Polyacrylamide Gel Electrophoresis Coupled with ... - ACS Publications

Dec 1, 1997 - In analogy to two-dimensional analysis, the mobility shift in native polyacrylamide gel electrophoresis (PAGE) due to a nucleotide subst...
7 downloads 0 Views 174KB Size
Anal. Chem. 1997, 69, 4899-4904

Polyacrylamide Gel Electrophoresis Coupled with Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry for tRNA Mutant Analysis Jing Wei and Cheng S. Lee*

Department of Chemistry and Ames Laboratory, USDOE, Iowa State University, Ames, Iowa 50011

In analogy to two-dimensional analysis, the mobility shift in native polyacrylamide gel electrophoresis (PAGE) due to a nucleotide substitution of a single-stranded transfer ribonucleic acid (tRNA) fragment serves as the first dimension for tRNA mutation analysis. Matrix-assisted laser desorption/ionization mass spectrometry (MALDIMS), as the second dimension, allows precise determination of the mass of the tRNA fragments resolved by native PAGE. Off-line combination of native PAGE with MALDI-MS is demonstrated for high-resolution analysis of tRNAval and its mutants, including a three-nucleotide deletion and 12 single-base substitutions. Three approaches, including direct extraction of tRNAs from gel into buffer solution, dissolution of membrane in the matrix solution, and direct desorption of tRNAs from the membrane, are studied for coupling native PAGE with MALDIMS. The membrane dissolution method is simple, and the resulting mixture is amenable to MALDI-MS analysis. In the membrane dissolution method, as little as 1 µg or 40 pmol of tRNA sample is loaded on a native gel, separated, capillary eluted onto a nitrocellulose membrane, and recovered using the matrix solution of 2,4,6trihydroxyacetophenone in acetone. Point mutations in deoxyribonucleic acid (DNA) are studied in the fields of mutagenesis, human genetics, and cancer genetics. Modern molecular methods for mutational analysis allow the detection of sequence alterations in genomic DNA without the need of sequencing the whole region of interest.1 These methods include single-strand conformation polymorphism analysis (SSCP),2,3 denaturing gradient gel electrophoresis (DGGE),4 chemical mismatch cleavage,5 ribonuclease cleavage,6 and heteroduplex analysis.7 The sequence dependence of DGGE results from secondary structures generated by chemically induced (partial) melting of double-stranded DNA. The combination of DGGE with (1) Gelfi, C.; Righetti, P. G.; Cremonesi, L.; Ferrari, M. Electrophoresis 1994, 15, 1506-1511. (2) Orita, M.; Iwahana, H.; Kanazawa, H.; Hayashi, K.; Sekiya, T. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 2766-2770. (3) Cotton, R. G. H. Mutat. Res. 1993, 285, 125-144. (4) Fisher, S. G.; Lerman, L. S. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 15791583. (5) Cotton, R. G.; Rodrigues, N. R.; Campbell, R. D. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 4397-4401. (6) Myers, R. M.; Larin, Z.; Maniatis, T. Science 1985, 230, 1242-1246. (7) Nagamine, C. M.; Chan, K.; Lau, Y. F. C. Am. J. Hum. Genet. 1989, 45, 337-339. (8) Fisher, S. G.; Lerman, L. S. Cell 1979, 16, 191-200. (9) Vijg, J. Bio/Technology 1995, 13, 137-139. S0003-2700(97)00725-7 CCC: $14.00

© 1997 American Chemical Society

size separation in a two-dimensional format enables large numbers of DNA fragments to be scanned for mutations.8,9 Bioanalytical methodologies for mutant detection of ribonucleic acid (RNA) appear to lag behind similar capacities that are being developed for DNA. RNA is a single-stranded polynucleotide with defined secondary and tertiary structures.10,11 In analogy to SSCP for DNA mutation analysis,2,3 single-base mutations in RNA should alter its secondary structure and electrophoretic mobility in nondenaturing gel. The use of native polyacrylamide gel electrophoresis (PAGE) is thus investigated in this study for highresolution analysis of single-base substitutions in transfer RNA for valine (tRNAVal ). In analogy to two-dimensional DNA analysis, the first dimension for tRNA mutation analysis is based on the mobility shift in native PAGE due to a nucleotide substitution of a single-stranded tRNA fragment. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), as the second dimension, allows precise determination of the mass of the tRNA fragments separated by native PAGE. The combination of native PAGE with size separation of MALDI-MS in a two-dimensional format provides a high-resolution approach for tRNA mutational analysis. Since the introduction of MALDI-MS in 1988,12 a number of research groups have reported the utility of MALDI-MS for analyzing various DNA and RNA samples.13-23 Various matrix compounds, including 3-hydroxypicolinic acid (3-HPA),14 picolinic acid (PA),16 2,4,6-trihydroxyacetophenone (2,4,6-THAP),18 and 2,3,4-trihydroxyacetophenone (2,3,4-THAP),21 were used for ef(10) Yue, D.; Kintanar, A.; Horowitz, J. Biochemistry 1994, 33, 8905-8911. (11) Liu, H.; Musier-Forsyth, K. Biochemistry 1994, 33, 12708-12714. (12) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. (13) Nordhoff, E.; Carmer, R.; Karas, M.; Hillenkamp, F.; Kirpekar, F.; Kristiansen, K.; Roepstorff, P. Nucleic Acids Res. 1993, 21, 3347-3357. (14) Wu, K. J.; Steding, A.; Becker, C. H. Rapid Commun. Mass Spectrom. 1993, 7, 142-146. (15) Kirpekar, F.; Nordhoff, E.; Kristiansen, K.; Roepstorff, P.; Lezius, A.; Hahner, S.; Karas, M.; Hillenkamp, F. Nucleic Acids Res. 1994, 22, 3866-3870. (16) Tang, K.; Taranenko, N. I.; Allman, S. L.; Chang, L. Y.; Chen, C. H. Rapid Commun. Mass Spectrom. 1994, 8, 727-730. (17) Liu, Y.-H.; Bai, J.; Liang, X.; Lubman, D. M. Anal. Chem. 1995, 67, 34823490. (18) Christian, N. P.; Colby, S. M.; Giver, L.; Houston, C. T.; Arnold, R. J.; Ellington, A. D.; Reilly, J. P. Rapid Commun. Mass Spectrom. 1995, 9, 10611066. (19) Shaler, T. A.; Wickham, J. N.; Sanners, K. A.; Wu, K. J.; Becker, C. H. Anal. Chem. 1996, 68, 576-579. (20) Dai, Y.; Whittal, R. M.; Li, L.; Weinberger, S. R. Rapid Commun. Mass Spectrom. 1996, 10, 1792-1796. (21) Zhu, Y. F.; Chung, C. N.; Taranenko, N. I.; Allman, S. L.; Martin, S. A.; Haff, L.; Chen, C. H. Rapid Commun. Mass Spectrom. 1996, 10, 383-388. (22) Koster, H.; Tang, K.; Fu, D.-J.; Braun, A.; van den Boom, D.; Smith, C. L.; Cotter, R. J.; Cantor, C. R. Nature Biotechnol. 1996, 14, 1123-1128. (23) Roskey, M. T.; Juhasz, P.; Smirnov, I. P.; Takach, E. J.; Martin, S. A.; Haff, L. A. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 4724-4729.

Analytical Chemistry, Vol. 69, No. 23, December 1, 1997 4899

Table 1. Wild-Type and Mutant tRNAVals tRNAVal and its mutantsa

expected MW

MW difference from wild type

wild type A6G A26G A41G A35C G45A U59G U64C A66G G45U U33C U34C C73G -CCA

24 547 24 567 24 567 24 567 24 523 24 527 24 586 24 546 24 567 24 508 24 546 24 546 24 587 23 553

+20 +20 +20 -24 -20 +39 -1 +20 -39 -1 -1 +40 -994

a Mutant tRNAs are designated by the letter and position number of the base in the wild-type sequence that has been altered, followed by the base that has been introduced, e.g., A6G has the adenine at position 6 replaced by guanine.

in acetone. As demonstrated by Liang et al.,24 the resulting mixtures containing the proteins electroblotted onto nitrocellulose were used directly as the interface between sodium dodecyl sulfate (SDS)-PAGE and MALDI-MS for protein characterization. In this study, three different approaches, including extraction of tRNAs from slab gel into buffer solution, dissolution of membrane (containing eluted tRNAs from gel) in the matrix solution, and direct application of the matrix solution onto the membrane, are employed and compared for coupling native PAGE with MALDIMS in two-dimensional analysis of tRNA mutants.

Figure 1. (A) Cloverleaf and (B) tertiary structures of Escherichia coli tRNAVal.

ficient desorption and ionization of oligonucleotides. Tang et al.16 succeeded in the detection of DNA fragments as large as 500 nucleotides using the matrix mixture of 3-HPA and PA. MALDIMS spectra of RNA up to 461 nucleotides were shown by Kirpekar and co-workers.15 A nitrocellulose film was used as a substrate for on-probe purification of DNA samples and for routine detection of double-stranded DNAs in an enzymatic digestion mixture ranging from 9 to 622 base pairs.17 The power of MALDI-MS for DNA sequencing was demonstrated by its high speed and its ability to identify sequences that were not readable in PAGE.22,23 Gel-separated tRNAs are not amenable to direct MALDI-MS characterization. Both nylon and nitrocellulose membranes are selected and evaluated as the substrates for capillary elution of tRNAs. The eluted tRNAs, free of buffers, salts, and other contaminants, can then be analyzed with MALDI-MS through membrane dissolution and/or direct laser desorption from a membrane support. Nylon membranes are widely used for efficient transfer and strong binding with oligonucleotides in Northern/Southern blotting. Additionally, nitrocellulose membranes are readily dissolvable using the matrix solution prepared (24) Liang, X.; Bai, J.; Liu, Y.-H.; Lubman, D. M. Anal. Chem. 1996, 68, 10121018.

4900 Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

EXPERIMENTAL SECTION Materials. Wild-type tRNAVal and its mutants, including a three-nucleotide deletion and 12 single-base substitutions (see Figure 1 and Table 1), were kindly provided by Dr. Jack Horowitz at Iowa State University. In vitro transcription and synthesis of wild-type tRNAVal and its mutants were described in detail elsewhere.25,26 A 40% acrylamide stock solution was obtained from Fisher (Fair Lawn, NJ) for the preparation of slab gels. All chemicals, including ethidium bromide, EDTA, sodium chloride, and sodium citrate were purchased from Sigma (St. Louis, MO). Acetonitrile, ammonium acetate, ammonium citrate, HPLC grade acetone, ethanol, 3-HPA, methanol, and 2,4,6-THAP were obtained from Aldrich (Milwaukee, WI) without further purification. Nitrocellulose (pore size of 0.45 µm) and nylon membranes (pore size of 0.22 µm) were purchased from Schleicher & Schuell (Keene, NH) and Amersham Life Science (Arlington Height, IL), respectively. Native PAGE, tRNA Extraction, and Capillary Elution. tRNA samples were applied and separated on 18% polyacrylamide gels as described elsewhere.27,28 For the extraction of tRNAs from slab gels,27 tRNA bands were excised with a sterile razor blade. Each gel slice was crushed in a sterile tube containing a solution of 0.5 M ammonium acetate and 1 mM EDTA at pH 8. The tubes were sealed and shaken vigorously overnight at 37 °C. The tRNAs (25) Sampson, J.; Uhlenbeck, O. C. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 10331037. (26) Chu, W.-C.; Horowitz, J. Nucleic Acids Res. 1989, 17, 7241-7262. (27) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning: A Laboratory Manual; Cold Spring Harbor Laboratory Press: New York, 1989. (28) Laemmli, U. K. Nature 1970, 227, 680-685.

Figure 2. (A) Resolution of wide-type (wt) tRNAVal from its mutants using native PAGE. (B) Capillary elution of tRNAs from slab gels onto the nitrocellulose membranes.

were collected in the supernatant by centrifugation and precipitated by the addition of ethanol. The precipitates were washed with a 70% ethanol solution, dried, and reconstituted in deionized water. The tRNAs were eluted from slab gels to the membranes in a transfer buffer of 1 M sodium chloride and 0.1 M sodium citrate at pH 8. The procedures were the same as those employed in Southern blotting and were described in detail by Sambrook et al.27 The gel-resolved tRNAs were stained with ethidium bromide. The membranes were rinsed with deionized water for removing high ionic strength buffer used during capillary elution. The membrane areas containing tRNAs eluted from slab gels were excised with a sterile razor blade for sample preparation in MALDI-MS. Matrix and Sample Preparation for MALDI-MS. The matrix of 2,4,6-THAP was dissolved with a concentration of 80 mg/mL in methanol for extracted tRNAs and direct membrane desorption experiments or in acetone for membrane dissolution measurements. Additionally, the matrix of 3-HPA was prepared with a concentration of 80 mg/mL in 30% acetonitrile and 70% distilled water for extracted tRNAs experiments. A standard solution of ammonium citrate (10 mg/mL) was prepared in 50% acetonitrile and 50% distilled water. One microliter of the solution mixture, containing matrix, ammonium citrate, and extracted tRNAs in a volume ratio of 1:1:1, was applied onto the probe tip and allowed to dry before the MALDI-MS analysis. Ammonium citrate suppressed the formation of alkaline metal adducts for improved mass resolution. For membrane dissolution measurements, the excised membranes containing tRNAs (∼1.5 × 3 mm2, corresponding to half the band on gel) were dissolved by the addition of 5 µL of matrix solution. As a result, a membrane concentration of 0.9 mm2/µL (1.5 × 3 mm2/5 µL) was obtained. Immediately after the addition of 2 µL of ammonium citrate, 1 µL of the mixture was applied twice onto the probe tip. For the direct desorption of tRNAs from the membranes, the excised membranes were mounted onto the probe tip by a double-sided adhesive tape. One microliter of solution mixture containing matrix and ammonium citrate in a volume ratio of 2:1 was applied twice onto the membranes. MALDI-MS. Experiments were performed using a ProFLEX linear time-of-flight instrument equipped with gridless delayed ion extraction and a flight tube of 1.3 m (Bruker Analytical Systems, Billerica, MA). Desorption was accomplished using a nitrogen laser at 337 nm. Ions were accelerated through an electric potential of 20 kV, with ion extraction occurring 200 ns (17.5 kV extraction voltage) after the laser pulse. Ions were analyzed by

a microchannel plate detector, and data were acquired at a digitization rate of 500 MHz. All of the spectra were obtained as 60 shot averages. All spectra were calibrated externally using proteins (equine cardiac cytochrome c, equine cardiac myoglobin, and bovine serum albumin) from Hewlett-Packard (Fullerton, CA) as standards. RESULTS AND DISCUSSION Resolution of tRNA Mutants Using Native PAGE. The tertiary structure of tRNAVal and the list of its mutants, including a three-nucleotide deletion and 12 single-base substitutions, are summarized in Figure 1 and Table 1, respectively. The wild type was successfully separated from each mutant tRNAVal using native PAGE (see Figure 2A). Single-base substitutions in tRNAVal resulted in changes in the secondary and tertiary structures of tRNA and contributed to their differences in electrophoretic mobility. In comparison with wild-type tRNAVal, all single-base substitution mutants, regardless of their increases or decreases in molecular weight (see Table 1), exhibited more compact structures and greater electrophoretic mobilities. There was no separation between wild type and singlenucleotide deletion mutant using native PAGE (data not shown). Resolution was obtained, however, between wild type and a threenucleotide deletion mutant from the 3′ end of tRNAVal. In comparison with singe-base substitution mutants, nucleotide deletion from the 3′ end of tRNA was not involved in base pairing (see Figure 1) and accounted for changes in molecular size more than in the secondary and tertiary structures of tRNA. Capillary Elution of Resolved tRNAs from Gels onto Membranes. Capillary elution of tRNAVal bands from slab gels onto the nitrocellulose and nylon membranes was investigated, following the procedures typically employed in Southern blotting.27 Nylon membranes have long been used as a blotting matrix for oligonucleotide immobilization in Northern and Southern blotting. Nitrocellulose membranes, on the other hand, have been widely used in protein blotting and as the interface between SDS-PAGE and MALDI-MS for the characterization of intact proteins.24,29 Furthermore, nitrocellulose membranes as a substrate for sample preparation allowed effective removal of salt contaminants by washing with deionized water or diluted acid for plasma desorption mass spectrometry of peptides and proteins30 and for MALDIMS detection of large DNA molecules.17 (29) Klarskov, K.; Roepstorff, P. Biol. Mass Spectrom. 1993, 22, 433-440. (30) Jonsson, G. P.; Hedin, A. B.; Hakansson, P. L.; Sundqvist, B. U. R.; Save, B. G. S.; Nielson, P. F.; Roepstorff, P.; Johansson, K.-E.; Kamensky, I.; Lindberg, M. S. L. Anal. Chem. 1986, 58, 1084-1087.

Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

4901

Figure 3. MALDI-MS analysis of wild-type tRNAVal extracted from native PAGE using 2,4,6-THAP as the matrix.

As shown in Figure 2B, the tRNAs were successfully transferred onto the nitrocellulose membranes without significant deterioration in resolution. Based on the intensities of tRNAs stained with ethidium bromide, high transfer efficiency of tRNAs was obtained when using the nitrocellulose membranes. Additionally, no significant difference in transfer efficiency was observed between the use of the nitrocellulose and nylon membranes. The membranes were then rinsed by deionized water for removing high ionic strength buffer used during capillary elution. MALDI-MS Analysis of Extracted tRNAs from Native PAGE. Three methodologies, including direct extraction of tRNAs from slab gel into buffer solution, dissolution of membrane in the matrix solution, and direct desorption of tRNAs from the membrane, were studied for coupling native PAGE with MALDIMS in two-dimensional analysis of tRNA mutants. Due to low extraction efficiency (less than 20%) of tRNAs from slab gel into buffer solution, 40 µg or 1.6 nmol of each wild-type and mutant tRNAVal was loaded onto the gel. The MALDI-MS spectrum of extracted wild-type tRNA (0.25 µg or 10 pmol onto the probe tip) with a mass resolution of 290 (full width at half-maximum, fwhm) was obtained using 2,4,6-THAP as the matrix (see Figure 3). The mass accuracy between the expected (see Table 1) and the measured molecular weights of wild-type tRNA was around 0.04%. The mass resolution and detection sensitivity of tRNAVal were slightly better than the results of tRNATyr and tRNASer analyzed by Gruic-Sovulj et al.31 The shoulder that appeared at m/z ) 24 843 (see Figure 3) was attributed to nonspecific elongation of an additional cytosine at the 3′ end during in vitro transcription using T7 polymerase.25,26,32,33 Wild-type tRNAVal (76 mer) sample was contami(31) Gruic-Sovulj, I.; Ludemann, H.-C.; Hillenkamp, F.; Weygand-Durasevic, I.; Kucan, Z.; Peter-Katalinic, J. Nucleic Acids Res. 1997, 25, 1859-1861. (32) Reyes, V. M.; Anbelson, J. Anal. Biochem. 1987, 166, 90-106. (33) Samuelsson, T.; Boren, T.; Johansen, T.-I.; Lusting, L. J. Biol. Chem. 1988, 263, 13692-13699.

4902 Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

nated with approximately 30% of 77 mer due to cytosine elongation. The amount of impurity was determined by total RNase T2 digestion and separation of digested products using twodimensional thin-layer liquid chromatography.25 Additionally, a three-nucleotide deletion mutant extracted from native PAGE was easily identified by MALDI-MS (see Figure 4). The heterogeneity in the 5′ end of this three-nucleotide deletion mutant caused by the variation in the number of phosphate groups25,26 together with 3′ end heterogeneity severely decreased the mass resolution. The 3′ end heterogeneity was due to both premature termination and unspecific nucleotide addition. The results shown in Figures 2-4 clearly illustrated the two-dimensional resolving power of native PAGE in combination with MALDI-MS for tRNA mutation analysis. 3-HPA was also employed as the matrix for the analysis of extracted wild-type tRNA in MALDI-MS (see Figure 5). A comparison of the results shown in Figures 3 and 5 reveals that the use of 2,4,6-THAP yielded better ionization intensity, mass resolution, and reproducibility than those of 3-HPA. The experimental results presented by Zhu et al.21 indicated that a matrix solution containing 2,4,6-THAP, 2,3,4-THAP, and ammonium citrate in a molar ratio of 2:1:1 gave the best detection of large DNA fragments in the mass range of interest. The resolution and shot-to-shot reproducibility of the THAP matrix were better than those of 3-HPA and PA matrices. Additionally, a substantial improvement in mass resolution of tRNATyr and tRNASer was achieved upon replacing 3-HPA with THAP as the matrix in MALDI-MS.31 As a result, the matrix of 2,4,6-THAP was selected and employed in further studies. Dissolution of Membranes Containing Gel-Separated tRNAs for MALDI-MS Analysis. Both the nylon and nitrocellulose membranes exhibited high transfer efficiency for capillary elution of tRNAs from gels onto membranes. Initial studies also indicated that both membranes were relatively clean and did not produce any significant background ions in the mass region of interested. Nitrocellulose membranes were readily dissolved

Figure 4. MALDI-MS analysis of tRNA-CCAVal extracted from native PAGE using 2,4,6-THAP as the matrix.

Figure 5. MALDI-MS analysis of wild type tRNAVal extracted from native PAGE using 3-HPA as the matrix.

using the matrix solution prepared in acetone. Due to the difficulty in the selection of suitable organic solvent for the dissolution of nylon membrane, only nitrocellulose membranes were employed for membrane dissolution studies. As reported by Liang et al.,24 the final membrane concentration in the matrix solution was a crucial factor in the analysis of SDSPAGE-separated proteins using MALDI-MS. A high nitrocellulose concentration resulted in a viscous solution and affected the incorporation of analyte molecules into the crystalline matrix. A low nitrocellulose concentration improved the crystallization of the matrix/analyte solution, but at the expense of detection sensitivity due to dilution. Based on these considerations and our experimental results, a membrane concentration of 0.9 mm2/µL

was selected for the analysis of native PAGE-separated tRNAs using MALDI-MS. Due to high transfer efficiency of tRNAs from gels onto membranes, 1 µg or 40 pmol of each wild-type and mutant tRNAVal was loaded onto the gel. The MALDI-MS spectrum of wild-type tRNA (0.29 µg or 11 pmol onto the probe tip) with a mass resolution of 100 was obtained using 2,4,6-THAP as the matrix (see Figure 6). The use of buffer extraction (see Figure 3) as the interface between native PAGE and MALDI-MS provided better ion intensity and mass resolution than those of the membrane dissolution method (see Figure 6). The dissolved nitrocellulose and ethidium bromide used for tRNA staining did not adversely affect the performance of MALDI-MS. As a result Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

4903

during the MALDI process was limited. As reported by Liang et al.,24 both the dissolution and the extraction methods were studied for the recovery of electroblotted proteins onto the nitrocellulose membranes. Again, the membrane dissolution method provided more efficient detection of peptides and proteins in MALDI-MS, especially the high molecular weight proteins. The adsorption of biomolecules onto the nitrocellulose membrane tends to be quite strong, making the desorption or extraction method less efficient. Direct MALDI-MS analysis of the membranes containing gel-separated tRNAs may be improved with the use of ultrathin membrane, similar to the preparation and the use of a nitrocellulose film substrate in MALDI-MS studies of DNA fragments.17

Figure 6. (A) MALDI-MS analysis of wild type tRNAVal loaded onto a native gel, separated, capillary eluted onto a nitrocellulose membrane, and recovered using the matrix solution of 2,4,6-THAP in acetone. (B) MALDI-MS analysis of wild-type tRNAVal loaded onto a native gel, separated, capillary eluted onto a nitrocellulose membrane, and desorbed from the membrane using 2,4,6-THAP as the matrix. (C) all conditions were the same as in (B) except for the use of a nylon membrane.

of low extraction efficiency, the required gel loading in buffer extraction, however, was 40 times that in the membrane dissolution method. Direct MALDI-MS Analysis of Membranes Containing GelSeparated tRNAs. At least 2 µg or 80 pmol of each wild-type and mutant tRNA was required for gel separation, capillary elution onto the membranes, and direct membrane desorption in MALDIMS. The mass spectra of wild-type tRNAVal (see Figures 7 and 8) were obtained by direct application of the matrix solution onto the nitrocellulose and nylon membranes. In comparison with the membrane dissolution method (see Figure 6), the MALDI process performed directly from the membranes apparently resulted in lower ion intensity due to incomplete desorption. The tRNAs were often adsorbed or embedded in the membranes, and the number of tRNAs accessible to the laser beam

4904 Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

CONCLUSION Off-line combination of native PAGE with MALDI-MS is demonstrated for two-dimensional analysis of tRNAVal and its mutants, including a three-nucleotide deletion and 12 single-base substitutions. Native PAGE resolves wide type from single-base substitution mutants on the basis of their differences in the secondary and tertiary structures. MALDI-MS, as the second dimension, clearly identify wild type against nonspecific nucleotide addition and nucleotide deletion mutants. Three approaches, including direct extraction of tRNAs from gel into buffer solution, dissolution of membrane in the matrix solution, and direct desorption of tRNAs from the membrane, are studied for coupling native PAGE with MALDI-MS. Buffer extraction provides the best ion intensity and mass resolution of tRNAs in MALDI-MS. However, the required gel loading for buffer extraction is significantly greater than those of membrane dissolution and membrane desorption. The membrane dissolution method is simple, and the resulting mixture of matrix, nitrocellulose, and tRNAs is amenable to MALDI-MS analysis. In the membrane dissolution method, as little as 1 µg or 40 pmol of tRNA sample is loaded onto a native gel, separated, capillary eluted onto a nitrocellulose membrane, and recovered using the matrix solution of 2,4,6-THAP in acetone. Direct MALDI-MS analysis of the membranes containing gel-separated tRNAs may be enhanced with the use of ultrathin membrane and IR laser due to the much larger ablation depth. ACKNOWLEDGMENT The authors would like to thank Dr. Jack Horowitz and Jack Liu of Iowa State University for their supply of tRNAVal and its mutants. Support for this work by the Microanalytical Instrumentation Center of the Institute for Physical Research and Technology at Iowa State University is gratefully acknowledged. C.S.L. is a National Science Foundation Young Investigator (BCS-9258652). Received for review July 7, 1997. Accepted September 22, 1997.X AC970725U X

Abstract published in Advance ACS Abstracts, November 1, 1997.