Direct Atmospheric Pressure Coupling of Polyacrylamide Gel

Electrophoresis to Mass Spectrometry for Rapid Protein Sequence. Analysis ... coupling mass spectrometry to polyacrylamide gel electrophoresis. In con...
0 downloads 0 Views 174KB Size
Download PDF
Direct Atmospheric Pressure Coupling of Polyacrylamide Gel Electrophoresis to Mass Spectrometry for Rapid Protein Sequence Analysis Joshua J. Coon,*,†,§ Heather A. Steele,‡,§ Philip J. Laipis,‡ and W. W. Harrison† Department of Chemistry, University of Florida, Gainesville, Florida 32611, Department of Molecular Biology and Biochemistry, University of Florida, Gainesville, Florida 32611 Received April 15, 2003

Using laser desorption-atmospheric pressure chemical ionization we describe a novel approach for coupling mass spectrometry to polyacrylamide gel electrophoresis. In contrast to other approaches, the method allows for the direct sampling of a polyacrylamide gel-embedded protein without the addition of any exogenous matrixes and is performed at atmospheric pressure. After electrophoresis and enzymatic digestion, the gel is analyzed at AP by photons that desorb neutral peptide molecules, followed by corona discharge ionization in the gas-phase, and subsequent mass analysis. Our experimental results demonstrate the method to (1) rapidly identify electrophoresed proteins via “peptide fingerprinting” using protein databases, (2) detect single-amino acid polymorphisms, and (3) has potential for sub-picomole sensitivity while still maintaining in situ gel desorption-ionization at ambient conditions. Keywords: ionization mass spectrometry • desorption ionization • polyacrylamide gels • proteins • ablation

Introduction Owing to its exceptional sensitivity, mass spectrometry stands at the forefront of an emerging subdiscipline of protein chemistry coined “proteomics”: the classification of the protein complement expressed by the genome of an organism.1 However, proteomics cannot be accomplished by use of the same analytical tools used in genomics, e.g., DNA chips, PCR, etc.; rather, it will require the development of new analytical methodologies.2-4 Protein sequencing by mass spectrometry was being performed already during the 1980s,5-7 but has increased in popularity and importance since, driven by the development of novel technology for coupling protein characterization methods to mass spectrometric analysis.8-12 Due to the sheer volume and compositional variation of expressed proteins, the sustained development of high-throughput mass spectrometric methods will be indispensable for characterization of the human proteome on a suitable time-scale. Complex mixtures of proteins require some degree of separation before introduction into the mass spectrometer, otherwise the major components can suppress the signal of the other components.13 Two-dimensional polyacrylamide gel electrophoresis (2-DE) remains the most widely employed protein separation technique because of its ability to separate thousands of proteins in a single analysis.14,15 Nonetheless, as with all methods, certain limitations remain, including dynamic * To whom correspondence should be addressed. † Department of Chemistry, University of Florida. ‡ Department of Molecular Biology and Biochemistry, University of Florida. § Present address: Department of Chemistry, University of Virginia, Charlottesville, Virginia 22901.

610

Journal of Proteome Research 2003, 2, 610-617

Published on Web 07/29/2003

range, mass range (6000-300 000),4 detection sensitivity, and incompatibility with hydrophobic proteins. Further, when 2-DE is combined with mass spectrometry, peptides/proteins typically require extraction, where yields of 1% or below are common.16 By far the most common method (“spot-picking”) to couple 2-DE (or PAGE) to mass spectrometry uses electrophoresis, chemical staining, excision of visualized spots, in-gel protease digestion, extraction, cleanup,17 and mixing with a matrixassisted laser desorption/ionization (MALDI) matrix for MALDIMS analysis.18 Despite the widespread application of this method, inherent problems limit its achievable sensitivity: (1) only stained spots are processed for MS analysis; (2) peptides must be extracted from the gel following digestion, so sensitivity cannot exceed that of chemical staining; and (3), the sheer volume of spots in a typical 2-DE experiment to excise and prepare for MS analysis typically requires some degree of automation.19 These limitations were recently highlighted in a study of yeast proteins by 2-DE, followed by LC-MS.20 An alternative to the so-called “spot picking” method described above is that of direct gel analysis via mass spectrometric imaging. An entire gel can be mass spectrometrically imaged, a process where sensitivity is not limited by chemical staining. Thus, we believe the direct MS analysis of PAGE (or 2-DE) has potential for both increased sensitivity and throughput as compared to current methods. The prevalent methodology for direct gel analysis, employed for nearly a decade, has not become a routine tool in proteomics research. Direct gel analysis has been carried out primarily by two methods: (1) electrophoresed proteins are transferred to a membrane, digested21-27 or left intact28-32, 10.1021/pr034031f CCC: $25.00

 2003 American Chemical Society

Coupling of PAGE Electrophoresis to MS

coated with a MALDI matrix, and inserted into vacuum for MALDI-TOF-MS analysis, (2) electrophoresed proteins are digested in situ (or left intact), the entire gel soaked in a MALDI matrix, and placed in a vacuum for UV-MALDI-TOF analysis.33-36 More recent work has suggested the gel material itself can be used as the MALDI matrix upon desorption with a tunable, mid-IR free-electron laser (FEL) operating at 5.9 µm; however, a widespread application of this approach will likely be hindered by the complexity and cost of the FEL.37 One of the most significant drawbacks in these approaches arises from the need to insert whole gels, or gel portions, into the vacuum system for mass analysis (resulting in a required step of dehydration or freezing prior to vacuum exposure). Alternatively, gel sampling and/or imaging under atmospheric pressure (AP) conditions requires no preparation, simplifying the process and increasing throughput. Furthermore, direct gel analysis could be performed with mass spectrometers having AP inlets, thus creating a single instrument capable of a variety of ionization techniques (e.g., electrospray ionization (ESI), AP chemical ionization (APCI), AP-MALDI, and laser desorption (LD)-APCI). Using a new method of AP laser desorption/ionization (LDAPCI),38 our laboratory has recently demonstrated the first direct gel analysis at AP of electrophoresed proteins.39 In contrast to MALDIswhere desorption and ionization are accomplished in a single stepsLD-APCI employs an IR laser pulse at 10.6 µm to effect the desorption of neutral molecules at AP, followed by ionization in the gas phase with a corona discharge.40 This technique draws on the fact that a substantially larger number of gas-phase neutral molecules, as opposed to ions, are produced during a desorption event.41 And because desorption is decoupled from ionization, the technique allows for individual optimization of each step with increased efficiency and selectivity. This benefit has lately been highlighted for the analysis of aqueous peptide solutions, where the use of the corona discharge enhanced the AP-IR-MALDI-generated signal of the analyte-protonated molecule by factors up to 1400.42 In this report, we describe the continued development of direct AP gel analysis with experimental results demonstrating that the method can (1) rapidly identify electrophoresed proteins via “peptide fingerprinting” using protein databases, (2) detect single-amino acid polymorphisms, and (3) achieve useful sensitivity ranges while still maintaining in situ gel desorption-ionization at ambient conditions.

Experimental Section Laser Desorption Interface. A detailed description of the LDAPCI source interface can be found elsewhere.40,42 Briefly, this source utilizes a heated capillary AP inlet (ThermoFinnigan, San Jose, CA) to transport AP generated ions into vacuum for mass analysis. The gel slices were applied directly to a 4 mm diameter stainless steel removable target located ∼3 mm axially from the heated capillary inlet and ∼2 mm offset from center. The target was held at an offset potential of +2 kV (model 205A, Bertan Associates, Hicksville, NY). The corona needle was positioned ∼3 cm from the inlet of the heated capillary with the tip axially aligned. A potential of +8.1 kV, from a standard ESI power supply (Analytica, Branford, MA), was used to generate the corona discharge. Laser desorption was achieved by irradiation of the target with a pulsed CO2 laser operating at 10.6 µm (µ-TEA, Laser Science Inc., Franklin, MA). The beam was focused to a spot

research articles diameter of ∼0.5 mm (∼2.0 × 107 W/cm2, assuming homogeneous distribution) using a 10 cm focal length zinc selenide lens (Laser Research Optics, Providence, RI). The laser pulsing was synchronized to coincide with the prescan period of the scan function, which occurred 1 ms before the ion injection period of each microscan. Mass Spectrometer. The mass spectrometer used in these studies, a modified quadrupole ion trap system (Finnigan GCQ, ThermoFinnigan, Austin, TX), was adapted to accept a twostage differentially pumped vacuum chamber and fitted with an Analytica ESI source manifold (Branford, MA). This manifold was further modified to accept a metal heated capillary AP inlet. To reduce solvation of the ions, the heated capillary was maintained at 200 °C; an offset potential of +130 V was applied to assist de-clustering. In certain cases, MALDI-TOF-MS analysis of tryptic peptides (extracted from in-gel digests) was performed. The MALDI-TOF mass spectrometer used in those cases was an Applied Biosystems Voyager-DE Pro (Framingham, MA) system, operated in the reflectron mode. Sample Preparation. Direct Analysis of Myoglobin-Containing Gel. Protein standards (cytochrome c, hemoglobin, and myoglobin,) were purchased from Sigma (St. Louis, MO) and used without further purification. Polyacrylamide gels were purchased from BioRad (15% Tris-HCl Ready Gels, Hercules, CA) and were stained with Coomassie brilliant blue G-250 after electrophoresis. Following staining, the spots were excised and washed for 30 min twice with 200 mM NH4HCO3, pH 8 washing buffer. Next, the gel slices were brought to dryness in a vacuum centrifuge (SVC100, Savant, Holbrook, NY). Using the difference in gel mass (before and after drying), the gel slice was rehydrated in trypsin solution (31.2-7.8 ng/µL trypsin in 25 mM aqueous ammonium bicarbonate) (Promega, Madison, Wisconsin, USA) containing 50 mM NH4HCO3 (typically 5-8 µL depending on gel slice size). Once rehydrated the gel slices were incubated at 37 °C for 20 h, after which the slices were placed on the LD-APCI target and directly analyzed. HCA II Single-Point Mutation Analysis. The human carbonic anhydrase II (HCAII) mutant was prepared using a bacterial expression vector optimized for site-specific mutagenesis and protein synthesis that expressed the wild-type form of HCAII.43 This vector was derived from the T7 expression vectors of Studier et al.44 and contained a bacteriophage f1 origin of replication for production of single-stranded DNA. The triple mutant was prepared using single-stranded DNA and mutating oligonucleotides in a stepwise fashion, so that the effect of individual mutations on catalytic activity could be determined. The triple mutant was assembled by ligation of appropriate restriction fragments from the individual constructs by standard recombinant DNA techniques. Protein expression from these vectors was in the range of 20-25 mg/L. All mutations were confirmed by DNA sequencing of the expression vectors. Purification of HCA II A23CL203CC205S was performed by affinity gel chromatography using p-(aminomethyl) benzenesulfonamide coupled to agarose beads.45 The concentration of HCA II was determined from the extinction coefficient 5.2 × 104 M -1 cm-1 at 280 nm.46 The mutant HCAII and wildtype were electrophoresed and digested as outlined above. In certain cases, however, reduction and alkylation were performed in the following manner: (1) after ammonium bicarbonate washes, gel slices were brought to dryness using a vacuum centrifuge and rehydrated with 150 µL of 10 mM 1, 4 dithiothreitol (DTT) in 100 mM aqueous Journal of Proteome Research • Vol. 2, No. 6, 2003 611

research articles ammonium bicarbonate (pH 8.9) for 30 min at 56 °C, (2) followed by the application of 150 µL of 55 mM iodoacetamide in 100 mM aqueous ammonium bicarbonate (pH 8.9) for 20 min at room temperature, (3) the gel slices were brought to dryness once again in a vacuum centrifuge and rehydrated with a trypsin solution as described above. Chemical Staining Study. Several polyacrylamide gels were loaded with 1000 pmol/lane of equine cytochrome c and electrophoresed. Afterward, each gel was stained using one of the following chemicals: (1) Coomassie brilliant blue G-250 (BioRad, Hercules, CA), (2) Coomassie brilliant blue G-250 Biosafe (BioRad, Hercules, CA), (3) Coomassie brilliant blue R-250 (BioRad, Hercules, CA), (4) silverstain without de-stain (SilverQuest, Invitrogen, Carlsbad, CA), (5) silverstain using kit de-stain (SilverQuest, Invitrogen, Carlsbad, CA), and (6) no stain. Numerous bands were excised from each gel, followed by digestion and analysis as outlined above. Thin-Gel Preparation. A Mini-Protean 3 gel casting system (BioRad, Hercules, CA) was employed to prepare polyacrylamide gels (15%) of 1.0 and 0.5 mm thickness. Lanes were cast with a width of 5.0 mm on the 1.0 mm gel (standard width); however, lane width was reduced to 2.5 mm on the 0.5 mm gels. All gels were cast using GelBond PAG film (Cambrex, East Rutherford, New Jersey, USA) to provide a supportive backings a process found to greatly simplify gel handling. Next, the 1.0 mm gels were loaded with 1000, 500, and 250 pmol of cytochrome c, while the 0.5 mm gels were loaded with 250, 125, and 65 pmol. Despite the lower amounts loaded on the thinner gels, the volume containing the protein was estimated to be identical due to reduced lane width and thickness. Following electrophoresis gels were processed and analyzed as outlined above.

Coon et al.

Figure 1. Direct LD-APCI-MS analysis of a polyacrylamide gel containing tryptic peptides of myoglobin. Mass spectrum represents the average of ∼75 single-shot mass spectra, 1000 pmol protein loaded. Table 1. Peptides Resulting from a Tryptic Digest of Horse Myoglobina theortical m/z

position

peptide sequence

1885 1854 1816 1607 1503

103-118 80-96 1-16 17-31 119-133

YLEFISDAIIHVLHSK GHHEAELKPLAQSHATK GLSDGEWQQVLNVWGK VEADIAGHGQEVLIR HPGDFGADAQGAMTK

1379 1272 791 749 736 709 662 650 631 470 409 397 310 294 284 275

64-77 32-42 57-63 134-139 97-102 51-56 57-62 148-153 140-145 99-102 43-45 48-50 146-147 46-47 97-98 78-79

HGTVVLTALGGILK LFTGHPETLEK ASEDLKKb ALELFR HKIPIK b TEAEMK ASEDLK ELGFQG NDIAAK IPIK FDK HLK YK FK HK KKb

Results and Discussion Our initial efforts to directly couple polyacrylamide gel electrophoresis to mass spectrometry demonstrated the value of the two-step LD-APCI approach.39 Unlike MALDI, which requires either the addition of an exogenous matrix or tuning of laser wavelength to coincide with the maximum absorbance of the gel, LD-APCI decouples desorption from ionization and allowed for the first direct AP mass spectrometric analysis of tryptic peptides from polyacrylamide gels. Specifically, the study demonstrated that ablation of a peptide-containing gel with 10.6 µm laser radiation at AP did not produce detectable peptide ions, but rather a substantial number of intact gasphase peptide molecules. By use of the corona discharge those gas-phase peptide molecules were ionized by proton transfers a gentle process, because there is extensive collisional cooling at ambient pressure. This report further explores the use of LDAPCI to directly couple mass spectrometry to PAGE with particular emphasis on applicability, effects of chemical staining, and sensitivity. Direct Analysis of Myoglobin-Containing Gel. The protein, horse myoglobin, was obtained and loaded at 1000 pmol/lane on a 15% polyacrylamide gel and electrophoresed. For analysis the spots were excised, washed, dried, and rehydrated in a trypsin-containing solution. The gel slices were then incubated at 37 °C for 20 h, after which the pieces were placed on the LD-APCI-MS target and directly analyzed. The resulting spectra (average of 75 single-shot mass spectra) are presented in Figure 1. Note that nearly every m/z detected results from a tryptic peptide of myoglobin. This likely occurs because in the gas-phase neutral molecule population pro612

Journal of Proteome Research • Vol. 2, No. 6, 2003

147

63, 78,79

K

a

Bold entries represent peptides detected during PAGE-LD-APCI-MS analysis. b Denotes detected missed cleavages. Solid lines mark scanned mass range.

duced after a laser desorption event, only those with the highest proton affinities will be ionized by the corona discharge, peptides in this case. For this protein, no oxidized methionines were detected; however, in our original report,39 a tryptic peptide containing methionine was observed in both oxidized and non-oxidized forms - presumably the oxidation occurred during electrophoresis. Table 1 presents the expected peptides from a tryptic digest of horse myoglobin. In fact, of the 16 tryptic peptides within the scanned mass range (instrument has an upper m/z limit of 1500), 13 were detected resulting in 43% protein coverage. Protein identification was accomplished by entering the observed masses (signal/background (S/B) > 3) into an online database (Mascot, peptide tolerance of 1 Da). The resulting search matched 13 peptides to horse myoglobin with the top ranking probability scores198. Further, because of the low background, the LD-APCI-MS technique has potential to identify smaller peptides normally obscured by the matrix background produced during MALDI-TOF-MS analysis. Perhaps the most inviting aspects of direct gel analysis using LD-APCI-MS are its simplicity and speed. For example, gel band mounting and MS analysis required ∼2-4 min for each

Coupling of PAGE Electrophoresis to MS

research articles

Figure 2. LD-APCI-MS and MALDI-TOF mass spectra of in-gel protein digests. (A) Direct in-gel LD-APCI sampling of human carbonic anhydrase II (HCA II) wild-type protein digest, (B) HCA II mutant, (C) HCA II mutant treated with iodoacetamide prior to digestion, and (D) MALDI-TOF analysis of HCA II mutant (iodoacetamide treated) in-gel digest extract. The LD-APCI-MS data represents the average of ∼100 single-shot mass spectra, while the MALDI-TOF-MS mass spectrum results from the accumulation of 100 single-shot mass spectra.

bandsbypassing many lengthy steps including extraction, cleanup, and loading for a typical MS analysis. And because sampling is performed at AP, the procedure can be easily automated, creating the possibility for whole gel mass spectrometric imaging. Detection of Single Amino Acid Point Mutations. The biological function of proteins can be altered by modifications carried out in the cell following synthesis (post-translational modification). These modifications include proteolytic cleavage and covalent modifications such as acetylation, glycosylation, hydroxylation, methylation, phosphorylation, and so forth. Additionally, up or down-regulation of post-translationally modified proteins can be associated with diseased or nondiseased cellular states. Amino acid changes are also observed, where a single amino acid is replaced by another in the protein’s sequence. These amino acid polymorphisms may or may not affect biological function, i.e., mutagenic or silent, respectively. Therefore, a method capable of rapidly monitoring protein modifications has potential value as an important tool for the early diagnosis of disease. In this section, we demonstrate the ability of the LD-APCI-MS methodology to detect single-amino acid polymorphisms in the protein human carbonic anhydrase II (HCA II). Mutant HCA II, with three single-point mutations, was prepared using methodology reported elsewhere.43 The

amino acid sequence of the HCA II was modified in the following manner: (1) residue 23, alanine to cysteine, (2) residue 203, leucine to cysteine, and (3) residue 205, cysteine to serine. 1000 pmol of both mutant HCA II and wildtype (unmodified) HCA II were loaded (separately) onto a polyacrylamide gel and electrophoresed. As described above, the protein bands were excised, enzymatically digested, and mass analyzed directly by LD-APCI-MS and indirectly by MALDI-TOF-MS. The results are presented in Figure 2. Unfortunately, two of the mutations were contained in a single tryptic peptide whose molecular weight (∼3500 daltons) was well beyond the mass range of our LD-APCI-MS instrument. However, the third mutation, located in the DFPIAK tryptic peptide (where A is replaced with C in the mutant), was observed in the LD-APCI-MS analysis of the wild-type HCA II (Figure 2A, m/z 690). Direct gel analysis of the mutant protein band did not produce signal at m/z 690, but a new mass was observed at m/z 722 (Figure 2B), albeit with poor S/B. A shift of +32 daltons corresponds to the replacement of A with C. Nonetheless, this identification could be considered tentative; therefore, to confirm the mutation, another protein band containing the mutant HCA II was treated with iodoacetamide prior to enzymatic digestion. This reagent selectively reacts with cysteine residues to produce carboxamidomethyl cysteine, Journal of Proteome Research • Vol. 2, No. 6, 2003 613

research articles

Coon et al.

Figure 3. Direct LD-APCI-MS analysis mass spectra of cytochrome c containing gels using various chemical staining methods. (A) Coomassie brilliant blues G-250, (B) Biosafe G-250, (C) and R-250, (D) no stain, (E) silver without destain, (F) silver destained. The above scales are fixed at 8.0 × 105 (arbitrary units) for all panels; b denotes detected cytochrome c tryptic peptides, whereas O indicates Coomassie fragment ions.

adding 57 daltons to the tryptic peptide residue of interest. LDAPCI-MS analysis of that band is shown in Figure 2C. Indeed, a new peak can be observed at m/z 779, a shift of +57 daltons. Six other tryptic peptides were also observed in the LD-APCIMS analysis of both the mutant and the wild-type HCA II protein bands. In this case, confirmation of the mutation identity and location could be made without tandem MS, due to selective modification of cysteine and the fact that no other amino acids are isobaric with alanine. For a comparison, the tryptic peptides from one of the mutant HCA II protein bands were extracted, following in-gel enzymatic digestion, and prepared for MALDI-TOF-MS analysis. Cleanup of the digestion extract was accomplished by use of a C18 stationary phase-containing pipet tip (ZipTip), followed by elution onto a MALDI target with a 15 mg/mL 4-hydroxy-R-cyanocinnamic acid matrix solution. The MALDITOF-MS analysis of the digest extract is presented in Figure 2D. Only two tryptic peptides were observed in the 100-1000 dalton mass range in the MALDI-TOF-MS analysis, neither of which corresponds to the mutant peptide. Moreover, a significant number of matrix background peaks were observed in the low mass range. To be fair, a number of other tryptic peptides were detected by MALDI-TOF-MS analysis in the mass range of 1000-2500; however, the tryptic peptide carrying the other two mutations was not. These data, we believe, show that the LD-APCI direct approach outperforms MALDI-TOF-MS analysis in the 1001000 dalton mass range. In this case, LD-APCI detected one of the three mutations, whereas MALDI-TOF-MS analysis detected none. Moreover, LD-APCI offers a more rapid and convenient method, compared to MALDI, with potential to become a valuable tool for both proteomics and the diagnosis of disease. But much work remains before the method can be sucessfully implemented for these proposed applications. Sensitivity and mass range are two aspects requiring further evaluation and improvement. The remaining sections of this paper discuss important parameters such as the effect of digestion volume 614

Journal of Proteome Research • Vol. 2, No. 6, 2003

and gel staining on sensitivity. Finally, the last section presents other strageties for enhancing sensitivity. Gel PreparationsDigestion. Most protocols written for ingel enzymatic digestion of proteins are designed to cover the entire gel slice with a large volume of enzyme-containing solution. In that procedure, the peptides, once produced by enzymatic cleavage of the protein, are free to migrate out of the gel and accumulate within the digestion solution. Following digestion the remaining peptides are extracted from the gel with a series of solvents. Those solutions are then combined, desalted, and concentrated for further MS analysis. In direct gel analysis, however, extraction of the peptides from the gel will result in a degradation of sensitivity. Therefore, a typical in-gel digestion protocol was modified such that small volumes of digestion solution were added to the gel slices to minimize peptide extraction. But this approach raised other concerns about the efficiency of the digestion process. For example, is an excess of solution necessary for proper enzyme function? To answer this question, a series of in-gel enzymatic digestions was performed on a mixture of two unresolved proteins (bovine hemoglobin and horse cytochrome c) with varying digestion volumes. The volumes of enzyme solution were 15, 25, and 35 µL; note the enzyme concentration was adjusted such that each volume delivered the same magnitude of enzyme. Gel bands were sized to eliminate excess digestion solution for the lowest volume, 15 µL, when rehydrated (following dehydration by vacuum centrifuge) with the digestion solution. After LD-APCI-MS analysis the number of detected tryptic peptides (S/B > 3) from the two proteins was calculated for each digestion volume: 15 µL - 18 peptides ( 0.6, 25 µL - 12 peptides ( 1.2, and 35 µL - 9 peptides ( 4.2. From these data, increased digestion volumes result in reduced peptide detection efficiencysa trend explained by increased peptide extraction. Gel PreparationsStaining. For all analyses presented above, the stain Coomassie brilliant blue (G-250) was used to locate protein bands for excision and digestion. Consequently, upon

Coupling of PAGE Electrophoresis to MS

research articles

Figure 4. LD-APCI-MS mass spectra resulting from the direct analysis of gel bands from either 1.0 mm thick (A-C, gel image G) or 0.5 mm thick (D-F, gel image H) gels, following in-gel protein digestion. Note the detectable amount of protein is lowered by reducing the volume in which it is contained.

LD-APCI-MS analysis the gel bands still accommodated the stain and its effect on the process was unknown. Initially, we surmised the presence of stain during the analysis was likely not beneficial. To examine this hypothesis, we compared LDAPCI-MS mass spectra of enzymatically digested proteincontaining gels stained with a variety of chemicals (Coomassie brilliant blues G and R-250, biosafe G-250, and silver stain with two types of de-staining) to those exposed to no stain. The results of this analysis are shown in Figure 3 (note relative abundance scale fixed for all panels). Surprisingly, sampling of the Coomassie stained gels (Figure 3A-C) generated the greatest number of detected tryptic peptides, producing higher-quality spectra than even the unstained gels. Overall, those gel bands produced similar spectra, but on average the S/B of the G-250 was improved compared to R-250 and the Biosafe formulation (composition proprietary). The silver stained gel showed some detectable peptides, even if with poor S/B, when the stain was left in the gel (Figure 3E); however, when the silver stain was removed(Figure 3F), prior to digestion using the MS sensitive protocol from the manufacturer, no detectable tryptic peptides were observed. The gels stained with Coomassie brilliant blues were bright blue in color, the silver stained bands (not de-stained) were gray, whereas the nonstained and the silver destained bands were clear. On the basis of the visual difference and mass spectra, we believe the stain molecules could be playing a role in the analyte desorption process. Specifically, absorbance by the matrix at the wavelength of the incident photons is critical for molecular desorption. Both Coomassie dyes possess a strong

absorption at 10.6 microns, which may explain the improved spectra in its presence. We attempted to characterize any absorbance differences between Coomassie stained gels and their nonstained counterparts by use of attenuated total reflectance-infrared spectroscopy (ATR-FTIR). Using this device, the sample was placed on a crystal surface and absorbance spectra were recorded for both stained and nonstained gel slices. However, because the gels were moist and the ATR provides minimal sample penetration, the resultant absorbance spectra were identical to the gel storage buffer and no useful comparison could be made. Difficulties in acquiring comparable absorbance data prevent comfirmation of our hypothesis that staining elevates the absorbance of the gel at 10.6 microns and thereby increases the peptide detection sensitivity. Future studies will explore desorption at a wavelength of ∼3.0 microns, where the gel itself has a significantly higher absorbance. Nevertheless, these data indicate sensitivity may be limited by the presence of stain molecules, suggesting that molecular desorption is strongly affected by a large excess of stain molecules, requiring large amounts of protein for direct gel analysis. Sensitivity. In our original report,39 no analyte signals were observed from gel bands where less than 500 pmol protein had been loaded; hence, sensitivity is a concern. In this report, we have demonstrated Coomassie blue staining beneficial, suggesting enhancement of gel absorption at 10.6 microns improves molecular desorption and thereby sensitivity. These observations indicate that desorption at 3.0 micronsswhere gels have a stronger absorptionswould lead to further imJournal of Proteome Research • Vol. 2, No. 6, 2003 615

research articles provements in sensitivity, and obviate the need for chemical stain-assisted desorption. We do not have, at present, a suitable laser to pursue this assumption. Another means to improve sensitivity is presented by use of thinner gels. Despite the deeper penetration depth of IR vs UV irradiation,47 inspection of the gel slices after sampling revealed the removal of very little material from the surface of the gel. Since the gels employed in this study were 1 mm thick, the majority of the loaded sample was not subjected to LD-APCIMS analysis, assuming a homogeneous protein distribution within the gel. To investigate this, polyacrylamide gels of 1.0 and 0.5 mm thickness were prepared. Lanes were cast with widths of 5.0 mm (1.0 mm gel) and 2.5 mm (0.5 mm gel) using GelBond PAG film to provide a supportive backing. Next, the 1.0 mm gels were loaded with 1000, 500, and 250 pmol of equine cytochrome c, whereas the 0.5 mm gels were loaded with 250, 125, and 65 pmol. Despite the lower amounts loaded on the thinner gels, the concentration of the protein was estimated to be identical due to reduced volume. Following electrophoresis, gels were processed and analyzed as outlined above. Representative LD-APCI-MS direct analysis spectra of the 1.0 mm gels are presented in Figure 4A-C, with Figure 4D-F displaying those from the 0.5 mm gels. As expected, a comparison of Figure 4A with 4C demonstrates that a 4-fold reduction of protein from 1000 to 250 pmol results in the questionable detection of only 3 tryptic peptides. In contrast, when the volume containing the 250 pmol is reduced by a factor of 4 (Figure 4D) the spectra produced are notably improved compared to Figure 4C, and more comparable to those containing 1000 pmol (Figure 4A). LD-APCI-MS analysis of gel bands containing 125 pmol (Figure 4E) also displayed good S/B, detecting several tryptic peptides. This study supports our hypothesis that LD-APCI-MS direct gel analysis is concentration, rather than mass sensitive. Further exploitation of this attribute holds real promise. A particularly well-suited approach may be that of ultrathin layer gel electrophoresis,48-50 using gels as thin as 25 µm in combination with isoelectric focusing (2D-E).51 Even without further optimization this combination could drop detectable amounts from ∼100 pmol at present, into the range of ∼1-5 pmol loaded protein. Besides altering the desorption wavelength and reducing gel size, there are several other approaches with considerable potential to improve sensitivity of the measurement. A recent study has reported a 20-fold improvement in AP-MALDI sensitivity by incorporation of a heated countercurrent gas in the ion source region to improve ion transport.52 Anticipated research in our laboratory will employ the use of multiple laser pulses per ion accumulation event, as opposed to the approach used hereinsa single laser pulse followed by mass analysis. And not to be underestimated, the interfacing of this source to a modern, commercial instrument is expected to provide significant improvements in detectable amount as well. Considering the results presented herein, and further development as described, we believe the novel method has potential to detect sub-picomole amounts of protein directly from polyacrylamide gels at AP.

Conclusions We have employed LD-APCI for the direct sampling of proteins from polyacrylamide gels at APsto our knowledge, the 616

Journal of Proteome Research • Vol. 2, No. 6, 2003

Coon et al.

only method capable of such analysis. The method was applied sucessfully both to identify protein sequence via peptide mass mapping and to detect single-amino acid polymorphisms. Additionally, these data demonstrate the method superior to MALDI for low mass peptide identification. Further, though not used here, tandem mass spectrometry could be easily incorporated to enhance sequence analysis information. Notwithstanding this success, other aspects, including sensitivity, require further improvement before the method will become analytically useful. On the basis of this work we propose that absorption at the desorption wavelength to be critical for successful LD-APCI direct gel analysis. In fact, the presence of Coomassie stain molecules was found to enhance the detection of tryptic peptides, indicating desorption at 3.0 µm (where the gel itself has a stronger absorbance) could both elminate this dependence and provide substantial improvements in detection sensitivity. We have also described the use of thinner gels, an approach with promise to push detectable amounts into useful sensitivity ranges. Mass range is another aspect of the LD-APCI method requiring further investigation. The mass spectrometer utilized in this worksa home-built ion trap systemshas a limited mass range (100-1500 Daltons); however, it cannot optimally inject ions across its entire mass range. Consquently, the upper mass range of the LD-APCI method remains in question and will be the subject of future research using other instrumentation. To summarize, the coupled gel electrophoresis-mass spectrometry technology we present has potential to make a significant impact on the speed and simplicity of protein identification methodology. This promise mainly stems from its unique ability to function under ambient conditions at AP, with attendant possibilities for automation and imaging in high throughput protein sequence analysis.

Acknowledgment. The authors would like to acknowledge Dr. Rick Yost for loan of the instrumentation used in these studies. J.J.C. thanks Drs. Nancy Denslow, Andy Ottens, Kevin McHale, Kathryn Williams, and Dave Powell for helpful discussions. Finally, J.J.C. gratefully recognizes support by a Superfund Fellowship from the Superfund Basic Research Program (Grant No. 2P42 ES07375). References (1) Wilkins, M. R.; Sanchez, J. C.; Golley, A. A.; Appel, R. D.; Humphery-Smith, I.; Hochstrasser, D. F.; Williams, K. L. Biotechnol. Genet. Eng. News 1996, 13, 19-50. (2) Haynes, P. A.; Yates, J. R. Yeast 2000, 17, 81-87. (3) Tyers, M.; Mann, M. Nature 2003, 422, 193-197. (4) Hancock, W. S.; Wu, S. L.; Shieh, P. Proteomics 2002, 2, 352359. (5) Hunt, D. F.; Yates, J. R.; Shabanowitz, J.; Winston, S.; Hauer, C. R. Proc. Natl. Acad. Sci. 1986, 83, 6233-6237. (6) Hunt, D. F.; Bone, W. M.; Shabanowitz, J.; Rhodes, J.; Ballard, J. M. Anal. Chem. 1981, 53, 1706-1708. (7) Lippstreu-Fisher, D. L.; Gross, M. L. Anal. Chem. 1985, 57, 11741180. (8) Hunt, D. F. J. Proteome Res. 2002, 1, 15-19. (9) Griffin, T. J.; Aebersold, R. J. Biol. Chem. 2001, 276, 45, 497-45, 500. (10) Gevaert, K.; Vandekerckhove, J. Electrophoresis 2000, 21, 1145. (11) Aebersold, R.; Mann, M. Nature 2003, 422, 198-207. (12) Lahm, H. W.; Langen, H. Electrophoresis 2000, 21, 2105. (13) Foret, F.; Priesler, J. Proteomics 2002, 2, 360-372. (14) Hamdan, M.; Galvani, M.; Righetti, P. G. Mass Spectrom. Rev. 2001, 20, 121-141. (15) Hille, J. M.; Freed, A. L.; Watzig, H. Electrophoresis 2001, 22, 40354052. (16) Rabilloud, T. Proteomics 2002, 2, 3-10.

research articles

Coupling of PAGE Electrophoresis to MS (17) Courchesne, P. L.; Patterson, S. D. BioTechniques 1997, 22, 244. (18) Williams, K.; LoPresti, M.; Stone, K. Optimization of In-Gel Digest Protocol in Techniques in Protein Chemistry VIII; Marshak, D., Ed.; Academic Press: New York, 1997. (19) Lopez, M. F. Electrophoresis 2000, 21, 1082-1093. (20) Gygi, S. P.; Corthals, G. L.; Zhang, Y.; Rochon, Y.; Aebersold, R. Proc. Natl. Acad. Sci. 2000, 97, 9390-9395. (21) Vestling, M. M.; Fenselau, C. Mass Spectrom. Rev. 1995, 14, 169178. (22) Vestling, M. M.; Fenselau, C. Anal. Chem. 1994, 66, 471-477. (23) Bienvenut, W. V.; Sanchez, J. C.; Karmime, A.; Rouge, V.; Rose, K.; Binz, P. A.; Hochstrasser, D. F. Anal. Chem. 1999, 71, 48004807. (24) Schleuder, D.; Hillenkamp, F.; Strupat, K. Anal. Chem. 1999, 71, 3238-3247. (25) Strahler, J. R.; Smelyanskiy, Y.; Lavine, G.; Allison, J. Int. J. Mass Spectrom. Ion Process. 1997, 169/170, 111-126. (26) Muller, M.; Gras, R.; Appel, R. D.; Bienvenut, W. V.; Hochstrasser, D. F. J. Am. Soc. Mass Spectrom. 2002, 13, 221-231. (27) Binz, P. A.; Muller, M.; Walther, D.; Bienvenut, W. V.; Gras, R.; Hoogland, C.; Bouchet, G.; Gasteiger, E.; Fabbretti, R.; Gay, S.; Palagi, P.; Wilkins, M. R.; Rouge, V.; Tonella, L.; Paesano, S.; Rossellat, G.; Karmime, A.; Bairoch, A.; Sanchez, J. C.; Appel, R. D.; Hochstrasser, D. F. Anal. Chem. 1999, 71, 4981-4988. (28) Schreiner, M.; Strupat, K.; Lottspeich, F.; Eckerskorn, C. Electrophoresis 1996, 17, 954-961. (29) Ogorzalek Loo, R. R.; Mitchell, C.; Stevenson, T. I.; Loo, J. A.; Andrews, P. C. Int. J. Mass Spectrom. Ion Processes 1997, 169/ 170, 273-290. (30) Zaluzec, E. J.; Gage, D. A.; Allison, J.; Watson, J. T. J. Am. Soc. Mass Spectrom. 1994, 5, 230-237. (31) Strupat, K.; Karas, M.; Hillenkamp, F.; Eckerskorn, C.; Lottspeich, F. Anal. Chem. 1994, 66, 464-470. (32) Eckerskorn, C.; Strupat, K.; Schleuder, D.; Hochstrasser, D.; Sanchez, J. C.; Lottspeich, F.; Hillenkamp, F. Anal. Chem. 1997, 69, 2888-2892. (33) Ogorzalek Loo, R. R.; Stevenson, T. I.; Mitchell, C.; Loo, J. A.; Andrews, P. C. Anal. Chem. 1996, 68, 1910-1917.

(34) Ogorzalek Loo, R. R.; Mitchell, C.; Stevenson, T. I.; Martin, S. A.; Hines, W.; Juhasz, P.; Patterson, D.; Peltier, J.; Loo, J. A.; Andrews, P. C. Electrophoresis 1997, 18, 382-390. (35) Ogorzalek Loo, R. R.; Cavalcoli, J. D.; VanBogelen, R. A.; Mitchell, C.; Loo, J. A.; Moldover, B.; Andrews, P. C. Anal. Chem. 2001, 73, 4063-4070. (36) Loo, J. A.; Brown, J.; Critchley, G.; Mitchell, C.; Andrews, P. C.; Ogorzalek Loo, R. R. Electrophoresis 1999, 20, 743-748. (37) Baltz-Knorr, M.; Ermer, D. R.; Schriver, K. E.; Haglund, R. F. J. Mass Spectrom. 2002, 37, 254. (38) Coon, J. J.; James, M. O.; Yost, R. A.; Harrison, W. W. 17th Asilomar Conf. Mass Spectrom. Pacific Grove, CA, 2001. (39) Coon, J. J.; Steele, H. A.; Laipis, P. J.; Harrison, W. W. J. Mass Spectrom. 2002, 37, 1163-1167. (40) Coon, J. J.; McHale, K. J.; Harrison, W. W. Rapid Commun. Mass Spectrom. 2002, 16, 681-685. (41) Van Breemen, R. B.; Snow, M.; Cotter, R. J. Int. J. Mass Spectrom. Ion Processes 1983, 49, 35-50. (42) Coon, J. J.; Harrison, W. W. Anal. Chem. 2002, 74, 5600-5605. (43) Kampmeier, J.; Dreisewerd, K.; Schurenberg, M.; Strupat, K. Int. J. Mass Spectrom. Ion Processes 1997, 169-170, 31. (44) Tanhauser, S. M.; Jewell, D. A.; Tu, C. K.; Silverman, D. N.; Laipis, P. J. Gene 1992, 117, 113-117. (45) Studier, F. W.; Rosenberg, A. H.; Dunn, J. J.; and Dubendorf, J. W. Methods Enzymol. 1990, 185, 60-89. (46) Khalifah, R. G.; Strader, D. J.; Bryant, S. H.; Gibson, S. M. Biochemistry 1977, 16, 2241-2247. (47) Engberg, P.; Millqvist, E.; Pohl, G.; Lindskog, S. Arch. Biochem Biophys. 1985, 241, 628-638. (48) Gerstner, A,; Csapo, Z.; Sasvari-Szekely, M.; Guttman, A. Electrophoresis 2000, 21, 834. (49) Guttman, A.; Ronai, Z. Electrophoresis 2000, 21, 3952. (50) Guttman, A.; Csapo, Z.; Robbins, D. Electrophoresis 2002, 2, 469. (51) Brumley, R. L.; Smith, L. M. Nucleic Acids Res. 1991, 19, 4121. (52) Miller, C. A.; Yi, D.; Perkins, P. D. Rapid Commun. Mass Spectrom. 2003, 17, 860-868.

PR034031F

Journal of Proteome Research • Vol. 2, No. 6, 2003 617