SDS-PAGE Focusing: Preparative Aspects - Analytical Chemistry

Publication Date (Web): October 10, 2007 ... 2007, 79, 821−827) on analytical SDS-PAGE focusing, a novel method is here reported for small-scale pre...
0 downloads 17 Views 440KB Size
Anal. Chem. 2007, 79, 8624-8630

SDS-PAGE Focusing: Preparative Aspects Gleb Zilberstein,† Leonid Korol,† Pier Giorgio Righetti,‡ and Shmuel Bukshpan*,†

Cleardirection Ltd., 4 Pekeris Street, Rehovot 76702, Israel, and Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Politecnico di Milano, Milano 20131, Italy

As a followup of our previous report (Zilberstein, G.; Korol, L.; Antonioli, P.; Righetti, P. G.; Bukshpan, S. Anal. Chem. 2007, 79, 821-827) on analytical SDSPAGE focusing, a novel method is here reported for smallscale prefractionation of complex protein mixtures, for subsequent proteome analysis, based on mass separation of SDS-protein micelles not in a gel matrix, but in liquid cationic polymers assembled in a multicompartment electrolyzer (MCE) in a stepwise fashion at discrete and increasing levels of positive charges (from 3 to 28 mM), the neighboring chambers being separated by neutral agarose membranes. Unlike conventional SDS-PAGE, in which separation by mass of SDS-laden polypeptide chains is obtained in constant concentration or porosity gradient gels, the present method of SDS-PAGE focusing exploits a “steady-state” process by which the SDSprotein micelles are driven to stationary zones along the migration path and trapped into different compartments of the MCE device via interaction (and subsequent charge neutralization) with cationic polymers of fixed (but increasing from chamber to chamber from cathode to anode) charge density. Minimization of migration of the liquid cationic polymers is obtained via use of low voltage and by arranging for a buffer conductivity gradient along the migration path. The present setup has the advantage of high protein recoveries (up to 90%) without any contamination from ungrafted monomers and catalysts, as occurring in proteins recovered by passive elution from gel matrixes. Additionally, resolution can be fine-tuned by selecting cationic polymers of varying charge density in microstep increments. The cationic polymers, of desired charge density and proper viscosity, are prepared by standard polymerization conditions, can be easily precipitated and washed free of monomeric contaminants, and stored in a dry form for subsequent use. Separation and purification of proteins in a preparative scale is customarily best performed by chromatographic techniques, since they permit much larger volume and sample handling and are in general extensively automated.1 By comparison, preparative electrophoretic methodologies (usually performed in gel matrixes) suffer from severe limitations, such as limited sample load ability * Corresponding author. E-mail: [email protected]. † Cleardirection Ltd. ‡ Politecnico di Milano. (1) Janson, J. C.; Ryde´n, L. Protein Purification; Wiley-VCH: New York, 1998; pp 1-695.

8624 Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

per unit of gel surface area (in general not exceeding 1-2 mg of protein cm-2)2 and inability of scaling up the electrophoretic chambers, due to poor heat dissipation in larger size instrumentation. Additionally, when proteins have to be extracted from a gel phase (typically a polyacrylamide gel), they are usually heavily contaminated by substances leached from the gel matrix, such as unreacted monomers (acrylamide and Bis are neurotoxins) and short-chain, un-cross-linked, soluble polymers.3 Even in preparative instrumentation designs, in which all components migrate the whole length of the gel and are collected in turn as they reach the end (usually by some form of elution or buffer flow system orthogonal to the field direction), major drawbacks are still apparent, since the later-eluting bands are subjected to bandbroadening due to diffusion and are collected in a relatively large volume of elution buffer. Under constant conditions, the resolution between bands and the bandwidth both increase in proportion to the square root of time, but the separation between the bands increases in direct proportion to time.2 Thus, there are diminishing returns as the gel length is increased. From this point of view, any electrophoretic technique operating in a free liquid phase offers distinct advantages for preparativescale separations, a concept well-ingrained in the continuous flow separators in liquid curtains, as developed by Strickler4 and Hannig.5 Best operating conditions for such instruments, though, were limited to whole cells and cellular organelles, due to their minute diffusion coefficients, limiting lateral spread of the various zones. Only the advent of focusing techniques permitted very high resolution of protein zones in free liquid systems.6 Among such preparative isoelectric focusing (IEF) devices, particularly convenient and user-friendly were the instruments developed by Bier’s group, such as the Rotofor7 and the recycling-IEF apparatus.8 Today, both the mini- and micro-Rotofor are well-ingrained as prefractionation tools in proteome analysis.9 A natural evolution (2) Rodbard, D.; Chrambach, A.; Weiss, G. H. In Electrophoresis and Isoelectric Focusing in Polyacrylamide Gels; Allen, R. C., Maurer, H. R., Eds.; de Gruyter: Berlin, 1974; pp 62-105. (3) Chiari, M.; Righetti, P. G.; Negri, A.; Ceciliani, F.; Ronchi, S. Electrophoresis 1992, 13, 882-884. (4) Strickler, A., Sep. Sci. 1967, 2, 335-355. (5) Hannig, K., In Electrophoresis: Theory, Methods and Applications; Bier, M., Ed.; Academic Press: New York, 1967; Vol. II, pp 423-471. (6) Righetti, P. G. Isoelectric Focusing: Theory, Methodology and Applications; Elsevier: Amsterdam, 1983. (7) Egen, N. B.; Bliss, M.; Mayersohn, M.; Owens, S. M.; Arnold, L.; Bier, M. Anal. Biochem. 1988, 172, 488-494. (8) Bier, M.; Egen, N. B.; Allgyer, T. T.; Twitty, G. E.; Mosher, R. A. In Peptides: Structure and Biological Function; Gross, E., Meienhofer, J., Eds.; Pierce Chemical Co.: Rockford, IL. 1979; pp 79-89. (9) Davidsson, P.; Paulson, L.; Hesse, C.; Blennow, K.; Nilsson, C. L. Proteomics 2001, 1, 444-452. 10.1021/ac701598y CCC: $37.00

© 2007 American Chemical Society Published on Web 10/10/2007

of the Hannig-type apparatuses was the Octopus, exploiting an upward liquid stream,10 today adopted, in the focusing mode, for either prefractionation in proteome analysis11 or directly as the first dimension of a peculiar 2D mapping procedure, by which all eluted fractions are directly analyzed by orthogonal sodium dodecyl sulfate-polyacrylamide gel electrophoresis12 (SDS-PAGE). With the advent of immobilized pH gradients13 (IPG), notwithstanding its exquisite resolution and load ability, it was soon realized that preparative procedures in gel phases, even with highly dilute matrixes,14 were bound for failure, for the same reasons discussed above. A turning point was thus the introduction of multicompartment electrolyzers (MCEs) equipped with isoelectric membranes,15,16 in which the gel phases (the isoelectric membranes) functioned as the active units in the separation process, whereas the proteins were collected in liquid-filled chambers delimited by said membranes. MCEs too were soon adopted, in a miniaturized version, for proteome prefractionation,17 together with a variant, called off-gel IEF, in which the proteins are collected in cups placed on the surface of an IPG strip.18 SDS-PAGE, perhaps the most popular electrophoretic technique for rapid assessment of protein purity and Mr values,19 also appears as a promising technique for microscale preparative runs. Disregarding past attempts at eluting protein zones via cutting out of stained zones, recent methodologies have focused on the possibility of direct electroelution of proteins from SDS-PAGE gels without gel sectioning.20-22 While such techniques could be applied to both 1D and 2D gels, a shotgun electroelution system, driving a multitude of proteins spots resolved in a 2D surface into a multichambered, gridlike layer composed of as many as 384 wells has been described, as a proteomic tool for large-scale analysis of resolved polypeptide chains at the end of a 2D mapping process.23 For prefractionation purposes in proteome analysis, it would be ideal if one could perform small-scale SDS-PAGE under focusing conditions and perhaps even in a liquid phase. Both requirements would appear to be quite utopian, and even against the grain of basic physical laws, which require, for a focusing process, amphotericity of the substance under fractionation, a condition certainly not met by SDS-protein micelles, well-known to behave as polyanions in which the amphoteric properties of (10) Kuhn, R.; Wagner, H. J. Chromatogr. 1989, 481, 343-350. (11) Burggraf, D.; Weber, G.; Lottspeich, F. Electrophoresis 1995, 16, 10101015. (12) Hoffmann, P.; Ji, H.; Moritz, R. L.; Connoly, L. M.; Frecklington, D. F.; Layton, M. J.; Eddes, J. S.; Simpson, R. J. Proteomics 2001, 1, 807-818. (13) Righetti, P. G. Immobilized pH Gradients: Theory and Methodology; Elsevier: Amsterdam, 1990. (14) Righetti, P. G.; Gelfi, C. J. Biochem. Biophys. Methods 1984, 9, 103-119. (15) Righetti, P. G.; Wenisch, E.; Faupel, M. J. Chromatogr. 1989, 475, 293309. (16) Righetti, P. G.; Wenisch, E.; Jungbauer, A.; Katinger, H.; Faupel. M. Chromatogr. 1990, 500, 681-696. (17) Herbert, B.; Righetti, P. G. Electrophoresis 2001, 21, 3639-3648. (18) Michel, P. E.; Reymond, P.; Arnaud, I. L.; Josserand, J.; Girault, H. H.; Rossier, J. S. Electrophoresis 2003, 24, 3-11. (19) Shapiro, A. L.; Vinuela, E.; Maizel, J. V., Jr. Biochem. Biophys. Res. Commun. 1967, 28, 815-829. (20) Chang, H. T.; Yergey, A. L.; Chrambach, A. Electrophoresis 2001, 22, 394398. (21) Buza`s, Z.; Chang, H. T.; Vieira, N. E.; Yergey, A. L.; Stastna, M.; Chrambach, A. Proteomics 2001, 1, 691-698. (22) Radko, S. P.; Chen, H. T.; Zakharov, S. F.; Bezrukov, L.; Yergey, A. L.; Vieira, N. E.; Chrambach, A. Electrophoresis 2002, 23, 985-992. (23) Antal, J.; Banyasz, B.; Buza`s, Z. Electrophoresis 2007, 28, 508-511.

polypeptide chains are swamped by the high negative charge of SDS micelles enveloping it. So much so that, for all practical purposes, the behavior of SDS-protein complexes, in free solution, is undistinguishable from that of DNA molecules, both requiring sieving matrixes for separation.24 The first dogma has been recently crushed by us, with the demonstration that indeed SDS-protein micelles could be focused, provided the positively charged counterions were grafted, gradient-wise, along the migration path of the analytes in an SDS-laden polyacrylamide gel.25 Steady-state conditions were demonstrated, since pattern stability could be guaranteed for at least 24 h of “focusing”. Our methodology appears to be a particular case of the general theory of Manning condensation,26 regarding the interaction between polymer brushes grafted onto a surface, on the one side, and oppositely charged polymers approaching them, on the other side. Although the novel method of SDS-PAGE focusing proposed by us had not been (and could have hardly been) predicted since the inception of SDS-PAGE up to the present, in a review published simultaneously with our report, Ivory27 suggested an idea that could, in a more general sense, have some resemblance with our approach, namely, the idea of “focusing in a gradient ion-exchange column”, although a mechanism of “focusing” appears to be more related to the creation of a conductivity gradient. In the present report, we hope we are providing enough evidence for smashing yet the other dogma, i.e., the impossibility of “focusing” SDS-protein micelles in a liquid phase. EXPERIMENTAL SECTION Chemicals and Materials. Basic Immobilines with various pK values (8.5, 9.3, 10.3, 12) were purchased from Amersham Pharmacia Biotech (Uppsala, Sweden). Fluka (Buchs, Switzerland) provided phosphoric acid, sodium hydroxide, acetic acid, Coomassie Brilliant Blue R, Tris, and Tricine. Acrylamide/bisacrylamide solution (Catalog No. 161-0156) was from Bio-Rad (Hercules, CA). Conventional SDS-Polyacrylamide Gel Electrophoresis. Electrophoresis of serum protein fractions separated by preparative SDS-PAGE focusing was performed in classical conditions using precast 10% commercial (Pierce) polyacrylamide gel plates. Cysteine alkylation was implemented by adding 10 mM free acrylamide in the presence of 5 mM fresh TBP, Three microliters of 2-fold diluted serum fractions was loaded per lane, and electrophoresis migration was performed at 200 V for 35 min. For MS identifications, the same samples were analyzed also by homemade SDS gels, cast as porosity gradients (8-18%T) in a discontinuous matrix in Laemmli28 sample buffer. Colloidal Coomassie Brilliant Blue G staining and destaining was achieved by using the method described by the supplier of reagents (Invitrogen Corp.). The migration plates were then stored dry after image acquisition with the Versa Doc system (Bio Rad). Preparative SDS-PAGE Focusing in Liquid Phases. For visualizing the method here described, one should refer to Figure 1, which gives a scheme of the MCE apparatus. This instrument (24) Andrews, A. T. Electrophoresis: Theory, Techniques and Biochemical and Clinical Applications; Clarendon Press: Oxford, 1986; pp 117-145. (25) Zilberstein, G.; Korol, L.; Antonioli, P.; Righetti, P. G.; Bukshpan, S. Anal. Chem. 2007, 79, 821-827. (26) Manning, G. S. Macromolecules 2001, 34, 4650-4655. (27) Ivory, C. Electrophoresis 2007, 28, 15-25. (28) Laemmli, U. K. Nature 1970, 227, 680-682.

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

8625

Figure 1. Upper drawing: scheme of the MCE instrument assembled with 10 chambers and (lower image) picture of the actual-size cell. The two terminal chambers are the electrodic reservoirs. The SDS-laden sample is typically applied in the first chamber next to the cathodic compartment, due to its anionic nature.

consists of 10 chambers, of which the two extremes represent the electrodic compartments. This first chamber next to the cathode is the sample loading chamber: here the sample is loaded in the presence of a 4% linear polyacrylamide solution and 100 mM Tris-Tricine buffer, pH 8.2, in 0.1% SDS. The solutions contained in the adjacent chambers are as follows: second chamber, liquid linear polyacrylamide (LPAA) copolymerized with 3 mM Immobiline of pK 10.3 in 70 mM Tris-Tricine buffer, pH 8.2, in 0.1% SDS; third chamber: LPAA with 8 mM Immobiline of pK 10.3 in 40 mM Tris-Tricine buffer, pH 8.2, in 0.1% SDS; fourth chamber, LPAA with 13 mM Immobiline of pK 10.3, dissolved in 25 mM Tris-Tricine buffer, pH 8.2, in 0.1% SDS; fifth chamber, LPAA with 18 mM Immobiline of pK 10.3, containing 15 mM TrisTricine buffer, pH 8.2, in 0.1% SDS; sixth chamber, LPAA with 23 mM Immobiline of pK 10.3, dissolved in 10 mM Tris-Tricine buffer, pH 8.2, in 0.1% SDS; seventh chamber, LPAA with 28 mM Immobiline of pK 10.3, in 2 mM Tris-Tricine buffer, pH 8.2, in 0.1% SDS (for a 9-chamber set-up). The distance between electrodes is ∼100 mm, and the apparatus is run at voltages not exceeding 10 V/cm for a period of 5 h. At the end of the run, the various fractions are harvested, and the protein is collected by trichloroacetic acid (TCA) precipitation and then analyzed by conventional SDS-PAGE. The bands excised from the analytical gel were digested with trypsin and subjected to MALDI-TOF MS analysis. MALDI-TOF MS Analysis. The protein zones of interest (see the boxed areas in Figure 3B) were excised from stained gels and subjected to in-gel trypsin digestion according to Shevchenko 8626

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

et al.29 The resulting peptides were loaded onto the target plate by mixing 1 µL of each solution with the same volume of a matrix solution (10 mg/mL R-cyano-4-hydroxycinnamic acid in 50% ethanol and 50% acetonitrile) and left to dry at room temperature. The MS analyses were performed by using a TofSpec-2E MALDITOF instrument (Micromass, Manchester, UK), equipped with a pulsed nitrogen laser (337 nm, pulse width 4 ns) and operated in reflectron mode with an accelerating voltage of 20 kV. Protein identification was performed by searching in the National Center for Biotechnology Information nonredundant database (NCBInr) by using the Mascot program (http://www.matrixscience.com). For positive identification, the score of the result had to be over the significance threshold (P < 0.05). RESULTS Figure 1A shows a drawing of the MCE chamber for preparative SDS-PAGE focusing, and Figure 1B is a picture of an actual size electrolyzer assembled. The various compartments are separated by septa made of 1% agarose, 1 mm thick and 10-mm diameter. These membranes are porous enough to allow migration of protein-SDS micelles, while impeding remixing of the LPAAImmobiline chains contained in the different sections. The electrolyzer can be assembled, accordion-wise, with a minimum of 3 chambers (the electrode vessels plus at least one sample compartment) up to 10 cells. Should the need arise for collecting a much large number of sample fractions, a device consisting of (29) Shevchenko, A.; Wilm, M.; Vorm, O.; Mann, M. Anal. Chem. 1996, 68, 850-858.

are arranged in order of their respective Mr values, in a cascade fashion, going from as high as 169 kDa down to as low as 14 kDa. This verifies the assumption that our MCE system operates on the basis of mass discrimination. The unique performance of our MCE device is also shown in Figure 4, which provides an additional example on the fractionation capability also on the “peptidome” range, i.e., in the Mr 10006000 Da, where SDS-PAGE in general fails. Here two commercial samples of insulin were fractionated, in two-separation chambers charged with pK 10 Immobiline polymers with a nonlinear concentration distribution optimized for the very low mass ranges 0-3 and 3-6 kDs. After fractionation, the content of the compartments was analyzed on a conventional (non-focusing) gel. As shown in Figure 4A and B, in both cases, the mass interval 0-3 kDa was neatly separated from the 3-6-kDa range. The results indicate that the insulin from Novo-Nordisk (A) is considerably different that the Russian sample (B). Figure 2. Drawing of a serpentine instrument comprising 49 cells for fraction collection. The shaded areas represent the agarose membranes.

as many as 49 cells, arranged in a serpentine fashion, has been built (Figure 2). In general, both chambers are operated at a maximum of 10 V/cm. This requirement stems from the observation that, at higher voltage gradients, the positively charged polyacrylamide chains would start migrating too in the electric field. In fact, whereas the SDS-protein micelles, once complexed with the LPAA-Immobiline chains of isocharge density, are expected to stop migration, due to mutual charge neutralization, it goes against the grain of basic physical laws that the liquid polyacrylamide chains, having a purely cationic surface, would remain stationary in an electric field. Quasi-stationary conditions are indeed achieved by a subtle interplay of driving forces in the electric field, frictional forces due to solution viscosity, and electrostatic repulsion forces among same-charge sign LPAAImmobiline polymers during polymer drift. Achieving steady-state conditions of the free polymer chains is also aided by arranging, at the start of the run, buffer solution conductivity gradients (from 100 to 2 mM Tris-Tricine; see Experimental Section), adjusted so that the field of the interface charge will be directed opposite to the external field. This mechanism enables the stationary distribution of the Immobiline polyvalent ions in the external electric field. Figure 3A shows the SDS-PAGE profiling of the six main fractions harvested from the MCE (see Figure 1) assembled with LPAA-Immobiline chains to capture specific Mr cuts in a human serum sample. In this image, one can see the ranges 1-10 (1), 10-20 kDa (2), followed by bands at 20-40 (3), 40-80 (4), 80120 (5), and >120 kDa (6). Such bands are seen as distinct, broad zones since the run has been performed in low %T polyacrylamide, in the absence of buffer discontinuities, in order to emphasize the discrete Mr cuts harvested. By using a homemade gel containing a porosity gradient (8-18%T) in the presence of buffer and gel discontinuities, much finer resolution was obtained when the various fractions were analyzed, as shown in Figure 3B. Some of the major bands (see boxed areas in Figure 3B) have been cut and subjected to identification via MALDI-TOF MS analysis. As shown in Table 1, it is seen that indeed the major zones identified

DISCUSSION General Experimental Aspects of the Present Technique. The novel method of preparative SDS-PAGE focusing in a liquid phase offers some unique aspects and some striking similarities with other recently introduced methodologies. We will first discuss the latter aspect: the MCE here developed seems to be fully complementary with the MCE operating under focusing conditions via isoelectric, buffering membranes. Figure 5 shows the modus operandi of the two systems: in classical MCE operating with the IPG technology, the membranes are the active units in the separation process, in that they act by titrating all species tangent to or crossing them, thus forcing all proteins, having isoelectric points (pI) comprised in between a set of two adjacent membranes, to be caught in the isoelectric trap defined by them. Thus, the cell delimited by a set of membranes acts only as a receptacle (or a trap) for species whose charge has been modulated by the walls delimiting them. Conversely, in our system, the membranes defining each compartment are not active in the separation process and are used solely to impede (or minimize) diffusion of the LPAA-Immobiline polymers contained in each cell. The active components driving the separation process are in fact the latter polymers, contained, in a stepwise gradient of charges, in each cell. Their action, as well-demonstrated by Zilberstein et al.25 and is based on the Manning condensation mechanism among polymers of opposite charge, leading to charge neutralization and thus to arrest of migration in the electric field via a peculiar type of “focusing” process. It is of interest, thus, to note that a single instrument can be used, according to the type of membranes and polymers utilized, either as a capturing process based on surface charge discrimination (pI) or on mass discrimination (Mr). The idea of discriminating and capturing SDS-protein micelles according to size in charged liquid polymers is novel and worth discussing in some detail. Per se, the concept of sieving in entangled liquid polymers is not novel and was amply demonstrated long ago by Bode.30 This principle was in fact fully adopted in capillary zone electrophoresis for size separation of DNAs31 as (30) Bode, H. J. Anal. Biochem. 1977, 83, 364-371. (31) Heiger, D. N.; Cohen, A. S.; Karger, B. L. J. Chromatogr. 1990, 516, 3348.

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

8627

Figure 3. (A) Analysis of six fractions collected in the MCE chamber via analytical SDS-PAGE in a Pierce precast gel. The chambers 1-6 of the MCE contained, respectively, 3, 8, 13, 18, 23, and 28 mM LPAA-Immobiline cationic polymers, selected for capturing the following Mr cuts: 1-10 (1), 10-20 kDa (2), followed by bands at 20-40 (3), 40-80 (4), 80-120 (5), and >120 kDa (6). (B) SDS-PAGE analysis in a 8-18% porosity gradient in discontinuous Laemmli buffer. The zones in the boxed areas were cut, digested with trypsin, and subjected to MALDI-TOF MS analysis. M: molecular mass markers. For identification of the major components in each analyzed fraction, see Table 1.

well as of SDS-protein micelles32 and is definitely preferred to gelling a firm polyacrylamide matrix into the capillary lumen. In (32) Karger, B. L.; Chu, Y. H.; Foret, F. Annu. Rev. Biophys. Biomol. Struct. 1995, 24, 579-610.

8628 Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

the latter case, in fact, distortion of the gel matrix would occur, leading to a rapid worsening of the separation process, also due to clogging of the pores caused by precipitated material at the gel/liquid interface. Conversely, entangled solutions of liquid

Table 1. MALDI-TOF MS Identification of the Main Components of Each Fraction Separated by Preparative SDS-PAGE Focusinga band

score

Z-score

hits

6 5 4 4 4 3 3 3 2 2 1

71,94 52,19 117.22 9.44 0.77 11.74 11.08 136.92 12.81 4.70 29.86

542,8 209,3 144.8 32.6 98.5 156.6 73.1 486.1 1083.9 316.1 1395.2

21 17 18 9 4 8 7 17 7 4 8

a

Accession number P01023 P00450 P01024 P02768 Q15642 P01009 P01024 P02647 P01024 P68871 P02766

protein ID

protein description

Mr (kDa)

pI

Cov (%)

A2MG_HUMAN (C_1) CERU_HUMAN (C_1) CO3_HUMAN (C_2) ALBU_HUMAN (C_1) CIP4_HUMAN (C_1) A1AT_HUMAN (C_1) CO3_HUMAN (C_7) APOA1_HUMAN (C_2) CO3_HUMAN (C_6) HBB_HUMAN (C_1) TTHY_HUMAN (C_1)

R-2-macroglobulin ceruloplasmin complement C3 β chain Serum albumin Cdc42-interacting protein 4 R-1-antitrypsin complement C3dg fragment apolipoprotein A-I(1-242) complement C3c fragment hemoglobin subunit β transthyretin

161 120 71 66 68 44 39 28 24 16 14

5.9 5.4 6.8 5.7 5.6 5.4 5.0 5.3 6.9 6.8 5.4

22 21 41 21 8 25 24 60 28 34 62

See Figure 3B.

Figure 4. Preparative fractionation of two insulin samples from Novo-Nordisk (A) and from the Russian manufacturer Ferane (B). The MCE was assembled with two separation chambers laden with Immobiline polymers with a nonlinear concentration distribution optimized for the very low mass ranges 0-3 and 3-6 kDa (0-8 mM non-linear gradient of pK 9.3 immobilines). The content of both chambers was analyzed by a SDS-PAGE focusing in a pK 9.3, 015 mM.

polymers can be refreshed after each run by expulsion from the capillary lumen under gentle operating pressure. Nobody, however, to our knowledge, has described entangled solutions of cationic polymers for a capture based not on size but on charge neutralization, and thus for a “focusing process”. Such a process has some distinct advantages for preparative purposes: first, the fact of using viscous polymer solutions permits high recoveries of the captured proteins, up to 90% (our data) as opposed to recoveries of 60-70% from gel matrixes.2 This is easily achieved by harvesting the solution in each chamber and precipitating the proteins in 10% TCA, a standard procedure for recovering proteins from cell extracts for removal of a variety of interfering substances.

Figure 5. Drawing of the two assemblies of a MCE device for (A) focusing of proteins according to their pI value with isoelectric Immobiline membranes; (B) focusing of proteins according to their Mr values with liquid cationic polymers.

Under these conditions, the bound SDS and the cationic polymers are released and remain in solution. The protein precipitate can the easily be resolubilized in, for example, TUC (thiourea-ureaCHAPS) buffer and processed in 2D maps or any other suitable 2D protocol, as routinely operated in proteome analysis.33 This procedure was used for estimating the recovery efficiency by testing with a single protein. In a number of separate experiments, the protein was recovered from its appropriate compartment and analyzed on a standard SDS gel and visualized with Commassie Blue. The total amount was estimated from the integrated optical density of the recovered sample measured on a laboratory scanner as compared with the initial amount. Typical recovery efficiency (33) Hamdan, H.; Righetti, P. G. Proteomics Today; Wiley: Hoboken, NJ, 2005: pp 341-405.

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

8629

in eight separate experiments was between 80 and 85% with maximum of 90%. The other advantage of working in soluble polymer solutions is the fact that such polymers can be easily prepared, precipitated in organic solvents, and washed free of all contaminating material typically contained in gel matrixes, such as ungrafted monomers and catalysts.34 In fact, one of the major problems in recovering proteins driven into gel matrixes (especially when performed by passive elution on crushed gel disks) is their contamination from such impurities, their level often vastly exceeding the amount of protein harvested. Additionally, the chain length (and thus viscosity) of these liquid polymers can be easily controlled by modulating the levels of catalysts and by temperature control (high levels of catalysts and progressively higher temperatures leading to shorter and shorter chains).35 Limits of SDS-PAGE Focusing. Some aspects of the present technique are worth discussing to a deeper extent. First, it might be argued if it is a genuine “focusing” technique or not, especially in regard to isoelectric focusing (or isopycnic centrifugation), where a true focusing process takes place, since the sample can be applied anywhere in the separation column, even uniformly distributed through it.6 The final steady state will be completely independent from the mode of sample application. Here, on the contrary, it should be underlined that the sample can only be applied at one point along the separation path, that is, at the point of minimum (positive) charge of the surrounding matrix (or liquid polymer): upon migration toward the other extreme of the column, i.e., toward the point of highest charge, each species in the sample will be captured and condensed at a point of “null net charge” of the complex. Thus, uniform sample loading or loading at the opposite extreme of the column will not be possible. From this point of view, perhaps, the term “focusing” might not be appropriate, as also pointed out by the reviewers, and in reality the technique should more correctly be called “arrested migration of an injected plug”. However, from the point of view of obtaining remarkably sharp zones and counteracting their diffusion, our method could be equated to a “focusing” process. The second point to be underlined is whether or not, in our preparative system, the charged liquid polymers (which could migrate into the electric (34) Righetti., P. G.; Gelfi, C. Anal. Biochem. 1997, 244, 195-207. (35) Barron, A.; Sunada, W. M.; Blanch, H. W. Electrophoresis 1996, 17, 597615.

8630

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

field if proper measures were not taken to keep them quasistationary) might be substituted with charged membranes, of appropriate charge density, just as done with isoelectric membranes in IPG-based MCEs.15,16 This, in principle, could be done, since here we work on stepwise charges (just like in IPG-based MCEs one does not work on a continuous, but on a stepwise pH gradient), but with a major limitation: the load capacity. If the sample ions were “arrested” and condensed onto such charged membranes, the load capacity would be limited by the total amount of positive charges present in such membranes. Once this total charge would be neutralized by the adsorbed sample ions, the excess protein would break through and spill over the following chambers and this would occur also for samples of lower charge density (as compared to that of the membrane), meant to be trapped and confined within a given chamber. Having the charged polymers present in the liquid phase within each chamber of the MCE ensures a much higher total charge and thus higher sample loads. CONCLUSIONS We believe that the present data on preparative SDS-PAGE focusing in liquid cationic polymer chains could represent a new tool in proteome prefractionation leading to improved discovery of low-abundance proteins, often masked (such as in sera, but in reality in most cell lysates as well) by the dominating presence of a minority of very high-abundance species. It goes without saying that our technique (both at the analytical and preparative scale) is performing very well also in DNA analysis, in amplifying the separation in compressed regions of DNA ladders as well as in purifying given DNA fragments contaminated by adjacent species (Zilberstein et al., manuscript in preparation). ACKNOWLEDGMENT P.G.R. is supported by grants from MURST (PRIN 2006), by Fondazione Cariplo and by the European Community (project Allergy card). We thank Dr. P. Antonioli for help with the SDSPAGE analysis and Dr. Marco Cantu` (Parco Tecnologico Padano, Lodi) for help with the MS analysis. Received for review July 30, 2007. Accepted August 29, 2007. AC701598Y