FTICR mass spectrometry for qualitative and quantitative bioanalyses
Jason S Page , Christophe D Masselon and Richard D Smith
Introduction
Mass spectrometry (MS) is playing an ever-increasing role in biological research, for example, by providing a basis for characterizing protein modifications and identifying novel biomarkers, pathogenic factors and protein interactions in networks and pathways. In particular, protein mixtures pose a significant analytical challenge owing to their complexity and range of relative abundances, as well as their possible modification states. The advent of suitable ionization techniques has made MS a primary tool for biochemical analysis, providing capabilities for rapid and sensitive identification and, increasingly, quantitation. The high resolution, large dynamic range and unmatched sensitivity of Fourier transform ion cyclotron resonance (FTICR)-MS currently provides the highest quality data for biological analyses.
This review highlights the most active and recent areas of bioanalytical research involving FTICR-MS, with an emphasis on proteomics applications. We cover `bottom-up' methods that are based upon peptide-level analyses of protein digests, and emphasize a methodology that increases the throughput of measurements at the whole proteome level using peptide accurate mass and time (AMT) tags as markers for proteins. We also highlight recent improvements to intact protein `top-down' characterization as well as improvements to dissociation techniques for both proteins or peptides that provide better sequence and structural information.
Characterization of biomolecules using accurate mass information from FTICR-MS
A recent article demonstrated the use of high magnetic field FTICR mass spectrometry to analyze a tryptic digest of bovine serum albumin [1.]. By averaging 100 spectra, 86 tryptic peptides were detected with an average mass measurement accuracy (MMA) of only 0.77 ppm. The peptides identified covered 100% of the protein's 583 amino acid residues. The complete coverage provides a basis for more complete identification of protein modifications, and the high mass accuracy provides confidence in the resulting assignments.
In contrast to analysis of a single protein, proteomics (i.e. the direct qualitative and quantitative analysis of the broad complement –– or subset thereof –– of the proteins present in an organism, tissue or cell under a given set of physiological or environmental conditions) presents much greater challenges. This is because of the large number of proteins (and the much greater number of peptides after proteolysis) and the wide range of relative abundances of potential interest. One approach to addressing these challenges has involved globally digesting the proteins to obtain distinctive peptides, which can be separated by liquid chromatography (LC) and identified by tandem mass spectrometry (MS/MS) in a relatively high-throughput and unbiased fashion compared with previous technologies based on two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) [2.]. The use of MS/MS constitutes a bottleneck for proteomics, however, because of the need to identify peptides or proteins one at a time. To circumvent this obstacle, a strategy has been developed that relies on the high mass accuracy of FTICR measurements, using AMT tags to eliminate the routine need for MS/MS in peptide or protein identification [3., 4. and 5.]. An AMT tag is a peptide with a sufficiently distinctive mass and LC elution time to act as a biomarker for a given protein. The process generally invoked for designation of an AMT tag is illustrated in Figure 1. Peptides from a global digest are identified by `shotgun' LC-MS/MS analyses [6.] and an accurate mass is calculated on the basis of the identified peptide sequence. A peptide detected at the calculated mass (typically within 1 ppm) and at the same LC elution time in an LC-FTICR analysis of a related sample establishes the peptide as an AMT tag for the peptide (and often the parent protein by inference) unless ambiguities arise. A database of such AMT tags applied to the LC-FTICR-MS data provides high confidence in identifications, increased sensitivity and higher throughput than methods that rely on repetitive MS/MS identification [4.].
(25K)
Figure 1. Experimental steps involved in the development of AMT tags. This example shows an AMT tag from the protein elongation factor Tu (EF-Tu). The cell extract is enzymatically digested and the digest analyzed by LC-MS/MS and LC-FTICR-MS. During MS/MS analysis, a peptide fragment from EF-Tu is isolated and dissociated. The peptide (potential mass tag; PMT) is subsequently identified using an automated search program. The exact mass of this tag is calculated and its elution time recorded. An AMT tag is established when a peptide eluting at the same time in the LC-FTICR experiment has a mass within 1 ppm of the calculated mass. This AMT tag then serves as a biomarker for the protein EF-Tu in future analysis of the organism by capillary LC-FTICR-MS.
The AMT tag approach was initially demonstrated in a global study of the bacterium Deinococcus radiodurans [3.]. In a set of ~50 000 detected species, ~7000 AMT tags were designated covering 1910 open reading frames –– 61% coverage of proteins predicted from the DNA sequence. Figure 2 shows a mass versus time display from an LC-FTICR-MS analysis of a D. radiodurans tryptic digest, illustrating the wealth of information obtainable. The ~22 000 detected putative peptides from this analysis are denoted by spots that represent molecular mass (y axis) and spectrum number (or elution time, x axis). The inset shows an expanded view of a small section of the plot and indicates several proteins identified with their corresponding AMT tags.
(72K)
Figure 2. A two-dimensional display from a capillary LC-FTICR-MS analysis of a tryptic digest of D. radiodurans proteins. Detected peptides are shown as `spots' and spot size is varied to indicate relative abundance. The y axis represents the mass of the species detected by FTICR-MS; the x axis is the consecutive spectra obtained during LC elution. A portion of the display is enlarged and annotated with the parent proteins identified by matching several of the eluting species with tryptic peptides present in the AMT tags database (the assigned open reading frame given is followed by a period and tXX giving the placement order of the tryptic peptide with respect to the whole protein). Note that the enlarged view generally shows sets of spots corresponding to a series of spectra for each peptide; the accuracy provided for the mass axis by FTICR (providing the high confidence in identifications) is too high to be communicated in such a display.
The sensitivity of proteomics measurements has been enhanced by nanoscale separation techniques [7. and 8.]. In one demonstration, nanoflow LC-FTICR analysis of 0.5 pg of a D. radiodurans proteome sample yielded confident identifications for the major proteins, indicating sufficient sensitivity for analysis of a single mammalian cell [8.]. Major impediments to such analyses presently lie in sample handling before LC-FTICR analysis. To further increase the throughput of the AMT tag approach, the first automated micro-capillary column LC-FTICR mass spectrometer was recently developed [9.]. This system combined a high field 9.4 tesla superconducting magnet with a commercial FTICR instrument that had a modified high-performance `front end', including tandem quadrupoles for ion manipulations, a dual electrospray ionization (ESI) emitter ion source/ion funnel for better mass calibration and sensitivity, and an automated very high pressure capillary LC system and autosampler for unattended 24/7 operation. With automation, the instrument demonstrated both improved reproducibility and robustness of operation.
Intact protein analysis
Although most proteome studies have conventionally relied on the use of 2D-PAGE, this technique is relatively slow and labor intensive, has significant limitations in sensitivity and dynamic range, and results in significant biases against hydrophobic proteins or proteins having extremes in molecular weight or pI. One alternative to this technique is capillary isoelectric focusing (CIEF) coupled to FTICR-MS, which increases throughput by detecting hundreds of proteins in a single ~1 h run [10., 11. and 12.]. In one example, up to 1000 putative proteins in the 2–100 kDa range were detected from only ~300 ng of an Escherichia coli cell lysate. This approach has also been shown to allow analysis of non-covalent complexes of proteins [12.]. In another study using CIEF-FTICR-MS, proteome-wide quantitation was demonstrated for protein expression in E. coli following Cd2+ stress, using a comparative analysis approach with a stable-isotope-labeled proteome (cells cultured in 13C-, 15N- and 2H-depleted media) [13.]. The abundance ratios determined for the 200 most abundant proteins ranged from <0.1 to 30 with high precision (~10%).
The significance of intact protein analysis for providing information on co-translational and post-translational modifications (PTMs) was demonstrated in a study of yeast ribosomal proteins [14.]. This investigation resulted in the identification of 42 of the 43 core large ribosomal subunit proteins and 58 (out of a possible 64) core large subunit protein isoforms, having unique masses in a single analysis. Also, for a complex protein mixture, using an acid-labile detergent during the electrophoresis step of continuous elution gel electrophoresis prior to LC and FTICR was demonstrated to enhance processing [15.].
`Top-down' protein characterization by FTICR
Typically, an accurate mass measurement for an intact protein, and particularly larger proteins, does not readily enable identification from DNA sequence information owing to sequence errors or PTMs. Once an intact protein has been identified, the location of PTMs generally remain undetermined. With the `top-down' approach, protein ions are dissociated and the resulting fragment masses can be used to obtain structural information [16., 17. and 18.]. For example, McLafferty and co-workers identified and quantitated glycosylation sites in a top-down fashion [19.] and, more recently, characterized secreted proteins from Mycobacterium tuberculosis containing extensive PTMs with this approach [20.]. Furthermore, Kelleher and co-workers [21.] implemented software to automate and increase throughput of such MS/MS analyses.
Dissociation methods
Central to the top-down approach is the need to effectively dissociate protein ions to gain structural information. Electron capture dissociation (ECD), a dissociation method presently unique to FTICR-MS, has attracted significant attention owing to the qualitative differences in dissociation products compared with other methods [22., 23., 24., 25. and 26.] and is reviewed in this issue by Zubarev and colleagues [27.]. In ECD, low energy electrons are captured by positive charge sites of multiply protonated proteins and/or peptide ions resulting in a localized deposition of energy, facilitating, for example, cleavage at a higher fraction of peptide and disulfide bonds. Thus, ECD provides more extensive sequence and PTM site information [22. and 25.], but at the expense of decreased sensitivity (primarily a result of the greater variety of fragments produced). At present, proteins as large as 45 kDa can be effectively studied. Glycoproteins are now also amenable to backbone fragmentation with ECD, while keeping the PTM intact [28.]. Additionally, ECD can be used to probe protein tertiary structures and folding [29. and 30.].
Multiplexing FTICR-MS/MS analyses
The high MMA and resolution of FTICR-MS has motivated the multiplexing of dissociation experiments to increase throughput. Digests of whole-cell lysates have been analyzed by FTICR-MS/MS in a multiplex mode, simultaneously identifying several peptides in each fragment spectrum [31. and 32.]. Figure 3 shows the results from a multiplexing experiment of a global tryptic digest from D. radiodurans, where ~13 000 peptide isotopic distributions were detected. One spectrum from the two-dimensional representation is highlighted to show the simultaneous assignment of four peptides based on their dissociation pattern. The top-down approach for protein identification has also been multiplexed [33.]. The authors demonstrated that four fragment ions are needed with a ±1 Da constraint to achieve >99% probability for identification in a set of 5000 proteins.
(39K)
Figure 3. An illustration of the ability to identify several peptides at one time through a multiplexing experiment in capillary LC-FTICR-MS. (a) A two-dimensional display of a tryptic digest from D. radiodurans. (b) Mass spectrum from the two-dimensional display in (a) at the area shown by the dashed line. The four labeled peaks (A–D) were subsequently isolated and dissociated at the same time to produce (c) the fragment mass spectrum. The box displays the sequence of the four peptides and their associated open reading frame numbers.
Extending the attributes of FTICR-MS
Intensive efforts are being devoted to improving FTICR-MS resolution, sensitivity and MMA. Recently, FTICR-MS was used to distinguish peptide phosphorylation and sulfation with a resolution of >400 000 [34.]. The sensitivity of FTICR-MS has been improved by the introduction of an electrodynamic ion funnel [35.] and multipole devices external to the ion cyclotron resonance cell, to perform ion selection, accumulation and fragmentation [36. and 37.]. Improvements to the accuracy of mass measurements have increased confidence in protein identification [38., 39., 40., 41., 42. and 43.]. To minimize the mass error, several strategies have been developed to allow calibrant ions to be introduced [38. and 39.]. Additionally, our laboratory has proposed a correction for space charge effects (the major component of mass errors in FTICR-MS) that relies on a deconvolution step that compares the masses of different charge states [40.]; we have also introduced a correction method for systematic errors not due to global space charge effects [41.].
Expanding the dynamic range of FTICR-MS
An increased FTICR-MS dynamic range is attractive for better coverage of the broad range of protein abundances, which are often of interest in biological analyses. This increase in dynamic range has been achieved in conjunction with LC separations by using `on-the-fly' selective dipolar excitation so that high abundance species detected in the previous spectrum do not contribute to the next spectrum [2., 44., 45. and 46.]. By allowing the selective accumulation of lower abundance species in complex proteomics mixtures, the overall dynamic range of the analysis is increased. This DREAMS (dynamic range enhancement applied to mass spectrometry) approach is illustrated in Figure 4. The use of DREAMS increased the number of isotopically labeled peptide pairs that were detected by 80% and provided quantitative information [44.]. In a similar fashion, automated gain control implemented with FTICR provides on-the-fly adjustments of ion accumulation time to decrease the amounts of high-abundance species and to increase the amounts of low-abundance species. By optimizing the use of the inherent dynamic range of the FTICR during separations, and by avoiding the accumulation of excessive ion populations that degrade mass measurement accuracy, the number of identified peptides is significantly increased [47. and 48.].
(22K)
Figure 4. The approach for dynamic range expansion in an LC-DREAMS-FTICR MS proteome tryptic digest analysis. DREAMS involves alternating subsequent spectra between (a) the normal and (b) the DREAMS acquisition modes. Before the DREAMS acquisition, a radio frequency (RF) waveform is applied to selectively eject the higher intensity components detected in the previous spectrum, allowing accumulation of lower intensity species to be detected in the next spectrum.
Conclusions
Recent advancements in FTICR-MS instrumentation have improved the characterization of intact proteins and enzymatic digests, increased throughput and improved sensitivity. Quantitation based on labeling with stable isotopes has been shown effective for determining relative abundances at both the peptide and intact protein levels [13.]. The use of nano-flow LC based on very small diameter capillaries to provide higher resolution separations with FTICR has been demonstrated, and the uniformly high ionization efficiencies obtained holds the potential for improved direct quantitation based on peak intensities without labeling [7. and 8.]. The power of FTICR is anticipated to be applied increasingly to intact protein level studies where the AMT tag approach should be particularly effective, providing improved characterization of PTMs. Although the use of intact protein AMT tags and stable-isotope labeling should provide extremely high throughput, sensitive and quantitative measurements, highly heterogeneous and/or very high molecular weight proteins are likely to remain problematic and best characterized at the peptide level.
相关文章：
1 R.D. Smith, G.A. Anderson, M.S. Lipton, L. Pasa-Tolic, Y. Shen, T.P. Conrads, T.D. Veenstra and H.R. Udseth, An accurate mass tag strategy for quantitative and high-throughput proteome measurements. Proteomics 2 (2002), pp. 513–523. | Abstract-MEDLINE
2 M.E. Belov, G.A. Anderson, M.A. Wingerd, H.R. Udseth, K. Tang, D.C. Prior, K.R. Swanson, M.A. Buschbach, E.F. Strittmatter, R.J. Moore and R.D. Smith, An automated high performance liquid chromatography-Fourier transform ion cyclotron resonance mass spectrometer for high-throughput proteomics. J Am Soc Mass Spectrom 15 (2004), pp. 212–232. Full Text + Links | PDF (1001 K)
