SEARCH

SEARCH BY CITATION

Keywords:

  • Capillary electrophoresis;
  • Capillary isoelectric focusing;
  • Peptides;
  • Proteins;
  • Technology;
  • Tutorial

Abstract

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

CZE and CIEF of proteins have preceded, and accompanied, the birth of proteomics. Although they might not be fully exploited in massive proteomic analyses (especially those projects aiming at a deep discovery of possibly the entire proteome of a cell or subcellular organelles or biological fluids), it still has interesting features and advantages, especially with samples of limited heterogeneity and in the field of purity checking for recombinant DNA proteins meant for human consumption. The purpose of this tutorial paper is to guide the reader through the history of the field, then through the main steps of the process, from sample preparation to analysis of proteins and peptides, while commenting on the constraints and caveats of the technique. The tutorial ends with an outlook on the future, which might be dominated by microchip electrophoresis, especially for ultrafast analyses of protein samples in a sieving mode, in presence of either sieving liquid polymers or firm gels polymerized within the microchannels. To this purpose, commercial instrumentation is already available on the market. This tutorial is part of the International Proteomics Tutorial Programme (IPTP 13).

Abbreviations
BGE

background electrolyte

CAs

carrier ampholytes

CO

Compton and O'Grady

ITP

isotachophoresis

rDNA

recombinant DNA

1 Historical background

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

CE has had a very long gestation period. Perhaps the ancestor of this methodology was Hiertén's PhD thesis [1], a corner stone of all electrokinetic methodologies, which should be mandatory reading in courses of Separation Science. His zone electrophoresis in a rotating tube was clearly a precursor of CE: everything was there, the theory, instrumentation, applications, only the capillary was missing! For that matter, also Virtanen's PhD thesis [2] came very close to it: his narrow-tube CE reached down to 200 μm id and his instrument was equipped even with a potentiometric detector. A full instrumental approach to electrophoresis was also reached with isotachophoresis (ITP) [3], in which both analytical (the Tachophor) and preparative (the Tachophrac) instruments were marketed [4], comprising already all features of future CE equipments: high-voltage power supply, online detectors (up to three, in fact a UV, a thermal, and a differential thermal signal), injection port, cooling cartridge, and the like. So, when Jorgenson and Lukacs [5] blew the horn on CE with their celebrated paper in Science, they only had to take the final step of introducing the silica capillary developed by Dandenau and Zerenner [6].

Why was there a need for developing CE? Chromatographers always boasted that their instrumentation was fully automated (since at least the mid-1970s) whereas in electrophoresis all methodologies were mostly manual and labor-intensive (especially 2-D map analyses). With the birth of CE, the coupling of the powerful separation mechanisms of electrophoresis with the instrumentation and automation concepts in chromatography took place. Since its inception, in the mid-eighties (and its full acceptance in the early nineties of last century, due to the instrument commercialization from several companies) CE has been the rising star of electrokinetic methodologies and has enjoyed constant growth over 20 years. Numerous symposia have been held on all aspects of CE, starting from the first one organized in Boston in 1989 [7], most of them published as special issues of Journal of Chromatography A. By the same token, many books have been published devoted to any possible application of CE. As this list would be terribly long, we will here quote just a few of them: the one by Sam Li (Prof. at NUS in Singapore) [8] because it was the very first one to appear, taking all users by surprise, and it was a genuine bible, collecting and organizing everything known up to that time in the field. A second one, by Khaledi [9] is a massive textbook comprising 31 chapters here too covering all aspects of CE and plenty of theory as well. A third one is also of interest [10], since it covers the applications of CE to analytical biotechnology and thus deals with several applications to peptides [11] and proteins. Other books deal with clinical and forensic applications of CE [12, 13], as well as with microchip technologies [14].

2 Basic concepts

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

In CE, narrow-bore fused-silica capillaries, typically with an inner diameter of 10–100 μm and usually filled only with buffer, are used. This type of capillary separation chamber has numerous advantages, particularly in respect to the detrimental effects of Joule heating. The low conductivity of the capillary chamber enables the application of high electric fields (100–500 V/cm) with negligible heat generation. In addition, the large ratio of surface area to volume of the capillary efficiently dissipates the heat that is generated. The use of high electric fields results in short analysis times and high efficiency and resolution. Separation efficiency, often in excess of 105 theoretical plates for proteins (but more than 106 for DNA), is due in part to the plug profile of the EOF—a phenomenon caused by the ionization of the silanols in the fused silica, which generates the bulk flow of solution in the capillary. This flow also enables the simultaneous analysis of all solutes regardless of their charge. Partly in common with HPLC and GC, CE has distinct advantages, such as automation, small sample size, minimal sample preparation, use of very small volumes of organic solvents and inexpensive chemicals, ease of buffer change and method development, and low cost of capillary columns.

CE has performed extremely well in DNA analysis, since these macromolecules, being negatively charged at most operative pH values, are repelled by the silica surface, also bearing negative charges (minimal at acidic, maximal at basic pH values) [15]. It is also an excellent tool for analysis of all sort of drugs and metabolites, even neutral and hydrophobic ones, thanks to the concepts of “micellar electrokinetic chromatography” developed by Terabe's group [16-18], by which all sorts of analytes, once incorporated into surfactant micelles, can migrate toward the detector window and be separated according to their residence times in the micelle. CE is also a good tool in chiral analysis, a field of importance in pharmaceutical products [19]. It must be stated, however, that there are still problems in protein fractionation by CE. These macromolecules are quite unwieldy to CE treatment, due to their amphoteric character and to their tendency to stick to the silica wall via interaction with silanols. In principle, binding to the wall could be minimized at the extremes of the pH scale, e.g. at pH 2.0 and pH 11–12, since under very acidic conditions the negative charge on the silica wall is almost completely abolished (the average pK of silanols having been variously estimated to range between pH 5 and pH 6.5) [20, 21], although peptides and proteins will bear maximal positive charge, which might still allow some residual binding to the wall. At the opposite extreme of the pH scale, both proteins and silica would bear maximum negative charge, so there could be coulombic repulsion. Yet, at real alkaline pH values (e.g. pH 11–12), the silica wall will begin to dissolve, thus destroying the separation channel. Over the years, a plethora of capillary coatings have been described acting via three main mechanisms: quenching the interaction via buffer pH and ionic strength or via addition of competing ions; adsorption of macro-ions to the wall or covalent binding of synthetic or natural polymers to the silanols [21-23]. Although a number of them seem to perform satisfactorily, no universal remedy has been found, so analysis of proteins via CE (especially in the case of complex proteomes) is still problematic. Nevertheless, some well-functioning protocols have been proposed and will be described below.

3 Capillary isoelectric focusing

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

CIEF, in presence of soluble carrier ampholytes (CAs), is a method of reasonable success in protein separation and analysis. The capillary has to be coated, of course, better if with a covalently affixed polymer, such as polyacrylamide or other suitable, hydrophilic coating. Nevertheless, under focusing conditions, proteins have minimal affinity for the silica wall, since, in the isoelectric state, they tend to form an inner salt [24]. One of the main applications of CIEF is the analysis of recombinant DNA (rDNA) products in search for the presence of deamidated species. Any deamidation event produces an extra species with a lower pI in respect to the parental molecule. This is used routinely in the biotech industry, since the analysis is fast, requires minute amounts of material, and the electropherogram can be easily quantitated via its UV absorption. Very many reports have appeared dealing with CIEF and it would be impossible to quote all of them here. In addition to some reviews [25-27], perhaps the most interesting article recently appeared is the one of Mack et al. [28], which represents the first systematic and thorough study in CIEF, aimed at defining and optimizing all experimental parameters critical to method reproducibility and robustness. Loss of the extreme pI CAs and sample components by the bidirectional isotachophoretic decay inherent to IEF [29, 30] was avoided by setting the concentration of the phosphoric acid anolyte to 200 mM and sodium hydroxide catholyte to 300 mM and by adding sufficient amounts of an acidic (pI < 3) and basic (10 < pI) sacrificial ampholyte to the CIEF sample solution. These “sacrificial ampholytes,” iminodiacetic acid and arginine, added to 1.7 and 40 mM, respectively, yielded a stable focused CA train that spanned the 4 < pH < 10 range without sacrificing either analyte resolution or sample throughput. The suggestions given here for monoclonal antibody analysis should be useful in most CIEF applications.

3.1 Tips and hints for success and survival

A number of experimental variables might prevent success in your attempts at CIEF.

3.1.1 Isoelectric precipitation

A vexing phenomenon in gel IEF, isoelectric precipitation (due to minimal charges at the pI and thus reduced hydration shell), could be devastating in CIEF. Precipitation and aggregation can generate particles that appear as artifactual spikes in the electropherogram. Precipitates may partially block the capillary, cause reduced or fluctuating currents, variable migration times and, in the worst scenario, hamper the current flux with consequent analysis failure. Large-size proteins such as Igs and membrane proteins are at high risk for aggregation in CIEF. Such a risk can be minimized by adding nonionic or zwitterionic surfactants (Triton-X 100, Brij, Tween, CHAPS), chaotropic agents (6 M urea), and/or organic modifiers (glycerol or propylene glycol) to the CA solution. If recovery of biological activity is a requirement, protein solubility in CIEF can be improved by addition of robust amounts of polyols (20% sucrose, sorbose, sorbitol) in combination with high concentrations of zwitterions (e.g. 200 mM taurine, 500 mM nondetergent sulphobetaines, 1 M bicine or CAPS) [31].

3.1.2 pI markers

In IEF, it is important to be able to map the pH gradient along the separation axis. In slab-gel IEF this can be obtained by running, in a separate track, a set of pI standards (proteins). This is not applicable in CIEF, since such pI standard proteins, being unstable and contaminated by impurities, variants, and isoforms, would confuse the sample signals, since they would have to be admixed to it. For internal standards, Shimura et al. have first proposed a set of 16 UV-absorbing tri- to hexapeptides covering the pH 3.38 to 10.17 range [32, 33] and then a set of 19 fluorescent peptides 4–13 residues long for use of pI markers via LIF detection [34]. On an alternate line of thinking, Slais and Friedl have proposed a series of UV-absorbing amino ethyl phenols as pI markers [35], followed by another set of fluorescent species [36].

3.1.3 Sample mobilization

Since IEF is a steady-state process, in principle the focused protein zones should remain immobile within that region of the capillary in which they found their pI conditions. Thus, ways have to be found to mobilize the train of bands past the detector, usually placed at the cathodic end of the capillary. In the classical setup, CIEF is seen as a two-step process, i.e. first a focusing process occurring in a well-coated capillary (i.e. devoid of any appreciable EOF) followed by either hydraulic mobilization obtained via pressure, vacuum, or gravity (as summarized in [25]) or by salt mobilization, as first proposed by Hjertén's group [37]. There is also an alternative method “single-step CIEF with EOF mobilization” by which IEF is allowed to proceed in uncoated (or poorly coated) capillaries. If the EOF flux is lower than the velocity of to focusing process, the entire train of proteins can reach steady state before emerging at the detector end [38, 39]. There could even be a third option: just to leave the focused train of proteins at the steady state and capture the image of the focused analytes via a whole capillary imaging system, as developed by Pawlyszyn et al. [40, 41]. The price to pay? A special CE equipment has to be built (see Fig. 1, which will give to the reader an idea of what a CE setup is, with the proviso that the detector is generally at one end of the capillary and that the capillary is coiled, since it can be quite long, up to 60 cm, and placed in a cartridge with circulating coolant). For this setup, in general the capillary is quite short (e.g. 5 cm) and its polyimide coating be completely peeled off. And there remains the problem that, even when the image is acquired, one cannot do any subsequent analysis of the focused species unless they are mobilized and forced to exit the capillary. To err on the safe side, we dare recommend the protocol worked out by Mack et al. [28] that consists on analyte mobilization past the detection window by replacing the catholyte (300 mM NaOH) with a 350 mM acetic acid solution and applying a 30 kV electrical potential for 30 min. Chemical mobilization avoids Taylor dispersion due to laminar flow occurring during any type of hydraulic mobilization, be it by pressure, vacuum, or gravity [42].

image

Figure 1. Block diagram of the imaging CIEF system. Reprinted by permission from International Scientific Communications Inc. (Shelton, CT, USA) and Convergent Bioscience Ltd. (Toronto, Canada).

Download figure to PowerPoint

3.1.4 Choice of CA

These compounds are a multitude of soluble, amphoteric buffers creating and maintaining the pH gradient in presence of an electric field. Their synthesis was first proposed by Vesterberg [43], the pupil of Svensson-Rilbe (in turn the pupil of the two Nobel laureates Svedberg and Tiselius) who laid the theoretical foundations of IEF [44]. In chemical terms, CAs are oligoamino, oligocarboxylic acids available from different suppliers under different trade names (typically Pharmalytes and Ampholines from GE Healthcare, Servalytes from Serva, and Biolytes from Bio Rad). Only recently the basic physicochemical parameters have been decoded, including their Mr values, the total number of individual chemicals and isoforms as well as their focusing behaviors (“good” or “poor” ampholytes) [45]. It should be noted that the critical parameter is not the total number of species (some brands comprise a concoction of close to 3000 different chemical species in the widest possible pH interval) but their ΔpK value across their pI, i.e. their ability to focus as sharp zones at the pI, a fundamental property, since it ensures the much needed buffering power and conductivity at their focusing position. The smaller the ΔpI (or, if you prefer, the pI-pKprox, or 1/2ΔpK) the higher their buffering power and conductivity. In order to appreciate these fundamental properties and grasp properly the concept of IEF and its subtleties, we encourage the readers to master the masterpiece of Rilbe [46] that summarizes a lifetime of efforts on electrokinetic methodologies. Guidelines on proper use of CAs can be found in [45].

3.1.5 Immobilized pH gradients

IPGs [47] would be the ideal choice for CIEF, since they would provide indefinitely stable pH gradients, a well-defined chemical milieu composed of no more than 10 Immobiline chemicals and prevent injection of soluble CAs into a hyphenated MS, where such soluble buffers would be detrimental to the MS signal. In fact IPGs are now routinely adopted as a first dimension in 2-D maps, both in narrow (e.g. 1 or 2 pH units) or extended (e.g. pH 3–10, especially in the nonlinear formulation) [48]. However, casting IPGs with a two-vessel gradient mixture in a capillary lumen that might contain barely 1 μL total volume or even less is not an easy proposition. Most scientists who attempted that failed. There appear to be only few reports on cIPG. In this respect, the authors do not attempt to prepare genuine IPGs by adopting the Immobiline chemicals, but rather they cast a gel within the capillary lumen, focus the soluble CAs, and then graft them onto the matrix. In a first report, the matrix was composed by two monomers, glycidyl methacrylate and ethylene glycol dimethacrylate [49]. In a second report, the matrix adopted was the well-known acrylamide and Bis couple [50]. In both cases, judging from the meager results obtained, this appears to be a difficult avenue to be explored. Additionally, having a firm gel or a monolith (albeit porous) inside a capillary might not be a good idea, since proteins would surely stick to the gel, either at the inlet port or all along the migration path, this producing the noxious problem of “carry-over.” A capillary filled just with plain buffer (or with sieving liquid polymers for size separations), which can be refreshed at each subsequent run is always the best choice.

3.1.6 Where and how should one apply the sample

This is also a nontrivial aspect, which has been recently theoretically and experimentally investigated by Takacsi-Nagy et al. [51]. Since one of the recent trends suggests to perform CIEF in presence of a concomitant EOF displacement, these authors have investigated two classical scenarios: analyte mixed with the CAs (as customarily done) or added as a short, individual sample plug either introduced before or after the CA solution filling the entire capillary, or as a plug inserted and sandwiched between zones of CAs. Simulation data revealed that sample application between two zones of CAs results in pH gradient disturbances and generation of conductance gaps (hot spots). Sample placed at the anodic side or at the anodic end of the initial CA zone is the most favorable configuration for CIEF exploiting electroosmotic zone mobilization. Do not be discouraged by the complex theory underlying this paper: the figures are a feast to the eye and have a great visual impact. Chapeau to the duo Mosher and Thormann, the only one perhaps able to perform “Visceral Research”!

3.2 Protein and proteomics applications of CIEF

We will now offer some examples of separations just to show the power of the technique, which provides short separation times, high resolution, and minimal sample consumption (the typical sample volumes injected in the capillary are of the order of a few nanoliters). Figure 2A shows separation of a mixture of four hemoglobins (Hbs), the normal human adult (A), together with the foetal (F) and two pathological ones (C and S) [52]. All of them are resolved to base line, minor impurities are also well separated, and the entire train of zones occupies a <1 min time window. Figure 2B shows separation of a mixture of Hb A and its glycated form (A1c) a separation that is routinely used in clinical chemistry to monitor for diabetes, normal Hb A1c levels being considered up to 6.2% and pathological levels being from 7% up to 10% and higher for severe diabetic cases. In a comparative study, Conti et al. [53] demonstrated very good agreement between the Hb A1c levels determined by CIEF with those obtained by gel slab IEF. Additionally, the same authors [54] reported very good separation and quantitation of the three main Hb components of umbilical cord blood (foetal, acetylated foetal, and adult Hbs; Hb F, Fac, A) in a pH 6–8 gradient. Due to their close pI values, baseline separations between Hb A and Hb F could only be achieved by adding a “separator,” i.e. a compound able to flatten the pH gradient in this pI region, in this particular case beta-alanine added at a 50 mM level.

image

Figure 2. (A) Separation of hemoglobin variants by CIEF in a 12 cm × 25 μm (id) coated capillary in presence of 3% pH 3–10 carrier ampholytes. Focusing and mobilization were carried out at constant voltages of 8 kV (from Zhu et al. [52] by permission). (B) Separation of normal human adult Hb from its glycated form (Hb A1c). The focusing mixture consisted of 5% Ampholine, pH 6–8, 0.5% Pharmalyte, pH 3–10, 3% short-chain liquid polyacrylamide, and an equimolar mixture of two “separators,” 0.33 M beta-alanine and 0.33 M 6-aminocaproic acid. The last two compounds flatten the pH gradient in the pI region of the two Hbs, permitting full resolution. Catholyte: 50 mM Lys (pH 9.7); anolyte: 50 mM acetic acid (pH 3.5) (from Conti et al. [53] by permission). Hb C: mutant Hb with a Glu [RIGHTWARDS ARROW] Lys substitution at the sixth amino acid of β-globin chains; Hb S: mutant Hb with a Val [RIGHTWARDS ARROW] Glu substitution at the sixth amino acid of β-globin chains; Hb F: foetal Hb; Hb A: adult Hb.

Download figure to PowerPoint

Since this tutorial is meant for a proteomics audience, it is highly desirable to see what CIEF can accomplish in this domain. Liquid-phase separations, often within capillaries, are indeed increasingly recognized as the best separation techniques for this approach. In particular CIEF, with a resolving power better than 0.01 pI and a concentrating power better than 500-fold seems particularly well suited for proteomics studies. Several examples have been provided by Smith and collaborators. For instance, in proteomic analysis of the microorganism Deinococcus radiodurans, at the intact protein level, a separation peak capacity of ∼800 and a sample concentration factor of ∼500 have been achieved [55-57]. The minimum discriminated ΔpI (peak width at half peak height) could be as small as 0.004. In the case of proteolytic tryptic digests (e.g. yeast cell lysate) the minimum baseline resolved ΔpI could be as low as 0.01 (excellent resolution, considering that peptides have a much higher diffusion coefficient than proteins!). This result corresponded to a separation peak capacity of ∼1000 and a sample concentration factor of ∼700 [58]. It should be noted that these results have been obtained with minute sample loads, typically of the order of 200–300 ng, a protein load that avoids the formation of precipitates. Such results are of course obtained by hyphenating CIEF with MS, in fact with the most powerful instrument available in those days, the FT-ICR machine, a mammoth, and very expensive instrument today largely substituted by the Orbitrap (or Velos or any of this family of machines). In should be noted that this setup, in principle, can be compared to the classical electrophoretic 2-D map analysis, since the first dimension is IEF and the second dimension (which is also used as a detector) is the MS, which separates proteins and peptides on the basis of the mass (in reality of the m/z ratio). Another way of obtaining a 2-D capillary separation system is by coupling CIEF with transient ITP and CE, as proposed by Mohan and Lee [59]. The two separation modes are joined by a microdialysis device containing acetic acid, serving as the anolyte for the first dimension CIEF separation and as the background electrolyte (BGE) for the second dimension ITP/CE separation. After the CIEF step, segments of the pH gradient are hydrodynamically injected into the ITP/CE capillary by gravity. The CAs serve as the leading electrolyte and acetic acid as the terminating electrolyte for the transient ITP step. The authors demonstrated a 2-D separation of tryptic peptides obtained from a three-component protein mixture and estimated a peak capacity of ∼1600, due to the fact that the two separation modes are largely orthogonal.

Warning. Notwithstanding this rosy picture, things with CIEF are not as smooth as here presented. To start with, the fiercest competitor of all electrokinetic methodologies is HPLC, a separation method that had been largely automated well before any electrophoretic methodology, still quite labor-intensive. Thus, although assessment of HbA1c for diabetic screening could be performed very well with CIEF, with results fully comparable with HPLC, it is a fact that in all clinical chemistry labs around the world this is routinely performed with dedicated HPLC miniaturized columns. And even though Smith's group has shown impressive results in proteomic analyses with hyphenated CIEF-FTICR, what he has presented in hyphenated RP-HPLC-FTICR is astonishing by comparison. If one were to look at his Fig. 7 in [57], one would immediately notice that, when screening a yeast cytosolic tryptic digest by this latter method, these authors could resolve more than one hundred thousand peptides, i.e. two orders of magnitude more than what could be seen by CIEF! And now the last blow. According to Graf and Waetzig [60], no matter how well a capillary is coated, protein adsorption to the silica wall can be minimized, but by all means not abolished. This is even truer in CIEF, since proteins reach steady state in a pH gradient that could easily span pH 3–10. Whereas at low pH values adsorption is almost abolished, since the silica wall is almost uncharged, as the pH gradient progressively increases toward pH 10, more and more silanols will be ionized and the risk of protein adsorption onto the wall will be maximized, notwithstanding that proteins, at their pI value, have a minimum of surface charge (and zero net charge, of course) while typically forming an inner salt.

4 CE of peptides

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

Peptide CE has much to offer and is perhaps the simplest and trouble-free aspect of capillary electrokinetic separations. It has a distinct advantage over RP-HPLC (still the method of election in peptide analysis) in that the latter usually fails in the separation of small similar polar peptides. Peptide analysis via electrophoresis has had a long gestation period, with the first equation correlating mobility (μ) to the charge/mass (Z/M2/3) ratio having been derived already in the sixties of last century by Offord [61]. As CE of peptides became more and more popular, this original equation was further refined and different models predicting with more accuracy the relationship of μ with Z/M were proposed by Grossmann et al. [62], Rickard et al. [63], Chen et al. [64], Compton and O'Grady (CO) [65], and Castagnola et al. [66]. All these models have been critically evaluated and amply discussed in a review by Castagnola et al. [67]. Perhaps the best model is the CO [65], as shown in Fig. 3, which demonstrates an excellent correlation between theoretical and experimental μ values for a set of 165 peptides (other models give a larger dispersion of data). This large body of peptides has been obtained by mixing trypsin digests of α and ß chains of Hb, of Met- and Leu-enkephalins and of a series of strictly related peptides of GGXA sequence. They range in length from 4 to 50 amino acids, in mass from about 400 to 5000 Da, and in charge from 2 to 10 (positive and negative) charges, although the most represented values are just up to 20 amino acids in length. Such excellent separations of selective endoprotease or chemical protein digests (peptide maps) represent one of the most powerful tools of biochemical analysis. Therefore, CZE can provide peptide maps suitable for sequencing purposes and offer valuable information about the structure and the physicochemical properties of peptides, such as their charge-to-radius ratio and its relationship with the composition of the solution, i.e. its ionic strength and pH [68-70]. In practice, one could compare such peptide maps as obtained from a wild-type protein and from potential point mutations (genetic variants); this has in fact been implemented by Ferranti et al. [71] and Ross et al. [72] for fast analyses of normal human Hb and of mutants such as HbE, HbS, and HbO-Arab, allowing for unequivocal variant determination. Additionally, Cobb and Novotny [73] compared the absolute sensitivity of microcolumn HPLC peptide mapping with CZE. They demonstrated that the separation of 4 pmol of a tryptic digest of ß-casein can be achieved by micro-HPLC, while the same separation by CZE can be performed with only 80 fmol of digest.

image

Figure 3. Correlation between the experimental mobilities (μexp) of 165 peptides and the theoretical values (μtheor) obtained by using the CO model with free parameters (from Castagnola et al. [67] by permission).

Download figure to PowerPoint

One is tempted to ask what happens with hydrophobic peptides and/or peptides that have almost identical charge/mass ration and differ solely by overall hydrophobicity (either minute or substantial). In such cases, one could hardly expect to achieve any reasonable separation and one would have to resort to RP HPLC, in which C18 resins have shown top performance. In those cases, however, CZE can perform very well by adopting the highly versatile MEKC approach described by Terabe's group [16-18]. A nice example as been provided by Liu et al. [74] who adopted buffer systems containing dodecyltrimethylammonium bromide, hexadecyltrimethylammonium bromide, SDS, and even two cyclodextrins as additives. These compounds, under different analytical circumstances, exhibit beneficial effects for peptides with similar net charges but different hydrophobicities. An interesting separation is indeed provided in Fig. 4, in which seven different enkephalin (Enk) analogues are separated by MEKC [75]. If one looks at panel A, though, this train of peaks does not look exciting at all, since only the first four are well separated. But here comes the “deus ex machina”: addition of barely 5% ACN to the BGE modulates the affinity of the various Enk analogues for the micelle of SDS and greatly ameliorates the patter (panel B, where now all seven compounds are fully separated and additional impurity peaks appear).

image

Figure 4. MEKC of enkephalin (Enk) analogues without (A) and with (B) addition of 5% ACN. Capillary: 65 cm × 75 μm, 100 mM Na borate, 100 mM SDS, pH 8.5, 15 kV, 25°C, λ = 200 nm. (1) metsulphoxide Enk; (2) methionine Enk; (3) [Ala2]methionine Enk; (4) leucine Enk; (5) leucine Enk amide; (6) leucine Enk-Arg; (7) pro-Enk (from Schwartz et al. [75] by permission).

Download figure to PowerPoint

4.1 Insider trading

If you want to bet your savings on CZE of peptides and be a winner, you might want to learn some basic tricks.

4.1.1 Mobility and the charge-to-mass ratio

To start with, one might wonder what is the basis of the excellent correlation between mobility μ and the charge-to-mass ratio shown in Fig. 3, a precious feature for performing, e.g. rapid screening of genetic variants. To start with, such separations are obtained at very acidic pH values, typically around 1.8, where two basic phenomena occur. The first one is the fact that the silica wall is essentially neutral and silanols all protonated, so that adsorption of peptides to the wall is minimized if not abolished tout court. The second one is that all peptides bear essentially only positive charges, since the carboxylic groups are extensively (although not quite fully) protonated. Due to that, peptides (which, if derived from trypsin digestion, should have at least two positive charges, namely the -NH2 terminus and either a Lys or an Arg, plus, where occurring, extra positive charges due to His) are not compact structures, but tend to stretch, thus denying the contribution of a third parameter that could affect the separation, namely the conformation. However, for rather long peptides (40–50 residues, see Fig. 3), if the number of positive charges is not sufficient, coulombic repulsion will be reduced and these peptides will tend to be more compact in solution, have a reduced frictional coefficient and thus deviate from the good linear relationship (μexp) (μtheor) observed in Fig. 3 (and in fact in all figures in ref. [67]).

4.1.2 Hydrophobic peptides

Other parameters have to be taken into account too. We have just seen that, for hydrophobic peptides of very similar mass and charge, one should resort to MEKC, a most powerful technique able to discriminate even on the basis of small hydrophobicity differences (see Fig. 4). Additionally, such separations can be further ameliorated by playing around with small amounts of organic modifiers (typically ACN, but also different types of alcohols) that can modulate the residence time of such hydrophobic analytes within the surfactant micelle.

4.1.3 Temperature dependence

An additional experimental parameter one could play with is the temperature, which also can profoundly affect the separation in the 20–60°C interval. On this last note, one could even perform CZE at subzero temperatures, in order to freeze cis-trans conformers of peptides. A spectacular case was reported by Ma et al. [76], which we are happy to report here as a tribute to Cs. Horvath, a scientist who shaped both the fields of chromatography and CE. Peptides containing Pro residues are known to exist in both the cis and trans conformation due to the rigidity of the peptidyl-proline bond. At temperatures near ambient, the relaxation time of the cis-trans isomerization is on the time scale of minutes, thus commensurate with the migration times in CZE under usual operating conditions. When the separation was conducted at temperatures as low as -17°C, two conformers of the heptapeptide Tyr-Pro-Phe-Asp-Val-Val-Gly-NH2 were fully resolved. In the case of Tyr-Pro-Phe-Gly-Tyr-Pro-Ser-NH2, all the four conformers, due to the presence of two peptidyl-Pro bonds, were also resolved. As an extra bonus, the hydrodynamic radii of the cis-trans conformers of two dipeptides, Phe-Pro, and Leu-Pro, could be estimated. In both cases, the trans isomers had 1.3 times greater Stokes radii than the cis conformers.

4.1.4 Peptides and metal ions

Some more tips. Other ways for modulating peptide mobility consist in adding metal ions originating from zinc or copper salts: they can interact with certain N, O, or S atoms of proteins and peptides. This causes the mobility of these species to decrease relative to species in which no complexation takes place and this could help in difficult separations [77, 78]. Other factors have to be taken into consideration too. For instance, whereas separations of tryptic peptides, as reported above, seems to perform quite well around pH 2.0–2.5, other pH values might have to be selected for improving the separation for, e.g. strongly acidic peptides, whose charge difference will be minimized at such low pH values, but certainly maximized at neutral to alkaline pH values [79]. In these last cases, peptide adsorption to the wall will be an ever present hazard, so that one will best work with a well-coated capillary and/or, at the very least, addition of oligoamines able to quench such wall interactions (e.g. diaminopentane, spermine, spermidine, tetramethylethylenediamine, pentaethylenehexamine, and the like) [80].

4.1.5 A race in the peptide corral

In proteomics, peptide analysis is very important, since it is at the basis of the so-called “bottom up” approach by which the global protein sample is treated with proteolytic enzymes (generally trypsin) and the resulting peptides fractionated and finally analyzed by MS. There is a vast literature on that and we are pleased to leave the stage to a series of surveys published almost biannually by Kašička [81-87], which can be used as the Arianna's thread to help the readers out of the labyrinth. With a last comment: peptide IEF, due to its high resolving power and ability for analyte concentration (by a factor of up to 500- to 1000-folds) can be successfully exploited, as shown by Shen et al. [58] and discussed above. Yet, one should not forget that the soluble CA buffers have Mr ranges (280 up to 1200) rather close to that of tryptic peptides, and also similar structures [45]. An anecdote is here worth recalling. When Vesterberg patented CAs in 1969 [43, 88], he stated that they should NOT contain peptide bonds. Over the years, plenty of companies tried to circumvent the claims, but to no avail: LKB Produkter AB brought them to court and they had to pay royalties on their own product, due to patent infringements. The last player that entered the arena (Pharmacia, now GE Healthcare) simply produced them with peptide bonds and there was nothing LKB could do against them. By the way, it turned out that these products (Pharmalyte) had in fact the best properties in terms of buffering power and conductivity and originated the best pH gradients. There is no denying, though, that Pharmalytes could represent a risk in proteomic analysis based on CIEF (or simply gel IEF) of peptides, since they would surely give MS signals strongly interfering with genuine peptide signals. Notwithstanding that, indeed IEF of peptides has turned out to be an excellent tool in proteomics, rivaling peptide separations via C18 columns. With a proviso, though, that it is performed NOT with soluble CA buffers, but in IPG, by a well-worked out technique dubbed IPG-IEF [89]. The results, especially in analysis of membrane proteins, could be quite extraordinary. For instance, Chick et al. [90], in the analysis of the rat liver membrane proteome, via IPG-IEF of tryptic peptides in the pH 3–10 range, have reported the identification of no less than 1549 nonredundant proteins, 42% being integral membrane species. Even more spectacular results have been obtained by Eriksson et al. [91] in the analysis of the microsomal proteome from the lung cancer cell line H69, by adopting IPG-IEF of peptides in narrow pH ranges. By MS analysis of the tryptic peptides focusing in a very narrow pH interval (pH 4.0–4.5) they could identify 3704 proteins. Although these data do not apply to CIEF per se, surely if IPGs in a capillary format could be properly developed, CIEF could become a valid tool in proteome analysis. At present the results have not been outstanding, but the field is under continuous development [92-97], so there might be hope in the future.

4.1.6 CA-based CE separations of peptides

This is an aspect that has been well exploited in the past, somehow fallen into oblivion but certainly worth resurrecting. The composition of the BGE essentially influences selectivity, separation efficiency, sensitivity, and speed of CZE. Hence, BGE composition is one of the key issues in the selection of experimental conditions for CZE separations and analyses of peptides as well as other types of analytes. In addition to classical acid/salt buffer-based BGEs, particular amphoteric compounds—so-called isoelectric buffers—are also employed as BGE constituents. The advantage of these isoelectric BGEs originates from their inherent property—low electric conductivity. This low conductivity allows application of high intensity electric field in the capillary at relatively low electric current and low input power and results in increased speed of analysis without negative effects of Joule heating on separation efficiency (peak broadening) [98]. Classical examples of isoelectric buffers are acidic and basic amino acids (Glu, Asp, His, Lys, Arg), used as free acid and bases, in their isoelectric state, as the sole BGE. The second group of isoelectric or, more exactly, quasi-isoelectric BGEs includes solutions of narrow pH cuts of CAs. These CA fractions of typical width of 0.1–0.15 pH unit can be obtained by preparative IEF, e.g. in thick granulated gels made of Sephadex beads or in a Rotofor apparatus. Although being composed of several components and containing both “good” and “poor” ampholytes, such narrow pH cuts of CAs proved to possess sufficient buffering [99] and loading capacity [100] to be used as BGEs in the so-called CA-based CE for the separation of peptides [101] and peptidomimetics [102].

4.2 2-D peptide electrophoresis

This is an interesting avenue worth exploring. The elective method for peptide separation is RP-HPLC, commonly in a C18 bonded phase on a wide-pore resin. The molecular basis of RP-HPLC selectivity is different from the separation principle of CZE, so there should be orthogonality in the coupling of the two methods. In fact, whereas the RP separation is based on polarity criteria, the CZE separation depends on the charge-to-mass ratio. Already in 1990, Bushey and Jorgenson [103] and subsequently Larmann et al. [104] connected the eluent from RP-HPLC narrow-bore columns with a CZE apparatus and demonstrated that the coupling offers a good chance for complex separations. It should be noted, in fact, that many peaks that coelute on the RP-HPLC column are resolved further by CZE. The peak capacity for the HPLC dimension has been estimated at 200, whereas for the CZE is 100. This translates into a 2-D peak capacity of 20 000, which is unmatched by any coupled-column system. This great resolving power has been demonstrated in practice: for instance, Hooker et al. [105] have shown that a 2-D separation of a tryptic digest of thyroglobulin generates more than 400 peptides, which could not possibly be resolved in a simple 1-D separation. These authors have even described a 3-D separation, combining RP-HPLC, CZE, and SEC. Although the separation capacity was impressive, their final appreciation was not. In their own words: “the increased peak capacity of this system may not be worth the extra effort and added complexity that is entailed.”

5 CE of proteins

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

Here the play becomes rough and calls for extraordinary measures. Early in the game it was realized that an “amour possible” easily develops among proteins and the wall silanols, correctly labeled by Choderlos des Laclos as “Les Liaisons Dangereuses.” Proteins can spontaneously bind to ionized silanols, especially at neutral to alkaline pH values, to the point (for basic species) of becoming irreversibly adsorbed. This “amour fou” is anathema in Separation Science, since a retentive chromatographic process contrasts with a rapid electrokinetic transfer, to the point of massively dissipating the analyte peak and ruining the separation. This was realized very early in the game; already in 1967, in his PhD thesis, in fact, Hjertén [1] recommended coating the wall with methyl cellulose to eliminate electroendoosmosis (EOF and with that protein adsorption, the two phenomena being intimately connected). With the advent of commercial instrumentation and the diffusion of CZE in many analytical labs around the word, the problem of protein adsorption unleashed a race on investigations on coating, in whose comparison the “Encierro de Pamplona” looked like a kids game. Hundreds of papers have been published on all possible remedies and cures to protein adsorption, which cannot possibly be quoted here (suffice to peruse through some of the major reviews covering this aspects [21-23], plus one, by the curious title: surfing silica surfaces superciliously [106]). This last review summarizes a saga on the development of an (and we do not hesitate to state that) extraordinary coating that had all prerequisites for entering the Pantheon of immortals, albeit fortuitous circumstances penalized it into almost oblivion. A whole family of such compounds was in fact developed [107-110], although the most peculiar (and best performing) one was the one depicted in Fig. 5, which we dubbed “skorpio,” namely the quaternarized piperazine (N-methyl, N-4-iodobutyl)-N′-methylpiperazine (M1C4I) (see the formula in the inset in Fig. 5). By serendipity (i.e. because a technician had forgotten to dissolve the compound in the running buffer, the same serendipity that helped Porath's group to discover molecular sieving, because they had forgotten to turn on the electric current in the column filled with cross-linked dextran; a cute anecdote narrated in [111]) we found out that, if the capillary had been preconditioned with M1C4I (at alkaline pH) and then the run continued in its absence, the EOF modulation would still be active. Such a behavior suggested that this compound had to be covalently affixed to the wall, via a reaction mechanism depicted in Fig. 5. We hypothesize that this molecule first docks onto the silica wall via the net positive charge of the quaternary nitrogen. Once hooked onto the wall, M1C4I is further pasted to it via hydrogen bonding occurring on the deprotonated tertiary nitrogen. Finally, given the close surface contact, the iodinated tail is able to “sting” and form a covalent bond with a neighboring silanol anion, thus inducing the alkylation of the surface through a siloxane bridge. The close contact arising from the relatively weak H-bond and ionic interaction enormously accelerates the nucleophilic substitution of iodine atom for entropic reasons.

image

Figure 5. Proposed mechanism of binding of the quaternarized piperazine M1C4I (formula in the upper left insert) to the silica wall. The hydrogen bond is formed by the tertiary amino group, the ionic bond by the quaternary nitrogen, and the covalent bond by the iodinated alkyl tail.

Download figure to PowerPoint

There is a unique valence for this compound and its related family: due to the fact that they are covalently bound to the wall, they are not needed any longer in the BGE, thus they do not interfere with subsequent MS analysis (most CZE instruments are hyphenated with MS instrumentation, with the analyte being directly inject in the mass analyzer). Conversely, plenty of physically adsorbed compounds (including polymers such as polyethylene imine) keep leaching out of the wall and are thus injected in the MS instrument, greatly damaging the analysis. That this is so has been fully confirmed by Elhamili et al. [112] (see Fig. 6; in this particular case another mono-quaternarized piperazine, 1-(4-iodobutyl) 4-aza-1-azoniabicyclo [2],[2],[2] octane iodide (M7C4I) has been adopted, behaving just like M1C4I). The surface thus obtained yields rapid separations (less than 5 min) of peptides and proteins at acidic pH values with high separation efficiencies (up to 1.1 × 106 plates/m for peptides and up to 1.8 × 106 plates/m for proteins) and no observed bleeding of the coating reagent into the MS. Another most important feature can be gleaned from Fig. 6: large intact proteins with molecular masses over 0.5 M Da could be easily separated. The coating showed good ability to handle these large proteins with high efficiency and retained peak shape as demonstrated by the separations of IgG1 (150 kDa) and thyroglobulin (669 kDa). No other coating reported in the literature seems to rival or even match these separations of very large macromolecules.

image

Figure 6. (A) CE analysis of IgG1 on a M7C4I-coated capillary. (B) CE analysis of thyroglobulin on M7C4I-coated capillary. Conditions: capillaries 50 μm id, 365 μm od, 50 cm total length, and 41.5 cm to the detection window: injection, 0.1 mg/mL was injected at 50 mbar for 30 s; applied negative voltage −22.5 kV, detection at 200 nm, BGE: 10 mM ammonium formate, pH 3. Abs: absorbance; mAU: milli-absorbance units.

Download figure to PowerPoint

One cannot close this tutorial without mentioning another important aspect of protein separations in a capillary format, namely the equivalent of SDS-PAGE performed in a capillary lumen. This is just about the only trouble-free separation of proteins, immune from noxious phenomena such as adsorption to the wall, skewed peaks, and the like, because in SDS proteins form mixed micelles with the surfactant, in which their amphoteric character is swamped by the negative charges of the detergent, so that their charge-to-mass ratio becomes constant, they behave as anionic macromolecules, i.e. just like nucleic acids, and therefore do not stick to the silica wall. Yet, SDS-CE did not materialize overnight. Originally, all scientists adopting size separations (either for proteins or nucleic acids) started out by polymerizing acrylamide in the capillary lumen, as polyacrylamide gels had been rightly considered as the best sieving matrix, ever since their adoption by Ornstein [113] and Davis [114]. Yet, this method soon turned out to be a nightmare. After barely a couple of runs, the capillary would clog, no current would any longer pass through it and the silica tube had to be disposed off and everything started all over again. It was soon realized that (i) plenty of material would precipitate at the interface BGE/polyacrylamide gel; (ii) this interface would soon shrink and be filled by tiny air bubbles impending the current flow. It was Karger's group [115] that first had the idea of lowering the cross-linking amount to barely 0.5% and then completely dispose of it and perform sieving runs in un-cross-linked linear polymers, a technique that has been termed size separation in sieving liquid polymers. In this system, the macromolecules are seen as migrating through “dynamic pores” formed by the fluctuating polymer chain network. Interestingly, this idea of exploiting liquid matrices for sieving had been put forward long before by Bode [116, 117] but it had fallen into oblivion. This method is the only one adopted today and it has two distinct advantages: (i) after each run, the matrix can be easily extruded from the capillary lumen, which can be filled up at each run with fresh matrix; (ii) as a result of that, no carry-over form run to run can ever occur. There has been one important, latest development in the field of size separations: since, for higher sensitivity, analyte monitoring should be performed at 214 nm, where also polyacrylamide absorbs UV light considerably, the next step was to introduce transparent polymers, notably dextran, PEG, pullulan [118], and any type of cellulose [119], although also replaceable agarose gels have been proposed [120]. An ample review on all these aspects has just been published by Zhu et al. [121], to which the readers are referred for further insight on these matters. With that, we feel we have closed the circle of this tutorial.

6 Future directions

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

In fact not. This tutorial cannot terminate without an outlook at recent developments and the future of CZE. This could very well be the field of microfabricated (or microchip) devices, which have been developed in the last few years with the aim of performing and integrating multiple analytical processes (e.g. sample pretreatment, solution distribution/mixing, separation, detection, etc.) on a chip platform [122, 123]. Although microchip electrophoresis can be carried out in free solution and also in the CIEF format, the development that has taken momentum is size separation either in sieving liquid polymers (SLPE) or in cross-linked gels (CGE). Due to the short column length and high separation efficiency, microchip sieving liquid polymers/CGE is generally fast, typically from a few seconds to a few minutes. Yao et al. [124] are recognized as the first who performed SDS-PAGE in a microfabricated channel, with separations carried out in less than 1 min. By combining an on-chip dye staining with an electrophoretic dilution step (similar to a destaining step), Bousse et al. [125] obtained excellent resolutions for microchip CGE of proteins. On the basis of this work, the first commercial microchip instrument was constructed by Agilent Technologies.

Here are some examples of interesting developments and applications. In 2004, Han and Singh [126] and Herr and Singh [127] applied an in-channel photopolymerization approach to prepare cross-linked gels inside a microchip channel for SDS-PAGE, with separation speeds of <30 s per run. These authors also prepared a gradient cross-linked gel for on-chip protein sizing [128] and successfully implemented sample preconcentration using these photopatterned gels [129]. Xu et al. [130] performed online electrokinetic supercharging preconcentration on a microchip to improve method sensitivity. Tsai et al. [131] tested simultaneous separations of both native and SDS-denatured proteins on a single microchip with 36 microchannels. Herr et al. [132] recently integrated saliva pretreatment (mixing, incubation, and enrichment) with subsequent quantitative immunoassays and measured the concentration of endogenous MMP-8 in saliva. More recently, He and Herr [133] photopatterned different gels inside a microfluidic chamber for protein immunoblotting. Their experimental setup is shown in Fig. 7. These few examples show that microchips might become the rising star in the nearest future and might turn out to be an interesting tool in protein analysis, although their worth in massive proteomic applications has still to be demonstrated.

image

Figure 7. Immunoblot chip. (A) Schematic design of the immunoblot chip for analysis of native proteins. The sample (2), sample waste (3), buffer (1, 4, 5, 6), and buffer waste (7, 8) reservoirs are indicated in sketch (not to scale). The middle region of the device (indicated as chamber) has a length of 1.5 mm, a width of 1 mm, and a depth of 20 μm. (B) Three separate zones inside the chamber dedicated to sample loading, separation, and blotting, respectively: a large pore-size protein loading gel on the top, a smaller pore-size protein separation gel on the bottom left and an antibody-functionalized blotting gel on the bottom right. Colored dyes were used to visualize the different gel regions. Reprinted from He and Herr [133] with permission.

Download figure to PowerPoint

7 Concluding remarks

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

Hopefully this tutorial will be of some help to beginners by permitting them to master CE in a rapid manner, while avoiding pitfalls and dead-end alleys (a powerpoint presentation on this tutorial is available in the Supporting Information). Peptide CE is definitely worth pursuing, since it is almost trouble-free and can give plenty of information on the size and charge properties of these analytes. Moreover, in combination with RP-HPLC, it can offer increased resolution of complex mixtures. SDS-CE is also a well-worked out technique and can be used for fast assessment of protein purity in, e.g. the biotech industry in following the purification process of rDNA products. Protein separations can also be of success, if proper coatings are adopted and interaction with the silica wall minimized. However, it must be borne in mind that analysis of very complex mixtures, such as sera and cell lysates, as most often involved in proteomics investigations, might not be easily tackled by CE, due to the fact that in such complex analytes plenty of interfering substances will be present, such as lipids and other metabolites that might foul the capillary, stick to its surface, form complexes with the proteins present, and totally ruin the separation. It is a fact that, in plenty of CE reports, especially those proving the quality of capillary coatings, only a set of well-behaved, highly purified protein standards are separated, generally yielding high plate numbers and sharp peaks. But when exploring complex biological matrices the picture could change dramatically. Focusing in a capillary format is also a well-worked out technique, although its use, so far, has been mostly in checking the charge purity of rDNA products (for instance, looking for deamidation or other PTMs, for which CIEF appears to be eminently suited).

8 References

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. 1 Historical background
  4. 2 Basic concepts
  5. 3 Capillary isoelectric focusing
  6. 4 CE of peptides
  7. 5 CE of proteins
  8. 6 Future directions
  9. 7 Concluding remarks
  10. Acknowledgments
  11. 8 References
  12. Supporting Information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
pmic7302-sup-0001-FigureS1.ppt80504KSupporting Information CZE Tutorial PPT Presentation

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.