SEARCH

SEARCH BY CITATION

Keywords:

  • combinatorial peptide libraries;
  • hexapeptide ligands;
  • low-abundance proteome;
  • mass spectrometry

Abstract

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information

The present review covers modern aspects of combinatorial peptide ligand libraries (CPLL), as used to analyze the “low-abundance proteome” in association with mass spectrometry. First, the capturing properties of baits of different lengths (from single amino acid to hexa-peptides) are described to show that a plateau is rapidly reached above a tetra-peptide in length, thus confirming the validity of having adopted hexapeptides for the considered application. The mechanism of interaction with proteins from very complex proteomes and the ability to decrease the dynamic concentration range is demonstrated with the help of mass spectrometry analysis. Examples are given on how treatment with CPLLs dramatically improves the detectability of peptides in mass spectrometry analysis, permitting detection of a very large number of proteins as compared with control, untreated samples. The use of complementary libraries is discussed with the aim to discover additional low-abundance species that escaped the first library. A discussion on the possibility to discover extremely rare gene products, and the quantitative aspect of the technology when associated with mass spectrometry is also provided. Some insights on the applications for hidden, low-abundance biomarkers are also presented. © 2008 Wiley Periodicals, Inc., Mass Spec Rev 27: 596–608, 2008


PROLOGUE

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information

Having mentioned the FortyNiners in the title, a brief digression on these legendary events is here due. We all associate the saga of the Old West to the caravans of white-topped Conestoga wagons hauled by oxen or horses that crossed the plains and the High Sierra to reach the golden land, California. Wrong. The first wave of FortyNiners was composed by “Argonauts,” who sailed from Boston or New York, and made the journey to the gold fields around Cape Horn—a frightful, 4-month long trip that took its death toll due to dysentery (they crossed twice the equator, where all food would spoil) and shipwrecks in the stormy waters of the Horn. Naively, the Argonauts headed for the Mother Load wearing silk top hats from Boston's revered Collins and Fairbanks and claw hammer coats from Brooks Brothers in New York!

Some brilliant entrepreneur had the idea of cutting the trip to one half, via a mixed ship-land trip. This novel itinerary, via the Isthmus of Panama, well before the cutting of the Channel, of course, constituted the second wave of Argonauts. The Argonauts, rather than making the entire circumnavigation via Cape Horn, would debark at Yankee Chagres on the Caribbean coast, hire native bongos up the Chagres river up to Gorgona or Cruces, and make the reminder of the trip (pompously named the Gold Trail) with the help of porters (estriveros), uncomfortably seated on a plank that the poor Indian had to carry on its back, till they reached Panama City. Here, another ship would take them to San Francisco. This short-cut became so popular that, by 1855, the Panama Railroad was built, between Aspinwall on the Atlantic side and Panama City on the Pacific. This railroad was the first transcontinental railroad in America, antedating as it did the Pacific Railroad across the Western states by 14 years. Some spectacular accidents occurred here too: the first one took place in the manufacturing facilities of Du Pont, in Wilmington, Delaware, where a tremendous explosion of the new explosive, nitroglycerine, destroyed all the facilities and even the countryside. Next year, a steamer at the Spinwall pier in the Isthmus exploded and laid the entire waterfront in ruins.

The caravans of Conestoga wagons that become lodged forever in the lexicon, legend, and consciousness of the American people represented in reality the third wave of the fortyniners, although they took place simultaneously with the first two waves, populated especially by those families who could not afford the cost of the other two transfers. These caravans too were besieged by troublesome events, such as the attacks from the plain Indians, as well as by the perilous crossing, all throughout the winter, of the High Sierras. The fact that the crossing of the Sierras was practically impossible for about half of the year, spurred a fourth attempt, driven by the army. This scheme was concocted by Jefferson Davis, head of the War Cabinet in Washington. The idea was to trace an itinerary in the South of the States, along the border with Mexico, a path that would be open all the year around. The Argonauts would have sailed to the tip of Florida, entered the Gulf of Mexico, debarked at the Rio Grande, taken a float of small boats to the Big Bend, and from there proceed by land, via New Mexico, Arizona, and the extreme south of California, to end in San Diego. This itinerary would have had the advantage of permitting colonization of the Southern lands, up to then scarcely populated. Additionally, it would be open all the year around, because no High Sierras had to be crossed. There was a big problem, though: the lack of water along the crossing, because the land was criss-crossed by deserts. Here Yankee ingenuity came to help: Davis bought twenty dromedaries from the khedive of Egypt, and in the autumn of 1858 sent lieutenant Miles in Texas, at the Pinery Station at the foot of the Guadalupe peak, to try to open a path and map the territory. What happened is shrouded in mystery: there are rumors that Miles moved at a brisk pace at least up to the Chiricahua Pass and met, and befriended, bands of Apache warriors. Some would say that he managed to cross the Sonora desert and the Death Valley, to meet the Pacific in San Diego. But nothing came out of it: Civil War, meanwhile, had been declared and the operation was dismantled. The same unverified rumors have it that the dromedaries ran away from the corral at the Pinery station and were last seen pasturing on the Chirichaua Mountains. Perhaps one day some paleontologist will find their bones and will derive a new theory about autochthonous American dromedaries, of solid capitalistic faith.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information

Thirty years have elapsed since our first, timid, and primitive attempt to perform two-dimensional (2D) maps of zeins (storage proteins in corn) (Righetti et al., 1977). Primitive as it was, our technique helped us to discover a polydispersity of zeins that hinted at the presence of a least a dozen different genes; however, our data were met with scorn, and were rejected by all dedicated journals of the epoch, who still lived with the stern credo of one gene—one zein (today, it is known that there are at least 20 genes coding for zeins). Since then, 2D maps have made a spectacular progress; first, with the work of O'Farrell (1975), and then with the introduction of immobilized pH gradients (IPG) (Bjellqvist et al., 1982) and of non-linear IPGs (Gianazza et al., 1985), amply used at present as the first dimension in 2D mapping. Today, the growth in proteome analysis has been striking, thanks to a three-pronged attack: separation methods (one should recall here also 2D chromatography, pompously called “multidimensional” at its inception) (Washburn, Wolters, & Yates, 2001), the spectacular progress in mass spectrometry (MS) (Steen & Mann 2004; Cornett et al., 2007; Köcher & Superti-Furga, 2007; Nesvizhskii, Vitek, & Aebersold, 2007; Wietze et al., 2007), and in bioinformatics (Hernandez, Binz, & Wilkins, 2007; Lisacek et al., 2007), that help to sift through the constant and huge outpouring of data.

Against this rosy picture, we admit that there are still quite a few aspects that elude us, such as (Wilkins & Appel, 2007):

  • a
    The separation and detection of all proteins in any given proteome remains a challenge. Low-abundance proteins continue to be elusive;
  • b
    De novo sequencing of proteins perdures being difficult;
  • c
    We cannot monitor changes in proteomes in real time;
  • d
    Proteomics is presently still a semi-quantitative, not a quantitative analysis;
  • e
    Our capacity to generate large volumes of proteomic data remains limited;
  • f
    Use of advanced statistics for experimental design and data analysis is still limited to a few laboratories;
  • g
    Extraction of knowledge from the large volume of data produced by analytical proteomics experiments such as Mass Spectrometry (MS) still represents a major bottleneck.

The present review could not possibly cover all these aspects, but will surely deal with the first point, namely a method suitable to detect the low-abundance proteome. In order to have access to the “deep proteome”, a number of pre-fractionation tools have been described (Herbert et al., 2007). Such tools comprise not only all possible chromatographic and electrophoretic methods (especially those based on isoelectric focusing), but also organelle pre-fractionation (Huber, Pfaller, & Vietor, 2003) and immuno-subtraction of most abundant species via antibody columns (up to 12 different antibodies mixed in a single column) (Echan et al., 2005). Other tools much in vogue today regard the capture of classes of proteomes based on their structure or function, such as the specific seizure of the glyco-proteome via lectin columns (Hirabayashi, 2004) or the capture of the phospho-proteome via metal chelator resins, affinity for titanium oxide (Larsen et al., 2005), or via anti-phosphoamino acid antibodies (Rush et al., 2005). Other techniques focus on charting protein–protein interactions, via two main methods, the yeast two-hybrid system, which maps binary or pair-wise associations (Li et al., 2004), or affinity capturing methods, coupled to MS identification, designed for characterization of protein complexes (Gavin, 2007).

Quite recently, the approach for the detection by two-dimensional electrophoresis (2D) and MS of rare proteins has taken a completely different turn: use of combinatorial ligand libraries to capture and “amplify” the low-abundance proteome (Righetti et al., 2006; Righetti & Boschetti, 2007; Boschetti et al., 2007a; Boschetti, Monsarrat, & Righetti, 2007), which constitutes ca. 50% of any proteome. The idea of libraries of millions of peptides, produced on the basis of a “one-bead, one-peptide” approach, dates back, as early as 1991, to a seminal paper by Lam et al. (1991). Its application to proteomics, though, is relatively recent, because the first report appeared only 3 years ago (Thulasiraman et al., 2005). A flurry of applications soon followed; for example, in urine (Castagna et al., 2005), serum (Guerrier et al., 2006; Sennels et al., 2007), human platelets (Guerrier et al., 2007a), red blood cell (Roux-Dalvai et al., 2008), bile (Guerrier et al., 2007b), and recombinant DNA product (Fortis et al., 2006; Antonioli et al., 2007) analyses. This technology is presently commercially available under the trade name of ProteoMiner.

In the present review, we will first deal with the physico-chemical properties of such ligand libraries that had been investigated only partially up to the present. We will underline the importance of the peptide length, and finally analyze the contribution of individual amino acids in capturing protein populations within a proteome with the objective of amplifying the signal of low-abundance species. How the ProteoMiner technology is synergistic with MS analysis and markedly improves the detection of proteins and peptides is also extensively highlighted below.

The General Behavior of Hexapeptide Libraries

The combinatorial peptide ligand library is a mixture of porous beads on which hexapeptides are chemically attached. Each bead carries a large number (billions) of copies of the same peptide bait; the beads are thus different from each other, and all combinations of hexapeptides are present. Depending on the number of amino acids used, a library contains a population of millions of different ligands (e.g., 11, 24, or 64 millions starting respectively from 15, 17, or 20 different amino acids). When a complex protein extract is exposed to such a ligand library in large overloading conditions, each bead with affinity to an abundant protein will rapidly become saturated, and the vast majority of the same protein will remain unbound. In contrast, trace proteins will not saturate the corresponding partner beads, but are captured in progressively increasing amounts as the beads are loaded with additional protein extract. Thus, a solid-phase ligand library enriches for trace proteins, while concomitantly reducing the relative concentration of abundant species. In theory each bead that carries a single peptide ligand should interact with proteins that share the same epitope complementary to the peptide bait. However, because in a number of cases peptide ligands can differ from each other by just one amino acid, similar interactions can be found with more than one protein—the single interactions are governed in this case by different dissociation constants. As experimentally demonstrated, in fact, the content of one single bead (single peptide structure) after contact with a crude protein extract, is constituted of few proteins—some dominant and other minor species (Boschetti et al., 2007a). Generally, the affinity between a hexapeptide ligand for a given protein might be considered as a rather weak binding event; however, experience with millions of hexapeptides has demonstrated that indeed such complexes can be very stable, sometimes requiring harsh conditions for the dissociation, such as 6 M guanidine hydrochloride.

Due to the complex involvement of different concurrent synergistic and/or antagonistic physico-chemical interactions, the capture of proteins is sensitive to changes of environmental conditions such as pH, ionic strength, and temperature. Displacement effects take also place as a consequence of the balance between the affinity and the concentration of proteins for a same ligand. As the amount of protein loading increases for a given volume of CPLL, the number of detectable species increases, thus delving deeper into the proteome. This fact, associated to the use of efficient and sensitive mass spectrometry equipment, allows one to approach the full composition of proteins of simple living cells.

In spite of the possibility of detecting proteins that are normally undetectable, there are species that are not visible any longer after the treatment of protein extracts with CPLL. As described, about 7% of abundant urine proteins (Castagna et al., 2005), about 5% of red blood cell proteins, and even 13% of platelets' extracted proteins, well detected before the treatment, are not detectable any longer in the eluted proteins from the beads. This situation suggests that either some proteins do not find a peptide partner to form a stable complex, or the protein-peptide complex is governed by a strong association constant preventing efficient recovery with current eluting agents.

Behavior of Oligopeptides of Variable Length, From Single Amino Acid to Hexa-Peptides

Although, up to the present, the preferential enrichment of low-abundance species by the use of peptide libraries was accepted as an act of faith, there are at least two important aspects that had not been investigated so far: (a) whether the length of a hexapeptide bait is optimal; (b) what are the fundamental rules that govern the capture phenomenon and the binding properties of these ligands. Such vital information would help to improve the quality of the commercial products, on the one side, and to optimize its use at the bench, on the other side. The first aspect has been recently probed by Simó et al. (2008). These authors have prepared combinatorial peptide beads that contain baits of variable lengths, from di- to hexa-peptides. In addition, 16 types of beads were made, each containing an individual amino acid, pooled, and used also as a capturing device. As a model, a total red blood cell (RBC) lysate was used, due to the fact that, in a previous work, Roux-Dalvai et al. (2007, 2008) had mapped this proteome, with combinatorial hexapeptides, to a very deep extent. Additionally, the RBC cytoplasmic proteome represented a unique challenge to explore and detect low-abundance species, considering that a single protein (hemoglobin, Hb) represents ca. 98% of its global assets. The results were striking and quite unexpected. To start with, the column that contained the pooled 16 different amino acids was able to capture a non-negligible portion of the RBC proteome that represented in fact ca. 50% of what was detectable (393 proteins vs. a total of 799). Secondly, as shown in Figure 1A, the number of species captured increased very rapidly with a di-peptide column, and the capture curve (as shown by the total number of spots counted in each 2D map) tended to plateau above a length of a tetra-mer. The 2D gel spot count trends were confirmed by the count trend of proteins identified by MS analysis from each eluate (Fig. 1B), by using nanoLC-MS/MS analysis. These facts suggest that not much was to be gained by increasing the length of the baits above a hexapeptide (this was about 10% of additional proteins). However, those data justified the adoption of this length for bait peptides because most novel species found were of relatively low molecular mass, a category quite difficult to amplify with short peptide chains. Whether longer baits (7–8 amino acids long and more) will be worth pursuing remains to be seen, due not only to the extra cost and synthesis complications, but also to the fact that the unavoidable stronger interactions obtained will render difficult the dissociation of proteins. In addition, progressively longer peptide chains (e.g., 10 amino acids or longer), might present structural problems due to formation of helices and/or collapsing of the chain onto itself in the case of purely hydrophobic amino acid stretches. Thus, it might very well turn out that longer baits would capture less, or at least not more, species. Outside the considerations of peptide length, the composition and the structure may play a role of even larger importance, a fact that is to date not yet reported.

thumbnail image

Figure 1. Number of proteins captured as a function of the peptide length. A: Plot of the number of spots counted in each relevant 2D map versus the length of the peptides bound to the beads, calculated for large 2D gels (18 × 20 cm), silver staining. B: Plot of the number of gene product identified by nanoLC-MS/MS as a function of peptide length (unpublished experiments with C. Simò).

Download figure to PowerPoint

Figure 2 shows another important finding: when comparing the “catch” of the 16 pooled amino acids versus that of the entire hexapeptide library, it can be appreciated that, even if both capture substantially the same number of species (a total of 393 vs. 396 respectively), the catch is particularly richer in the 10–40 kDa region for the hexapeptide baits. This finding suggests that interaction with proteins already begins with a single amino acid, but bio-specificity requires baits of a proper length, at least above four amino acids. This statement is also reinforced by the following observation: each column (starting already with the single amino acids) can strongly modify the relative protein composition of the loaded RBC lysate, by a probable displacement effect and hence dramatically reduce the level of Hb in the recovered eluates. If the capture mechanism were of a pure ion-exchange interaction, given the nature of the baits, then Hb, by far the dominant species, would have quickly saturated all available binding sites and displaced most of the other proteins in the RBC lysate.

thumbnail image

Figure 2. Mr distribution (in 10 kDa groups up to 70 kDa) of the proteome of the RBC lysate, as captured by the 1-mer and 6-mer library beads. 1-mer: 393 total species; Ctrl library: 396 total proteins. The differential analysis shows that species of low mass are more numerous in Ctrl than in 1-mer column compared to other mass intervals (unpublished experiments with C. Simò and A. Bachi).

Download figure to PowerPoint

From Multimer Libraries to the Individual Amino Acids

The unique behavior of the pooled amino acid column, as well as the influence of the terminal amino acid of the peptide sequence, as discussed above, spurred Bachi et al. (2008) to investigate the action and properties of the 16 amino acids, taken individually. In this respect, their findings have been quite unique and not really predictable from the early work of Porath's group (Porath & Fryklund, 1970; Porath & Fornsted, 1970), who first reported the use of affinity columns laden with individual amino acids (just β-Ala and Arg) for the capture of proteins. Among the described findings, a dichotomy appeared between a class of “Grand Catchers” (Arg, His, Ile, Lys, Phe, Trp, Tyr, and Val), all able to bind from 300 up to 450 unique gene products with large overlaps (data not shown), and the “Petite Catchers” (Asp, Asn, Gln, Glu, Gly, Pro, Ser, and Thr), which bind in general half as much, with the notable exception of Glu, which seems to be unable to bind red blood cell proteins. If the individual eluates are pooled and compared to a library of mixed amino acids, then the 2D maps are practically identical, as expected (see Fig. 3, panels A and B). However, each individual amino acid has an unambiguous role in capturing exclusively some proteins (Fig. 3, panel C, numbers located at the extremities of the scheme). By confronting homogeneous classes of proteins (e.g., the basic, the hydrophobic aromatic, neutral hydrophilic, etc.) it was found that, in general, half as many proteins were in common among the members of each family. In a 16-way comparison, only 72 proteins (less than 10% of the total catch, amounting to 800 unique, non-redundant proteins) appear to be the common catch of all 16 amino acids (see Fig. 3, panel C). These data suggest that such proteins might have general structural motifs recognized by any generic bait.

thumbnail image

Figure 3. Proteins from RBC lysate captured by single amino acids attached on beads. A,B are 2D gel electrophoresis of respectively the pool of all single amino acids eluates and the eluate from a column containing an equivalent mixture of all amino acid beads mixed together (called also 1-mer library). First dimension: linear IPG pH 3–10; second dimension: 8–18%T polyacrylamide gel. Staining with colloidal Coomassie Blue. Protein loading for each gel was 80 µg. The number of spots counted was similar. C: represents the gene product count after nanoLC-MS/MS of each amino acid eluate. At the periphery of Venn diagram (petal extremity) are the numbers of exclusive proteins found from single eluates; at the center of the graph is the number of species that are common to all eluates (adapted from Bachi et al., 2008; Simó et al., 2008).

Download figure to PowerPoint

By far, it would appear that the strongest interactions, and thus the strongest catch, occur with the three aromatic moieties of Phe, Tyr and Trp, all able to capture a practically identical number of proteins. Ionic interactions, which in principle should be the strongest ones, appear to behave in a peculiar way: they are quite strong with the three basic amino acids (Arg, Lys, and His) but rather weak with their acidic counterparts (Glu, Asp). Based on these facts, Bachi et al. (2008) suggested a peculiar mechanism of interaction: upon formation of the ion-pair, the complex between the protein and the bait is stabilized by the hydrophobicity of the side chain (a butyl in the case of Lys). If that is true, then this factor would explain why Glu and Asp exhibit such a poor binding ability. This notion is also reinforced by the fact that generally the best eluting agent from the ligand libraries is a solution of 7 M urea, 2 M thiourea, and 2% CHAPS, a cocktail well-known to break hydrophobic (as well as hydrogen-bonding) interactions. The amino acids taken individually and even as a pool (see the previous paragraph) are, however, not ideal for the described application because a large number of proteins present in the initial extract are ignored. However, it appeared quite clear that they unexpectedly induced an unambiguous effect on the increase of signal of a number of undetectable proteins. With more in-depth investigations, they might also teach how to create customized libraries to address how to catch specific groups of proteins.

The Complementary Effect of Modified Libraries

Current hexapeptide libraries carrying the primary amine at the distal side of the chain might not comprise an exhaustive number of baits and structure variations for all proteins present in the sample. In a number of protein extracts, we described the absence of a certain number of proteins after treatment that were found in the initial sample by mass spectrometry. In this context, it appeared that the sequential use of a carboxylated library captured additional species that escaped from the first, amino-terminus, library. This result was obtained from platelet extracts (Guerrier et al., 2007a), from human bile (Guerrier et al., 2007b), as well as from red blood cell lysates (Roux-Dalvai et al., 2008). As an average, the additional proteins caught by the second library represented between 20% and 30% of the total number. At the same time, a large redundancy was evidenced between the first and the second library. The properties of proteins captured by either the first or the second library did not change significantly in terms of molecular-mass interval or hydrophobicity-index dispersion, except for the isoelectric point. In spite of a relatively large overlap, the primary amino terminal libraries captured more acidic proteins than the carboxyl terminal libraries. This result was shown in spite of the presence of physiological ionic strength buffer, where pure ion-exchange effects should be minimized with the peptide beads. It has been anticipated that the difference should be attributed to the modulation of the terminal chemical group of the peptide chain. This finding was not completely a surprise when considering that the terminal and the more exposed amino acid plays an important role in the capture of proteins, as demonstrated by Simó et al. (2008). Figure 4, extracted from various experiments, shows the importance of a complementary library at least when mass spectrometry is used for the identification of the entire collection of species from various proteomes.

thumbnail image

Figure 4. Overlapping diagrams of proteins found with two different libraries and for four different samples. At the left of each diagram are data from amino-terminal hexapeptide library and on the right are numbers from the terminal carboxylated form of the same hexapeptides. It is noticeable that libraries are most of the time complementary, increasing thus the ability of capturing more proteins from crude extracts. A: Crude extract from human platelets; (B) delipidated human bile; (C) proteins from human red blood cell lysate; (D): redox—involved proteins from an extract of S. cerevisiae (unpublished data).

Download figure to PowerPoint

The Effects of Ligand Libraries on Mass Spectrometry Analysis

It is of great interest, from a point of view of MS analysis, to see how a proteomic eluate from a ligand library behaves as compared with the untreated control sample. From data obtained by SDS–PAGE and 2D maps, all proteomes investigated appeared as much more complex than the control. This difference is due to the amplification of signal of low-abundance proteins that are generally undetectable. For an optimal detection and thus identification of species present after treatment, it is thus advisable to fractionate as much as possible. This fractionation can be performed at two different levels: first, during the protein desorption from the CPLL, and secondly, by slicing out more fractions from the SDS–PAGE prior to trypsin digestion. In addition, the recovered proteins after CPLL treatment can be fractionated by more classical methods such as liquid chromatography or isoelectric separations. Each time the CPLL was applied to protein extracts, the number of proteins detected was significantly larger than in the control, and spanned from a factor of two to a factor of five. Referring to serum proteins, Sennels et al. (2007) described an increase of protein spots in 2D map analysis from 115 to 790; for platelet extracts, the number of gene products identified by LC-MS/MS jumped from 197 before treatment to 435. As a last example, relatively old though, in human urines the number of proteins found by FT-ICR-MS before treatment was 134 and 385 after treatment with CPLL (Castagna et al., 2005).

In the case of the RBC cytoplasmic proteome, a very large amount of lysate was loaded on 1 mL of CPLL (5.73 g total protein). Due to this massive loading, it was expected to get a very large number of proteins to justify a fractionated elution and even a double library placed in series. From about 5,730 mg of initial loaded proteins, the six collected fractions taken together amounted to only 8.14 mg, representing 0.14% of the initial charge. Elution from both libraries was performed at first by a thiourea-urea-CHAPS solution, followed by an acidic urea treatment; finally, the beads were stripped with a hydro-organic alkaline elution solution. These fractions mixed together and compared to the initial RBC extract showed a very large difference in 2D gel analysis, where the number of protein spots jumped from 81 before treatment to more than 900 after treatment (see Fig. 5).

thumbnail image

Figure 5. 2D analysis of proteins from a lysate of pure human red blood cells. The initial extract is represented on the left and shows about 80 spots on a load of 1,300 µg of total proteins. The mixture of all eluates from the CPLL column is shown on the right with a number of counted spots of about 950 on a load of 640 µg of total proteins. First dimension: linear IPG pH 3–10; second dimension: 8–18%T polyacrylamide gel. Mr = mass markers. Gels are stained with silver (unpublished experiments with C. Simò).

Download figure to PowerPoint

Each fraction separated by SDS–PAGE was sliced into 20 parts, and each slice was treated with trypsin prior nanoLC-MS/MS analysis; a large number of proteins were identified. When analyzing MS data in a comparative manner, the number of peptides found for the same protein before and after the treatment with ProteoMiner, was much reduced for the high-abundance proteins and much increased for the low-abundance species (Roux-Dalvai et al., 2007, 2008) demonstrating thus the mechanism of reduction of the concentration of one category of proteins while increasing the concentration of another.

It is moreover observed that, after sample treatment with CPLL, the most numerous new MS signals are found within the zone of low masses between 1 and 10 kDa (see examples in Fig. 6). Beyond this simple observation, it may be of interest to exploit this peculiar property to discover new peptidomics species within biomarker discovery investigations. It is known, in fact, that fragments of proteins are produced in some pathological situations (Tammen et al., 2007). The activation of general and specific proteases, a phenomenon that frequently happens in pathological situations, ends up degrading proteins and hence producing signatures of diagnostic interest. Although this idea is still a speculation, the use of CPLL might open a new road in the discovery of specific situations.

thumbnail image

Figure 6. Mass spectrometry analysis using SELDI MS and ProteinChip arrays of different biological samples before (left column) and after treatment with a hexapeptide library (right column). The spectra window focus on relatively low masses. From (AD) are respectively rat serum proteins (acidic urea eluate and CM10 array); human cerebrospinal fluid (acidic urea eluate and CM10 array); human saliva proteins (acidic urea eluate and CM10 array); human urine proteins (thiourea-urea-CHAPS eluate and IMAC-Cu++ array). m/z = mass over charge.

Download figure to PowerPoint

The use of ligand libraries amplifies the signal of low-abundance proteins and has another important effect on mass spectrometry analysis. This effect is when a new MS peak of interest is found as a result of an expression difference. This difference is what is generally named “biomarker.” Peptide libraries would be used in order to increase first the concentration of such a target species before its purification or enrichment, prior to formal identification by either peptide mass fingerprinting or by peptide sequencing. This approach has been successfully demonstrated by Guerrier and Boschetti (2007) for the identification of a signal at 12.4 kDa (after analyzing serum samples with ProteoMiner), which was identified as a truncated form of prothrombin precursor activation peptide fragment 2. Such an approach will likely be generalized as new protein signals of fundamental or diagnostic interest are detected by mass spectrometry.

Associating CPLL and mass spectrometry has a tremendous interest also in the determination of trace impurities on natural and recombinant DNA biopharmaceuticals. Fortis et al. (2006) and Antonioli et al. (2007) demonstrated that the detection of impurities is far more sensitive after amplification of very low-abundance species (the impurities) than by a direct analysis by any type of analytical approach. For these applications CPLL, 2D, and MS were used synergistically because of (i) the large amplification of impurity signals (2D and MS), (ii) easy formal identification of evidenced species (MS), and (iii) unambiguous differentiation between an entire entity and its fragments.

Jumping Genes and Stakhanovite Genes

The proper association of peptide ligand libraries and mass spectrometry represents a tandem with multiple effects. One, of utmost importance, is the capability to approach the very deep proteome with the possibility to reach the limits of the protein composition of a cell; in other words, the deconvolution of protein compositions that are at the basis of a survival. This goal appears as reachable—at least for simple cells—not only from prokaryotes but also from eucaryotes, such as for instance red blood cells. In a recent presentation, Roux-Dalvai et al. (2008) showed that the total count of gene products, that represents the 2% minority RBC cytoplasmic proteome, reached 1,578 unique proteins; U: that is a truly outstanding count, considering that a RBC is a much simplified cell (no nucleus, no other organelles) than any other human cell lines, believed to comprise a genetic asset of >12,000 unique gene products (a mere estimate, because at present, no study has been able to explore to full proteome of any human cell). What is the meaning of such a large body of proteins in the RBC cytoplasm? Although these findings are yet to be published, that catalog of proteins goes well beyond the belief that RBC cell function could simply be optimized around oxygen and carbon dioxide as well as proton transport, plus some additional, very important enzymatic functions with pathways involved in particular disorders. In this respect, we should mention glycolysis and nucleotide metabolism disorders, collectively called chronic (or hereditary) non-spherocytic hemolytic anemias (Eber, 2003; Vulliamy & Luzzatto, 2003) or defects that involve glucose-6-phosphate dehydrogenase (G6PD) deficiency, pyruvate kinase, glucose phosphate isomerase, pyrimidine 5′-nucleotidase, triosophosphate isomerase, and phosphofructokinase (PFK) deficiencies.

Due to the large number of proteins identified (1,578 vs. a few hundred up to this investigation) (Kakhniashvili, Bulla, & Goodman, 2004; Pasini et al., 2006; Bhattacharya, Mukhopadhyay & Chakrabarti, 2007), it is questionable whether all of them are genuine components of mature RBCs, or if some are incompletely degraded products from the reticulocytes. Although the degradation process is initiated in reticulocytes, it is perpetuated up to the transformation process in RBCs. Another possibility is that many non-essential proteins are not necessarily fully degraded inside RBC, because this step would cost energy that might be better used to accomplish fundamental functions of RBCs. All of these possibilities are speculations that could be elucidated by deciphering certain pathways that are approachable only by preferentially enriching low-abundance proteins.

Notwithstanding the uncertainties on the roles of all the proteins detected for the first time, some unique insight emerges. It is worth recalling here the story of Barbara McClintock and her “jumping genes.” McClintock showed that genes could transpose within chromosomes; that they could move around (the so-called “jumping genes”). This discovery was obtained through the investigation of maize (corn) genetics via careful hybridization. However, when, in 1948, she published her first paper that showed that the chromosome-breaking locus did something hitherto unknown for any genetic locus (i.e., it moved from one chromosomal location to another, a phenomenon she called transposition), her results were met with scorn by the scientific community. In 1983, 35 years after publication of the first evidence for transposition, McClintock was awarded the Nobel Prize for Physiology or Medicine, precisely “for her discovery of mobile genetic elements.”

An illustration of a related fact is summarized in Table 1, where some unique results are apparent from the RBC study. After ProteoMiner treatment, no fewer than eight different globin chains could be detected. The first two are well-known components of adult Hb, which is a tetramer composed of two α- and two β-chains. The two following ones (γ- and ε-globins) are known as fetal chains, whose genes were believed to be silenced a few months after birth. Our data demonstrated that, in reality, they continue to produce globin chains, at a level well below 1% as compared to the α- and β-chains, all along the entire life span. Nobody, however, had ever reported or even suspected that also the other four globin chains (ζ, θ, δ, µ, called embryonic chains) would survive in adult life, because then genes were thought to be silenced already in the switch from embryo to a developed foetus. These findings pose the intriguing question: if genes are truly silenced or are kept in a “state of alert” so as to be ready for any emergency, should the need arise. We thus move from the “jumping genes” of McClintock to what we here have dubbed “Stakhanovite genes” that, like the Russian working hero, never quite stop working. It would be nice to see if this hypothesis would be verified by ProteoMiner treatment of other cell lysates, in search of the expression of “silent genes.”

Table 1. Stakhanovite genes in an RBC lysateThumbnail image of

As a conclusion of the above, when merging sample treatment via peptide ligand libraries, such as ProteoMiner, with mass spectrometry, one could expect outstanding discoveries that would not be possible by using current tools. The combination of these two technologies, associated with other investigation means, such as group separation of treated samples, isoelectric-based separations or even just liquid chromatography, should give rise to even more complete information not only in regard to the protein composition of a cell, but even more importantly in regard to molecular interactions and finally to unknown metabolic pathways.

The ProteoMiner and Differential Proteomics: Myth or Reality?

The aspect of differential proteomics is a very important issue that has not been fully addressed up to present time. If the reduction of dynamic concentration range with peptides libraries was to produce a full leveling of all protein concentrations, no differential proteomic studies could be performed to evidence up- and down-regulated protein expressions (although the library would have still utility for the amplification of protein fragment traces as well as truncated forms). In this last scenario, the concentration of individual proteins would be so much modified that any quantitative data would be meaningless, particularly for differential analysis. Differential analysis is, in fact, a fundamental technique in the field of proteomics, because it would allow one to discover biomarkers of disease and to find out important proteins that would be up- or down-regulated (or switched on or silenced) in a host of pathologies, in drug treatment, and the like (Hamdan & Righetti, 2005).

A mode of making a differential analysis of biological samples treated with CPLL can be performed using 2D gel analysis. Although not quantitative, it allows one to show differences in the number of spots between the control and the pathological sample before and after CPLL treatment. Naturally, many more spots are visible after treatment, but what is important in this area is the difference in the number and positioning of spots. Figure 7 illustrates, under a simplified manner, the differential analysis of an acute coronary syndrome case. As derived from 2D analysis, it is anticipated that DIGE technology would benefit from CPLL-treated samples as mentioned by Davidsson in late 2007 (personal communication) for the research of biomarkers related to plaque rupture in the cardiovascular area. The data suggested that CPLL, associated with various technologies including DIGE, enhanced the signals of new peptide/protein biomarker candidates.

thumbnail image

Figure 7. Expression difference two-dimensional maps before (native) and after treatment with hexapeptide library (CPLL-treated). Samples are normal human serum (Ctrl) and Acute Coronary Syndrome (ACS). It is noticeable that a large number of protein spots of low molecular mass (few thousand Da) is evidenced in ACS serum and particularly after amplification of the signal of low abundance proteins by CPLL. First dimension: linear IPG pH 3–10; second dimension: 8–18%T polyacrylamide gel. Staining with colloidal Coomassie Blue. ACS serum by courtesy of Dr O. Meilhac, INSERM U-698, Paris, France; 2D gel analysis from Dr S. Freeby, Bio-Rad lab., Hercules, CA.

Download figure to PowerPoint

The hexapeptide ligand library was involved in a very early study performed within the frame of ovarian cancer pathology with promising discoveries. Subtle differences have been demonstrated with ovarian cancer sera. When the serum was treated with a hexapeptide ligand library, differences in signals were evidenced. Other differences have also been shown after fractionating the CPLL-treated serum by anion exchange chromatography. Experimental data conducted with several dozen of samples and recently presented (Vergote, van der Zee, & Fung, 2005), showed the presence of few new biomarkers of low abundance as possible candidates related to this kind of pathology. Several of them seem to correlate with surgery; several others seem to correlate with the outcome. Figure 8 is an extract of this study, where the separation of long survivors versus short survivors is shown, based on the use of two biomarkers discovered with this novel technology.

thumbnail image

Figure 8. Results of research of markers of diagnostic interest by means of CPLL. Initial biological sample was human serum of patients with ovarian cancer. Separation of two populations from ovarian cancer (long and short survivors respectively ‘l’ and “s”) determined by analyzing data from two selected protein biomarkers (2,498 and 11,183 Da). Courtesy of Dr Eric Fung (Ciphergen Biosystems, Inc., Fremont, CA, USA).

Download figure to PowerPoint

Another important question correlated with biomarker discovery is the quantitation aspect when a hexapeptide ligand library is used to treat samples. When adding the question of quantitation to differential expression, one might consider two distinct aspects: the intrinsic value of quantitation due to (i) mass spectrometry analysis performance, and (ii) to the intervention of peptide ligand library. The former question has been addressed a number of times and cannot have other utility than as a differential approach. In this domain, quantitative methods are generally based on stable isotope-labeling and ended up to essentially three conceptual approaches that involved MS-based strategies: biological or metabolic incorporations (SILAC), enzymatic incorporations, and chemical tag incorporations with 2 or more labels. Although still very recent, these methods can be used in conjunction with various separation techniques like mono-, two-dimensional electrophoresis and liquid chromatography. This concept applies at least to differential expression meant to discover biomarkers related to specific states or diseases. Having stated that, the question is around the applicability of treated samples with hexapeptide libraries as a mean to compress the difference in protein concentration. As discussed above, high-abundance proteins reach the saturation of bead-binding capacity during the capture stage, and are impossible to quantify because they attain a certain value at saturation that does not move, even if the amount of sample loaded on the beads is changed. Nevertheless, for all proteins that have not reached the saturation, the phenomenon of compression in dynamic range still continues. For very low-abundance proteins, it is probable that the saturation may never be reached due to the thermodynamics of the system (balance between the concentration of given species and their dissociation constant). In a recent attempt, Paulus et al. (2008) added serum amyloid A protein (SAA) to serum and plasma samples to demonstrate the linearity of quantitation in response to increasing amounts of such low-abundance proteins. The spiking covered the following interval: 1, 5, 10, 20, 40, 80, and 160 ng/µL SAA. The recovery and linearity of the signal of SAA was assessed with 2-D gel electrophoresis and western blot analysis. The authors used surface-enhanced laser desorption/ionization (SELDI) mass spectrometry to confirm western blot data, and concluded that the ProteoMiner treatment did indeed preserve the proportionality and hence the ability to quantitate low- and medium-abundance proteins in serum and plasma as suitable for biomarker discovery.

In another series of trials, Roux-Dalvai et al. (2008) described the addition of 100, 300 or 1000 pmol of yeast alcohol dehydrogenase (ADH) to 680 mg of red blood cell lysate, prior to processing with a 0.5 mL of a CPLL column. After extensive washing, the captured proteins were eluted with a 9 M urea solution, acidified with citric acid to pH 3.3. Three replicate treatments were performed for each ADH amount. The same amount of protein from the nine eluates was loaded on a 12% SDS–PAGE gel, and a slice was cut in each lane at the ADH molecular mass. After tryptic digestion, the sample was analyzed by nanoLC-MS/MS. The results demonstrated a linear increase of signal as the amount of protein spiked in the lysate increased, as shown either on an individual parent ion assigned to ADH, or on the global average value calculated from all the different peptides and MS injections. Moreover, this value showed very little variation among three different replicate experiments, and showed the good reproducibility of the peptide library treatment. Thus, the relative abundance ratio of a protein between two different untreated samples appeared well-conserved, even after peptide library treatment of the two samples.

The two-mentioned examples, of course, old true for all proteins that do not saturate their respective baits on the combinatorial ligand library; that should apply to most, if not all, low-abundance proteins. These data also suggest that the reduction of dynamic concentration range for low-abundance proteins is operated as a compressed spiral of reduction of differences in concentrations. When this operation is performed simultaneously to a control, it should be possible to apply also all the above-mentioned quantitative MS-related technologies, such as ICAT, SILAC, and iTRAC.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information

We hope that this review will offer a glimpse at the unique potential of combinatorial ligand libraries in capturing and preferentially enriching the low-abundance proteome associated to the powerful analytical method of mass spectrometry. The examples given here, especially the deep exploration of the RBC cytoplasmic proteome, are quite emblematic in exhibiting the extraordinary potential of this sample treatment.

In spite of the fact that the technology is still at its early stage of development, a number of other investigations are currently underway, among which the exploration of the cerebrospinal fluid (where thousands of proteins are being discovered), of egg white (where plenty of previously unreported proteins are being detected), as well as a search of hidden allergens in food stuff, such as corn. It is also to be expected that soon this technology will be adopted by several laboratories around the world, and this expansion surely will greatly enrich the knowledge of all proteomes, mammalians, plants, bacterial, viral, in all domains of living organisms. The studies presented here have shed light on unpredicted aspects of the capture mechanisms and will surely help in devising new ligand libraries endowed with more powerful properties than the present ones especially dealing with specific protein categories. Finally, recent studies that show the practically quantitative recovery of spiked proteins, as well as good correlations of signal intensity over a broad range of spiked protein concentrations, suggest that the present methodology could be exploited in differential proteomic studies, to allow investigations aimed at biomarker discovery. With an extra bonus: the “amplification” of very low-abundance species should allow one to detect biomarkers at the very early stages of a disease, when such biomarkers are released in body fluids at extremely low levels. It is additionally hoped that the users of hexapeptide libraries will have a much easier life than that of the “FortyNiners” mentioned in the prologue.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information

The authors are grateful to Dr. F. Roux-Dalvai, Dr. A. Gonzales de Peredo, Dr. B. Monsarrat (Toulouse, France), Dr. L. Guerrier, Dr. F. Fortis, Dr. N. Le Moan (Paris, France), Dr. C. Simó, Dr. A. Bachi, Dr. M. Masseroli (Milan, Italy) for help with the experiments described in the present review. PGR is supported by a grant from Fondazione Cariplo, Milano.

REFERENCES

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information
  • Antonioli P, Fortis F, Guerrier L, Rinalducci S, Zolla L, Righetti PG, Boschetti E. 2007. Capturing and amplifying impurities from purified recombinant monoclonal antibodies via peptide library beads: A proteomic study. Proteomics 7: 16241633.
  • Bachi A, Simó C, Cattaneo A, Guerrier L, Fortis F, Boschetti E, Masseroli M, Righetti PG. 2008. Performance of combinatorial peptide libraries in capturing the low-abundance proteome. II: Behaviour of resins containing individual amino acids. Anal Chem (in press).
  • Bhattacharya D, Mukhopadhyay D, Chakrabarti A. 2007. Haemoglobin depletion from red blood cell cytosol reveals new proteins in 2-D gel-based proteomics study. Proteomics Clin Appl 1: 561564.
  • Bjellqvist B, Ek K, Righetti PG, Gianazza E, Görg A, Westermeier R, Postel W. 1982. Isoelectric focusing in immobilized pH gradients: Principle, methodology and some applications. J Biochem Biophys Methods 6: 317339.
  • Boschetti E, Lomas L, Citterio A, Righetti PG. 2007a. Romancing the “hidden proteome”, Anno Domini two zero zero seven. J Chromatogr A 1153: 277290.
  • Boschetti E, Monsarrat B, Righetti PG. 2007. The “Invisible Proteome”: How to Capture the Low Abundance Proteins Via Combinatorial Ligand Libraries. Curr Proteomics 4: 198208.
  • Castagna A, Cecconi D, Sennels L, Rappsilber J, Guerrier L, Fortis F, Boschetti E, Lomas L, Righetti PG. 2005. Exploring the hidden human urinary proteome via ligand library beads. J Proteome Res 4: 19171930.
  • Cornett DS, Reyzer ML, Chaurand P, Caprioli RM. 2007. MALDI imaging mass spectrometry: Molecular snapshots of biochemical systems. Nature Methods 4: 828833.
  • Eber SW. 2003. Disorders of erythrocyte glycolysis and nucleotide metabolism. In: HandinRI, LuxSE, StosselTP, editors. Blood: Principles and practice of hematology, Chapter 53. Philadelphia: Lippincott Williams & Wilkins. pp. 18871920.
  • Echan LA, Tang HY, Ali-Khan N, Lee KB, Speicher DW. 2005. Depletion of multiple high-abundance proteins improves protein profiling capacities of human serum and plasma. Proteomics 13: 32923303.
  • Fortis F, Guerrier L, Areces LB, Antonioli P, Hayes T, Carrick K, Hammond D, Boschetti E, Righetti PG. 2006. A new approach for the detection and identification of protein impurities using combinatorial solid phase ligand libraries. J Proteome Res 5: 25772585.
  • Gavin AC. 2007. Protein-protein interactions. In: WilkinsMR, AppelRD, WilliamsKL, HochstrasserDF, editors. Proteome research: Concepts, technology and applications. Berlin: Springer. pp. 169192.
  • Gianazza E, Giacon P, Sahlin B, Righetti PG. 1985. Non-linear pH courses with immobilized pH gradients. Electrophoresis 6: 5356.
  • Guerrier L, Boschetti E. 2007. Protocol for the purification of proteins from biological extracts for identification by mass spectrometry. Nature Protoc 2: 831837.
  • Guerrier L, Thulasiraman V, Castagna A, Fortis F, Lin S, Lomas L, Righetti PG, Boschetti E. 2006. Reducing protein concentration range of biological samples using solid-phase ligand libraries. J Chromatogr B 833: 3340.
  • Guerrier L, Claverol S, Fortis F, Rinalducci S, Timperio AM, Antonioli P, Jandrot-Perrus M, Boschetti E, Righetti PG. 2007a. Exploring the platelet proteome via combinatorial, hexapeptide ligand libraries. J Proteome Res 6: 42904303.
  • Guerrier L, Claverol S, Finzi L, Paye F, Fortis F, Boschetti E, Housset C. 2007b. Contribution of solid-phase hexapeptide ligand libraries to the repertoire of human bile proteins. J Chromatogr A 1176: 192205.
  • Hamdan M, Righetti PG. 2005. Proteomics today: Protein assessment and biomarkers using mass spectrometry, 2-D electrophoresis and microarray technology. Hoboken: Wiley-VCH. pp. 127147.
  • Herbert BR, Righetti PG, Citterio A, Boschetti E. 2007. Sample preparation and pre-fractionation techniques for electrophoresis-based proteomics. In: WilkinsMR, AppelRD, WilliamsKL, HochstrasserDF, editors. Proteome research: Concepts, technology and applications. Berlin: Springer. pp. 1540.
  • Hernandez P, Binz PL, Wilkins MR. 2007. Protein identification in proteomics. In: WilkinsMR, AppelRD, WilliamsKL, HochstrasserDF, editors. Proteome research: Concepts, technology and applications. Berlin: Springer. pp. 4168.
  • Hirabayashi J. 2004. Lectin-based structural glycomics: Glycoproteomics and glycan profiling. Glyconj J 21: 3540.
  • Huber LA, Pfaller K, Vietor I. 2003. Organelle proteomics: Implications for subcellular fractionation in proteomics. Circulation Res 92: 962968.
  • Kakhniashvili DG, Bulla LA Jr, Goodman SR. 2004. The human erythrocyte proteome. Analysis by ion-trap mass spectrometry. Mol Cell Proteomics 3: 501509.
  • Köcher T, Superti-Furga G. 2007. Mass-spectrometry-based functional proteomics: From molecular machines to protein networks. Nature Methods 4: 807815.
  • Lam KS, Salmon SE, Hersh EM, Hruby VJ, Kazmierski WM, Knapp RJ. 1991. A new type of synthetic peptide library for identifying ligand-binding activity. Nature 354: 8284.
  • Larsen MR, Thingholm TE, Jensen ON, Roepstorff P, Jorgensen TJ. 2005. Highly selective enrichment of phosphorylated peptides from peptide mixtures using titanium oxide microcolumns. Mol Cell Proteomics 4: 873886.
  • Li S, Armstrong CM, Bertin N, Ge H, Milstein S, Boxem M, Vidalain PO, Han JD, Chesneau A, Hao T, Goldberg DS, Li N, Martinez M, Rual JF, Lamesch P, Xu L, Tewari M, Wong SL, Zhang LV, Berriz GF, Jacotot L, Vaglio P, Reboul J, Hirozane-Kishikawa T, Li Q, Gabel HW, Elewa A, Baumgartner B, Rose DJ, Yu H, Bosak S, Sequerra R, Fraser A, Mango SE, Saxton WM, Strome S, Van Den Heuvel S, Piano F, Vandenhaute J, Sardet C, Gerstein M, Doucette-Stamm L, Gunsalus KC, Harper JW, Cusick ME, Roth FP, Hill DE, Vidal M. 2004. A map of the interactome network of the metazoan C. elegans. Science 303: 540543.
  • Lisacek F, Hoogland C, Bougueleret L, Bairoch A. 2007. Data integration in proteomics. In: WilkinsMR, AppelRD, WilliamsKL, HochstrasserDF, editors. Proteome research: Concepts, technology and applications. Berlin: Springer. pp. 145168.
  • Nesvizhskii A, Vitek O, Aebersold R. 2007. Analysis and validation of proteomic data generated by mass spectrometry. Nature Methods 4: 787797.
  • O'Farrell PH. 1975. High resolution two-dimensional electrophoresis of proteins. J Biol Chem 250: 40074021.
  • Pasini EM, Kirkegaard M, Mortensen P, Lutz HU, Thomas AW, Mann M. 2006. In-depth analysis of the membrane and cytosolic proteome of red blood cells. Blood 108: 791801.
  • Paulus A, Thulasiraman V, Academia K, Freeby S, Roth S, Wehr T, Liu N, Smith K. 2008. Quantification of serum and plasma proteins after enrichment of low-abundance proteins with a combinatorial peptide library. BioTechniques (in press).
  • Porath J, Fornsted N. 1970. Group fractionation of plasma proteins on dipolar ion-exchangers. J Chromatogr 51: 479489.
  • Porath J, Fryklund L. 1970. Chromatography of proteins in dipolar ion adsorbants. Nature 226: 11691170.
  • Righetti PG, Boschetti E. 2007. Sherlock Holmes and the proteome—A detective story. FEBS J 274: 897905.
  • Righetti PG, Gianazza E, Salamini F, Galante E, Viotti A, Soave C. 1977. Zein: Macromolecular properties, biosynthesis and genetic regulation. In: RadolaBJ, GraesslinD, editors. Electrofocusing and isotachophoresis. Berlin: W. de Gruyter. pp. 199211.
  • Righetti PG, Boschetti E, Lomas L, Citterio A. 2006. Protein equalizer technology: The quest for a “democratic proteome”. Proteomics 6: 39803992.
  • Roux-Dalvai F, Gonzalez de Peredo A, Guerrier L, Boschetti E, Righetti PG, Simó C, Citterio A, Zanella A, Burlet-Schiltz O, Monsarrat B. 2007. Study of hidden erythrocyte proteome by means of combinatorial peptide ligands and LC-MS/MS with LTQ Orbitrap. Mass spectrometry and proteomics conference, Pau, France, September 17–20.
  • Roux-Dalvai F, Gonzalez de Peredo A, Simó C, Guerrier L, Bouyssie D, Zanella A, Citterio A, Burlet-Schiltz O, Boschetti E, Righetti PG, Monsarrat B. 2008. Extensive analysis of the cytoplasmic proteome of human erythrocytes using the peptide ligand library technology and advanced mass spectrometry. Mol Cell Proteomics (submitted).
  • Rush J, Moritz A, Lee KA, Guo A, Goss VL, Spek EJ, Zhang H, Zha XM, Polakiewicz RD, Comb MJ. 2005. Immuno-affinity profiling of tyrosine phosphorylation in cancer cells. Nature Biotech 23: 94101.
  • Sennels L, Salek M, Lomas L, Boschetti E, Righetti PG, Rappsilber J. 2007. Proteomic analysis of human blood serum using peptide library beads. J Proteome Res 6: 40554062.
  • Simó C, Bachi A, Cattaneo A, Guerrier L, Fortis F, Boschetti E, Podtelejnikov A, Righetti PG. 2008. Performance of combinatorial peptide libraries in capturing the low-abundance proteome. I: Behavior of mono- to hexa-peptides. Anal Chem (in press).
  • Steen H, Mann M. 2004. The ABC's (and XYZ's) of peptide sequencing. Nature Rev Mol Cell Biol 5: 699711.
  • Tammen H, Peck A, Budde P, Zucht HD. 2007. Peptidomics analysis of human blood specimens for biomarker discovery. Expert Rev Mol Diagn 7: 605613.
  • Thulasiraman V, Lin S, Gheorghiu L, Lathrop J, Lomas L, Hammond D, Boschetti E. 2005. Reduction of the concentration difference of proteins in biological liquids using a library of combinatorial ligands. Electrophoresis 26: 35613571.
  • Vergote I, van der Zee A, Fung ET. 2005. Proteomic discovery of biomarkers for prognosis in ovarian cancer. 14th International Meeting of the European Society of Gynecological Oncology (EGSO) Istanbul, Turkey, September 25–29.
  • Vulliamy TJ, Luzzatto L. 2003. Glucose-6-phosphate dehydrogenase deficiency and related disorders. In: HandinRI, LuxSE, StosselTP, editors. Blood: Principles and practice of hematology. Philadelphia: Lippincott Williams & Wilkins. Chapter 54 pp. 19211950.
  • Washburn MP, Wolters D, Yates JR III. 2001. Large-scale analysis of the yeast proteome by multidimensional protein identification technology. Nature Biotechnol 19: 242247.
  • Wietze ES, Old WM, Resing KA, Ahn NG. 2007. Mapping proteins post-translational modifications with mass spectrometry. Nature Methods 4: 798806.
  • Wilkins MR, Appel RD. 2007. Ten years of the proteome. In: WilkinsMR, AppelRD, WilliamsKL, HochstrasserDF, editors. Proteome research: Concepts, technology and applications. Berlin: Springer. pp. 113.

Biographical Information

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information

Prof. Pier Giorgio Righetti earned his Ph.D. in Organic Chemistry from the University of Pavia in 1965. He spent 3 years as a Post. Doc. at MIT and 1 year at Harvard (Cambridge, MA, USA). He is currently full professor of Biochemistry at the Milan's Polytechnic (Italy). He is in the Editorial Board of several International scientific journals and has co-authored the book Hamdan & Righetti, Proteomics Today, Wiley-VCH, Hoboken, 2005. He has published >700 scientific articles in the international literature. He has patented a number of methodologies in Separation Science, such as isoelectric focusing in immobilized pH gradients, multicompartment electrolyzers with isoelectric membranes, temperature-programmed capillary electrophoresis.

Biographical Information

  1. Top of page
  2. Abstract
  3. PROLOGUE
  4. INTRODUCTION
  5. CONCLUSIONS
  6. Acknowledgements
  7. REFERENCES
  8. Biographical Information
  9. Biographical Information

Egisto Boschetti Internationally recognized as an expert in protein separation by liquid chromatography, he currently serves as Scientific Development Manager at Bio-Rad, LS Division, after having served for 6 years as VP R&D of Ciphergen Biosystems. With a degree in biochemistry from the University of Bologna (Italy) and an MBA, (Paris), Dr. Boschetti is one of the co-founders of Biosphere Medical, Inc. His extensive experience in the proteins field is witnessed by over 200 scientific publications and the application of more than 60 patents. Major accomplishments in proteomics are the reduction of dynamic concentration range of proteins and non-redundant fractionation prior to mass spectrometry and 2D electrophoresis analysis.