Tandem Mass Spectrometry of Peptides with N-Terminal Glutamine Studies on a Prion Protein Peptide

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

M. A. Baldwin,* A. M. Falick, B. W. Gibson, S. B. Prusiner, N. Stahl, and A. L. Burlingame Departments of Pharmaceutical Chemistry,Neurology, Biochemistry and Biophysics, and the Liver Center, University of California, San Francisco, California, USA

This paper reports the unusual collision-induced fragmentation of peptides having N- terminal glutamine. One of these glutamine-containing peptides was isolated from an endoproteinase Lys-C digest of the scrapie prion protein (PrP 27-30) with the predicted sequence QHTVTITTK. Daughter ion mass spectra show predominant N-terminal a,, b,, and, to a lesser extent, c,, sequence ion series at 17 u less than the predicted masses. This is interpreted as arising from an ionic fragmentation that accompanies the peptide-backbone cleavage, glutamine losing ammonia to give pyroglutamic acid, a reaction that parallels the commonly observed solution-phase process. This behavior is less evident when strongly basic residues near the C-terminus cause the C-terminal fragments (G, y., and z,,) to predominate. (J Am Sot Mass Spectrom 1990, 1, 258-264)

e development of rapid and sensitive soft ion- ing the collision-induced dissociation (CID) technique ization methods such as fast atom bombardment to peptides is the considerable knowledge base con- T (FAB) [l] and liquid secondary ion mass spec- cerning their fragmentation and the consequent high trometry (LSIMS) [2, 31, based on kiloelectronvolt par- success rate for interpretation of the data. The major- ticle bombardment of nonvolatile samples dissolved in ity of ions arise from (1) characteristic side-chain losses a liquid matrix, has led to the widespread use of mass from amino acids; (2) the formation of immonium ions spectrometry in peptide and protein analysis [4-61. indicative of individual amino acids; (3) cleavages of Protonated molecular ions are readily sputtered from the peptide backbone, giving N-terminal ions (a,, b,, solution for the majority of peptides in the mass range and c,) and C-terminal ions (xn, y”, and z,) indicative 500-4000 u, allowing molecular masses to be deter- of the sequence; and (4) further breakdown of the se- mined for peptides produced by chemical or enzymatic quence ions by side-chain losses (d,, vn, and w,). The digestion of proteins. However, sputtered ions pro- latter ions reinforce the subsequent data and can fur- duced by such “soft ionization” techniques normally ther identify amino acids, in particular leucine versus have relatively little internal energy, resulting in a lim- isoleucine [lo, 111. ited degree of fragmentation. Although the fragmen- The interpretation of peptide tandem mass spectra tation processes of peptides ionized by FAB or LSIMS (MS/MS) is sufficiently straightforward to have encour- have been described [5, 71, at the subnanomole level aged the development of computerized spectral pre- the spectra show either very small fragment ion peaks diction and interpretation [12, 131, but for such ap- or none at all and thus yield little or no sequence in- proaches to be successful it is important to catalog any formation. systematic divergences from the expected behavior. In Collisional activation by either high- or low-energy studies of the structure of the scrapie prion protein collisions of the ions with neutral gas atoms (or a solid isolated from the brains of hamsters [ 141, we observed surface) can be used to increase the internal energy that a peptide with an N-terminal glutamine residue of the parent ions, leading to greatly increased frag- exhibited such a divergence, and we have since estab- ment ion abundance, giving extensive amino acid se- lished that this is a general phenomenon. quence information [S, 91. A major advantage of apply- Analysis of the PrPsc peptide QHTVTITTK (residues 185-193), which can be generated by endo- ‘Permanentaddress: School of Pharmacy, University of London, proteinase Lys-C digestion of the scrapie protein, is Brunswick Square, London WClN IAX, United Kingdom. of considerable interest because of its possible role Address reprint requests to Dr. A. M. Falick, Depz-tment of Pharmaceuti- cal Chemistry, University of California-San Francisco, San Francisco, CA in determining scrapie incubation times. Scrapie is a 94143-0446. transmissible neurodegenerative disease, and PrP% is

@ 1990 American Society for Mass Spectromehy Received September 28, 1989 104-0305/90/$3.50 Accepted January 4, 1990 JAm SotMass Spectrom199LJ,l, 258-264 MS/MS OF N-TERMINAL GLUTAMINE PEPTIDES 259

a host-encoded protein that forms a component of the 0 putative infectious agent. Both hamsters and mice with P H2N--C-C+ *c-c+ the above PIP sequence have short incubation times, -NH, whereas mice with long incubation times have two (342 f ,C”2 1 amino acid substitutions in the PrP gene, one of which H,ii-C H HN-C’H is a valine at amino acid 189 to give QHTWTTTK [15]. \ \ Long incubation times are dominant in mice that are Scheme I heterozygous for the PrP gene, and it is of great interest to measure the respective amounts of the scrapie prion protein produced from each allele. Although C- Glu-HTVITITK (MH+ peak at m/z 999.4) and in Fig- DNA (amplified by polymerase chain reaction) indi- ure lb for QHTVmTK (MI-I+ peak at m/z 1016.4). cates that the transcription of both alleles is similar The spectrum ln Figure la is relatively simple With [16], there is no information on the relative amounts few abundant peaks. Apart from peaks due to loss of of the two PrP forms that are converted to PrPsc during 45 u from the MH+ ion and the histidine immonium the course of the disease. In this laboratory, mass spec- ion at m/z 110, the major peaks correspond to early trometry and MS/MS are playing an important role in members of the b, ion series arising from cleavage of characterizing the PrP structures and resolving ques- the peptide backbone at the amide bonds with charge tions such as this. retention on the N-terminal fragments. Examples are b2 (m/z249) and b3 (m/z 350). Early members of the a,, Results and Discussion immonium ion series are also seen, giving somewhat weaker peaks for a2 (m/z221) and G (m/z421). The Endoproteinase Lys-C digestion of the scrapie prion higher mass region shows some ion formation from protein (PrP 27-30) gave a series of peptides that were the C-terminal fragments, but those peaks are much separated by reverse-phase high-performance liquid smaller despite the presence of a basic lysine residue. chromatography (HPLC), individual fractions being This may be explained by competition with the histi- collected for amino acid analysis and LSIMS. Two ad- dine residue close to the N-terminus. jacent peaks in the chromatogram gave identical amino In some respects the spectrum of the glutamine- acid compositions, but by LSIMS the peaks due to containing analogue in Figure lb is similar in that the their protonated molecular ions were found to dif- N-terminal ions are predominant. However, it is more fer, being observed at m/z 1016 and 999, respectively. complex, and the anticipated mass shifts of 17 u for The former was identified as the mass of a peptide the a, and b, ions occur only partially. In addition to QHTVTTTTK predicted from the gene sequence for the a, and b, ions at the expected masses, there are PrP protein [17, 181, which was also consistent with the associated peaks of greater abundance occurring 17 u amino acid analysis. The other component was iden- lower, that is, at the same masses as the a,, and b,, ions tified as the same peptide but with an N-terminal py- from the pyroglutamic acid-containing peptide. Thus roglutamic acid residue arising from an internal nucle- the spectrum in Figure lb appears to be a composite ophilic attack with expulsion of ammonia. The reaction of the anticipated spectrum plus a substantial contri- is illustrPted in Scheme 1. For proteins with N-terminal bution from that of Figure la, both a, and b, ion series glutamine, this is a known in vivo posttranslational being accompanied by a, -17 and b, - 17 ion series. We modification [19], but it is also a chemical process that rationalize these results as a partial loss of NH3 from occurs in peptides [20], as discussed by Carr and Bie- the molecular ion and/or the N-terminal fragment ions. mann [21]. They state that it is most likely to occur As this is a tandem spectrum, the ions observed can in acid media, which is consistent with our experience arise only from the parent ion selected in the MSI, and that the pyro-Glu and Gln forms can be completely therefore there is no possibility that this phenomenon resolved by HPLC and shown to be pure, but storage is due to loss of ammonia from the neutral molecule. in O.l”% trifluoroacetic acid (TFA) often results in sig- Thus the internal energy imparted by collision is suf- nificant conversion of Gln to pyro-Glu. ficient to break several bonds, with backbone cleavage This work was being carried out as part of a project being accompanied by ammonia elimination. to map the amino acid sequence and posttranslational For clarity, only the N-terminal ions are labeled modifications of the PrP proteins, and this peptide cor- in Figure lb, but a portion of this spectrum is ex- responded to a portion of the sequence not previously panded in Figure lc. This shows that C-terminal ions observed. Consequently it was important to confirm are formed and that virtually every peak can be as- the sequence, and this was undertaken by MS/MS. signed as either an N-terminal or a C-terminal frag- It was anticipated that the existence of both the N- ment. terminal glutamine and N-terminal pyroglutamic acid This analysis could be complicated by the fact that peptides would aid the interpretation because the N- c, series ions are 17 u higher than the bn series, mak- terminal fragments would exhibit 17-u shifts whereas ing Gln-peptide c,, - 17 ions coincident with pyre- the C-terminal fragments would have identical masses. Glu-peptide b, ions, as illustrated in Figure lc. The The MS/MS spectra are shown in Figure la for pyro- c ions give relatively weak peaks in most cases; how- 260 BALDWIN ET AL. J Am Sot Mass Spectrom 19!W,l, 258-264

100

50 a,

80 (b)

70 a-l7 60

0 1 1 60 2lio

i 12-1‘7 100

90 b a0 (4 7 70 C 7 60 b,-17 50 I 40

30 a,-17 Figure 1. Collision-induced dis- 2. ya b,-17 sociation (CID) daughter ion mass w,d, aa spectra of the hfH* ions of (a) 1O b, =a . pyre-Glu-HTVTRE (m/r 999.4) and (b) QHTVllTK (m/z 1016.4); 0 I (c) an expansion of the rna~s range 200 400 450 500 ZOO-500u from Figure lb. Mass ever, to clearly establish that the loss ofI 17 u in the free ammo groups, and derivatization of both of these ions arises by elimination of ammonia from the N- basic sites could be anticipated to have relatively lit- terminal glutamine, we carried out an N-acetylation on tle effect on the product ion distribution in the MS/MS QHTVTTTTK. This example was selected because both spectrum. This spectrum, which is illustrated in Figure the N-terminal glutamine and C-terminal lysine have 2 and in which only the N-terminal ions are identified, J Am Sot Mass Spectrom1990,1,2%?-264 MSiMS OF N-TERMINAL GLUTAMlN@ i=%lTILlES 261

Fii 2. CID daughter ion mass J L spectrumof the MH+ ion of N,N’- 00 dmtylated QfrW4TKE (m/z Mass X100.5). shows that, as with the unacetylated compound, the residues, although it is observed to some degree for major fragments are attributable to the b, series ions. the b, series right through to blo with no exceptions. There are no peaks arising from losses of 17 u, con- Another V8 peptide with N-terminal glutamine is fii the role of the free N-terminal NH2 group in the pentapeptide QLARE obtained by digestion of col- attacking the amide and eliminating NHB. Losses of 42 icin Ia [23]. The strongly basic arginine residue near and 59 u from the b, and a,, ions can be interpreted the C-terminus has a profound effect on its daughter as arising through sequential losses of ketene and am- ion mass spectrum illustrated in Figure 3, transferring monia. most of the ion current to C-terminal fragments. One It is noteworthy that the daughter ion spectra of of the more abundant N-terminal fragment ions is the QHTVTITTK and the related compounds illustrated a2 fragment whose peak appears at m/z 214, which is in Figures 1 and 2 are relatively unusual in that a large accompanied by a larger a, - 17 peak at m/z 197, but fraction of the ion current is carried by a few peaks the % peak at m/z 441 has no significant corresponding in the lower mass (~300 u) region. As stated above, a,, - 17 peak at m/z 424. Similarly, the relatively mi- histidine is able to stabilize the charge, but it also ap- nor b, series ions show little evidence of 17-u losses. pears that the repeating threonine region is relatively One further indication of loss of 17 u can be seen at unstable and tends to fragment. m/t 84, the peak for the al immonium ion for pyroglu- Two further examples of peptides containing N- tamic acid, even though the glutamine immonium ion terminal glutamine were studied. Digestion of a pro- peak at m/z 101 is of greater height. However, the m/z tein with StaphyfococcusRUWS V8 can give peptides 84 peak is normally observed in daughter ion spectra without basic residues, as cleavage normally occurs of peptides containing Gln regardless of its position in on the C-terminal side of glutamic or aspartic acid the sequence [12]. residues. An example with an N-terminal glutamine is The N-glutamine compounds studied here gener- the undecapeptide QAGGDATENFE, which was ob- ally do not show major peaks for loss of 17 u from tained by digestion of trypsin-solubfiized rabbit cy- MH+ , even though these are stable ions correspond- tochrome b 5 [ZZ], and for which the pyroglutamic acid- ing to MH+ of the pyroglutamic acid compounds. containing form was also isolated. The daughter ion (Ions corresponding to loss of 17 u can also arise by mass spectra are illustrated in Figures 3a and 3b for processes other than loss of NH, from glutamine.) It the pyro-Glu and Gln peptides, respectively. There is has been shown that the b, ions from the pentapep- no basic residue at the C-terminus, and the sequence tide Leu-enkephalin are formed by relatively complex, peaks arise largely from the N-terminus, particularly high-energy processes compared with the y,, ions (241. the b, ions, although the C-terminal ion peaks as rep- If this is a general phenomenon, it suggests that NH, resented by the y, series are of significant abundance. loss from MH’ occurs only for ions of high internal For clarity and convenience, only the a,,, b,, a,-17, and energy, which then undergo further breakdown. Hunt b, - 17 peaks are identified in Figure 3b. The majority et al. [25] noted previously that low-energy CID spec- of the remaining peaks can be ascribed to other frag- tra of peptides that have amino group side chains can mentations as described in the introduction, plus inter- also display the usual fragment types with further loss nal sequences. From these two spectra it is again clear of NH,. However, the findings reported here demon- that the a,, and b, ions for the glutamine-containing strate that if a peptide contains N-terminal Gln, the peptide are accompanied by a,, - 17 and b, - 17 series. N-terminal sequence ions can be dominated by the cor- This is most pronounced for ions of up to about five responding ions 17 u lower. 262 BALDWIN ET AL 1 Am Six Mass Spechvm 1990, 1,258-264

1.5

40 X 30 ‘a b, I 20

0 100 200 300 40

90 04 80

60 &-l-17

Figure 3. CID daughter ion mass spectra of the MH’ ions of (a) pyre-Glu-AGGDATENFE (m/z 1121.5) and (b) QAGGDATENFE (m/z1138.5).

c 100

Figure 4. CID daughter ion mass spectrum of the MH+ ion of QLARE (m/z616.3). J Am kc Mass Spectrom 1990,1.258-264 MS/MS OF N-TERMINAL GLUTAMINE PEF’I-IDES 263

Conclusions obtained from S. uureus VS digests of trypsin- solubilized rabbit cytochrome b5 [22] and colicin Tandem mass spectrometry can play an important role Ia [23], respectively. All peptides were isolated by in the sequencing of peptides at the picomole level. reverse-phase HPLC using either C, or Cl* columns In comparison with Edman sequencing, it offers the (Vydac) with water-acetonitrile-0.1% TFA gradients. advantages of being able to cope with blocked N- N-Acetylation of QHTVTTTTK was carried out by terminal peptides and other posttranslational modi- addition of 5 ILL of acetic anhydride to a dry aliquot fications such as glycosylation, phosphorylation, sul- (~0.5 nmol) of the peptide at room temperature. The fation, ester&cation, and disulfide bridge formation. reaction was allowed to proceed for 1 min, after which However, successful spectra1 interpretation relies on the excess reagent was removed by vacuum centrifu- being able to relate the peaks in the mass spectra to the gation. amino acid sequences in a consistent and predictable Tandem mass spectrometry was carried out on a way. Many rules for peptide fragmentation by CID are Kratos Concept IIHH four-sector instrument equipped already understood, and we are currently building up with an array detector [28]. Parent ions were gener- a database that will further aid interpretation and pre- ated from a glycerol-thioglycerol matrix in an LSIMS diction of spectra. source [29] employing an 18-keV Cs’ primary beam. It was already known that the presence of an N- The accelerating voltage in MS1 was 8 kV, and the col- terminal Gln residue in a peptide formed by digestion lision energy for CID was 6 keV. The collision gas (he- of a larger peptide or a protein could result in NH3 lium) was used at the pressure sufficient to suppress elimination in solution with formation of a pyroglu- the parent ion beam to about 30% of its initial value. tamic acid residue. This can be identified by the pres- The instrument was controlled and data were acquired ence of two closely spaced chromatographic peaks, with a DS-90 data system. The spectra shown in the which give MH+ separated by 17 u by FAB or LSIMS. figures are from single acquisitions using the array de- A similar effect can occur with the ions in the CID spec- tector. An array acquisition consists of a single series trum, shifting the N-terminal fragment ion series by 17 of adjacent l-s exposures, each covering a mass range u. This effect is most pronounced for peptides that fa- of f 2% of the central mass. The amounts of material vor N-terminal ions and is least apparent for peptides used to produce the daughter ion spectra could not having basic residues at or close to the C-terminus. be measured accurately owing to uncertainties in the It has been noted that a similar cyclization can occur degree of digestion and losses resulting from isolation in peptides that contain N-terminal carboxymethylcys- and purification; however, we estimate the amounts as teine in solution with the elimination of water [26]. It 100-200 pm01 per spectrum. may prove interesting to determine whether a similar loss of 18 u from the N-terminal ions is observed under Acknowledgments CID conditions. This work was supported by National Institutes of Health grant no. 603132 (S.B. kusiner), NIH Divisionof Research Resources Experimental grant no. RR01614 (A.L. Burlingame), and National Science Foun- dation Biological Instrumentation Program grant no. DIR8700766 The peptides studied were obtained from enzyme di- (A.L. Burlingame). We gratefully acknowledge the technical as- gests of various proteins. QHTV’RTTK was produced sistance of F.C. Walls. We also wish to thank the referees for helpful comments. in an endoproteinase Lys-C digest of the scrapie protein PrP. The protein was purified and precipitated with ethanol as described previously [27]. The ethanol- References precipitated protein pellets were solubilized in 6 M 1. Barber, M.; Bordoli, R. S.; Sedgwick, R. D.; Tyler, A. N. J. guanidine HCl, 50 mM Tris Cl (pH 8.2-8.45), and 5 mM Chem. Sot., Chmn. Commun. 1981,325-327. EDTA. The protein was reduced with 2 mM dithio- 2. Aberth, W.; Straub, K.; Burlingame, A. L. &al. Chem. 1982, threitol, alkylated with 6 mM iodoacetic acid, and 54, 2029-2034. precipitated with lo-15 volumes of ethanol overnight 3. Aberth, W.; Burlingame, A. L. In Ion Fwmafion in Organic at -20 “C. The pellet was collected by centrifuga- Solids; Benninghoven, A., Ed.; Springer Verlag: New York, 1983; pp 167-171. tion, washed with ethanol to remove the last traces 4. Burlingame, A. L.; Maltby, D.; Russell, D. H.; Holland, P. of guanidine HCl, resuspended in 0.1% SDS, 50 mM T. Anal. Chem. 1988, 60, 294R-342R. Tris Cl (pH 8.45), and 1 mM EDTA, and boiled for 1 5. Biemann, K.; Martin, S. A. Mass Spectmm. R~u. 1987, 6, l-75. min. This preparation was digested overnight at 37 “C 6. McNeal, C. J., Ed. The Analysis ofPeptides and Proteins by Muss with 5-10 pg of endoproteinase Lys-C (sequencing Spectrometry; Wiley: Chichester, 1988. grade, Boehringer Mannheim). The sample was incu- 7. Roepstorff, I’.; Fohlmann, J. Biomed. Mass Spectrum. 1984, II, bated with 0.3 unit of PIPLC (phosphoinositol-specific 601. 8. McLafferty, F. W., Ed. Tandem Mass Spectromety; Wiley: New phospholipase-C) for 3 h at 37 “(3 in the presence of York, 1983. the protease inhibitor 1 mM PMSF (cr-toluenesulfonyl 9. Biemann, K.; Scoble, H. A. Stience 1987, 237, 992-998. fluoride). 10. Johnson, R. S.; Martin, S. A.; Biemann, K. Int. J. Mass Spec- The peptides QAGGDATENFE and QLARE were tmm. ion Processes 1988, 86, 137-154. 264 BALDWINET AL. I Am Sot Mass Spectrom 1!390,1.258-264

11. Johnson, R. S.; Martin, S. A.; Biemann, K.; Stilts, J. T.; 21. Carr, S. A.; Biemaxm, K. In Methods in Ewymolqy; Vol. 106, Watson, J. T. Anal. Chem. 19137,59, 2621-2625. Weld, F.; Moldave, K., Eds.; Academic: Orlando, FL, 1984; 12. Johnson, R.; Biemann, K. Biomed. Environ. Mass Spectrom. pp 29-58. 1989, 18, 945-957. 22. Gibson, B. W.; FaIick, A. M.; Lipka, J.; Waskell, L. Biorhim. 13. Hines, W.; Hermann, J.; Gibson, B. W.; Burlingame, A. L., Biophys. Acta, in press. unpublished studies, 1989. 23. Falick, A. M.; Mel, S. F.; Stioud, R. M.; Burlingame, A. L. 14. Prusiner, S. 8. Ann. Rev. Microbic% 1989, 43, 345-374. In Techniques in Protein Chemistry; HugIi. T., Ed.; Academic: 15. Westaway, D.; Goodman, I’. A.; Mirenda, C. A.; McKinley, Orlando, FL, 1989; pp 152-159. M. P.; Carlson, G. A.; Pmsiner, S. B. CelE 1987, 51,651-662. 24. Alexander, A. J.; Boyd, R. K. Int. J. Mass Spectrom. ion Pro- 16. Mirenda, C.; Westaway, D.; Prusiner, S. B., unpublished re- cesses 1989, 90, 211-240. suits, 1989. 25. Hunt, D. F.; Yates, J. R., III; Shabanowitz, Jo; Winston, S.; 17. Basler, K.; Oesch, B.; Scott, M.; Westaway, D.; WaIchli, M.; Hauer, C. R. Pmt. N&l. Acad. Sci. U.S.A. 1986, 83,6233-6237. Groth, D. F.; McKinley, M. I’.; Prusiner, S. B.; Weissmann, 26. Bradbury, A. F.; Smyth, D. G. Biochem. 1.1973, 131,637-642. C. Cell 1986, 46, 417-428. 27. Stahl, N.; Borchelt, D. R.; Hsiao, K.; F’rusiner, S. 8. Cell 18. Oesch, B.; Westaway, D.; Wakhli, M.; McKinley, M. P.; Kent, 1987, 51, 229-240. S. B. H.; Aebersold, R.; Bany, R. A.; Tempst, I’.; Teplow, D. 28. Walls, F. C.; Baldwin, M. A.; F&k, A. M.; Gibson, 8. W.; B.; Hood, L. E.; Pmsiner, S. 8.; Weissmann, C. Cell 1985, GiIIece-Castro, B. L.; Kaur, S.; Maltby, D. A.; Medzihradsky, 40, 735-746. K. F.; Evans, S.; Burlingame, A. L. In BioEogicnlMuss Spec- 19. Orlowski, M.; Meister, A. The Enzymes, Vol. 4; Academic: tromety; Burlingame, A. L.; McCIoskey, J. A., Eds.; Elsevier: New York, 1971; pp 123-151. Amsterdam, in press. 20. Weitzerbii-Falzspan, J.; Das, 8. C.; Gros, C.; Petit, J.-F.; Led- 29. Falick, A. M.; Wang, G. H.; Walls, F. C. Anal. Chmt. 1986, erer, E. Eur. J. Biochm 1973, 32, 525-532. 58, 1308-1311.