450
Anal. Chem. 1991, 63,450-453
Novel Method for Matrix-Assisted Laser Mass Spectrometry of Proteins Shankai Zhao, Kasi V. Somayajula, Andrew G . Sharkey, and David M. Hercules*
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260 Franz Hillenkamp, Michael Karas, and Arndt Ingendoh
Institut fur Medizinische Physik, Universitat Munster, 4400 Munster, FRG
A new method is demonstrated for obtaining laser mass spectra of proteins by using a nltrobenzyl alcohol matrix (NBA), consisting of nitrobenzyl alcohol, methanol, and water in a ratio of 3:2:1, and by using fibrous material as the substrate. Selected proteins were measured up to m a s 97 400. The detection limit when using the NBA matrix Is demonstrated to be 5 fmol for a single-shot spectrum of albumin. Mixtures of up to six proteins were analyzed and their molecular weights estimated by centroid analysis. Supresslon effects were observed when pepsin is present in a mixture of peptides.
INTRODUCTION Mass spectrometry of proteins in the mass range above m / z 20 000 has been practical only for the last few years. By use of a resonantly absorbing matrix (matrix laser desorption ionization (LDI)),it has been possible to obtain mass spectra of proteins as large at 274000 daltons (1-7); spectra of proteins have been obtained by using only femtomoles of sample (8, 9). These results clearly point out the possible development of mass spectrometry as a practical tool for measuring the molecular weights of a variety of proteins and other high molecular weight compounds. Matrix-assisted UV laser mass spectrometry of nonvolatile compounds was introduced by the Hillenkamp group in 1987 (IO), using nicotinic acid and other matrices as the UV absorber. The first use of this matrix for proteins was reported in 1988 (3); a spectrum of albumin (MW = 66000) was obtained by averaging 100 single-shot spectra on the LAMMA1000. T o date, the largest protein measured has been three subunits of Jack Bean Urease (MW = 274000) ( I ) ; nicotinic acid has been the preferred substrate for proteins. Beavis and Chait (11, 12) have continued studies along the line of Hillenkamp's work and investigated more than 20 matrices to define properties necessary for matrix enhancements. We have reported (13) the use of a nitrocellulose membrane to obtain the laser mass spectra of liquids. New matrix compounds now enable matrix LDI a t higher wavelengths, 337 and 355 nm (24). The present paper introduces a new method for obtaining matrix-assisted laser desorption mass spectra of proteins. The liquid matrix used consists of a mixture of nitrobenzyl alcohol (NBA), methanol, and water applied to a fibrous material as substrate. We report here the analysis of proteins up to 97 400 daltons; spectra have been obtained by using femtomole samples of proteins. Additionally, mixtures of proteins have been analyzed successfully although interference by pepsin in a mixture has been observed. EXPERIMENTAL SECTION Laser mass spectra were obtained by using the LAMMA-1000 laser mass spectrometer at the Institute of Medical Physics,
University of Munster, West Germany. The LAMMA-1000 is a reflectron-type time-of-flight microprobe mass spectrometer. Ion formation was induced by a Q-switched, frequency-quadrupled Nd-YAG laser emitting at 265 nm. The laser was focused to spot sizes of 10-50-pm diameter on the sample surface; irradiances were between IO7 and 108 W/cm2. The LAMMA-1000 laser microprobe mass spectrometer has been described elsewhere (15). A postacceleration of 30 kV is used before detection of the ions. In a limited number of cases, peaks with known mass in the same spectrum (usually Na+ = 23, K+ = 39, and a fragment peak at m/z 80) were used for internal recalibration. After calibration, the M+ ion peak centroid was used. Spectra were obtained by using a nitrobenzyl alcohol matrix, consisting of m-nitrobenzyl alcohol, methanol, and water in a ratio of 2 3 1 on fine fibrous paper (Kimwipe)as a substrate. The paper was cleaned by boiling in doubly distilled water three times and drying in an oven. Samples for analysis were prepared by first dripping 0.5 pL of NBA onto a 1-mm2 area on fibrous paper followed by 0.5 pL of the sample solution. The paper was attached to the sample holder with double-sided sticky tape. Samples were M in aqueous solution and prepared in concentrations of stored in glass vials. Following direct preparation of M solutions, the solutions were diluted three times to lo-* M. The following proteins were investigated: insulin (bovine pancreas), cytochrome C (horse heart), lysozyme (chicken egg) 0-lactoglobulin (bovine milk), trypsinogen (bovine pancreas), ovalbumin (chicken egg), albumin (bovine serum), catalase (bovine liver), and phosphorylase b (rabbit muscle). They were obtained from Sigma Chemical Co. (St. Louis, MO) and were used without further purification. SDS-MW-Marker-17and -70 kits were also obtained from Sigma and were used without purification.
RESULTS AND DISCUSSION Laser mass spectra were obtained for seven proteins in the molecular weight range of 5759 to 97 400. Those studied were (1) insulin (MW = 5733), (2) cytochrome C (MW = 12 384), (3) lysozyme (MW 14300), (4) @-lactoglobulin (MW = lS277), ( 5 ) trypsinogen (MW i= 24000), (6) ovalbumin (MW = 45000), (7)albumin (MW = 66000); and (8) phosphorylase (MW i= 97400). Figure 1shows spectra for four of the proteins studied (2, 5 , 6, 7). These are single-shot spectra. Spectra of proteins having molecular weights below 20 OOO always show cluster ions up to 3M+. Proteins of higher molecular weight generally show only M2+and/or 2M+. The multiply charged and cluster ions are aids in molecular weight determination. As reported for matrix laser desorption ionization using solid matrices, it is possible to obtain multiple spectra from a given sample preparation from different locations or multiple shots a t the same spot. This effect is even more pronounced with NBA. This is useful in averaging spectra to reduce the noise. Figure 2 shows the 1st (A) and the 100th (B) spectra of @lactoglobulin taken from the same spot. It is clear that the quality of the spectrum was degraded very little during the data acquisition process both with regard to the signal-to-noise ratio and the intensity of the molecular ion peak relative t o matrix signals in the low mass range.
0003-2700/91/0363-0450$02.50/00 1991 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 63, NO. 5, MARCH 1, 1991
20000
40000
451
1 H,/l
50000
100000
n,/z D
M+
I
I
u L
\ z
2-
M++
4 U
-
1-
U
C
20000
40000
60000
r 100000
zdoooo
H,/Z
Figure 1. Laser desorption mass spectra obtained by using nitrobenryl alcohol on fibrous material substrate. (A) Cytochrome C (horse heart), (B) trypsinogen (bovine pancreas); (C) ovalbumin (chicken egg), (D) albumin (bovine serum). Y I
A
M+
C
.3-
20000
40000
6000080000 W,/Z
20000
40000 60000 80000
$12 Flgure 2. Laser desorption mass spectrum of @-lactoglobulinin NBA on fibrous material substrate. (A) First laser shot of the spot; (B) 100th
laser shot of the spot. The results from Figure 2 clearly indicate that it is possible to keep the sample on the surface in the vacuum for considerable time (>lo0 spectra) by using the NBA matrix and fibrous material (NBA-FM). Another interesting observation is that the matrix signals only slightly exceed or may be of lower intensity than the molecular ion signal, even given the fact that the matrix molecules are in excess by a factor of lo4. Laser mass spectra obtained by using the NBA-FM matrix are similar to those obtained by using the nicotinic acid matrix (10). Molecular ions, cluster ions, and doubly charged ions
are all seen, but not fragment ions. Liquid matrix laser mass spectrometry reported here is clearly a soft ionization method. One can replenish sample molecules on the surface for a considerable time to improve both sensitivity and reproducibility. Detection Limits. UV laser mass spectrometry is capable of detecting femtomole quantities of proteins when using the nicotinic acid matrix (8) and signal averaging 10-30 shots. Thus, it is important to compare the NBA fibrous material matrix with these results. For the present discussion, the detection limit will be d e f i e d as that amount of protein which will reproducibly give a signal in a single-shot spectrum that is 2 times the rms noise. Three proteins were used for detection limit measurements lysozyme, ovalbumin, and albumin. They were diluted to three concentrations from a stock solution: 5 X 5X and 1x M. Direct preparation of M solution was followed by diluting 3 times to make lo* M. Figure 3 shows single-shot spectra obtained from sample quantities of 250,25, and 5 fmol of albumin on the fibrous paper. Only a small fraction (>1%) of the material added is consumed in obtaining a mass spectrum. Similar results were observed for the other two proteins studied. Because at least several hundred single-shot spectra can be obtained from such a preparation, the actual amount consumed to obtain a spectrum is more in the attomole range. The above results clearly indicate that low detection limits for proteins are possible by using the NBA-FM paper matrix. More than an order of magnitude improvement in detection limit should readily be achieved by reducing the sample volume (and thus the sample area) and averaging spectra from several hundred shots. Mixtures. We also have investigated use of the NBA-FM matrix for measuring mixtures of proteins. Two mixtures were used as examples: SDS-MW-Marker-17 and -70 kits, which are used as calibrants for molecular weight determinations. Figure 4 shows the spectrum obtained from Marker-17. The spectrum presented is the average of 10 shots. It included myoglobin fragments I11 (MW = 2510), I1 (MW = 6210), I (MW = 8160), and I + I1 (MW = 14440) and myoglobin
ANALYTICAL CHEMISTRY, VOL. 63,NO. 5, MARCH 1, 1991
452
A
s /I i
24250 46777
I
I
100000
50000
200000
100000 W,/Z
>
:I
:: I
50000
c
w
100000
50000
100000 150000 M,/Z
Figure 5. Laser desorption mass spectrum of MW-Marker-70 kit obtained by using the NBA-FM substrate. See text for details. (A) LD spectrum of the kit including pepsin (average of 30 shots); (B) LD spectrum of the kit without pepsin (average of 10 shots).
I
100000
200000 My/?
Figure 3. Laser desorption mass spectrum of albumin using the NBA-FM substrate. Spectra were obtained using different amounts of protein deposited on the substrate. (A) 250 fmol, (6)25 fmol, (C) 5
fmol.
8138 rl
u
i
I
2Mo
I
‘2 > U
“ :1
U
C d
10000
20000
M,/z Figure 4. Laser desorption mass spectrum of MW-Marker-17 kit obtained by using the NBA-FM substrate.
backbone (MW = 16950) as listed by the supplier. A molecular ion peak from each of these proteins is clearly seen in the figure. The estimate of the molecular weight from centroid measurement is shown at the top of each peak as an example of molecular weight determination; the centroid measurements can be improved by careful internal recalibration. The peak a t m / z 10614 is from a constituent not included in the list of components by the distributor of the kit. The makeup of myoglobin molecular weight markers has been the subject of a recent investigation which found that the bands assigned by the companies are incorrect (16). These investigators also found the peak corresponding to mlz 10614 reported here. The correct assignments proposed are 2560,
6412, 8300, 10 866, 14 710, and 17 280. These are all within 3-4% of the results reported here. Figure 5A shows the spectrum obtained from the Marker-70 kit; it is averaged over 30 shots. The kit includes six proteins: lysozyme (MW 14300), 6-lactoglobulin (MW = l8400), trypsinogen (MW = 24000), pepsin (MW = 34700), ovalbumin (MW = 45000), and albumin (MW = 66000) (all the molecular weights given are those from the supplier). The intensities are rather weak, and molecular ions can be observed only for lysozyme, @-lactoglobulin,trypsinogen, and ovalbumin. Figure 5B shows a spectrum obtained from a mixture of the proteins in Marker-70 without pepsin. The amounts of proteins are identical with those in Marker-70, and the spectrum is an average of 10 shots. Molecular ion peaks are seen for all components; the peak at 33 468 is probably the M2+peak from albumin. The mass corresponding to the centroid of each peak without internal recalibration is marked on the figure. It is clear from the above that the presence of pepsin has a negative effect on the laser mass spectrum of a protein mixture. A spectrum of pepsin alone cannot be obtained by using the NBA-FM matrix, and only low intensity signals are observed with the nicotinic acid matrix. Spengler and Cotter did not observe the suppression with the matrix they used (6). The supressing effect of pepsin needs to be investigated further. A method for matrix-assisted, laser desorption mass spectrometry of high molecular weight proteins has been developed. By use of a liquid matrix on a fibrous substrate with a large surface area, it is possible to generate molecular ions of high molecular weight proteins in a highly reproducible manner.
LITERATURE CITED (1) Hillenkamp, F.; Karas, M. Proceedings of the 37th ASMS Conference on Mass Spectrometry and Allied Topics, Miami Beach, FL, 1989; p 1168. (2) Hillenkamp, F. Proc. of 11th Int. Mass Spectrom., Bordeaux. 1988; Advances in Mass SDec.: Hevden and Sons: Chichester. U.K.. 1989: VOl. 22, p 354. (3) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 6 0 , 2299. (4) Salehpour, M.; Perera, I.; Kjellberg, J.; Hedin. A,; Islamian, M. A,; Hakansson, P.; Sundquist, B. U. R. Rapid Commun. Mass Spectrom. 1989. 3 , 259.
Anal. Chem. 1991, 63,453-456
453
(14) Hillenkamp, F.; Karas, M.; Ingendoh, A,; Stahl, B. Proceedings of the 2nd Int. Symp. on M s s Spectrom. in Health and Life Sciences; Burlingame, A,, McCloskey, J. A,, Eds.; Elsevier: Amsterdam, in press. (15) Heinen, H. J.; Meier, S.;Vogt, H.; Wechsung, R. Int. J . Mass Spectrom. Ion Phys. 1983, 4 7 , 19. (16) Kratzin, H. D.; Wiltfang, J.; Karas, M.; Neuhoff, V.; Hilschmann, N. Anal. Biochsm. 1989, 183, 1.
(5) Beavis, R. C.; Chait, B. T. Rapid Commun. Mass Spectrom. 1989, 3 , 233. (6) Spengler, B.; Cotter, R. J. Anal. Chem. 1990, 62,793. (7) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.;Yoshida, Y.; Yoshida, T. Rapid Commun. Mass Spectrom. 1988, 2 , 151. (8) Karas, M.; Bahr, U.; Ingendoh, A.; Hillenkamp, F. Angew. Chem., Int. Ed. Engl. 1989, 28,760. (9) Karas, M.: Ingendoh, A.; Bahr, U.;Hillenkamp, F. Biomed. Environ. Mass Spectrom. 1989, 18,841. (10) Karas, M.; Bachman, D.; Bahr, U.;Hillenkamp, F. Int. J. Mass Spectrom. Ion Proc. 1987, 78,53. (1 1) Beavis, R. C.; Chait, B. T. Proceedings of the 37th ASMS Conference on Mass Spectrometry and Allied Topics, Miami Beach, FL, 1989; p 1186. (12) Beavis, R. C.;Chait, B. T. R8pid Commun. Mass Spectrom. 1989, 3 , 432. (13) Zhao, S.;Somayajula, K. V.; Sharkey, A. G.; Hercules, D. M. Z . Anal. Chem. 1990, 338. 588.
RECEIVED for review July 16, 1990. Revised manuscript received November 16,1990. Accepted December 7,1990. This work was supported by a grant from the Gustavus and Louise Pfeiffer Research Foundation, the National Science Foundation, the Deutsche Forschungsgemeinschaft, and the Bundesministerium fur Forschung and Technologie.
Determination of Benzo[ a Ipyrene Sulfate Conjugates from Benzo[ a ] pyrene-Treated Cells by Continuous-Flow Fast Atom Bombardment Mass Spectrometry Yohannes Teffera, William M. Baird, and David L. Smith*
Department of Medicinal Chemistry and Pharmacognosy, Purdue University, West Lafuyette, Indiana 47907
The level of certain water-soluble hydrocarbon conjugates, such as benzo[a]pyrene sulfates (BP-SO,), Is a direct measure of carclnogenlc polycyclic aromatic hydrocarbon metabolism and an indication of exposure. A new method, based on continuous-flow high-resolution fast atom bombardment mass spectrometry, has been developed for the analysis of BP-SO, in the medium of cell cultures treated with benro[alpyrene. An organic solvent extract of medium from cultures of the human hepatoma cell line (HepG2) was fractionated by reversed-phase SEP-PAK chromatography and microbore high-performance liquid chromatography (HPLC). The HPLC fraction containing BP-SO, was collected, dried, and injected into a stream of acetonitriie/water/giycerol that was continuously flowing to the tip of the sample probe which was being bombarded continuously by a beam of high-energy xenon atoms. Molecular anions of BP-SO, ( m / z 347) desorbed from the liqoid were analyzed by a high-resolution ( m / A m5000) mass spectrometer and recorded as a function of time. As little as 1.5 pg of BP-SO, could be detected with a S I N ratio of 8. The mass spectrometer response was linear with respect to the quantity of BP-SO, injected over the range from 15 to 625 pg. The results obtained with this method show that the HepG2 cultures metabolized 3% of the benzo[a]pyrene into the BP-SO, conjugate in 24 h. This procedure, which was used to detect and quantify directly BP-SO, in culture medium without the use of a radiolabeledprecursor, should be generally applicable for analyses of sulfated conjugates resulting from the metabolism of different hydrocarbons.
INTRODUCTION The carcinogen benzo[a]pyrene (BP) is metabolized to reactive metabolites that bind to macromolecules. This interaction of reactive metabolites with cellular DNA is believed
to be the initial step in tumor induction ( I , 2). However, a large proportion of BP is metabolized to water-soluble products. These water-soluble conjugates are formed by enzymatic conjugation of phenols with glucuronic acid and sulfate or conjugation of BP epoxides with glutathione (3-8). The predominance of either of the conjugates depends on the type of the cells used. It is reported that the major conjugates found in rodent cells are glucuronides,while in cultured human tissues the predominant products are sulfate esters and glutathione conjugates ($16). In the human hepatoma cell line (HepGB), sulfate conjugates are one of the major water-soluble metabolites (17). Conjugation can be very important in removal of proximate and ultimate carcinogenicBP metabolites. It has been shown that inhibition of conjugative enzymes results in an increase in the formation of mutagenic B P derivatives (18)and an increase in the extent of covalent binding of BP to DNA (11). Determination of conjugates can be used to study the efficiency of the conjugative enzymes in an organism. Furthermore, the detection of conjugates could be used as an indication of exposure of an organism to BP. The metabolism of B P to form water-soluble benzo[a]pyrene sulfates (BP-S04)in cell culture was first reported by Cohen et al., who used a combination of analytical methods, including radiolabeled precursors (3H BP, and 35SNa2S04), synthetic model compounds, and high-performance liquid chromatography (HPLC) to identify the metabolite as BP3-S04 (4). In subsequent investigations of the conversion of BP into water-soluble conjugates in cells, the BP-S04 metabolites have usually been detected only as radiolabeled substances with chromatographic and chromophoric properties similar to those of synthetic standards. Analytical methods that can be used to detect BP-S04with high specificity and sensitivity are needed to advance our understanding of the mechanisms of carcinogenesis of hydrocarbons. Although radioisotopic data demonstrate that some portion of a metabolite is derived from BP, they convey no direct structural information. Alternatives to radiotracer
0003-2700/91/0363-0453$02.50/0 0 1991 American Chemical Society
