Dynamic Labeling during Capillary or Microchip Electrophoresis for

The analysis of proteins under denaturing conditions is routinely performed with SDSrpolyacrylamide gel elec- trophoresis. The automated capabilities ...
0 downloads 0 Views 81KB Size
Download PDF
Anal. Chem. 2001, 73, 4994-4999

Dynamic Labeling during Capillary or Microchip Electrophoresis for Laser-Induced Fluorescence Detection of Protein-SDS Complexes without Preor Postcolumn Labeling Lian Ji Jin,† Braden C. Giordano,† and James P. Landers*,†,‡

Department of Chemistry, University of Virginia, McCormick Road, Charlottesville, Virginia 22901, and Department of Pathology, University of Virginia Health Science Center, Charlottesville, Virginia 22901

The analysis of proteins under denaturing conditions is routinely performed with SDS-polyacrylamide gel electrophoresis. The automated capabilities of CE, use of nongel sieving matrixes, and on-line optical detection by either ultraviolet (UV) absorption or laser-induced fluorescence (LIF) promise to revolutionize this method. While direct on-line detection of proteins is possible as a result of their intrinsic ability to absorb light in the UV part of the spectrum (detection sensitivity comparable to Coomassie Blue staining of gels), LIF provides more powerful detection but requires pre- or postcolumn fluorescence labeling of the proteins. The development of a protocol analogous to that used for double-stranded DNA analysis, where fluorescent intercalating dyes are simply included in the separation medium, would simplify sizebased protein analysis immensely. This would avoid the complications associated with covalent modification of the proteins but still exploit the sensitivity of LIF detection. We demonstrate that this is possible with CE and microchip detection by incorporating, into the run buffer, a fluorescent dye that interacts hydrophobically with proteinSDS complexes. Key to this is a dye that fluoresces significantly when bound to protein-SDS complexes but not when bound to SDS micelles. Comparison of electropherograms from CE-based denaturing protein analysis with UV and LIF detection indicates that the presence of the fluor does not alter separation of the proteins. Moreover, comparison with electropherograms generated from microchip electrophoresis with LIF detection shows that equivalent patterns can be obtained. Despite the unoptimized nature of this separation system, a dynamic labeling protocol that allows for LIF detection for proteins is attractive and has the potential to circumvent the tedious labeling steps typically required. Analysis of complex protein mixtures by sodium dodecyl sulfate-polyacrylamide electrophoresis (SDS-PAGE), a sizebased separation mode, is common practice. The past decade has * Corresponding author: (phone) 804-243-8616; (fax) 804-924-3048; (e-mail) [email protected]. † Department of Chemistry, University of Virginia. ‡ Department of Pathology, University of Virginia Health Science Center.

4994 Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

seen miniaturized electrophoresis, such as capillary electrophoresis (CE), be explored for denaturing protein analysis with the discrete advantages being automation and on-line optical detection. More recently, the microfabricated counterpart of CE, microchip electrophoresis, has been investigated for the purpose of protein analysis due to the inherent high-speed and high-throughput capabilities of this technique. Gel electrophoresis is the most widely used analytical technique in the life sciences, particularly for nucleic acid and protein analysis. Although the two-dimensional gel electrophoresis for protein analysis demonstrated by O’Farrell1 in 1975 has provided the greatest resolving power, one-dimensional gel electrophoresis of proteins is still the most frequently used form as a result of the simple methodology and the parallel processing capability for comparative multiple sample analysis.2 The advent of CE, a miniaturized form of electrophoresis, has revolutionized conventional electrophoresis in terms of rapid separation and automation, as well as parallel processing capability.3 As an extension of this innovation, microchip electrophoresis has been pioneered by the efforts of the Manz,4 Harrison,5 Ramsey,6 and Mathies7 groups. This is currently attracting widespread interest, not only due to its potential for parallel sample processing (analogous to slab gel electrophoresis) but also due to the high-speed separations demonstrated with this platform and the potential for automation.7 Ultraviolet (UV) absorption is a common detection mode with commercial CE instrumentation, whereby analytes can be detected with adequate sensitivity via the analyte-specific absorption of light at certain wavelengths. This also applies to protein-SDS complexes, which can be detected at several specific wavelengths in the UV part of the spectrum. This includes 280 nm (absorption of light by aromatic side groups of amino acids), 220 nm (absorption of light by peptide bond between amino acids) and, with greatest sensitivity, 200 nm. The problem associated with (1) O’Farrell, P. H. J. Biol. Chem. 1975, 250, 4007-4021. (2) Hames, B. D., Ed. Gel Electrophoresis of Proteins, A Practical Approach, 2nd ed.; Oxford University Press: New York, 1998. (3) Lu, X. D.; Yeung, E. S. Appl. Spectrosc. 1995, 49, 605-609. (4) von Heeren, F.; Verpoorte, E.; Manz, A.; Thormann W. Anal. Chem. 1996, 68, 2044-2053. (5) Chiem, N. H.; Harrison, D. J. Electrophoresis 1998, 19, 3040-3044. (6) Jacobson, S. C.; Ramsey, J. M. Electrophoresis 1995, 16, 481-486. (7) Huang, X. H. C.; Quesada, M. A.; Mathies, R. A. Anal. Chem. 1992, 64, 967-972. 10.1021/ac010590z CCC: $20.00

© 2001 American Chemical Society Published on Web 09/13/2001

these lower wavelengths (i.e., 220 and 200 nm) is that the sieving matrixes (high molecular weight polymers) used to separate the proteins also absorb light in this range. Consequently, detection sensitivity is reduced (or even completely impeded) at 200 and 220 nm, leaving detection at 280 nm as the only option. Unfortunately, detection sensitivity is reduced almost 10-fold at 280 nm in comparison with absorption at 200 nm. Solving the background UV absorption problem inherent with polymers has been avoided primarily by pre- or postcolumn labeling with fluorescent dyes and laser-induced fluorescence detection (LIF). While this enhances the sensitivity of protein detection immensely, it is timeconsuming and often requires a special device or skills to perform, not to mention being hampered by problems with incomplete labeling (which creates a microheterogeneous population of protein species). The difficulties associated with pre- or postcolumn fluorescent labeling can be avoided using a technique analogous to one used for sized-based separations of DNA fragments, where a fluorescent dye added to the separation buffer displays enhanced fluorescence when intercalated into doublestranded DNA. Double-stranded (ds) DNA can be made to fluoresce efficiently as a result of a dynamic interaction with a variety of fluorescence probes. Included in the group are ethidium bromide,8 YO-PRO,9 and TO-PRO,10 all of which intercalate into the (ds)DNA. With LIF detection, sensitivities several orders of magnitude higher than that attainable with UV absorbance detection, capillary and microchip platforms have made great strides in supplant the slab gel format. Of special interest to those researchers in life sciences is the prospect of the microchip platform to perform protein-SDS separations with all the advantages inherent to miniaturization. Although pre- or postcolumn labeling of proteins followed by LIF detection has been predominant in microchip separations,11,12 a simple method for performing LIF detection of proteins is desired. It is clear that a similar labeling approach applicable to proteins would be invaluable, i.e., a method that allows for a fluor to rapidly and noncovalently bind to proteins for LIF detection. Work in this arena was first presented by Mouradian et al.13 and by our group14 at a recent conference. The work of Mouradian et al.13 was further described in a paper by Bousse et al.15 Using a proprietary dye added to the separation buffer during electrophoresis, proteinSDS complexes are labeled and can be detected after some onchip manipulation. While incompatible with capillary-based electrophoresis, this work provides a significant advance in proteinlabeling technology for microchip analysis. The current report discusses a simpler approach whereby the hydrophobic nature of a commercially available protein fluor is (8) Ulfelder, K. J.; Schwartz, H. E.; Hall, J. M.; Sunzeri, F. J. Anal. Biochem. 1992, 200, 260-267. (9) McCord, B. R.; McClure, D. L.; Jung, J. M. J. Chromatogr., A 1993, 652, 75-82. (10) Rampal, S.; Lui, M. S.; Chen, F. T. A. J. Chromatogr., A 1997, 781, 357365. (11) Yao, S.; Anex, D. S.; Caldwell, W. B.; Arnold, D. W.; Smith, K. B.; Schultz, P. G. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 5372-5377. (12) Liu, Y. J.; Foote, R. S.; Jabobson, S. C.; Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 2000, 72, 4606-4613. (13) Mouradian, S.; Bousse, L.; Minalla, A.; Yee, H.; Williams, K.; Dubrow, R. Oral presentation at HPCE 2001, Boston, MA. (14) Jin, L. J.; Giordano, B. C.; Landers, J. P. Poster presentation at HPCE 2001, Boston, MA. (15) Bousse, L.; Mouradian, S.; Minalla, A.; Yee, H.; Williams, K.; Dubrow, R. Anal. Chem. 2001, 73, 1207-1212.

Figure 1. Mechanism for dynamic fluorescent labeling of proteinSDS complexes. As protein-SDS complexes migrate through the capillary, free dye binds to the complex and undergoes a conformational change.

exploited for dynamic labeling of SDS-saturated protein, without the need for additional steps (i.e., dilution) prior to detection. A commercial fluorogenic dye (NanoOrange) used for quantitative detection of proteins is ideal for the dynamic, noncovalent labeling of proteins. If the dynamic binding capabilities of this fluor could be exploited on CE and, further, on the microchip platform, protein-SDS separations would be simplified accelerated and tedious pre- or postcolumn labeling steps circumvented. A dynamic labeling technique extends the LIF detection repertoire for both capillary and microchip-based electrophoresis to denaturing protein analysis in a simplistic manner analogous to (ds)DNA separation and detection. The mechanism for dynamic labeling of SDS-protein complexes is proposed in Figure 1, which shows a hypothetical schematic for protein detection based on dynamic fluorescent labeling. The key feature governing the success of this mechanism is that the fluor does not generate substantial signal unless bound to protein. A modified version of this has been recently demonstrated by Bousse et al.15 using a red laser-excitable dye added to the separation buffer. However, a dilution step prior to detection is required, presumably to reduce the background signal generated by the fluor. While ideal for the microchip, the on-chip dilution makes this labeling technology less compatible with existing capillary-based electrophoresis. It is clear that the optimal fluorescent dye system for dynamic labeling of proteins would produce the minimum background fluorescence when included in the separation buffer, would generate a large signal when bound to protein-SDS complexes, and would not significantly effect protein-SDS complex mobility throughout the separation. It is likely that a number of dyes bearing these characteristics exist. NanoOrange is a fluorescein-like dye that, upon binding to the protein-SDS complex surface or hydrophobic domain of nondenatured proteins, undergoes significant enhancement in fluorescence due to comformational changes induced by binding. Legendre et al.16 have demonstrated a comparable mechanism for the noncovalent binding of near-infrared dyes (diethylthiatricarbocyanine iodide and IR-125) to proteins. In free solution, these (16) Legendre, B.; Soper, S. Appl. Spectrosc. 1996, 50, 1196-1202.

Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

4995

dyes showed minimal background fluorescence yet, upon binding to proteins, enter a favorable hydrophobic environment that enhances fluorescence.16 We hypothesized that incorporation of this dye into a CE separation buffer specific for SDS-denaturing protein electrophoresis would not contribute significantly to background fluorescence, bind to protein-SDS complexes, and allow for their detection by argon ion LIF. EXPERIMENTAL SECTION Apparatus. Beckman P/ACE 5510 (Beckman Instruments, Inc.) with both UV and LIF detectors was used to obtain capillary electrophoresis data. The microchip electrophoresis system was assembled in-house. A laser beam of 488 nm was emitted from an argon ion laser source (Laser Physics) and reflected to the beam splitter (505DRLP02; Omega Optical) set 45° to the incident beams and collected onto the channel of electrophoretic microchip by a microscope objective (16×/numerical aperture, 0.32; Melles Griot). Two mirrors (Thor Labs, Inc.) positioned parallel to each other were set after the laser source to level up the laser beam to the beam splitter. Fluorescence emitted by the sample was collected by the same microscope objective and focused by a 150mm lens onto a Tu-Can PMT (Hamamatsu) filtered by a 590-nm band-pass filter (590DF30, Omega Optical) set before PMT. The data collection was processed via a program written in Labview. Microchip Fabrication. The microchip was fabricated by using standard photolithography and wet chemical etching. A film mask containing the microchip design was prepared on a negative film using an image setter. The microchip design, a traditional cross-T type, was transferred onto the glass wafer (Nanofilm) with positive photoresist by UV exposure. The channels were etched with HF solution. The etched plate was thermally bonded to a drilled cover plate in a programmable furnace (Ney Dental Inc.). The configuration of a single channel microchip is 7.5 cm from injection cross to the outlet and 0.5 cm from injection cross to inlet and sample and sample waste. The separation channel is 30 µm deep and 120 µm wide at half-height. The injection channel is of the same depth as the separation channel and 150 µm wide at half-height. Reagents. A Bio-Rad protein separation kit was obtained from Bio-Rad Laboratory. A SDS-denatured protein sample was prepared according to the procedure provided with the kit. Respective proteins used for quantitation study were obtained from Sigma (St. Louis, MO). NanoOrange dye was purchased from Molecular Probes. NanoOrange was operated according to the product instruction manual. NanoOrange stock solution (500×) was added to the separation buffer at 1× concentration. Microchip Separation. Before microchip separation, the channel was conditioned by rinsing successively with 1 M NaOH, H2O, and separation buffer for 5 min each. After channel conditioning, the microchip was set to the stage and channel alignment was performed followed by sample loading. An injection voltage of 400 V was applied to the sample and sample waste reservoirs for 60 s to migrate the sample across the cross-channel section while the inlet and outlet were left floating. A separation voltage of 4200 V was subsequently applied to the inlet and outlet for 240 s for electrophoresis. RESULTS AND DISCUSSION Capillary Electrophoresis-UV Detection of SDS-Denatured Proteins. Capillary electrophoretic separation and UV detection 4996

Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

of a SDS-denatured protein size standard was first carried out using a commercial separation buffer that had dual functionality. The hydrophilic polymer-based sieving matrix allowed for proteinSDS complexes to be separated based on their molecular size with the added bonus that its chemical character also allowed it to function as a dynamic coating agent. The polymer and concentration has been optimized for the analysis of protein size in the molecular weight (MW) range of approximately 14 000-200 000. The pattern for the capillary-based separation of the eight protein mixture following SDS denaturation and electrophoresis is shown in Figure 2 and demonstrates the ability to separate proteins based on size. Plotting the log of the molecular weight against the reciprocal of the migration time (1/MT) illustrates the capabilities for estimating molecular size over a reasonably large mass range (Figure 2 inset). The ability to perform on-line detection with UV absorbance (0-15 mAU) in combination to the reasonably large range (14 000-200 000 mass) clearly points to the potential for replacing slab gel electrophoresis. Capillary Electrophoresis-LIF Detection of SDS-Denatured Proteins. In addition to the protein-SDS complexes being detectable by LIF upon addition of the fluorecent dye to the separation buffer, a key requirement of the mechanism proposed in Figure 1 is that dye binding does not alter electrophoretic mobility; i.e., electrophoretic patterns similar to that observed with UV absorbance detection should be observed with LIF detection. As seen in Figure 3, capillary electrophoretic SDS-denaturing separation of the protein size standard with argon ion LIF detection was found to be similar to that observed with UV detection (compare with Figure 2). This provided evidence that the electrophoretic mobility of the protein standards did not change significantly with the binding of fluorescent dye. This is further demonstrated with the Figure 3 inset, where the log MW versus 1/MT plot is virtually the same as presented in Figure 2, illustrating that the presence of the fluorescent dye in the separation buffer does not appear to affect protein-SDS complex mobility or the efficiency of the separation. It is noteworthy to point out that the actual concentration of NanoOrange in the run buffer is not known, and consequently, dye concentrations are quoted relative to stock solution. The potential problems associated with an electrophoretic mobility shift due to the presence of a dye are not new. Liu et al.12 and Harvey et al.,17 in their respective work with this dye as a potential buffer additive have addressed electrophoretic mobility differences due to the protein-fluor interaction. Liu et al.12 focused on using this dye as a postcolumn labeling reagent in the separation of model proteins. The authors indicated that the presence of the dye in the run buffer would adversely affect mobility. However, with the dye introduced to the separation channel at a mixing tee near the detection point, dye binding to the protein did not have time to affect electrophoretic mobility significantly. Indeed, Harvey et al.17 demonstrated quite clearly that the presence of this dye in the run buffer for CZE separations did alter electrophoretic mobility. It is noteworthy to point out that in these studies,12,17 the proteins separated were not denatured. (17) Harvey, M. D.; Bablekis, V.; Banks, P. R.; Skinner, C. D. J. Chromatogr., B 2001, 754, 345-356.

Figure 2. Capillary electrophoresis of protein-SDS size standard with UV detection (214 nm). Conditions: capillary, 50-µm-i.d. bare silica; effective/total length, 19/26 cm; buffer, Bio-Rad CE-SDS run buffer; sample concentration, 2 mg/mL in total for eight proteins; injection, 20 s at 10 kV; separation, 15 min at 15 kV. Peak identity: (1) lysozyme (14.4 kDa), (2) trypsin inhibitor (21.5 kDa); (3) carbonic anhydrase (31.0 kDa); (4) ovalbumin (45.0 kDa); (5) serum albumin (66.2 kDa); (6) phosphorylase B (97.0 kDa); (7) β-galactoside (116 kDa); (8) myosin (200 kDa).

Figure 3. Capillary electrophoresis of protein-SDS size standard with laser-induced fluorescence detection (488/590 nm). Electrophoretic conditions were the same as in Figure 2 except that capillary effective/total length was 20/27 cm and NanoOrange stock solution (500×) was added into CE-SDS run buffer at 1× concentration.

The problem with altered electrophoretic mobility is avoided in CE separations under SDS-denaturing conditions because the mobility of the protein complex is governed primarily by SDS, resulting in a separation based solely on molecular size. The dye used in this separation does not appear to contribute to the overall

charge-to-size ratio of the complexes. While the procedures for UV absorbance and argon ion LIF detection of protein-SDS complexes were identical, there is a slight discrepancy in migration time between the UV absorbance and LIF detection profiles. This is attributable to differences in the lengths of capillary used Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

4997

Figure 4. Microchip electrophoresis of protein-SDS size standard with laser-induced fluorescence detection (488/590 nm). CE-SDS run buffer and sample concentration as stated in Figure 3. Electrophoretic conditions as described in the text.

in the respective analyses as well as day-to-day variations typically observed in CE-based separations (data not shown). Overall, these results were encouraging given that the fluorescent dye was simply added to a commercially available CE-SDS separation buffer with no separation or signal optimization carried out. Microchip-LIF Detection of SDS-Denatured Proteins. Electrophoretic microchip devices have been demonstrated to be capable of executing (ds)DNA separations with fluorescent intercalating dyes as part of the sieving matrix. The capability of microchips for doing denaturing protein separations via a dynamic labeling has been shown in this study. Not surprisingly, the electrophoretic conditions used for CE-LIF of protein-SDS complexes were also found to be effective for microchip-LIF detection. The electrophoretic pattern obtained from microchip electrophoresis is concordant with the CE-LIF results (Figure 4). It is interesting that a nearly identical peak pattern was generated by the microchip system, a phenomenon that has been observed in the separations of (ds)DNA and other analytes. This raised the question as to whether it would be possible to perform protein quantitation with this system. While the fluorescent dye used in this study is utilized in protein quantitation, it was not clear that the dye would bind to proteins in a predictable fashion during electrophoresis. Quantitative Capabilities of Dynamic Labeling for Microchip Electrophoresis with LIF. To evaluate the quantitative nature of the binding of the NanoOrange dye to protein during CE separations, the quantitation of carbonic anhydrase (CA, peak 3 in the protein size standard) by CE-LIF was investigated using lysozyme (peak 1) as the internal standard (50 µg/mL). Solutions of CA ranging in concentration from 0 to 200 µg/mL were injected electrokinetically for 20 s in a field of 370 V/cm (∼10 nL injected) and electrophoresed in a capillary containing the fluorescent dyeaugmented sieving polymer and peak area (PA) determined from 4998

Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

LIF. The ratio of the CA-to-lysozyme peak area was plotted against protein concentration to obtain a calibration curve. Despite the fact that this system is completely unoptimized (i.e., choice of polymer, polymer concentration, buffering agent, ionic strength, SDS concentration, etc.), there was a reasonably linear relationship in the 0-200 µg/mL concentration range (R2 ) 0.993). This is an encouraging sign of the potential for this method to be quantitative. However, this will only be born out through optimization of the separation with respect to polymer, buffer, pH, and surface coating. One interesting observation in these data is that the peak area for the internal standard and peak area for a CA concentration of 100 µg/mL were nearly identical. On the basis of what is known about protein-SDS complex formation, NanoOrange is expected to bind on a “per-mass” basis. Regardless of protein size, the peak areas should be same provided that the same mass of an individual protein is injectedsthis is not supported by the data. It is premature to speculate on whether there is a sound basis for this phenomenon or whether it is simply an artifact of a completely unoptimized separation. Limit of detection (LOD) is, obviously, one important metric by which the value of this assay will be judged. It is not realistic to estimate the sensitivity level at this point because the separations described in this report were proof of concept and certainly not optimized. As found with other fluorescent dyes, Nano Orange binds avidly to the SDS micelles present in the CE-SDS buffer media (that is to non-protein-bound SDS)) and generates considerable background fluorescence (130-160 RFU) at a 1× concentrationsthis contributes significantly to the noisy baseline. Fluorometric study showed that the presence of 0.1% SDS (below the cmc) in 100 mM Tris-HCl buffer does not enhance the fluorescence of NanoOrange at all. With an optimal concentration of SDS in the buffer media, background noise will be reduced and signal intensity enhanced, all without affecting the denatur-

ation of proteins. Only then it will be possible to fully address the LOD issue. CONCLUSIONS This study has shown that dynamic labeling of protein-SDS complexes can be accomplished using NanoOrange, making argon ion LIF detection a simple and easy approach for both capillary and microchip electrophoresis. Simple addition to the separation buffer of an appropriate fluorescent dye that binds protein-SDS complexes results in a method that allows for LIF detection with both capillary and microchip electrophoresis platforms. While the quantitative potential of this method clearly exists, the detection sensitivity is not as high as will be needed for real-world

applications. This is largely due to the commercial sieving matrix used for CE-SDS separations that was chosen for initial proof-offeasibility studies. However, current and future efforts evaluating separation parameters such as buffer system, buffer pH, ionic strength, and polymer type will allow optimization of the method to fully exploit the potential of a particular dye for LIF detection of protein-SDS complexes under dynamic fluorescent-labeling conditions.

Received for review May 29, 2001. Accepted August 3, 2001. AC010590Z

Analytical Chemistry, Vol. 73, No. 20, October 15, 2001

4999