner R, Goodlett DR, Aebersold R and Hood L (2001) Integrated genomic and proteomic analyses of a systematically perturbed metabolic network. Science (Wash DC)  292: 929-934[Abstract/Free Full Text].
• Jensen ON, Wilm M, Shevchenko A and Mann M (1999) Peptide sequencing of 2-DE gel-isolated proteins by nanoelectrospray tandem mass spectrometry. Methods Mol Biol  112: 571-588[Medline].
• Ji J, Chakraborty A, Geng M, Zhang X, Amini A, Bina M and Regnier F (2000) Strategy for qualitative and quantitative analysis in proteomics based on signature peptides. J Chromatogr B  745: 197-210.
• Jung E, Heller M, Sanchez J-C and Hochstrasser DF (2000) Proteomics meets cell biology: The establishment of subcellular proteomes. Electrophoresis  21: 3369-3377[Medline].
• Kodadek T (2001) Protein microarrays: prospects and problems. Chem & Biol  8: 105-115[Medline].
• Knotts TA, Orkiszewski RS, Cook RG, Edwards DP and Weigel NL (2001) Identification of a phosphorylation site in the hinge region of the human progesterone receptor and additional amino-terminal phosphorylation sites. J Biol Chem  276: 8475-8483[Abstract/Free Full Text].
• Knowles SR, Uetrecht J and Shear NH (2000) Idiosyncratic drug reactions: The reactive metabolite syndromes. Lancet  356: 1587-1591[Medline].
• Krone JR, Nelson RW and Williams P (1996) Mass spectrometric immunoassay, in Ultrasensitive Biochemical Diagnostics. SPIE Proceedings (Cohn GE, Soper SA and Chen CHW eds) vol 2680, pp 415-421 (paper 2680-50); SPIE-The International Society for Optical Engineering, Bellingham, WA.
• Kyranos JN, Cai H, Wei D and Goetzinger WK (2001) High-throughput high-performance liquid-chromatography/mass spectrometry for modern drug discovery. Curr Opin Biotechnol  12: 105-111[Medline].
• Kuwata H, Yip TT, Yip CL, Tomita M and Hutchens TW (1998) Bactericidal domain of lactoferrin: detection, quantitation, and characterization of lactoferricin in serum by SELDI affinity mass spectrometry. Biochem Biophys Res Commun  245: 764-773[Medline].
• Lahm HW and Langen H (2000) Mass spectrometry: A tool for the identification of proteins separated by gels. Electrophoresis  21: 2105-2114[Medline].
• Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, et al. (2001) Initial sequencing and analysis of the human genome. International Human Genome Sequencing Consortium. Nature (Lond)  409: 860-921[Medline].
• Lazarou J, Pomeranz BH and Corey PN (1998) Incident of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA  279: 1200-1205[Abstract/Free Full Text].
• Link AJ, Eng J, Schieltz DM, Carmack E, Mize GJ, Morris DR, Garvik BM and Yates JR, 3rd (1999) Direct analysis of protein complexes using mass spectrometry. Nat Biotechnol  17: 676-682[Medline].
• Lopez MF, Kristal BS, Chernokalskaya E, Lazarev A, Shestopalov AI, Bogdanova A and Robinson M (2000) High-throughput profiling of the mitochondrial proteome using fractionation and automation. Electrophoresis  21: 3427-3440[Medline].
• Mann M and Pandey A (2001) Use of mass spectrometry-derived data to annotate nucleotide and protein sequence databases. Trends Biochem Sci  26: 54-61[Medline].
• March R (2000) Pharmacogenomics: the genomics of drug response. Yeast  17: 16-21[Medline].
• Merrick BA, Zhou W, Martin KJ, Jeyarajah S, Parker CE, Selkirk JK, Tomer KB and Borchers CH (2001) Site-specific phosphorylation of human p53 protein determined by mass spectrometry. Biochemistry  40: 4053-4066[Medline].
• McLeod HL and Evans WE (2001) Pharmacogenomics: Unlocking the human genome for better drug therapy. Annu Rev Pharmacol Toxicol  41: 201-121.
• Miura K (2001) Imaging and detection technologies for image analysis in electrophoresis. Electrophoresis  22: 801-813[Medline].
• Muddiman DC, Gusev AI and Hercules DM (1995) Application of time-of-flight mass spectroscopy for the quantitative analysis of biological molecules. Mass Spectrom Rev  14: 383-429.
• Munchbach M, Quadroni M, Miotto G and James P (2000) Quantitative and facilitated de Novo sequencing of proteins by isotope N-terminal labeling of peptides with a fragmentation-directing moiety. Anal Chem  72: 4047-4057[Medline].
• Nelson RW, McLean MA and Hutchens TW (1994) Quantitative-determination of proteins by matrix-assisted laser-desorption ionization time-of-flight mass-spectrometry. Anal Chem  66: 1408-1415.
• Nelson RW, Krone JR, Bieber AL and Williams P (1995) Mass-spectrometric immunoassay. Anal Chem  67: 1153-1158[Medline].
• Nelson RW, Jarvik JW, Taillon BE and Tubbs KA (1999) BIA/MS of epitope-tagged peptides directly from E. coli lysate: multiplex detection and protein identification at low-femtomole to subfemtomole levels. Anal Chem  71: 2858-2865[Medline].
• Nelson RW and Krone JR (1999) Advances in surface plasmon resonance biomolecular interaction analysis mass spectrometry (BIA/MS). J Mol Recognit  12: 77-93[Medline].
• Nelson RW, Nedelkov D and Tubbs KA (2000) Biosensor chip mass spectrometry: A chip-based proteomics approach. Electrophoresis  21: 1155-1163[Medline].
• Nemeth JF, Hochensang GP, Jr, Marnett LJ and Caprioli RM (2001) Characterization of the glycosylation sites in cyclooxygenase-2 using mass spectrometry. Biochemistry  40: 3109-3116[Medline].
• Neubauer G, King A, Rappsilber J, Calvio C, Watson M, Ajuh P, Sleeman J, Lamond A and Mann M (1998) Mass spectrometry and EST-database searching allows characterization of the multi-protein spliceosome complex. Nat Genet  20: 46-50[Medline].
• Neubauer G and Mann M (1999) Mapping phosphorylation sites of gel-isolated proteins by nanoelectrospray tandem mass spectrometry: Potentials and limitations. Anal Chem  71: 235-242[Medline].
• Niwa T (1997) Mass spectrometry in the search for uremic toxins. Mass Spectrom Rev  16: 307-332[Medline].
• Norris AJ, Whitelegge JP, Faull KF and Toyokuni T (2001) Analysis of enzyme kinetics using electrospray ionization mass spectrometry and multiple reaction monitoring: fucosyltransferase V. Biochemistry  40: 3774-3779[Medline].
• Oda Y, Nagasu T and Chait BT (2001) Enrichment analysis of phosphorylated proteins as a tool for probing the phosphoproteome. Nat Biotechnol  19: 379-382[Medline].
• Pandey A and Mann M (2000) Proteomics to study genes and genomes. Nature (Lond)  405: 837-846[Medline].
• Pandey A, Podtelejnikov AV, Blagoev B, Bustelo XR, Mann M and Lodish HF (2000) Analysis of receptor signaling pathways by mass spectrometry: Identification of Vav-2 as a substrate of the epidermal and platelet-derived growth factor receptors. Proc Natl Acad Sci USA  97: 179-184[Abstract/Free Full Text].
• Pasa-Tolic L, Jensen PK, Anderson GA, Lipton MS, Peden KK, Martinovic S, Tolic N, Bruce JE and Smith RD (1999) High throughput proteome-wide precision measurements of protein expression using mass spectrometry. J Am Chem Soc  121: 7949-7950.
• Posewitz MC and Tempst P (1999) Immobilized gallium(III) affinity chromatography of phosphopeptides. Anal Chem  71: 2883-2892[Medline].
• Qian X, Zhou W, Khaledi MG and Tomer KB (1999) Direct analysis of the products of sequential cleavages of peptides and proteins affinity-bound to immobilized metal ion beads by matrix- assisted laser desorption/ionization mass spectrometry. Anal Biochem  274: 174-180[Medline].
• Rabilloud T (2000) Detecting proteins separated by 2-D gel electrophoresis. Anal Chem  72: 48A-55A[Medline].
• Rudiger AH, Rudiger M, Carl UD, Chakraborty T, Roepstorff P and Wehland J (1999) Affinity mass spectrometry-based approaches for the analysis of protein-protein interaction and complex mixtures of peptide-ligands. Anal Biochem  275: 162-170[Medline].
• Siuzdak G (1996) Mass Spectrometry for Biotechnology. Academic Press, San Diego, California.
• Spahr CS, Susin SA, Bures EJ, Robinson JH, Davis MT, McGinley MD, Kroemer G and Patterson SD (2000) Simplification of complex peptide mixtures for proteomic analysis: Reversible biotinylation of cysteinyl peptides. Electrophoresis  21: 1635-1650[Medline].
• Steen H, Kuster B, fernandez M, Pandey A and Mann M (2001) Detection of tyrosine phosphorylated peptides by precursor ion scanning quadrupole TOF mass spectrometry in positive ion mode. Anal Chem  73: 1440-1448[Medline].
• Stoerker J, Mayo JD, Tetzlaff CN, Sarracino DA, Schwope I and Richert C (2000) Rapid genotyping by MALDI-monitored nuclease selection from probe libraries. Nat Biotechnol  18: 1213-1216[Medline].
• Thomas JJ, Bakhtiar R and Siuzdak G (2000) Mass spectrometry in viral proteomics. Acc Chem Res  33: 179-187[Medline].
• Tomer KB (2001) Separations combined with mass spectrometry. Chem Rev  101: 297-328[Medline].
• Tubbs KA, Nedelkov D and Nelson RW (2001) Detection and quantification of -2-microglobulin using mass spectrometric immunoassay (MSIA). Anal Biochem  289: 26-35[Medline].
• Vener AV, Harms A, Sussman MR and Vierstra RD (2001) Mass spectrometric resolution of reversible protein phosphorylation in photosynthetic membranes of Arabidopsis thaliana. J Biol Chem  276: 6959-6966[Abstract/Free Full Text].
• Wang R, Sweeney D, Gandy SE and Sisodia SS (1996) The profile of soluble amyloid beta protein in cultured cell media. Detection and quantification of amyloid beta protein and variants by immunoprecipitation-mass spectrometry. J Biol Chem  271: 31894-31902[Abstract/Free Full Text].
• Washburn MP, Wolters D and Yates JR, 3rd (2001) Large-scale analysis of the yeast proteome by multidimensional protein identification technology. Nat Biotechnol  19: 242-247[Medline].
• Westergren-Thorsson G, Malmstrom J and Marko-Varga G (2001) Proteomicsthe protein expression technology to study connective tissue biology. J Pharm Biomed Anal  24: 815-824[Medline].
• Williams KL (1999) Genomes and proteomes: toward a multidimensional view of biology. Electrophoresis  20: 678-688[Medline].
• Wilm M, Neubauer G and Mann M (1996) Parent ion scans of unseparated peptide mixtures. Anal Chem  68: 527-533[Medline].
• Worrall TA, Schmeckpeper BJ, Corvera JS and Cotter RJ (2000) Allele-specific HLA-DR typing by mass spectrometry: An alternative to hybridization-based typing methods. Anal Chem  72: 5233-5238[Medline].
• Yates JR, 3rd (2000) Mass spectrometry: from genomics to proteomics. Trends Genet  16: 5-8[Medline].
• Zhou H, Watts JD and Aebersold R (2001) A systematic approach to the analysis of protein phosphorylation. Nat Biotechnol  19: 375-378[Medline].
• Zuo X and Speicher DW (2000) A method for global analysis of complex proteomes using sample prefractionation by solution isoelectrophoresis prior to two-dimensional electrophoresis. Anal Biochem  284: 266-278[Medline].