Use of ethanolamine for sample stacking in capillary electrophoresis

2440 M. Louwagie, T. Rabilloud and J. Garin Electrophoresis 1998, 19, 2440-2444

Mathilde Louwagie' "hierry Rabilloud' JbrSme Garin'

'Laboratoire de Chimie des ProtBines, 'Laboratoire de Biobnergbtique Cellulaire et Pathologique, CEAIGrenoble, DBpartement de Biologie Molbculaire et Structurale, Grenoble, France

Use of ethanolamine for sample stacking in capillary electrophoresis

Capillary zone electrophoresis (CZE) in the presence of ethanolamine was used in a micropreparative mode. Sample volumes up to 1 pL could be loaded onto a 100 pm diameter capillary without loss in resolution. Coupled to narrow-bore reversed-phase high-performance liquid chromatography, ethanolamine-CZE allowed the collection of sufficient amounts of pure peptidic material to perform amino acid sequence analysis.

In recent years, the technical platform in the field of protein microanalysis has evolved greatly. Protein chemists can now play with orthogonal techniques like 2-D gel electrophoresis, reversed-phase high-performance liquid chromatography (RP-HPLC), amino acid sequence analysis and mass spectrometry (for a review, see [l]). For peptide purification, RP-HPLC is the most popular choice owing to its ability to provide the resolution needed to identify peptides, as well as for its robustness and reproducibility. Besides, the use of capillary zone electrophoresis (CZE) for peptide separation was limited, mainly because of the low volumes which can be loaded onto the capillary (typically, a few nanoliters). In the absence of stacking methods, CZE has poor detection limits compared to RP-HPLC. As exemplified [2], CZE without preconcentration techniques can be used as a micropreparative step for sequencing studies when dealing with peptide solutions in the millimolar range. For dilute samples, on-line sample preconcentration techniques such as stacking [3-7] and isotachophoresis [8] have had a profound effect on the utilization of capillary electrophoresis in research and clinical laborator- ies. Methods employing a hybrid capillary consisting of an inlet section filled with a chromatographic sorbent, typically reversed-phase type, and a standard CZE bare fused-silica separation capillary were also described [9-lo]. Here, we report the use of ethanolamine for sample stacking in the standard CZE operating mode. As much as 1 pL samples could be loaded onto a 100 pm capillary without loss in resolution. The interest of this approach was illustrated by using it as an easy additional micropreparative step after narrow-bore RP- HPLC. Starting with an 'in-gel' tryptic digest of bovine serum albumin (50 picomoles), peptides were purified by narrow-bore RP-HPLC. Subsequently, ethanolamine- CZE allowed the separation and collection of peptides unresolved by RP-HPLC. Enough CZE-purified peptide material could be collected for successful amino acid sequence analysis.

Correspondence: Dr. M. Louwagie, CEA/Grenoble, DCpartement de Biologie MolCculaire et Structurale, Laboratoire de Chimie des Pro- teines, 17 rue des Martyrs, F-38054 Grenoble, France (Tel: +33-4-7688- 9657; Fax: +33-4-7688-9808; E-mail: [email protected])

Abbreviations: CZE, capillary zone electrophoresis; MALDI-TOF-MS, matrix assisted laser desorption/ionization-time of flight-mass spectrometry

Keywords: Capillary electrophoresis / Sample stacking / Microse- quence analysis

Ethanolamine, citric acid and sodium citrate (trisodium salt) were from Sigma (St. Louis, MO, USA). Trifluoroa- cetic acid was from Pierce (Rockford, IL, USA) and acetonitrile from Merck (Darmstadt, Germany). All reagents were used as received. Fused-silica capillaries of 100 pm ID and 57 cm length were purchased from Beckman (Fullerton, CA, USA). CZE separation was carried out on Beckman P/ACE System 5510 equipped with a UV detector and interfaced with a Dell 466 computer uti- lizing System Gold software for instrumental control and data collection. Electrophoresis was performed in 40 mM citrate, pH 2.5 (run buffer). The buffer reservoirs were changed every day. As test peptides the following basic, neutral, and acidic peptides were used: P1, YFTL- KIRGRKRFEMFRE (pl 11.0); P2, LIQLRHVLTKAY (pl 10.05); P3, WTLKIRGRKRFEMFRELNEALELKD

MYANYAIGK (PI 7.2); P6, VGFEDHIAAITRNY (pl 5.4); P7, KLEDGLVFPIPTNIQV (pl 4.2). For sample preparation, synthetic peptides were dissolved in 0.1 O/o TFA, and ethanolamine was added at a final concentration of 40 mM, which allowed the sample pH to be raised to 9.5. For BSA tryptic peptides collected after narrow-bore HPLC, ethanolamine was added at the same final concentration. In the concentrating operating mode, samples were loaded onto the capillary by applying a reduced pressure to the cathodic electrode reservoir for a specified time while the anodic end of the capillary was immersed in the sample solution (20 pL). Unless otherwise specified, loading time was set at 60s, corre- sponding to a 1 pL load. After sample application, the anodic end of the capillary was returned to the run buffer. A small volume of the run buffer was then loaded by applying a reduced pressure to the cathodic reservoir for 2 s, after which electrophoresis was performed at 30 kV. The temperature was maintained at 23°C for all experiments. For amino acid sequence analysis, CZE eluted fractions were collected electrophoreti- cally by changing the outlet vials, which were filled with 20 pL of the run buffer. Amino acid sequence analysis was performed on a Procise 392A sequencer (Perkin- Elmer, Applied Biosystems Division, Foster City, CA, USA) according to the manufacturer's recommenda- tions.

(PI 9.7); P4, HKYQTLAGIVC (PI 8.2); P5, HSEVSNQ-

Efficient sample stacking protocols in CZE have already been described for peptide solutions. One of the best- documented studies was the pioneer approach by Aeber- sold and Morrison [6]. They achieved simple and efficient electrophoretic stacking of the injected peptides at the interface between a high-pH sample application solu-

0 WILEY-VCH Verlag GmbH, 69451 Weinheim, 1998 0173-0835/98/1414-2440 $17.50+.50/0

Electrophoresis 1998, 19, 2440-2444 Sample stacking in capillary electrophoresis 2441

Y

f E

40 I

I t = Q h 3 Q

40 4 I t = l h I

B

4 0

2 0

0 I I I I

4 0 - 1 I

20

0 I I I I 1

4 6 8

Time (minutes) Time (minutes)

Figure 1. Sample stacking: a calibrated mixture of seven synthetic peptides (see text) was prepared (A) at a final concentration of 45 picomoles/pl by adding NH,OH (10 mM) as described in [S] or (B) at a final concentration of 15 picomoles/pL by adding TFA and ethanolamine (0.1% and 40 mM, respectively). Serial injections were performed at different time points after sample preparation (A: 0, 0.5 and 1 h; B: 0, 1.5 and 3 h). Peptides were loaded onto the capillary by applying reduced pressure to the cathodic end of the capillary, (A) for 20s, and (B) for 60s, before analysis.

tion (20 mM NH,OH, pH > 10) and a low-pH electrophoresis buffer (10 mM citrate buffer, pH 2.5). This approach was successfully used for the analysis of peptide fractions collected from a narrow-bore HPLC system. However, as shown in Fig. ZA, the high volatility of ammonia can be a major drawback: upon repetitive injections of a seven-peptide test mixture, peak broad- ening and loss in resolution were observed. High resolution could not be restored easily by re-adding ammonia because excess ammonia led to poor resolution (not shown). This quick deterioration, due to the significant pH decrease caused by ammonia evaporation, makes this procedure hardly usable as a preparative step before amino acid sequencing when repetitive injections and separations on the capillary are requested. Therefore, we looked at other buffers characterized by a high boiling temperature to avoid evaporation and high pK, values. Such buffers were tested for their capacity to induce sample stacking. Among them, only ethanolamine and diethylethanolamine behaved satisfactorily. The presence of ethanolamine in the sample was shown to be the major parameter for efficient stacking to occur because it raises the pH sample to 9.5. The concentration of citrate ions in the run buffer was found to be optimum around 40 mM. The resolving power of this ethanolamine stacking procedure (ethanolamine-CZE) was tested by look- ing at the separation of the test mixture. In Fig. lB, sep-

aration of the peptides in the above-described stacking system is shown. When peptides were solubilized into the run buffer, 1 pL injections led to poor resolution (not shown). In the presence of ethanolamine, volumes up to 1 pL could be injected with an adequate resolution while ammonia sample volumes were not to exceed 350 nL in order to obtain a similar resolution (Fig. 1A). Furthermore, as expected and contrary to samples prepared with ammonia, only minor modifications in the electropherogram could be observed when the sample was loaded several hours after preparation (Fig. 1B). This reproducibility makes repetitive injections possible, as well as collection for amino acid sequence analysis.

The lower concentration limit for peptide analysis by ethanolamine-CZE was evaluated. Therefore, three synthetic peptides (peptides P2, P5 and P7) were prepared at concentrations ranging from 0.1 VM to 10 VM. From Fig. 2, it appears that as little as 200 femtomoles of peptides loaded in 1 pL can be detected, with a signal-to-noise ratio superior to 10. Thus, the minimal peptide concentration required for analysis by ethanolamine-CZE is in the 0.2-0.5 VM range.

We checked whether peptides that were collected after separation by ethanolamine-CZE were suitable for direct Edman amino acid sequence analysis. First, we tested


: ; Time (minutes)

Figure 2. Sensitivity of ethanolamine-CZE. A calibrated mixture of three synthetic peptides (P2, P5 and P7) was prepared at a final concentration of 0.2 WM by adding TFA and ethanolamine (0.1% and 40 mM, respectively). Sample was loaded onto the capillary by applying reduced pressure to the cathodic end of the capillary for 60s.

our mixture of peptides loaded onto the capillary at a concentration of 15 pmoles/pL. Ethanolamine-CZE-separated peptides were collected into the run buffer and loaded onto microsequence glass fiber disks pretreated with Polybrene (Applied Biosystems). A noisy baseline was observed on the PTH chromatograms exhibiting several major contaminating peaks. These peaks could be attributed to the citrate present in the CZE-collected peptide samples. Thus, an attempt was made to elimi- nate the citrate adsorbed onto the microsequence glass fiber disks: the disks were washed with 0.1% TFA (5 mL) before being loaded onto the sequencer. For all peptides that were sequenced, the resulting initial sequence yield was low (less than 5%) pointing to the washing out of the peptides from the microsequence disks during the cleaning step. This difficulty could be avoided by loading the ethanolamine-CZE separated peptides onto PVDF disks inserted into microtubes (ProSorb, Applied Biosys- tems) instead of glass-fiber disks. The PVDF membrane was then washed with 0.1% TFA (3 X 50 pL), and Poly- brene was added before loading onto the sequencer [ll]. Multiple sequence analyses on three different ethanolamine-CZE purified peptides (P2, P5 and P7) showed good initial and repetitive yields (52% * 10% and 94% k 2%, respectively). Note that attempts were made to simplify this methodology by separating and collecting the peptides from the capillary into volatile buffers like trifluoracetic acid (0.1 Yo or 1 Yo) or hydrochloric acid (0.1 N) instead of citrate buffer. This procedure was aban- doned, however, since the resolution of the separation was dramatically decreased.

Ethanolamine-CZE was applied to the purification of proteolytic peptides from a test protein. Peptides derived from an ‘in-gel’ tryptic digestion of 50 picomoles of bovine serum albumin were separated by narrow-bore RP-HPLC (Fig. 3A). MALDI time of flight(T0F)-MS analysis was carried out on UV absorbing peaks eluted from the reversed-phase column. We found that most peaks consisted of peptide mixtures. Thus, the material

I I

- 7 0 l-

X Y

0 0 N

e a n U

60 80 100 120

Time (minutes)

U I 2 4 6 8 10 12

Time (minutes)

Figure 3. Narrow-bore analysis of an in-gel tryptic digest, followed by CZE analysis. BSA (50 picomoles) was in-gel digested with trypsin (0.5 pg). Resulting peptides were extracted using a mixture of acetonitrile and hydrogenated Triton X-100 (0.1%), as described in [13], and applied to a narrow-bore RP-HPLC column (C18, 5 yrn, 150 X 0.5 mm, Perkin Elmer). Peptides were eluted with a 0-40% linear gra- dient of acetonitrile containing 0.1% TFA, at a flow rate of 5 pL/min. (A) Eluting peptides, as detected by UV absorption at 215 nm, were collected manually. (B) The marked peak, (HPLC peak 88 min) was subjected to ethanolamine-CZE analysis after the addition of ethanolamine (40 mM, in a final volume of 20 pL).

present in these UV peaks was not suitable as such for direct amino acid sequence analysis. Therefore, an additional purification step was mandatory: ethanolamine was added to these RP-HPLC fractions, and 1 pL samples were injected onto the CZE system. Figure 3B dis- plays the ethanolamine-CZE electropherogram obtained from one RP-HPLC fraction, which was shown, by MALDI-TOF-MS analysis, to contain a mixture of three different tryptic peptides. Starting with this fraction, four successive 1 pL injections made it possible to collect enough material for amino acid sequence analysis (Fig. 4); the major UV absorbing peak on the CZE profile was shown to be pure and easily identified as peptide 143- 149 of bovine serum albumin.

Multiple strategies are now available for the identifica- tion of proteins from one-dimensional or 2-D gels. Fol- lowing “in-gel” digestion with trypsin, a protein can, for example, be identified by the precise measurement of masses of the generated tryptic peptide masses by

Electrophoresis 1998, 19, 2440-2444 Sample stacking in capillary electrophoresis 2443

minut 8s

Figure 4. Amino acid sequence analysis. One pL fractions of the collected RP-HPLC peak (elution time, 88 min) were subjected to ethanolamine-CZE separation. CZE peak eluting at retention time 8.3 min was collected. Material resulting from four consecutive injections were pooled and loaded onto a Prosorb tube for amino acid sequence analysis (initial yield, 4 picomoles).

MALDI-TOF-MS [12]. The “peptide-mass map” that is obtained is matched against a sequence database whose entries have been transformed to a set of masses by means of selectable parameters like the specificity of the enzyme. When no entry is found in the databases, amino acid sequence information is mandatory in order to clone and sequence the gene of interest. This can be achieved by tandem mass spectrometry (MSIMS) or by Edman degradation. In that case, obtaining proteolytic peptides pure enough for amino acid sequence analysis may be difficult when using a single RP-HPLC step. In our hands, this problem is acute when dealing with high molecular mass proteins. Because such proteins produce

a high number of peptides when digested with trypsin or other proteolytic enzymes, ethanolamine-CZE can be an interesting additional step in a purification scheme. Our work shows that it is possible to combine narrow- bore RP-HPLC directly with CZE in order to obtain peptides which are pure enough for amino acid sequence analysis.

The authors thank Agnds Chapel for providing them with the synthetic peptides. This work was supported by grants from the Direction des Sciences du Vivant of C.E.A.

Received March 4, 1998


References [l] Patterson, S. D., Aebersold, R., Electrophoresis 1995, 16,

[2) Camilleri, P., Okafo, G. N., Southan, C., Brown, R., Anal. Biochem.

[3] Burgi, D. S., Chien, R. L., Anal. Chem. 1991, 63, 2042-2047. [4] Chien, R. L., Burgi, D. S . , Anal. Chem. 1992, 64, A489-496. [S] Shihabi, Z. K., in: Landers, J. P. (Ed.), Handbook of Capillary Elec-

[6] Aebersold, R., Morrison, H. D., J. Chromatogr. 1990, 516,

[7] Schwer, C., Lottspeich, F., J. Chromatogr. 1992, 623, 345-355.

1791-1814.

1991, 198, 36-42.

trophoresis, CRC Press, Boca Raton, FL 1997.

79-88.

[8] Foret, F., Szoko, E., Karger, B. L., Electrophoresis 1993, 14,

[9] Strausbauch, M. A., Madden, B. J. , Wettstein, P. J. , Landers, J. P.,

[lo] Beattie, J. H., Self, R., Richards, M. P., Electrophoresis 1995, 16,

[ll] Werner, W. E., Grimley, C., Yuan, P. M., Anal. Biochem. 1996, 237,

[12] Shevchenko, A., Wilm, M., Vorm, O., Jensen, 0. L., Podtelejnikov, A. V., Neubauer, G., Shevchenko, A,, Mortensen, P., Mann, M., Biochem. SOC. Transact. 1996, 28, 893-896.

[13] Adessi, C., Chapel, A., Vincon, M., Rabilloud, T., Klein, G., Satre, M., Garin, J., J. Cell Sci. 1995, 108, 3331-3337.

4 17-42 8.

Electrophoresis 1995, 16, 541-548.

322-328.

146-149.

Documents

Use of ethanolamine for sample stacking in capillary electrophoresis