Dr Thomas Letzel is Head of Analytical Research at Munich Technical University.

Protein and Peptide Analysis by LC–MS

Experimental Strategies

By Thomas Letzel

The Royal Society of Chemistry

Copyright © 2011 The Royal Society of Chemistry
All rights reserved.
ISBN: 978-1-84973-182-9

Contents

Contributors, xix,
Chapter 1 Top Down and Bottom Up Analysis of Proteins (Focusing on Quantitative Aspects) Friedrich Lottspeich, 1,
Chapter 2 How to Couple and Handle Liquid Chromatography with Mass Spectrometry Thomas Letzel, 11,
Chapter 3 Expression and Purification of Bioactive Proteins/Peptides with Conventional Liquid Chromatography Takayuki Ohnuma and Tamo Fukamizo, 26,
Chapter 4 Liquid Chromatography-Mass Spectrometry of Intact Proteins Nicolas L. Young and Benjamin A. Garcia, 38,
Chapter 5 LC-MS(/MS) of Trypsin-Hydrolysed Proteins Seronei C. Cheison and Ulrich M. Kulozik, 56,
Chapter 6 On-line Protein Digestion in Combination with Chromatographic Separation and Mass Spectrometric Detection S. Johannes Hoos and Wilfried M.A. Niessen, 71,
Chapter 7 Bioinformatic Tools for the LC-MS/MS Analysis of Proteins and Peptides Christian Webhofer and Michael Schrader, 87,
Chapter 8 Quantitative LC-MS of Proteins Gabriele Stöhr and Andreas Tebbe, 104,
Chapter 9 LC-MS for the Identification of Post-Translational Modifications of Proteins Boris Macek, 123,
Chapter 10 LC-MS for the Determination of the Enzymatic Activity of Proteins Romy K. Scheerle and Johanna Graßmann, 133,
Chapter 11 Functional Analysis of Proteins, Including LC-MS and Special Freeware Michael Krappmann and Thomas Letzel, 142,
Chapter 12 Industrial Standards and Strategies in LC-MS Analysis of Proteins Rene Wissiack, 156,
Subject Index, 168,

CHAPTER 1

Top Down and Bottom Up Analysis of Proteins (Focusing on Quantitative Aspects)

FRIEDRICH LOTTSPEICH

Max Planck Institute of Biochemistry, Protein Analysis, Am Klopferspitz 18, 82152 Martinsried, Germany

1.1 Introduction

One key focal point in proteome research is the determination of changes in protein expression and their modifications. In the early years of proteomics the field was dominated by protein chemists and the main approach was 2D-PAGE where differential maps revealed protein pattern differences. The detailed analysis of the different protein spots only became feasible after the introduction of mass spectrometry. However, 2D-PAGE was difficult to reproduce, was not automated and had several limitations with important subsets of proteins (e.g. hydrophobic, very basic, very large or very small proteins). Furthermore, the quantification of the proteins was usually performed by image analysis following several staining methods, which exhibit different signal intensities with different proteins. The dynamic range of detection spans only about 2–3 orders of magnitude, resulting in the visualization of only relatively highly abundant proteins. Additionally, image analysis in principle cannot deal with protein mixtures in a single spot, which, due to the complexity of a proteome, is the common case. Finally, enzymatic cleavage of the protein in the gel matrix suffered from low peptide recovery.

All these limitations encouraged mass spectrometric experts to develop alternative strategies for proteome analyses. Mass spectrometry was used to work with small molecules and therefore it was tempting to cleave the very heterogeneous and unpleasant protein complexity of a proteome enzymatically into small peptides (see also Chapter 5) with much more favourable properties concerning hydrophobicity, diversity and accessibility for multidimensional chromatographic separations and mass spectrometry (see also Chapter 6). Soon mass spectrometry together with informatics and protein databases were developed and optimized to handle complex peptide mixtures, allowing peptide identification by high-throughput MS-MS (see also Chapter 7). The mainstream of proteome research followed this track, called ‘bottom up’ or ‘shotgun’ proteomics (Figure 1.1).

Unfortunately, the bottom up approach also has several severe and fundamental limitations. First, by cleaving the proteome into peptides the complexity increases by a factor of about 40, producing hundreds of thousands of peptides. This is a number that swamps even the most modern mass spectrometers. In consequence, only a fraction of these peptides can be analysed in detail (‘under-sampling’ effect), and it is difficult to assure that identical peptides are analysed from each sample, which is essential to unravel a quantitative fluctuation in the amounts of certain proteins. Second, and even more serious, the context of a protein and the derived peptides is destroyed. A certain peptide may be derived from different proteins or from different forms of a certain gene product, such as post-translational modified, processed or truncated protein species, or from proteins having in common major amino acid sequence stretches like splicing variants or protein isoforms. A single gene will almost always produce an unpredictable multiplicity (tens or even hundreds) of different protein species, which are composed predominantly of identical peptides. Consequently, the quantitative analysis of a peptide monitors only the sum of all proteins that contain this particular peptide. Unfortunately, usually it is not known which protein species are expressed at a certain proteome state. Since in proteome analysis the sequence coverage usually is far less than 50%, the probability of missing the nature of the diversity or the modifications is rather high. In conclusion, the quantity of a peptide determined by a bottom up approach does not necessarily reflect the quantity of a protein of interest. This is completely different in a protein-based, i.e. ‘top down’, approach (see also Chapter 4). The molecular structure and the nature of an intact protein are well defined by molecular properties like the molecular mass and the position in a separation space (isoelectric point, chromatographic position, etc.). If, for example, one compares two proteome states, where a single protein is processed by a cleavage after a lysine, this is easily seen on a protein-based (top down) analysis, like 2D-PAGE or even 1D-PAGE. However, with a bottom up approach, in the two situations all the peptides appear exactly the same and the biological difference will not be detectable (Figure 1.2).

Despite these obvious limitations, bottom up approaches are widely used. The relative technical simplicity and the enormous instrumental development tailor-made towards peptide-based approaches, on the separation as well as on the mass spectrometry side, forced the proteomics field to go mainly with bottom up strategies. However, in recent years the importance of the protein diversity caused by post-translational modifications (PTM; see also Chapter 9), degradations and processing events has become evident to proteome scientists. Still, major technical hurdles have to be vanquished but the awareness of the potential and the advantages of top down proteomic approaches is significantly increasing.

1.2 Quantitative Proteomics

1.2.1 Quantitative Proteomics by Label-Free Techniques

Mass spectrometry per se is not an absolutely quantitative technique. Sequence-dependent peptide ionization efficiencies and suppression of neighbouring signals by dominant peptides results in a low correlation between peptide mass signal intensity and the amount of the peptide (see also Chapter 2). Especially with highly complex mixtures, as commonly achieved in proteomes studies, this inhibits an easy and direct quantification of peptides by signal intensity. However, recently label-free LC-MS quantification methods have been described to determine relative abundances of proteins between multiple conditions. ‘Spectral counting’ methods based on the number of spectra for a certain protein found in a proteome analysis does correlate quite well with the protein amount and thus may provide an estimation on the relative protein amount. With highly abundant proteins the response in spectral increasing protein amount is saturable. With proteins of low abundance the data at low spectral count are noisy and the sensitivity for fold changes decreases.

More accurate is the modification of the spectral count approach by Silva et al. The authors found that a protein’s abundance could be well estimated from the average mass spectrum peak intensity of its three best-detected peptides, assuming that the signal intensity of a fully ionized peptide is roughly dependent on the protein amount.

All label-free proteomics approaches published so far are based on bottom up proteomics strategies and are successful on relatively non-complex samples with almost no sample preparation applied. However, in a more complex situation like human tissues or body fluids multidimensional sample preparation is the key issue to reduce the complexity to a tolerable level. With label-free proteomics a separate analysis has to be performed for each proteome state and no multiplexing is possible; the whole separation and analysis workflow has to be performed for each sample individually. We may expect it to be extremely difficult to reach the required quantitative reproducibility, at least with top down proteomics approaches.

1.2.2 Quantitative Proteomics by Isotopic Labelling Techniques

For many years, metabolic studies have used non-radioactive isotope labelling combined with mass spectrometry as a powerful tool for quantification. Analogues of the metabolites to be tested were synthesized containing 13C, 15N or 2H, and were spiked in defined amounts into the biological sample. As isotopic variants of all molecules behave identically and exhibit the same ionization behaviour during an experiment, quantification by signal intensity of isotopologues is highly accurate. This successful concept was transferred to proteomics. Two or more protein or peptide samples are differentially labelled, one with an isotopically ‘light’ and the others with isotopically ‘heavy’ tags. The samples are then combined, thereby ‘freezing’ the relative amount of proteins or peptides. The complexity of the samples is then reduced by using one or more separation steps. After reduction of complexity and enzymatic cleavage, peptides resulting from corresponding proteins of both samples retain the same chemical properties despite being differentially labelled. A certain peptide from different proteome states can be detected as a mass pair or a mass multiplet (two or more proteome samples) during mass spectrometry, differing only by the masses introduced by the isotopic labels. Corresponding peak heights or areas are then compared to calculate the relative abundance of corresponding peptides of the different samples.

The crucial difference between different labelling strategies is the time point of incorporation of the isotopic label. Prior to the labelling step, samples have to be processed independently in parallel. However, any reaction or handling step not performed under isotopic control may result in quantification errors.

The approaches most often at present are summarized in Figure 1.3. Subsequently their advantages and limitations will be discussed briefly. Additionally, an exemplary isotopic labelling experiment can be found in Chapter 7 with a detailed method description.

1.2.2.1 Introduction of the Isotopic Label at the Level of Living Cells (Metabolic Labelling)

The main advantage of the metabolic labelling strategy is that the label is introduced into living cells by in vivo incorporation of amino acids containing stable isotopes. Therefore, cells from different states, following differential labelling, can be mixed before lysis. Subsequent steps of fractionation and purification do not affect the accuracy of quantification. Consequently, stable isotope labelling by amino acids in cell culture (SILAC) has become one of the most widely used strategies in quantitative proteomics. Two or more cell populations are simply grown in different media, each containing a light version or one or more heavy versions of a suitable amino acid. Several amino acids are described as being used in the SILAC approach. Labelling of arginine and lysine, followed by tryptic digestion, results in labelling of almost every peptide except the C-terminal of each protein. The use of other amino acids such as tyrosine or methionine has also been described. SILAC is mainly used for cell culture-based proteomics approaches. An example of a SILAC experiment with detailed descriptions can be found in Chapter 8.

The in vivo incorporation of stable isotopes has been demonstrated even in animals. One significant limitation of SILAC is that it cannot easily be used for samples which are not grown in culture. Samples obtained from patients (e.g. tissues) can only be analysed by the addition of an artificial internal standard composed of a relevant mixture of cell lines that somehow can resemble the protein content of the actual tissue. Samples from body fluids can only be routinely quantitatively analysed by use of chemically introduced isotopic tags.

1.2.2.2 Introduction of Chemical Tags to Complex Protein Mixtures

When using a chemical-based labelling approach, stable isotope-bearing reagents react with the reactive sites (SH- or amino groups) of a protein. In 1999, Gygi and colleagues introduced this new approach based on chemical labelling using isotope-coded affinity tags (ICAT) directed to cysteine residues. However, cysteine is a rather rare amino acid. Therefore, after enzymatic cleavage only a few peptides carry the isotopic label, i.e. the quantitative information. Thus, with this technique the sequence coverage remained marginal and it is very little used now. A more extended, robust and complete labelling was obtained with the amino group directed ICPL label, which has become the predominant reagent for isotope-labelled top down proteomic approaches, especially since dedicated software (ICPLQuant) has been developed covering the whole workflow and the automated quantitative data analysis. The ICPL method allows for up to fourfold multiplexing and provides highly accurate and reproducible quantification, high protein sequence coverage, including PTMs and isoforms, and is compatible with all commonly used protein and peptide separation techniques. Two or more protein mixtures obtained from different proteomic states are individually reduced and alkylated to denature the proteins and to ensure easier access to free amino groups which are subsequently derivatized with the 12C (light), 2H (medium), 13C (heavy) and 13C2H (ultraheavy) variants of the ICPL reagent. After combining the mixtures, any separation method can be applied to reduce the complexity of the sample on the protein level. Isoelectric focusing by OffGEL or 2D gel electrophoresis may be used as high-resolution separation technologies, where especially protein isoforms can be well distinguished. After significant reduction of complexity the protein fractions are enzymatically digested, preferably using a double enzyme approach. The resulting peptides are quantified by mass spectrometry. Identical peptides derived from the differently labelled protein samples differ in mass and thus can be assigned to the corresponding proteomic state. Each lysine-containing peptide will appear as a multiplet in the acquired MS spectra. The ratios of the ion intensities of these sister peptide multiplets allow for the determination of the relative abundance of their parent proteins in the original samples. After relative quantification only differently regulated proteins have to be identified either by peptide mass fingerprint (PMF) or CID.

1.2.2.3 Introduction of Chemical Tags in Complex Peptide Mixtures

Isotopic labelling methods have also become popular for bottom up strategies, to achieve a more accurate quantification.

Amino group directed non-isobaric reagents such as ICPL are well suited for this purpose. However, the concept of these reagents is based on the relative quantification of stable isotope-labelled peptides prior to MS-MS analysis. Since with bottom up proteomics approaches the complexity is already increased significantly by the enzymatic cleavage of the proteome, a further increase in complexity is caused by the different isotopic derivatives of the proteomics states. As a consequence, many peptides coelute during chromatography of complex biological samples, causing signal suppression in mass spectrometry and making quantitative interpretation and identification difficult. To avoid a further increase in peptide complexity and at the same time allow for higher multiplexing, an approach using isobaric isotope reagents was recently introduced.

1.2.3 Isobaric Tags for Relative and Absolute Quantification

The core of this methodology is a multiplexed set of up to eight isobaric reagents (iTRAQ, AB-Sciex; TMT, ThermoScientific). The labels consist of an N-hydroxysuccinimide moiety, reacting with amino groups, and two isotope- coded regions, a balance group and a reporter group. The latter is released during MS-MS, yielding MS signals at 113–121 Da (ITRAQ) or 126–131 Da (TMT). Corresponding isobaric peptides of all the proteome states coelute during chromatography and are indistinguishable in MS, but exhibit low-mass MS-MS signature ions (reporter ions) that support relative peptide quantification. Since for quantification, MS-MS spectra of each peptide are needed, tens of thousands of MS-MS spectra per analysis are necessary. Furthermore, for a correct quantification no coeluting isobaric peptides should be present, which is hard to achieve without reduction of complexity. Therefore, these approaches show promise mainly with proteomic samples of rather low complexity.

1.2.4 Absolute Quantification in Proteomics with Targeted Proteomics

Most proteomics projects so far have been performed using relative quantification, i.e. only monitoring changes in the level of a large number of proteins. However, for a deeper understanding of biological situations and modelling purposes, as in systems biology, it is often necessary to know the absolute amount of certain proteins. This is in most cases only possible with targeted proteomics techniques. Unfortunately, an internal labelled reference protein standard for the protein(s) of interest is usually not available. To circumvent this, peptides contained in a protein may serve as surrogate markers for the protein itself. Peptides can easily be synthesized with isotopic amino acids or can be reacted with isotopic reagents to serve as internal standards. Therefore, today targeted proteomics approaches are typical bottom up proteomics approaches with all the limitations described above. Thus, multiplexed isotopic labelled peptide-based approaches offer the possibility of performing absolute quantification by using one label for the peptide mixture of defined amounts of synthetic peptides contained in the proteins of interest. Several elegant methods have been proposed for the cost-effective and accurate production of the standard sample.

(Continues…)Excerpted from Protein and Peptide Analysis by LC–MS by Thomas Letzel. Copyright © 2011 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.