Anal. Chern. 1988, 60, 1086-1088
1086
Table I. Results of Control Experiments To Determine Extent of Fractionation of SF, on the Gas Chromatographic Column expt no. 1 2 3 4
sample size, pmol
633S.%O
634S,k
~[34s(0.505) - 33S),%O
6
-0.22
4.41
0.01
8
-0.17
0.04
10 19
-0.16
-0.25 -0.37 -0.26
-0.18
0.02 0.05
Table 11. Isotopic Results of Fluorination of Various Sulfur Species
molecule
sample size, pmol
633S;%o
634S,"%o
633S/634S
10.01 9.82 10.32 10.50
0.514
10
5.15 5.08 5.26 5.39
25 30 25 30
10.30 10.52 8.21 8.47
20.07
20.51
0.513 0.513
16.03 16.48
0.512 0.514
SF,
15
SF4 SF, SF4
20
S2Fl" S2FIO
SFSCI SF6CI a
5
Against standard SFRdefined as 634S= 6%
0.518
0.510 0.513
= 0.00%0).
reproducibility of the mass spectrometer. Fractionated SFs samples were prepared, ranging from +5%0 to +40%0. These fractionated samples then define the isotopic mass fractionation line.
lection processes is negligible, and the column is the major source of sample fractionation. Although the gas chromatographic separations can introduce scatter of up to f0.4k, in 6%, the separation is essential for obtaining pure samples for mass spectrometric analysis. The gas chromatogram is useful in that it provides an indication of other possible sulfur species which may be present due to incompleteness of the fluorination process. Isotopic composition data for the fluorination of SF4,SF5C1, and S2Fl0are shown in Table 11. The sample sizes and 633S/6"4Sratios are included. The 633S/634Sratio for SF4, SF5C1,and S2FIofalls within the range 6% = (0.505 f 0.01)634S as predicted. The average value of the sulfur isotopic com~ 634S= 20.3 f position for SF4 is 634S= 10.2 f 0 . 2 7 ~S2FIo 0.2%, SF5Cl 634S= 16.2 = 0 . 2 % ~The reproducibility of the entire fluorination and separation, therefore, is f0.2%0 for 634S,similar to that obtained by Puchett et al. (6) for meteoritic mineral fluorination and purification. The chromatographic procedure reported in this paper differs from ref 6 in the use of Porpak QS as the column packing material which is required for these particular separations involving different S,-F, molecules from SF6and to permit determination of the fluorination yields. A molecular sieve column would be more suitable for fluorination and quantitative separation of sulfur-bearing minerals. The technique described here is preferred for fluorination of gaseous sulfurfluorine molecules and their subsequent purification and mass spectrometric analysis. Registry No. SF,, 1783-60-0; SF5C1, 13780-57-9; S2Flo, 5114-22-1;BrF5, 7789-30-2;33S,14257-58-0;34S,13965-97-4;32S, 13981-57-2.
-
RESULTS AND DISCUSSION The data for the control gas chromatographic fractionation experiments are given in Table I. There is a slight isotopic fractionation which occurs on the chromatograph and is consistent with the relationship 6?3 = 0.505(6%), as derived for SF, by Hulston and Thode (1). This fractionation ratio depends on the specific molecule and its mass. for 633S/634S The ratio derives from the dependency of the molecular partition function ratios upon the normal vibrational frequencies which are a function (in part) of the molecular mass (10). For both diatomic and polyatomic sulfur molecules, the relationship 633S= (0.505 f 0.01)634S holds, provided the extent of fractionation is not large (1). The magnitude of fractionation, due to the entire transfer and collection process on the fluorination line and gas chromatographic separation, was determined by a control experiment. The control SF6 sample yielded 633S= -0.2017~ and 634S= -0.305%0,which is similar to the fractionation on the column. This implies that the isotopic fractionation due to the transfer and col-
LITERATURE CITED Hulston, J. R.; Thode, H. G. J . Geoohvs. Res. 1965, 70, 3475-3484. Clayton, R. N.; Grossman, L.; Mayeda. T. K. Science (Washington, D . C . ) 1973, 182, 485-468.
Thiemens, M. H.; HeMenreich, J. E. Science 1983, 219, 1073-1075. Heidenreich, J. E.; Thiemens, M. H. J . Chem. Phys. 1985, 8 4 , 2 129-2-136. -~ ~
.
Bains-Sahota, S.K.; Thiemens, M. H. Meteoritics 1987, 22(4), 7. Puchett, H.; Sabels, B. R.; Hoering, T. Geochim. Cosmocbim. Acta 1973. 35. 625-628. Thode, H.'G.;Rees, C. E. Lunar Planet. Sci. 1979, 10, 1629-1636. Janssen, F. J. J. G. K e r n Sci. Tech. Rep. 1984, 2(2), 9-18. Clayton, R. N.; Mayeda, T. K. Geochim. Cosmocbim. Acta 1983, 2 7 , 43-52. Matsuhisa, Y.; Goldsmlth, J. R.; Clayton, R. N. Geochim. Cosmocbim. Acta 1978, 4 2 , 173-162.
RECEIVED for review September 25,1987. Accepted January 14,1988. NASA (Swroop Bains-GSRP#23376) and the donors of the Petroleum Research Fund, administered by the American Chemical Society, (18189-AC2,5) are thanked for financial support.
Plasma Desorption Mass Spectrometry of Peptides Adsorbed on Nitrocellulose from a Glutathione Matrix Ian Jardine,* Gale F. Scanlan, and Anthony Tsarbopoulos Department of Pharmacology, Mayo Clinic, Rochester, Minnesota 55905
Daniel J. Liberato Department of Drug Metabolism, Hoffmann-La Roche, Inc., Nutley, New Jersey 07110 Two significant advances in peptide sample preparation for analysis by plasma desorption time-of-flight mass spectrometry (PDMS) (2-3) were recently reported (4, 5).
In the first advance, nitrocellulose films were used as sample stage backings to adsorb peptides and proteins for subsequent PDMS analysis. This method conveniently allows the ap-
0003-2700/88/0360-1086501.50/00 1988 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 60, NO. 10, MAY 15, 1988
plication of very small peptide sample sizes from aqueous solutions (picomole to nanomole amounts of sample in 1-5 pL); allows for effective elimination of salt contaminants in the biomolecular film by washing, thus improving the molecular ion yields; produces relatively narrow and sharp peaks in the mass spectrum; and, finally, generates useful multiply charged molecular ions. In the second advance, peptides were dissolved and electrosprayed (6) in solutions containing reduced glutathione (l:l, mo1e:mole) onto PDMS sample targets. Relative to electrospraying without glutathione, this procedure produced increased molecular ion signal, reduction of base-line noise, narrowing of peak widths, and an increase in multiply charged ions. It was suggested that these effects may be caused by enhanced peptide folding and aggregation in glutathione solution, as well as by the possibility of lowering the samplesubstrate binding energy during desorption. Attempts to use the more convenient nitrocellulose technique in our laboratory for the PDMS analysis of peptides were often successful; but often failure resulted, particularly for peptides of >5000 molecular weight. We, therefore, decided to try and combine the nitrocellulose and the glutathione approaches described above. We have found that applying peptides in a 1:l solution of reduced glutathione directly to nitrocellulose targets, with subsequent washing of the adsorbed sample with 0.1 % trifluoroacetic acid solution, results in a higher success rate of peptide analysis by PDMS.
EXPERIMENTAL SECTION Instrument. Plasma desorption mass spectra were recorded with a BIO-ION Nordic (Uppsala, Sweden), BIN-1OK californium-252 time-of-flight mass spectrometer using an accelerating voltage of 20 kV and a flight tube length of 15 cm. This instrument has been described in ref 7 . The spectra of peptides of >loo00 molecular weight were accumulated over a 6 9 h period. Spectra for peptides of lower molecular weight were generally acquired over much shorter time periods of less than 1 h. Peptides. Synthetic growth hormone releasing factor and recombinant interleukin-2 from E. coli were obtained from Hoffmann-La Roche, Inc. (Nutley, NJ). Recombinant human C5a complement proteins were obtained from Abbott Laboratories (Abbott Park, IL). Recombinant thioredoxin from E. coli was purchased from Calbiochem (San Diego, CA). Recombinant human growth hormone from E. coli was from Lilly Research Laboratories (Indianapolis,IN). All disulfide-containingpeptides were obtained and used in the oxidized form. Sample Preparation. Aluminum sample foils were coated with a nitrocellulose film by using the electrospray method (6, 8). Peptide samples were prepared as a 1:l (mo1e:mole) aqueous solution in reduced glutathione (5). One to five microliters of solution was applied to the center of the nitrocellulose-coated sample foil and allowed to stand for 5 min to allow peptide adsorption to the nitrocellulose. The solvent was then removed by rapidly spinning the sample foil at approximately 2000 rpm on a bench top laboratory centrifuge fitted with a sample foil holder. The sample on the nitrocellulose was then washed by applying approximately 0.1-1.0 ml of a 0.1% trifluoroacetic acid solution to the surface, followed by removal of the wash solution by centrifugation, and then insertion in the mass spectrometer. RESULTS AND DISCUSSION Among the peptides which could not be successfully analyzed by applying them to nitrocellulose targets without glutathione, but which were subsequently successfully analyzed after adsorption to the nitrocellulose target from a glutathione solution, were growth hormone releasing factor (MW 5041) (9),human C5a complement proteins (MWs approximately 8500) (IO),thioredoxin from Escherichia coli (MW 11673) (11, 12), interleukin-2 (MW 15547) (13), and human growth hormone (MW 22125) (14). The PDMS spectra of growth hormone releasing factor, thioredoxin, and interleukin-2 are shown in the attached figure. All spectra
1087
a Growth hormone releasing factor MW 5039.7
I
5-1
0 10
MfZ Figure 1. Positive ion californium252 plasma desorption timeof-flight mass spectra of 800-pmol samples of growth hormone releasing factor. (a) The sample was prepared in aqueous solution, applied to a nltrocellulosacoated target, washed with 0.1 % trifluoroacetic acid, and dried before PDMS anaysis. (b) The sample was prepared as a 1:l (mo1e:mole)aqueous solution with reduced glutathione, applied to a nitroceliulosacoated target, washed with 0.1 % trlfluoroacetic acid, and dried before PDMS analysis.
3i
Thioredoxin MW 11673.4
MH:'
I MH'
2
Interleukin-2
31
1
MH'
0 4000
10000 MI2
20000
Flgure 2. Positive Ion californium252 plasma desorption time-of-flight mass spectra of thioredoxin and lnterieukin-2. Peptide samples were prepared as 1:1 (mole:mole) aqueous solutions with reduced glutathione, applied to nitrocellulose-coated targets, washed with 0.1 % trifluoroacetic acid, and dried before PDMS analysis.
have been automatically background subtracted by the data system. The PDMS spectra of 800-pmol samples of growth hormone releasing factor obtained by desorption from nitrocellulose both without and with application to the target in a 1:l glutathione solution are presented in parts a and b of Figure 1 for comparison. The enhancement of the PDMS spectrum of this peptide when reduced glutathione is included in the analysis is clear from these spectra. Similar enhancements were oberved for the PDMS analysis of the other peptides reported. That is, esentially no peaks were observed above background before glutathione treatment.
ANALYTICAL CHEMISTRY, VOL. 60, NO. 10, MAY 15, 1988
1088 Table
I
growth
MH+ obsd (calcd)
MH?' obsd (calcd)
5040.9 (5040.7)
2520.8 (2520.9)
MHS3+obsd (calcd) - (1680.9)
hormone
releasing factor thioredoxin 11671.0 (11674.4) 5833.4 (5827.7) 3893.7 (3892.1) interleukin-2 15531.4 (15548.1) 7755.1 (7774.6) 5174.9 (5183.4) The spectrum of thioredoxin shown in Figure 2 was obtained on 100 pmol of peptide. Spectra of interleukin-2 obtained by using 200 pmol, 2 nmol, and 20 nmol were essentially identical with each other, including the absolute as well as relative intensities of the MH+, MH$+, and M H l + ions. The spectrum shown is of the 2-nmol sample which had a slightly more clearly defined MH+ ion than that of the 200-pmol spectrum. The mass measurement accuracy for the observed ions of these peptides as determined by PDMS is shown in Table I and is generally always better than 0.2%. This is an acceptable level of accuracy at these masses for the simple time-of-flight maw spectrometer of short flight path (15 cm) used in these experiments. In conclusion, for PDMS analysis of peptides at the 1-nmol level, we recommend preparing peptide samples in a 1:1 (mo1e:mole) aqueous solution of reduced glutathione, applying this solution to nitrocellulose coated targets, washing the adsorbed peptide with a 0.1 % trifluoroacetic acid solution, and drying, before PDMS analysis.
ACKNOWLEDGMENT We thank George Carter of Abbott Labs for the samples
of recombinant human C5a complement proteins, John Occolowitz of the Lilly Research Laboratories for the sample of recombinant human growth hormone, and Ronald D. Macfarlane of Texas A&M University for the suggestion of spin-drying the PDMS samples. Registry No. Glutathione, 70-18-8;nitrocellulose, 9004-70-0; growth hormone-releasing factor, 9034-39-3.
LITERATURE CITED Torgerson, D. F.; Skowronskl, R. P.; Macfarlane, R. D. Eiochem. Bophys. Res. Commun. 1974, 6 0 , 616. Macfarlane, R. D.; Torgerson, D. F. Scknce (Washington, D . C . ) 1978, 797, 920. Sundqvlst, B. V. R.; Macfarlane, R. D. Mess Spectrom. Rev. 1985, 4 , 421. Jonsson, G. P.; Hedin, A. B.; Wkansson, P. L.; Sundqvist, 8. V. R.; Sdve, B. G. S.; Nielson, P. F.; Roepstorff. P.; Johansson, K.-E.; Kamenbky. I.; Lindberg, M. S. L. Anal. Chem. 1988, 58, 1064. Alai, M.; Demlrev, P.; Fenselau, C.; Cotter, R. J. Anal. Chem. 1986, 58, 1303. McNeal. C. J.; Macfarlane, R. F.; Thurston, E. L. Anal. Chem. 1979, 57, 2036. Sundqvist, B.; Kamensky, I.; HAkansson, J.; Kjeliberg. J.; Salehpour, M.; Wkkllyasekera, S.; Fohlman, J.; Peterson, P. A,; Roepstorff, P. Anal. Chem. 1984, 7 7 , 242. Chait, B. T.; Fleld, F. H. Ebchem. Eiophys. Res. Commun. 1988, 734,420. Gulllemln, R.; Brazeau, P.; Bohlen, P.; Esch, F.; Ling, F.; Wehrenberg, W. B. Science(Washlngton, D.C.) 1982, 278,585. Fernandez, H. N.; Hugli, T. E. J . Bioi. Chem. 1978, 253,6955. Holmgren, A. Eur. J . Eiochem. 1968, 6 , 475. H%g, J.-0.; Von Bahr-Llndstrom, H.; Josephson, S.; Wallace, B. J.; Kushner, S. R.; Jtknvall, H.; Holmgren, A. Eiosci. R e p . 1984, 4 , 917. Robb, R. J. Methods Enzymol. 1985, 776,493. Niall, H. D. Nature (London), New Eiol. 1971, 230,90.
RECEIVED for review May 11,1987. Accepted February 1,1988. This work was supported by NIH Grants GM 32928 and RR 02682.
