Anal. Chem. 2007, 79, 6840-6844
Thermoresponsive MALDI Probe Surfaces as a Tool for Protein On-Probe Purification Meiling Li,† Ganga Fernando,† Lorraine G. van Waasbergen,‡ Xuanhong Cheng,§ Buddy D. Ratner,§,# and Gary R. Kinsel*,†
Department of Chemistry and Biochemistry, Southern Illinois University Carbondale, Carbondale, Illinois 62901-4409, Department of Biology, The University of Texas at Arlington, Arlington, Texas 76019, Department of Bioengineering, The University of Washington, Seattle, Washington, 98195, and Department of Chemical Engineering, The University of Washington, Seattle, Washington, 98195
Thin film depositions of rf plasma polymerized N-isopropylacrylamide (ppNIPAM) show a phase transition temperature below which the polymer surface is hydrophilic, and protein nonadsorptive, and above which the polymer surface is hydrophobic, and protein-retentive. Results presented here demonstrate that this thermoresponsive plasma polymer can be coated on the surface of a MALDI probe and subsequently used for on-probe biomolecule cleanup. Specifically, a contaminated biomolecule can be applied to the ppNIPAM coated MALDI probe surface at a temperature above the phase transition temperature, washed using solvent also held above the phase transition temperature, and then analyzed by reducing the probe temperature to room temperature before adding the MALDI matrix. With the use of this approach, it is demonstrated that cytochrome c contaminated with 0.3% SDS, which yields only a very weak MALDI ion signal as directly deposited, can be purified on-probe using the thermoresponsive plasma polymer to improve significantly the ion signal. It is further shown that the decontamination of whole cell protein extracts from cyanobacteria is augmented through the use of the ppNIPAM coated MALDI probe. Matrix assisted laser desorption/ionization mass spectrometry (MALDI MS) has become a critical tool for high throughput proteomic applications. The tolerance of MALDI to a relatively high level of contaminants in the sample1 makes it particularly attractive for the direct analysis of complex biological mixtures. However, samples that contain salts and surfactants, which are commonly used reagents in protein sample workup, are often poorly ionized in a MALDI experiment.2-7 Sodium dodecylsulfate (SDS), one of the most commonly used detergents, has been * Author to whom correspondence should be sent. E-mail: gkinsel@ chem.siu.edu. † Southern Illinois University Carbondale. ‡ The University of Texas at Arlington. § Department of Bioengineering, The University of Washington. # Department of Chemical Engineering, The University of Washington. (1) Patterson, S. D.; Aebersold, R. Electrophoresis 1995, 16, 1791-1814. (2) Brockman, A. H.; Dodd, B. S.; Orlando, R. Anal. Chem. 1997, 69, 47614720. (3) Kallweit, U.; Bornsen, K. O.; Kresbach, G. M.; Widmer, H. M. Rapid Commun. Mass Spectrom. 1996, 10, 845-850.
6840 Analytical Chemistry, Vol. 79, No. 17, September 1, 2007
shown to be particularly detrimental to the MALDI signal.4-7 Several on-probe washing procedures have been shown to be simple and effective strategies for sample purification prior to MALDI MS analysis.8-11 Many of these approaches make use of the natural affinity of proteins for hydrophobic surfaces to keep the proteins adsorbed on the surface during the washing procedure.8,9,11 Poly(N-isopropyl acrylamide) shows a lower critical solution temperature (LCST) of 31 °C in an aqueous environment.12 When grafted onto solid surfaces by rf plasma polymerization of N-isopropyl acrylamide, the resultant plasma polymer (ppNIPAM) gives a “smart” surface physical property that can be controlled by an external stimulus, namely, temperature.13-15 Below the LCST, the ppNIPAM-grafted surface is hydrophilic and protein nonadsorptive. As the temperature increases above the LCST, the grafted polymer chains collapse and the surface becomes hydrophobic and protein-retentive. Thus, the hydrophobicity of this surface can be switched on and off by changing temperature. In this study we make use of this surface property to make a functional MALDI probe for the on-probe decontamination of biomolecules prior to MALDI MS analysis. Specifically, the contaminated biomolecule is applied to the ppNIPAM modified probe surface at high temperature, washed at the same temperature, and the purified biomolecule is then released, by decreasing the surface temperature, for subsequent MALDI MS analysis. The (4) Rosinke, B.; Strupat, K.; Hillenkamp, F.; Rosenbusch, J.; Dencher, N.; Kruger, U.; Galla, H. J. Mass Spectrom. 1995, 30, 1462-1468. (5) Jeannot, M. A.; Jing, Z.; Li, L. J. Am. Soc. Mass Spectrom. 1999, 10, 512520. (6) Puchades, M.; Westman, A.; Blennow, K.; Davidsson, P. Rapid Commun. Mass Spectrom. 1999, 13, 344-349. (7) Galvani, M.; Hamdan, M. Rapid Commun. Mass Spectrom. 2000, 14, 721723. (8) Blackledge, J. A.; Alexander, A. J. Anal. Chem. 1995, 67, 843-848. (9) Worrall, T. A.; Cotter, R. J.; Woods, A. S. Anal. Chem. 1998, 70, 750-756. (10) Jensen, C.; Haebel, S.; Andersen, S. O.; Roepstorff, P. Int. J. Mass Spectrom. Ion Processes 1997, 160, 339-356. (11) Hung, K. C.; Ding, H.; Guo, B. Anal. Chem. 1999, 71, 518-521. (12) Kubota, K.; Hamano, K.; Kuwahara, N.; Fujishige, S.; Ando, I. Polym. J. (Tokyo, Jpn.) 1990, 22, 1051-1057. (13) Pan, Y. V.; Wesley, R. A.; Luginbuhl, R.; Denton, D. D.; Ratner, B. D. Biomacromolecules 2001, 2, 32-36. (14) Ratner, B. D.; Cheng, X.; Wang, Y.; Hanein, Y.; Bohringer, K. F. Polym. Prepr. 2003, 44, 198-199. (15) Cheng, X.; Wang, Y.; Hanein, Y.; Bohringer, K. F.; Ratner, B. D. J. Biomed. Mater. Res. 2004, 70, 159-168. 10.1021/ac0711236 CCC: $37.00
© 2007 American Chemical Society Published on Web 08/08/2007
potential advantage of this approach, over the use of other commercial polymers, is that the material protein binding properties can be reduced during the critical MALDI matrix deposition step, leading to enhancement of the protein MALDI ion signal.16-21 EXPERIMENTAL SECTION Materials. Ninety-seven percent N-isopropyl acrylamide (NIPAM) and hexafluoropropylene oxide (C3F6O) were purchased from Aldrich (Milwaukee, WI) and used as received. Cytochrome c, R-cyano-4-hydroxy-cinnamic acid (CHCA), and phosphate buffered saline (PBS) were purchased from Sigma (St. Louis, MO). Tris-glycine-SDS was purchased from Fisher Scientific (Pittsburgh, PA). Extraction of protein from cyanobacteria (Synechocystis PCC 6803) was achieved using B-Per in 20 mM Tris‚ HCl from Pierce (Rockford, IL). Plasma Polymer MALDI Probe Modification. Preparation of the plasma polymer modified MALDI probe was achieved in two steps. In the first step, an approximately 22 mm diameter aluminum disk was placed in an rf plasma reactor and C3F6O was introduced at a monomer pressure of 112 mTorr and a flow rate of 4.2 mL/min (STP). The pulsed plasma was initiated and operated at a power of 300 W and a duty cycle of 10 ms on/100 ms off for 1 min and 10 ms on /200 ms off for 9 min. The resulting fluorocarbon coated aluminum disk exhibited a water contact angle of 75°. In the second step, the aluminum disk was positioned inside a mask having 12 circular openings of 3 mm diameter and the assembly was placed into an rf plasma reactor. Vapor-phase NIPAM was introduced to the plasma reactor at a pressure of 140 mTorr and polymerized as described previously.22 In brief, the deposition process included an 80 W methane deposition, followed by ppNIPAM deposition with stepwise decreases in power from 80 to 1 W. After removal of the aluminum disk from the mask, the ppNIPAM-grafted surface was rinsed with deionized water to remove uncrosslinked monomer before use. The resulting ppNIPAM-grafted surface gave a water contact angle of 13° at room temperature. MALDI MS Experiments. All MALDI MS experiments were conducted using a Bruker Autoflex MALDI TOFMS. The Autoflex target was machined to introduce circular cutouts suitable for loading the plasma polymer modified aluminum disks into the sample stage. The MALDI matrix for all experiments was CHCA dissolved in 50:50 methanol/water at a concentration of 12 mg/ mL. All MALDI mass spectra were acquired using a nitrogen laser operating at 377 nm and adjusted to an intensity just above threshold for ion formation. The mass spectra shown in Figures 1-3 represent the average of 20 laser shots, and the mass spectra shown in Figures 4 and 5 represent the average of 50 laser shots. (16) Walker, A. K.; Wu, Y.; Timmons, R. B.; Nelson, K. D.; Kinsel, G. R. Anal. Chem. 1999, 71, 268-272. (17) Walker, A. K.; Qiu, H.; Wu, Y.; Timmons, R. B.; Kinsel, G. R. Anal. Biochem. 1999, 271, 123-130. (18) Chen, K.; Walker, A. K.; Wu, Y.; Timmons, R. B.; Kinsel, G. R. J. Mass Spectrom. 1999, 34, 1205-1207. (19) Walker, A. K.; Land, C. M.; Kinsel, G. R. J. Am. Soc. Mass Spectrom. 2000, 11, 62-68. (20) Zhang, J.; Kinsel, G. R. Langmuir 2002, 18, 4444-4448. (21) Zhang, J.; Kinsel, G. R. Langmuir 2003, 19, 3531-3534. (22) Pan, Y. V.; Wesley, R. A.; Luginbuhl, R.; Denton, D. D.; Ratner, B. D. Biomacromolecules 2001, 2, 32-36.
Figure 1. MALDI mass spectra of cytochrome c on conventional stainless steel and ppNIPAM modified MALDI probes using CHCA matrix: (A) conventional MALDI mass spectrum on a stainless steel probe; (B) MALDI mass spectrum of a sample deposited on the ppNIPAM modified probe held at 40 °C after washing with 40 °C water and adding CHCA at 40 °C; (C) MALDI mass spectrum of a sample deposited on the ppNIPAM modified probe initially held at 40 °C and washed using 40 °C water after reducing the probe to 23 °C and adding CHCA at 23 °C; (D) MALDI mass spectrum of a sample deposited on the ppNIPAM modified probe held at 23 °C after washing with 23 °C water and adding CHCA at 23 °C.
Figure 2. MALDI mass spectra of the first washing from the ppNIPAM modified probe on a stainless steel probe using CHCA matrix: (A) the first 40 °C washing from the probe held at 40 °C; (B) the first 23 °C washing from the probe held at 23 °C.
RESULTS AND DISCUSSION Tests of the Thermoresponsive Surface Using Pure Protein. The initial group of experiments performed was designed to confirm that thermal cycling of the ppNIPAM modified MALDI target, resulting in binding and release of surface deposited proteins, could be observed in MALDI mass spectra acquired from the target. To perform this test, a series of experiments were performed using cytochrome c (0.02 mg/mL in deionized (DI) water) as the test protein. A 1.5 µL aliquot of this solution deposited on the ppNIPAM modified MALDI target spread over an area of approximately 10 mm2 resulting in a surface concentration of cytochrome c of about 3 ng/mm2. On the basis of an Analytical Chemistry, Vol. 79, No. 17, September 1, 2007
6841
Figure 3. MALDI mass spectra of cytochrome c prepared in 0.3% SDS using CHCA matrix: (A) conventional MALDI mass spectrum on a stainless steel probe; (B) MALDI mass spectrum of a sample deposited on the ppNIPAM modified probe initially held at 40 °C and washed using 40 °C water after reducing the probe to 23 °C and adding CHCA at 23 °C.
Figure 4. MALDI mass spectra of a cyanobacteria protein extract containing B-PER nonionic surfactant using CHCA matrix: (A) conventional MALDI mass spectrum on a stainless steel probe; (B) MALDI mass spectrum of the protein extract deposited on the ppNIPAM modified probe initially held at 40 °C and washed using 40 °C water after reducing the probe to 23 °C and adding CHCA at 23 °C.
estimated cross-sectional area for cytochrome c of 7.1 × 10-12 mm2/molecule,23 this surface concentration corresponds to approximately one monolayer on a microscopically smooth surface. This surface concentration was selected since any surface modulated binding/release of protein can only be expected to influence those molecules which are in direct contact with the surface itself. Initially, a conventional MALDI mass spectrum of cytochrome c deposited on a stainless steel MALDI probe was acquired. Specifically, a 1.5 µL aliquot of the cytochrome c solution was deposited on the probe followed by a 1.0 µL aliquot of 0.1% formic acid and a 2 µL aliquot of the CHCA MALDI matrix solution. (23) Gao, J.; Whitesides, G. M. Anal. Chem. 1997, 69, 575-580.
6842 Analytical Chemistry, Vol. 79, No. 17, September 1, 2007
Figure 5. MALDI mass spectra of a cyanobacteria protein extract containing B-PER nonionic surfactant and 0.4% SDS using CHCA matrix: (A) conventional MALDI mass spectrum on a stainless steel probe; (B) MALDI mass spectrum of the protein extract deposited on the ppNIPAM modified probe initially held at 40 °C and washed using 40 °C water after reducing the probe to 23 °C and adding CHCA at 23 °C.
Figure 1A shows the mass spectrum obtained from this sample. The two peaks at 6116 and 12 232 Da correspond to doubly and singly charged cytochrome c, respectively. In the second experiment, a 1.5 µL aliquot of the cytochrome c solution was deposited on the ppNIPAM modified MALDI probe, held at 40 °C by a block heater, and allowed to stand for about 30 min. Additional DI water was added to the sample spot during this time, as needed, to avoid complete evaporation of the liquid. Subsequently, the sample was washed three times using 2 µL aliquots of 40 °C DI water. Finally, a 1.0 µL aliquot of 0.1% formic acid and a 2 µL aliquot of the CHCA MALDI matrix solution, both held at 40 °C, were added to the sample spot. The MALDI mass spectrum obtained from this sample is shown in Figure 1B, where it is apparent that essentially no signal for the cytochrome c is observed. In the third experiment, the same procedure as in experiment two was followed, except the probe was removed from the block heater and allowed to equilibrate to room temperature (23 °C) before adding the formic acid and MALDI matrix solutions, also at room temperature. The MALDI mass spectrum obtained from this sample is shown in Figure 1C, where it is apparent that the reduction of the temperature of the probe and matrix solutions leads to the recovery of the signal for the cytochrome c. Finally, in a fourth experiment, the same procedure as in experiment two was again followed, except that the temperature of the probe was reduced to 23 °C before washing the probe with 23 °C DI water and adding the formic acid/MALDI matrix solutions, also held at 23 °C. The MALDI mass spectrum obtained from this sample is shown in Figure 1D, where it is observed that washing the probe at room temperature leads to the disappearance of the cytochrome c ion signal. The interpretation of the data shown in Figure 1 is consistent with expectations based on the thermoresponsive surface cycling between protein nonadsorptive and protein retentive states with changes in the temperature of the probe and sample preparation conditions. For the conditions used to acquire Figure 1B, it is
expected that excess cytochrome c deposited on the probe will be washed away by the 40 °C DI water. The remaining surface retained cytochrome c could be available for MALDI analysis but since the MALDI matrix is added at high temperature, the cytochrome c remains bound to the ppNIPAM surface and is unable to cocrystallize with the MALDI matrix. By reduction of the probe and formic acid/MALDI matrix solution temperature after washing the probe, the surface retained cytochrome c is released from the ppNIPAM polymer and becomes available for matrix cocrystallization leading to a signal for cytochrome c as observed in Figure 1C. However, if the probe temperature is reduced before the washing procedure, the majority of the cytochrome c is washed from the probe and is therefore unavailable to produce a signal in the MALDI mass spectrum, as is seen in Figure 1D. To provide additional support to this interpretation of the data, MALDI mass spectra were acquired from the first washing solutions obtained from experiments two (Figure 1B) and four (Figure 1D) by depositing the wash solutions on a conventional metal probe and adding the CHCA MALDI matrix to these samples. Figure 2A is the MALDI mass spectrum of the first 40 °C washing solution obtained from experiment 2 above. Only a weak signal for cytochrome c is observed in this mass spectrum. This result may be compared with the MALDI mass spectrum of the first 23 °C washing solution obtained from experiment four above (Figure 2B) where a somewhat more intense signal for the cytochrome c is observed. The results of the analyses of the washing solutions shown in Figure 2 are consistent with the interpretation of the data shown in Figure 1. Specifically, a smaller signal for the cytochrome c is expected when washing the probe at high temperature as only excess cytochrome c is expected to be washed away, and a significant fraction of the cytochrome c is expected to be retained by the ppNIPAM coated MALDI probe. In contrast, a larger signal for the cytochrome c is expected when the washing the probe at room temperature because of the protein nonadsorptive properties of the ppNIPAM polymer at this lower temperature. These experiments provide support for the assertion that the thermoresponsive ppNIPAM surface retains protein to a greater extent above the LCST and releases protein below the LCST. Application of the Thermoresponsive Surface as a Desalting Device. Additional studies were performed to determine if the thermoresponsive ppNIPAM surface could be used to augment cleanup of proteins prior to MALDI MS analysis. Specifically, it was expected that washing a contaminated protein sample at temperatures above the surface LCST, where the surface should strongly retain the proteins, should result in less loss of the protein to the washing solution and consequently more protein available for subsequent MALDI MS analysis performed at temperatures below the LCST. The results of a pair of experiments using 1.5 µL aliquots of a 0.02 mg/mL cytochrome c solution in 0.3% SDS are shown in Figure 3. Figure 3A is the MALDI mass spectrum of the cytochrome c in 0.3% SDS as deposited on a conventional metal probe surface. The deleterious effect of the SDS is clearly evident upon comparing the weak signal for the cytochrome c observed in this mass spectrum with the analogous spectrum of cytochrome c prepared in DI water shown in Figure 1A. Subsequently, the
SDS contaminated cytochrome c was deposited on the ppNIPAM coated MALDI probe held at 40 °C and then washed using 40 °C DI water to remove the SDS, prior to reducing the probe temperature to 23 °C before adding the MALDI matrix at 23 °C. Figure 3B shows the MALDI mass spectrum of the washed sample wherein a much stronger signal for the cyctochrome c is now evident. This result suggests that the thermoresponsive surface can be effectively used in the cleanup of contaminated protein samples. To further test the utility of the ppNIPAM coated MALDI probe, an analysis of a crude protein extract from cyanobacteria was performed. Figure 4A shows a MALDI mass spectrum of the proteins extracted from cyanobacteria using the commercial mild, nonionic detergent B-Per in 20 mM Tris‚HCl as deposited on a conventional metal probe surface. Figure 4B is the same protein extract as deposited on the ppNIPAM coated MALDI probe and after washing using the procedure described for cytochrome c. A comparison of parts A and B of Figure 4 reveals that the mass spectral quality is improved following washing of the sample. This improvement is especially evident for the higher molecular weight species at 13 116.9 and 18 164.7 Da. The effect of SDS contamination and subsequent cleanup, using the thermoresponsive MALDI probe, on the cyanobacteria protein extract MALDI mass spectra was also investigated. In these experiments the cyanobacteria protein extract was deliberately contaminated with 0.4% SDS and then analyzed by MALDI MS, both as is on a stainless steel MALDI probe and following washing on the ppNIPAM coated MALDI probe. Figure 5A is the MALDI mass spectrum of the SDS contaminated cyanobacteria protein extract as deposited on a conventional metal probe and following addition of the CHCA matrix. Essentially no useful ion signals above 7000 Da are observed in this mass spectrum. Figure 5B is the MALDI mass spectrum of the SDS contaminated cyanobacteria protein extract as deposited on the ppNIPAM coated MALDI probe and after washing using the procedure described for cytochrome c. The MALDI mass spectrum is dramatically improved following this washing procedure, both in comparison with the SDS contaminated sample (Figure 5A) as well as in comparison with the mass spectra obtained from the samples without SDS contamination (parts A and B of Figure 4). Why adding the SDS to the protein extract and then removing it through washing should lead to a general improvement in the bacterial protein ion signals is not entirely clear. It is conceivable that the SDS serves to denature the proteins, allowing more efficient MALDI ionization, but this effect does not seem apparent in the analysis of the cytochrome c, as revealed by comparing Figure 1C with Figure 3B. Alternatively, it is possible that the SDS addition and removal serves to eliminate a broader spectrum of cellular contaminants that are acting to inhibit the crystallization and/or desorption/ionization of the bacterial proteins. It has also been suggested in several recent publications that the SDS may itself serve to enhance the desorption/ionization efficiency in a MALDI experiment if the concentration of the SDS is at, or near, the critical micellar concentration.24,25 While our goal in these experiments was to remove the SDS in the washing procedure, (24) Tummala, R.; Limbach, P. A. Rapid Commun. Mass Spectrom. 2004, 18, 2031-2035. (25) Tummala, R.; Green-Church, K. B.; Limbach, P. A. J. Am. Soc. Mass Spectrom. 2005, 16, 1438-1446.
Analytical Chemistry, Vol. 79, No. 17, September 1, 2007
6843
the possible retention of some residual SDS within the protein sample cannot be discounted. Regardless, it is clear that the thermoresponsive ppNIPAM coated MALDI probe can serve as an effective stage for cleanup of bacterial protein extracts, leading to improved MALDI mass spectra. CONCLUSION For the direct modification of the surface of MALDI targets, rf plasma polymer deposition offers an attractive approach. This approach to the deposition of polymer thin films is relatively simple and offers unique opportunities to introduce a wide variety of chemical functionalities, both generally across the surface of the MALDI probe and in the form of patterned arrays. In this report a functional MALDI probe, having temperature controlled protein retentive properties, is constructed via the patterned deposition onto a MALDI target of a thermoresponsive polymer thin film resulting from rf plasma polymerization of NIPAM. The results of a group of control experiments, performed using cytochrome c as the analyte, demonstrate that the protein retentive properties of the ppNIPAM coated probe vary as expected in response to the temperature of the probe and the solvents used to deposit and wash the protein sample. More importantly, in this work it is shown that this thermoresponsive polymer coating, in combination with control of the probe and solvent temperatures, can be used effectively to clean
6844
Analytical Chemistry, Vol. 79, No. 17, September 1, 2007
up contaminated protein samples prior to MALDI analysis, thus minimizing the deleterious effect of common protein processing contaminants such as SDS. This utility is particularly evident in the analysis of SDS contaminated bacterial protein extracts where washing of the sample as deposited on the ppNIPAM coated MALDI probe significantly improves the mass spectral quality. While the use of this procedure for sample cleanup does involve a modest amount of temperature control of the MALDI target and the solutions used, the temperature cycles used are certainly not particularly extreme (23-40 °C) and could be easily incorporated in a conventional sample preparation scheme. Furthermore, the rf plasma approach to deposition of the ppNIPAM coating offers considerable flexibility, allowing the incorporation of this sample cleanup stage as arrays on conventional MALDI targets and in microfluidic sample manipulation devices. ACKNOWLEDGMENT G.R.K. acknowledges NSF Grant CHE-0317073. B.D.R. acknowledges NSF-Engineering Research Center Grant No. ERC-9529161.
Received for review May 29, 2007. Accepted May 31, 2007. AC0711236
