Electrophoresis 1991,12, 55-58
Pier Giorgio Righetti Claudia Ettori Marcella Chiari Department of Biomedical Sciences and Technologies, University of Milano, Italy
CZE of acrylamido-buffers for IEF
55
Analysis of acrylamido-buffers for isoelectric focusing by capillary zone electrophoresis Immobilized pH gradients use a series of weak acrylamido acids and bases (Immobiline) to create a pH gradient along the separation axis. These buffers can be degraded in water by two mechanisms: (i) hydrolysis of the amido bond, with generation of free acrylic acid and either an amino acid or a diamine; (ii) autopolymerization to oligomers and/or n-mers. In order to check for these degradation products, different capillary zone electrophoresis systems for analysis of all Immobilines have been devised. The acidic compounds are resolved in 100 mM acetate, pH 4.0, whereas the alkaline Immobilines are separated in 50 mMphosphate buffer, pH 7.7 (or pH 7.2 for the weaker species). Polymers of alkaline Immobilines are resolved in 50 mM phosphate buffer, pH 2.5, in 1 % Ficoll-400. All lmmobilines are detected underivatized, by their adsorption at 2 14 or 254 nm. A calibration curve has been constructed for quantification of acrylic acid contamination. As little as 1 mol% of acrylic acid contamination in Immobiline solutions can be detected, with a sensitivity limit below 0.2 mM (at the injection port).
1 Introduction Isoelectric focusing (IEF) in immobilized pH gradients (IPG) represents perhaps the most powerful development in electrokinetic separations, with an unrivaled resolving power and a high load capacity in preparative runs [ 11. The power and the precision of IPG relies on the quality of the buffers used to generate and maintain the pH gradient in the electric field. Unlike conventional IEF, where the pH gradient is obtained by electrophoretic sorting of a vast number of soluble amphoteric buffers, called carrier ampholytes (for a review see [21), the IPG technique uses a set of few, well-defined chemicals available commercially as crystalline powders or liquids. Wehaverecently decodedthe structure andgiven theformulas of the acidic [3] and basic [4] Immobiline chemicals. Furthermore, over the years, we have proposed a number of additional chemicals to expand the fractionation ability of IPGs: both more acidic [5, 61 and more alkaline [7] compounds have been produced in our laboratory. We have also synthesized analogues of the weakest Immobiline bases (the niorpholino derivatives, with pK values of 6.2 and 7.0) by introducing a thiomorpholino ring; the pK values of these compounds were increased to pK 6.6 and 7.4, respectively, thus offering additional species buffering around neutrality, i. e. in a region which normally lacks suitable buffering groups and where the bulk water conductivity reaches a minimum [81. A new, hydrophilic Immobiline with a pK of 8.05 has also been synthesized recently, in order to close the gap in the pH 7.0-8.5 region [ 91. Thus, as the family ofacrylamido-buffers is expanding (we have described now a total of 14 monoprotic compounds and there is a report on a biprotic species, itaconic acid) [ 101, there is a need for a rapid and sensitive screening test to check their purity and potential degradation with time. In an extensive investigation, we found that Immobilines, to varying degrees, were subjected to two degradation pathways: (i) hydrolysis of the amido bond, producing free acrylic acid and an amino acid(for the acidic species) or adiamine (for the alkaline compounds); (ii) spontaneous autopolymerizaCorrespondence:Prof. P. G. Righetti, University of Milano, Via Celoria 2, Milano 20133, Italy Abbreviations:CTAB, cetyltrimethylammonium bromide; CZE,capillary zone electrophoresis; HPLC, high performance liquid chromatography; IEF, isoelectric focusing; IPG, immobilized pH gradients 0VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 199 J
tion, producing a number of oligomers up to n-mers, able to aggregate and precipitate large proteins I 111. Both phenomena were largely abolished by proposing a new formulation in solution: the alkaline Immobilines are dissolved in npropanol, while the acidic species are prepared in water laced with low concentrations of an inhibitor [ 111. Yet a need for a rapid and sensitive analysis method still exists since these acrylamido-buffers have to be transferred to water for gel preparation and are often stored as water solutions in laboratories handling large amounts of gels. Due to its high resolving power and speed of operation, capillary zone electrophoresis (CZE) seemed to be the technique of choice (for a recent collection of research articles on CZE, see [12]). We report here the CZE analysis of all presently available Immobilines.
2 Materials and methods 2.1 Materials Commercial Immobilines were purchased from Pharmacia-LKB Biotechnology, Bromma, Sweden. Noncommercial acrylamide weak acids and bases were synthesized in our laboratory as reported in the appropriate references [ 5-91. Polyethylene glycol (PEG)-35 000 was from Fluka (Buchs, Switzerland), while hydroxymethyl cellulose (HMC), Ficoll400 (400 000 D a average mass) and polyvinylpyrrolidone 360 (360 000 Da average mass) were from Sigma (St. Louis, MO). These additives were used as 0.1 % (HMC), 1 % (Ficoll) or 5 % (PEG-35) solutions in CZE. Acrylic acid was from Fluka and was distilled just prior to use. Cetyltrimethylammonium bromide (CTAB) was from Aldrich (Steinheirn, Germany). 2.2 Methods 2.2.1 CZE CZE was performed in a Beckman (Palo Alto, CA) instrument (P/ACE System 2000) equipped with a 50 cm long capillary of 50 pm diameter. All runs were performed at 25 OC in a thermostatted environment. Five types of runs were performed: (i) for the acidic Immobilines, 100 mM acetate buffer, Ol73-0835/9l/OlOl-0055 %3.50+.25/0
56
EIeclrophoresis 1991,12,55-58
P. G. Righetti efal.
pH 4.0; conditions: 20 kV, 23 PA; (ii) for the commercial basic Immobilines: 50 mM phosphate buffer, pH 7.7; conditions: 13 kV, 80 PA; (iii) for the weakly basic Immobilines: 50 mMphosphate buffer, pH 7.0; conditions: 20 kV, 100 PA; (iv) for acrylic acid contamination in alkaline Immobiline species: 50 mM phosphate buffer, pH 7.0, plus 2 mM CTAB; run: 20 kV, 100 PA; for the acrylic acid calibration curve: 100 mM acetate, pH 4.0, + 2 mM CTAB; run: 20 kV, 23 PA; (v) for polymer analysis: 50 mM phosphate buffer, p H 2.5, with addition of 1 % Ficoll-400; conditions: 25 kV, 74 PA. In all cases (except for Figs. 4 and 5 ) the migration direction was toward the negative electrode, which means that the acidic Immobilines are transported there by electroosmosis, as they migrate electrophoretically toward the positive electrode. The sample was injected into the capillary by pressure from a nitrogen tank (approximately 80-85 psi), usually for 5 s. The calibration curve for acrylic acid was constructed with the Beckman integration system Gold.
odic transport strongly competes with the cathodic electroosmotic flow. To a reduced extent, this is also true for pK 3.1. In fact, these two compounds show fronting, as opposed to a small degree of tailing, in the other peaks. Figure 2 shows the separation of the 4 commercially available basic Immobilines in phosphate buffer, pH 7.7. In this case, both the electrophoretic migration and electroosmotic flow are in the same direction, and the peaks are sharp, with a much reduced separation time (less than 5 min, as opposed to 15 min in Fig. 1). However (see Fig. 9, when running a mixture of acidic and basic Immobilines, the polarity is reversed and CTAB is added to suppress electroosmosis. As we have reIP'
7.0
2.2.2 Preparation of Immobiline oligomers Oligomers of alkaline Immobiline buffers were prepared by incubating 0.2 M water solutions with 0.4 % ammonium persulfate. The reaction was allowed to proceed overnight at room temperature in an air-equilibrated vessel [ 131.
3 Results Figure 1 shows the separation of 5 acidic Immobilines (pK values 4.6,4.4,3.6,3.1 and 1.0) and of a contaminant, acrylic acid, added to the mixture. It is seen that the conditions used are able to fully separate these species, and are clearly optimal for the weaker acids (pK 4.6,4.4, 3.6 and acrylic acid). Due to its high mobility, the pK 1.0 (2-acrylamido-2-methylpropanesulfonic acid) peak is highly skewed because its an-
,
D K ..I
, 90
Tbme ( m i " )
A -
9
E: Figure 2. Separation of alkaline Immobilines by CZE. A mixture of 2.5 m M each of the Immobilines pK 9.3, 8.5, 7.0 and 6.2 was run in the P/ACE System 2000 in 50 mM phosphate buffer, pH 7.7, at 13 kV and 80 pA. All other conditions as in Fig. 1. Cathodic migration.
*... 0
PI,..
,
.pl(
7.0
E
0
N
a E
-
P
N U
?
l 9
L
!
l
L
L
~
~
~
~
~
-
I
'
I
~
I
~
I
L
I
-
I
L
I
-
I
Time (mi")
FigureI. AnalysisofacidicImmobilinesbyCZE.Amixtureof2.5 mMeach of the Immobilines pK 4.6,4.4,3.6,3.1 and f .O was introduced by pressure (5 s)in a50 pmcapillary, 50cmlong,intheBeckmanP/ACESystem2000. Therunwasat25 T i n 100 mMacetatebuffer,pH4.0,at20 kVand23 pA. Detection was by UV absorption at 2 14 nm. The sample was contaminated with 2.5 mM acrylic acid (Acr. A.). Migration toward the cathode by electroosmosis. Here and in the following figures the different peaks have been identified by injecting the Immobiline species one at a time.
Figure 3. Separation of weakly basic Immobilines by CZE. A mixture of 2.5 mM each of the Immobilines pK 7.4, 7.0, 6.6 and 6.2 was run in the P/ACE System 2000 in 50 mM phosphate buffer, pH 7.0, at 20 kV, 100 pA. The pK 7.4 and 6.6 buffers are the thiomorpholino derivatives of the commercial morpholino species (pK 7.0 and 6.2, respectively).All other conditions as in Fig. 1. Cathodic migration.
Electrophoresis 1991, 12, 55-58
CZE of acrylamido-buffers for IEF
57
cently synthesized two analogues of the weak bases pK 6.2 and pK 7.0 (with a thiomorpholino substituting for the morpholino ring) it was of interest to see the separation of these compounds as well. By lowering the operative pH, this analysis is easily accomplished (Fig. 3). Note that, as the pK values are regularly spaced at 0.4 pH unit intervals, so is the peak distance in the electropherogram. For laboratory tests, one needs to know the lowest amount of acrylic acid which can be detected and quantified in Immobiline preparations. For this purpose, we have constructed a calibration curve by injecting serial dilutions of acrylic acid, ranging from 10-0.2 mM. As shown in Fig. 4, good linearity and goodcorrelation werefound(curve:y= 1 . 6 1 3 +0.0677; ~ coefficient of determination: 0.986). On this basis, a solution of Immobiline containing 1 mol % of acrylic acid was analyzed in CZE. When a solution of20 mMImmobiline pK 6.2 (con-
N
!I--- --___
I
.L_L
>__(_ I
I _ _ L L I 2 - L - L L - I - 1
9
L L - . I _ I - U I I _ L - _ I - Y
T l m e (rnnn)
Figure 6 . Monitoring for polymer formation in Immobiline chemicals. Control (lower panel) and polymerized pK 8.5 Immobiline (upper profile) were analyzed by C Z E in 50 mM phosphate buffer, pH 2.5, supplemented with 1 'YOFicoll-400. The run was at 25 OC,25 kV and 74 FA. UV readings at 254 nm. Cathodic migration.
__t___-
~
1
0
3
2
6
5
(I
Rrea
Figure4. Calibration curveofacrylic acidinCZE. Serialdilutionsoffreshly distilled acrylic acid (from 10to 0.2 mM) were run in P/ACE system 2000in 100mMacetate buffer,pH4.0,supplementedwith2mMCTAB.Integration and linear fit obtained with the Beckman system Gold developed for HPLC. Anodic migration (reversed polarity).
taining 1 mol %, i. e. a total of 0.2 mM acrylic acid) is injected, a spike corresponding to the acrylic acid position can be clearly seen (Fig. 5 ) . If the zone corresponding to acrylic acid is magnified (by a factor of 10 on the absorbance axis, and by a factor of five on the time axis) the spike can be clearly seen as a peak (see inset in Fig. 5). Similar results have been obtained with all other basic Immobilines (not shown). The problem of spotting oligomers was more complex. It can easily be done by polymerizing a gel in the capillary [ 141, but this procedure is particularly delicate because minute air bubbles could be trapped, thus breaking the electric circuit. We have preferred to fill the capillary with a liquid polymer, as suggested in 1151. As shown in Fig. 6, the oligomers can be seen as retarded peaks. In the absence of a liquid polymer, it was impossible to separate the monomer from the oligomers (not shown), suggesting that the mobility in free phase is almost identical (only a peak broadening is observed). In the lower part of Fig. 6, the profile of the control, unpolymerized pK 8.5 Immobiline, is visible.
4 Discussion 4.1 On the analysis of acrylic acid traces in Immobilines .JL*+
I
& L ,
-
9
-
~
.
.
-
~
.
~
-
- - . Tt me (m
~ I n)
-
~
-
-
~
~
-
-.-, 9
Figure 5. Monitoring for acrylic acid contamination in alkaline Imrnobilines. A 20 mM solution of pK 6.2 Immobiline was supplemented with 0.2 mMqrylic acid andinjectedintoP/ACEsystem2000(5sbypressure). Run in 50 mM phosphate, p H 7.0, and 2 mM CTAB at 20 kV, 100 FA. The insert shows the acrylic acid (Acr. A,) spike magnified by a factor of 10 on the y-axis (from 100to 10 milli-abs) and by a factor o f 4 on the x-axis (from 20 to 5 min). The spike was integrated with the aid ofthe calibration curve in Fig. 4. Anodic migration (reversed polarity).
The C Z E methodology appears to be valuable for checking the purity and stability of Immobiline chemicals that are either commercially available or produced in the laboratory. Given the exquisite resolution and the extreme reproducibility of the IPG technique, it is worth while having a fast and reliable technique at hand te check any anomalies in the Immobiline chemicals, as they are the ultimate cause of trouble in IPG runs. Originally, when checking for degradation products in Immobiline chemicals, Giveby et al. [ 1 11 had proposed two techniques: (i) gas chromatography for quantitation of traces of acrylic acid and (ii) high performance liquid chromatography (HPLC, in the gel permeation mode) for
58
Ekctrophoresis 1991,12,55-58
P. G. Righetti el al.
oligomer and n-mer detection. CZE compares favorably with both methods. For instance, in the case of acrylic acid analysis, Giveby et al. [ 111 have detected 0.76 mol % as a well-visible peak and 0.10 mol% as a minute spike in the chromatogram of different Immobilines. In the former case, when the amount of Immobiline injected was 0.2 M,the total amount present in the capillary was 1.52 mM; in the latter case, the lowest amount detected was 0.2 mM, i. e. the same order of magnitude as our detection limit. In reality, CZE could offer a much higher detection limit if one were to resort to sample derivation and detection by laser-induced fluorescence. In the latter case, detection limits as low as 500 attomoles have been reported [ 161. In terms of analysis times, CZE also compares favorably with gas chromatography: whereas Ghveby et al. [ l l ] had reported times of 20-30 min for gas chromatography of different Immobilines, in our case running times as short as 5-10 min are sufficient (even for complex mixtures of Immobilines). 4.2 On the analysis of polymers in Immobilines When we first reported the presence of polymers in Immobiline solutions in 1987, the data met with skepticism and incredulity. However, this finding solved one ofthe main problems of IPGs, namely the disappearance of large proteins from analytes. As a model, we proposed a cross-linking of these large proteins by the Immobiline oligomers, with production of a large precipitating reticulum, as in antigen-antibody reactions [171. Our data were then fully confirmed by Giveby et al. [ 111 and by the fact that the removal of such oligomers completely prevented protein precipitation [ 161. For the analysis of oligomers and n-mers in Immobiline solutions, Giveby et al. [ 111 proposed gel permeation chromatography in HPLC columns: as seen in their Figs. 5 and 6, excellent separations are obtained, with polymer detection at 2 10 nm (because double bonds would no longer be contributing to the absorbance). In our report (see Fig. 6), it is seen that polymer separation can be achieved by incorporating asoluble,neutral polymer in the background electrolyte, as already reported by Zhu et al. [ 151. However, we propose a different polymer, i. e. Ficoll-400, a highly soluble polysaccharide well known for cell-fractionations in both rate-zonal and isopycnic separations [181. We had also tried another soluble polymer, i. e. oolyvinylpyrrolidone 360, but its high absorbance at 214 nm (2.7 for a 1 % solution) prevented any meanigful reading. Also, Ficoll-400 has an appreciable absorbance at 214 nm (0.340 for a 1 % solution) but only 0.170 at 254 nm (in addition, this absorbance refers to a 1 cm path length, whereas the optical channel in CZE is only 50 pm). Performing molecular sieving in the absence of a gel (typically polyacrylamide) is quite convenient in CZE, since the polymerization of a gel in such a narrow bore (e, g., 25-50 pm) is inconvenient and does not yield reproducible results. The idea that soluble polymers could introduce molecular sieving (a concept usually associated only with gel matrices) has had a long gestation period and was first proposed in 1980 by Bode [191, who suggested that polyacrylamide gels behave like an “extended viscoelastic continuum” and a “viscosity emulsion”. In this analogy, sieving in gels was seen as due to the effects of elastic
repulsion, which, in turn, are dueto spontaneous reorientation of the polymer chains which can fluctuate in the surrounding space. The same author has provided a considerable body of evidence showing molecular sieving in mixtures of linear, uncrosslinked polymers and gels, such as polyethyleneglycol in cellulose acetate [201 or polyacrylamide in agarose (211. There are still some minor peaks in the electropherograms which are not yet identified. Their analysis and characterization is in progress.
Supported in part by grants from Consiglio Nazionale delle Ricerche (Roma), Progetto Finalizzato Chimica Fine II, and by Minister0 della Pubblica Istruzione. W e thank Beckman Italy (and in particular Drs. R . Montini and S . Di Biase)for the kind loan of the equipment and for support and valuable suggestions. Received June 20,1990
5 References [ 11 Righetti, P. G., ImmobilizedpHGradients: TheoryandMethodology,
Elsevier, Amsterdam 1990. [21 Righetti, P. G.. Isoelectric Focusing: Theory, Methodology and Applications, Elsevier, Amsterdam, 1983. [31 Chiari, M., Casale, E., Santaniello, E. and Righetti, P. G., ppl. Theor. Elecfrophoresis 1989, 1, 99-102. [41 Chiari, M., Casale,E., Santaniello,E. and Righetti, P. G.,Appl. Theor. Electrophoresis 1989, I , 103-107. [51 Gianazza, E., Celentano, F., Dossi, G., Bjellqvist, B. and Righetti, P. G., Electrophoresis 1984,5,88-97. [61 Righetti, P. G., Chiari, M., Sinha, P. K. and Santaniello, E., J. Biochern. Biophys. Methods 1988,16, 185-192. [71 Gelfi, C., Bossi, M. L., Bjellqvist, B. and Righetti, P. G.,J. Biochem. Biophys. Methods 1987, I5,41-48. [Sl Chiari, M., Righetti, P. G., Ferraboschi, P., Jain, T. and Shorr, R., Electrophoresis 1990, 11, 617-620. (91 Chiari, M., Pagani, L., Righetti, P. G., Jain, T., Shorr, R. and Rabilloud, T., J. Biochem. Biophys. Methods 1990,21, 165-172. 101 Charlionet, R., Sesbouk, R. and Davrinche, C.,Electrophoresis 1984, 5, 176-178. 111 Giveby, B. M., Pettersson, P., Andrasko, J., Ineva-Glygare, L., Johannesson, U., Gorg, A., Postel, W., Domscheit, A., Mauri, P. L., Pietta, P., Gianazza, E. and Righetti, P. G., J. Biochem. Biophys. Methods 1988,16, 141-164. 121 Karger, B. (Ed.), Proceedings of the lst International Symposium on High Performance Capillary Electrophoresis, J. Chromatugr. 1989, 480, 1-435; 1990,516, 1-298. I131 Rabilloud, T., Pernelle, J. J., Wharmann, P., Gelfi, C. and Righetti, P. G., J. Chromatogr. 1987,402, 105-1 13. (141 Karger, B. L., Cohen, A. S. andGuttman, A.,J. Chromatogr. 1989, 492,585-614. (151 Zhu, M., Hansen, H. S. and Gannon, F.,J. Chromatogr. 1989,480, 3 11-319. [161 Chung, Y . F. and Dovichi, N. J., Science 1988,242, 562-564. [171 Righetti, P. G., Gelfi, C., Bossi, M. L. and Boschetti, E., Electrophoresis 1987,8, 62-70. [ 181 Rickwood, D. (Ed.), Centrifugation: A Practical Approach, IRL Press, Oxford 1984, pp. 3 1-32. [191 Bode, H. J., in: Radola, B. J. (Ed.), Electrophoresis ’79, de Gruyter, Berlin 1980, pp. 39-52. [201 Bode, H. J., Anal. Biochem. 1977,83,204-210. [211 Bode, H. J., Anal. Biochem. 1977,83,364-371.
Pier Giorgio Righetti Claudia Ettori Marcella Chiari Department of Biomedical Sciences and Technologies, University of Milano, Italy
CZE of acrylamido-buffers for IEF
55
Analysis of acrylamido-buffers for isoelectric focusing by capillary zone electrophoresis Immobilized pH gradients use a series of weak acrylamido acids and bases (Immobiline) to create a pH gradient along the separation axis. These buffers can be degraded in water by two mechanisms: (i) hydrolysis of the amido bond, with generation of free acrylic acid and either an amino acid or a diamine; (ii) autopolymerization to oligomers and/or n-mers. In order to check for these degradation products, different capillary zone electrophoresis systems for analysis of all Immobilines have been devised. The acidic compounds are resolved in 100 mM acetate, pH 4.0, whereas the alkaline Immobilines are separated in 50 mMphosphate buffer, pH 7.7 (or pH 7.2 for the weaker species). Polymers of alkaline Immobilines are resolved in 50 mM phosphate buffer, pH 2.5, in 1 % Ficoll-400. All lmmobilines are detected underivatized, by their adsorption at 2 14 or 254 nm. A calibration curve has been constructed for quantification of acrylic acid contamination. As little as 1 mol% of acrylic acid contamination in Immobiline solutions can be detected, with a sensitivity limit below 0.2 mM (at the injection port).
1 Introduction Isoelectric focusing (IEF) in immobilized pH gradients (IPG) represents perhaps the most powerful development in electrokinetic separations, with an unrivaled resolving power and a high load capacity in preparative runs [ 11. The power and the precision of IPG relies on the quality of the buffers used to generate and maintain the pH gradient in the electric field. Unlike conventional IEF, where the pH gradient is obtained by electrophoretic sorting of a vast number of soluble amphoteric buffers, called carrier ampholytes (for a review see [21), the IPG technique uses a set of few, well-defined chemicals available commercially as crystalline powders or liquids. Wehaverecently decodedthe structure andgiven theformulas of the acidic [3] and basic [4] Immobiline chemicals. Furthermore, over the years, we have proposed a number of additional chemicals to expand the fractionation ability of IPGs: both more acidic [5, 61 and more alkaline [7] compounds have been produced in our laboratory. We have also synthesized analogues of the weakest Immobiline bases (the niorpholino derivatives, with pK values of 6.2 and 7.0) by introducing a thiomorpholino ring; the pK values of these compounds were increased to pK 6.6 and 7.4, respectively, thus offering additional species buffering around neutrality, i. e. in a region which normally lacks suitable buffering groups and where the bulk water conductivity reaches a minimum [81. A new, hydrophilic Immobiline with a pK of 8.05 has also been synthesized recently, in order to close the gap in the pH 7.0-8.5 region [ 91. Thus, as the family ofacrylamido-buffers is expanding (we have described now a total of 14 monoprotic compounds and there is a report on a biprotic species, itaconic acid) [ 101, there is a need for a rapid and sensitive screening test to check their purity and potential degradation with time. In an extensive investigation, we found that Immobilines, to varying degrees, were subjected to two degradation pathways: (i) hydrolysis of the amido bond, producing free acrylic acid and an amino acid(for the acidic species) or adiamine (for the alkaline compounds); (ii) spontaneous autopolymerizaCorrespondence:Prof. P. G. Righetti, University of Milano, Via Celoria 2, Milano 20133, Italy Abbreviations:CTAB, cetyltrimethylammonium bromide; CZE,capillary zone electrophoresis; HPLC, high performance liquid chromatography; IEF, isoelectric focusing; IPG, immobilized pH gradients 0VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 199 J
tion, producing a number of oligomers up to n-mers, able to aggregate and precipitate large proteins I 111. Both phenomena were largely abolished by proposing a new formulation in solution: the alkaline Immobilines are dissolved in npropanol, while the acidic species are prepared in water laced with low concentrations of an inhibitor [ 111. Yet a need for a rapid and sensitive analysis method still exists since these acrylamido-buffers have to be transferred to water for gel preparation and are often stored as water solutions in laboratories handling large amounts of gels. Due to its high resolving power and speed of operation, capillary zone electrophoresis (CZE) seemed to be the technique of choice (for a recent collection of research articles on CZE, see [12]). We report here the CZE analysis of all presently available Immobilines.
2 Materials and methods 2.1 Materials Commercial Immobilines were purchased from Pharmacia-LKB Biotechnology, Bromma, Sweden. Noncommercial acrylamide weak acids and bases were synthesized in our laboratory as reported in the appropriate references [ 5-91. Polyethylene glycol (PEG)-35 000 was from Fluka (Buchs, Switzerland), while hydroxymethyl cellulose (HMC), Ficoll400 (400 000 D a average mass) and polyvinylpyrrolidone 360 (360 000 Da average mass) were from Sigma (St. Louis, MO). These additives were used as 0.1 % (HMC), 1 % (Ficoll) or 5 % (PEG-35) solutions in CZE. Acrylic acid was from Fluka and was distilled just prior to use. Cetyltrimethylammonium bromide (CTAB) was from Aldrich (Steinheirn, Germany). 2.2 Methods 2.2.1 CZE CZE was performed in a Beckman (Palo Alto, CA) instrument (P/ACE System 2000) equipped with a 50 cm long capillary of 50 pm diameter. All runs were performed at 25 OC in a thermostatted environment. Five types of runs were performed: (i) for the acidic Immobilines, 100 mM acetate buffer, Ol73-0835/9l/OlOl-0055 %3.50+.25/0
56
EIeclrophoresis 1991,12,55-58
P. G. Righetti efal.
pH 4.0; conditions: 20 kV, 23 PA; (ii) for the commercial basic Immobilines: 50 mM phosphate buffer, pH 7.7; conditions: 13 kV, 80 PA; (iii) for the weakly basic Immobilines: 50 mMphosphate buffer, pH 7.0; conditions: 20 kV, 100 PA; (iv) for acrylic acid contamination in alkaline Immobiline species: 50 mM phosphate buffer, pH 7.0, plus 2 mM CTAB; run: 20 kV, 100 PA; for the acrylic acid calibration curve: 100 mM acetate, pH 4.0, + 2 mM CTAB; run: 20 kV, 23 PA; (v) for polymer analysis: 50 mM phosphate buffer, p H 2.5, with addition of 1 % Ficoll-400; conditions: 25 kV, 74 PA. In all cases (except for Figs. 4 and 5 ) the migration direction was toward the negative electrode, which means that the acidic Immobilines are transported there by electroosmosis, as they migrate electrophoretically toward the positive electrode. The sample was injected into the capillary by pressure from a nitrogen tank (approximately 80-85 psi), usually for 5 s. The calibration curve for acrylic acid was constructed with the Beckman integration system Gold.
odic transport strongly competes with the cathodic electroosmotic flow. To a reduced extent, this is also true for pK 3.1. In fact, these two compounds show fronting, as opposed to a small degree of tailing, in the other peaks. Figure 2 shows the separation of the 4 commercially available basic Immobilines in phosphate buffer, pH 7.7. In this case, both the electrophoretic migration and electroosmotic flow are in the same direction, and the peaks are sharp, with a much reduced separation time (less than 5 min, as opposed to 15 min in Fig. 1). However (see Fig. 9, when running a mixture of acidic and basic Immobilines, the polarity is reversed and CTAB is added to suppress electroosmosis. As we have reIP'
7.0
2.2.2 Preparation of Immobiline oligomers Oligomers of alkaline Immobiline buffers were prepared by incubating 0.2 M water solutions with 0.4 % ammonium persulfate. The reaction was allowed to proceed overnight at room temperature in an air-equilibrated vessel [ 131.
3 Results Figure 1 shows the separation of 5 acidic Immobilines (pK values 4.6,4.4,3.6,3.1 and 1.0) and of a contaminant, acrylic acid, added to the mixture. It is seen that the conditions used are able to fully separate these species, and are clearly optimal for the weaker acids (pK 4.6,4.4, 3.6 and acrylic acid). Due to its high mobility, the pK 1.0 (2-acrylamido-2-methylpropanesulfonic acid) peak is highly skewed because its an-
,
D K ..I
, 90
Tbme ( m i " )
A -
9
E: Figure 2. Separation of alkaline Immobilines by CZE. A mixture of 2.5 m M each of the Immobilines pK 9.3, 8.5, 7.0 and 6.2 was run in the P/ACE System 2000 in 50 mM phosphate buffer, pH 7.7, at 13 kV and 80 pA. All other conditions as in Fig. 1. Cathodic migration.
*... 0
PI,..
,
.pl(
7.0
E
0
N
a E
-
P
N U
?
l 9
L
!
l
L
L
~
~
~
~
~
-
I
'
I
~
I
~
I
L
I
-
I
L
I
-
I
Time (mi")
FigureI. AnalysisofacidicImmobilinesbyCZE.Amixtureof2.5 mMeach of the Immobilines pK 4.6,4.4,3.6,3.1 and f .O was introduced by pressure (5 s)in a50 pmcapillary, 50cmlong,intheBeckmanP/ACESystem2000. Therunwasat25 T i n 100 mMacetatebuffer,pH4.0,at20 kVand23 pA. Detection was by UV absorption at 2 14 nm. The sample was contaminated with 2.5 mM acrylic acid (Acr. A.). Migration toward the cathode by electroosmosis. Here and in the following figures the different peaks have been identified by injecting the Immobiline species one at a time.
Figure 3. Separation of weakly basic Immobilines by CZE. A mixture of 2.5 mM each of the Immobilines pK 7.4, 7.0, 6.6 and 6.2 was run in the P/ACE System 2000 in 50 mM phosphate buffer, pH 7.0, at 20 kV, 100 pA. The pK 7.4 and 6.6 buffers are the thiomorpholino derivatives of the commercial morpholino species (pK 7.0 and 6.2, respectively).All other conditions as in Fig. 1. Cathodic migration.
Electrophoresis 1991, 12, 55-58
CZE of acrylamido-buffers for IEF
57
cently synthesized two analogues of the weak bases pK 6.2 and pK 7.0 (with a thiomorpholino substituting for the morpholino ring) it was of interest to see the separation of these compounds as well. By lowering the operative pH, this analysis is easily accomplished (Fig. 3). Note that, as the pK values are regularly spaced at 0.4 pH unit intervals, so is the peak distance in the electropherogram. For laboratory tests, one needs to know the lowest amount of acrylic acid which can be detected and quantified in Immobiline preparations. For this purpose, we have constructed a calibration curve by injecting serial dilutions of acrylic acid, ranging from 10-0.2 mM. As shown in Fig. 4, good linearity and goodcorrelation werefound(curve:y= 1 . 6 1 3 +0.0677; ~ coefficient of determination: 0.986). On this basis, a solution of Immobiline containing 1 mol % of acrylic acid was analyzed in CZE. When a solution of20 mMImmobiline pK 6.2 (con-
N
!I--- --___
I
.L_L
>__(_ I
I _ _ L L I 2 - L - L L - I - 1
9
L L - . I _ I - U I I _ L - _ I - Y
T l m e (rnnn)
Figure 6 . Monitoring for polymer formation in Immobiline chemicals. Control (lower panel) and polymerized pK 8.5 Immobiline (upper profile) were analyzed by C Z E in 50 mM phosphate buffer, pH 2.5, supplemented with 1 'YOFicoll-400. The run was at 25 OC,25 kV and 74 FA. UV readings at 254 nm. Cathodic migration.
__t___-
~
1
0
3
2
6
5
(I
Rrea
Figure4. Calibration curveofacrylic acidinCZE. Serialdilutionsoffreshly distilled acrylic acid (from 10to 0.2 mM) were run in P/ACE system 2000in 100mMacetate buffer,pH4.0,supplementedwith2mMCTAB.Integration and linear fit obtained with the Beckman system Gold developed for HPLC. Anodic migration (reversed polarity).
taining 1 mol %, i. e. a total of 0.2 mM acrylic acid) is injected, a spike corresponding to the acrylic acid position can be clearly seen (Fig. 5 ) . If the zone corresponding to acrylic acid is magnified (by a factor of 10 on the absorbance axis, and by a factor of five on the time axis) the spike can be clearly seen as a peak (see inset in Fig. 5). Similar results have been obtained with all other basic Immobilines (not shown). The problem of spotting oligomers was more complex. It can easily be done by polymerizing a gel in the capillary [ 141, but this procedure is particularly delicate because minute air bubbles could be trapped, thus breaking the electric circuit. We have preferred to fill the capillary with a liquid polymer, as suggested in 1151. As shown in Fig. 6, the oligomers can be seen as retarded peaks. In the absence of a liquid polymer, it was impossible to separate the monomer from the oligomers (not shown), suggesting that the mobility in free phase is almost identical (only a peak broadening is observed). In the lower part of Fig. 6, the profile of the control, unpolymerized pK 8.5 Immobiline, is visible.
4 Discussion 4.1 On the analysis of acrylic acid traces in Immobilines .JL*+
I
& L ,
-
9
-
~
.
.
-
~
.
~
-
- - . Tt me (m
~ I n)
-
~
-
-
~
~
-
-.-, 9
Figure 5. Monitoring for acrylic acid contamination in alkaline Imrnobilines. A 20 mM solution of pK 6.2 Immobiline was supplemented with 0.2 mMqrylic acid andinjectedintoP/ACEsystem2000(5sbypressure). Run in 50 mM phosphate, p H 7.0, and 2 mM CTAB at 20 kV, 100 FA. The insert shows the acrylic acid (Acr. A,) spike magnified by a factor of 10 on the y-axis (from 100to 10 milli-abs) and by a factor o f 4 on the x-axis (from 20 to 5 min). The spike was integrated with the aid ofthe calibration curve in Fig. 4. Anodic migration (reversed polarity).
The C Z E methodology appears to be valuable for checking the purity and stability of Immobiline chemicals that are either commercially available or produced in the laboratory. Given the exquisite resolution and the extreme reproducibility of the IPG technique, it is worth while having a fast and reliable technique at hand te check any anomalies in the Immobiline chemicals, as they are the ultimate cause of trouble in IPG runs. Originally, when checking for degradation products in Immobiline chemicals, Giveby et al. [ 1 11 had proposed two techniques: (i) gas chromatography for quantitation of traces of acrylic acid and (ii) high performance liquid chromatography (HPLC, in the gel permeation mode) for
58
Ekctrophoresis 1991,12,55-58
P. G. Righetti el al.
oligomer and n-mer detection. CZE compares favorably with both methods. For instance, in the case of acrylic acid analysis, Giveby et al. [ 111 have detected 0.76 mol % as a well-visible peak and 0.10 mol% as a minute spike in the chromatogram of different Immobilines. In the former case, when the amount of Immobiline injected was 0.2 M,the total amount present in the capillary was 1.52 mM; in the latter case, the lowest amount detected was 0.2 mM, i. e. the same order of magnitude as our detection limit. In reality, CZE could offer a much higher detection limit if one were to resort to sample derivation and detection by laser-induced fluorescence. In the latter case, detection limits as low as 500 attomoles have been reported [ 161. In terms of analysis times, CZE also compares favorably with gas chromatography: whereas Ghveby et al. [ l l ] had reported times of 20-30 min for gas chromatography of different Immobilines, in our case running times as short as 5-10 min are sufficient (even for complex mixtures of Immobilines). 4.2 On the analysis of polymers in Immobilines When we first reported the presence of polymers in Immobiline solutions in 1987, the data met with skepticism and incredulity. However, this finding solved one ofthe main problems of IPGs, namely the disappearance of large proteins from analytes. As a model, we proposed a cross-linking of these large proteins by the Immobiline oligomers, with production of a large precipitating reticulum, as in antigen-antibody reactions [171. Our data were then fully confirmed by Giveby et al. [ 111 and by the fact that the removal of such oligomers completely prevented protein precipitation [ 161. For the analysis of oligomers and n-mers in Immobiline solutions, Giveby et al. [ 111 proposed gel permeation chromatography in HPLC columns: as seen in their Figs. 5 and 6, excellent separations are obtained, with polymer detection at 2 10 nm (because double bonds would no longer be contributing to the absorbance). In our report (see Fig. 6), it is seen that polymer separation can be achieved by incorporating asoluble,neutral polymer in the background electrolyte, as already reported by Zhu et al. [ 151. However, we propose a different polymer, i. e. Ficoll-400, a highly soluble polysaccharide well known for cell-fractionations in both rate-zonal and isopycnic separations [181. We had also tried another soluble polymer, i. e. oolyvinylpyrrolidone 360, but its high absorbance at 214 nm (2.7 for a 1 % solution) prevented any meanigful reading. Also, Ficoll-400 has an appreciable absorbance at 214 nm (0.340 for a 1 % solution) but only 0.170 at 254 nm (in addition, this absorbance refers to a 1 cm path length, whereas the optical channel in CZE is only 50 pm). Performing molecular sieving in the absence of a gel (typically polyacrylamide) is quite convenient in CZE, since the polymerization of a gel in such a narrow bore (e, g., 25-50 pm) is inconvenient and does not yield reproducible results. The idea that soluble polymers could introduce molecular sieving (a concept usually associated only with gel matrices) has had a long gestation period and was first proposed in 1980 by Bode [191, who suggested that polyacrylamide gels behave like an “extended viscoelastic continuum” and a “viscosity emulsion”. In this analogy, sieving in gels was seen as due to the effects of elastic
repulsion, which, in turn, are dueto spontaneous reorientation of the polymer chains which can fluctuate in the surrounding space. The same author has provided a considerable body of evidence showing molecular sieving in mixtures of linear, uncrosslinked polymers and gels, such as polyethyleneglycol in cellulose acetate [201 or polyacrylamide in agarose (211. There are still some minor peaks in the electropherograms which are not yet identified. Their analysis and characterization is in progress.
Supported in part by grants from Consiglio Nazionale delle Ricerche (Roma), Progetto Finalizzato Chimica Fine II, and by Minister0 della Pubblica Istruzione. W e thank Beckman Italy (and in particular Drs. R . Montini and S . Di Biase)for the kind loan of the equipment and for support and valuable suggestions. Received June 20,1990
5 References [ 11 Righetti, P. G., ImmobilizedpHGradients: TheoryandMethodology,
Elsevier, Amsterdam 1990. [21 Righetti, P. G.. Isoelectric Focusing: Theory, Methodology and Applications, Elsevier, Amsterdam, 1983. [31 Chiari, M., Casale, E., Santaniello, E. and Righetti, P. G., ppl. Theor. Elecfrophoresis 1989, 1, 99-102. [41 Chiari, M., Casale,E., Santaniello,E. and Righetti, P. G.,Appl. Theor. Electrophoresis 1989, I , 103-107. [51 Gianazza, E., Celentano, F., Dossi, G., Bjellqvist, B. and Righetti, P. G., Electrophoresis 1984,5,88-97. [61 Righetti, P. G., Chiari, M., Sinha, P. K. and Santaniello, E., J. Biochern. Biophys. Methods 1988,16, 185-192. [71 Gelfi, C., Bossi, M. L., Bjellqvist, B. and Righetti, P. G.,J. Biochem. Biophys. Methods 1987, I5,41-48. [Sl Chiari, M., Righetti, P. G., Ferraboschi, P., Jain, T. and Shorr, R., Electrophoresis 1990, 11, 617-620. (91 Chiari, M., Pagani, L., Righetti, P. G., Jain, T., Shorr, R. and Rabilloud, T., J. Biochem. Biophys. Methods 1990,21, 165-172. 101 Charlionet, R., Sesbouk, R. and Davrinche, C.,Electrophoresis 1984, 5, 176-178. 111 Giveby, B. M., Pettersson, P., Andrasko, J., Ineva-Glygare, L., Johannesson, U., Gorg, A., Postel, W., Domscheit, A., Mauri, P. L., Pietta, P., Gianazza, E. and Righetti, P. G., J. Biochem. Biophys. Methods 1988,16, 141-164. 121 Karger, B. (Ed.), Proceedings of the lst International Symposium on High Performance Capillary Electrophoresis, J. Chromatugr. 1989, 480, 1-435; 1990,516, 1-298. I131 Rabilloud, T., Pernelle, J. J., Wharmann, P., Gelfi, C. and Righetti, P. G., J. Chromatogr. 1987,402, 105-1 13. (141 Karger, B. L., Cohen, A. S. andGuttman, A.,J. Chromatogr. 1989, 492,585-614. (151 Zhu, M., Hansen, H. S. and Gannon, F.,J. Chromatogr. 1989,480, 3 11-319. [161 Chung, Y . F. and Dovichi, N. J., Science 1988,242, 562-564. [171 Righetti, P. G., Gelfi, C., Bossi, M. L. and Boschetti, E., Electrophoresis 1987,8, 62-70. [ 181 Rickwood, D. (Ed.), Centrifugation: A Practical Approach, IRL Press, Oxford 1984, pp. 3 1-32. [191 Bode, H. J., in: Radola, B. J. (Ed.), Electrophoresis ’79, de Gruyter, Berlin 1980, pp. 39-52. [201 Bode, H. J., Anal. Biochem. 1977,83,204-210. [211 Bode, H. J., Anal. Biochem. 1977,83,364-371.
