protein and peptide analysis by mass spectrometry

1 Mass Spectrometry in the Analysis of Peptides and Proteins, Past and Present Peter Roepstorff When the editor, John Chapman, asked me to write the introductory chapter to this volume and told me that it would be dedicated to the late Michael “Mickey” Barber, I felt very honored and also humbled because I have always considered Mickey to be one of the most outstanding pioneers in the field of mass spectrometry (MS) of protems. Most younger scientists associate Mickey’s name with the invention of ionization of nonvolatile compounds by fast-atom bombardment (FAB) m 1981 (I). It is also true that this invention had a major impact on the practical possibilities for mass spectrometric analy- sis of peptides and proteins. However, only a few of the present generation of scientists involved m MS of peptldes and proteins know that MS of peptides was already an active field 30 years ago and also that Mickey’s career in many ways reflects the development m the field through all these years. In the 196Os, a number of groups had realized and actively investigated the potential of MS for peptlde analysis. The only ionization method avallable was electron impact (EI), which required volatile denvatlves. This necessitated extensive derivatization of the peptides prior to mass spectrometric analysis. The followmg groups were pioneers in investigating MS: the group headed by M. Shemyakin at the Institute for Chemistry of Natural Products of the USSR Academy of Sciences, which worked with acylated and esterlfied peptides (2); K. Biemann’s group at Massachusetts Institute of Technology, which used acylation followed by reduction of the peptides to ammo alcohols followed by trimethylsilylatlon and gas chromatography (GC)/MS (3); and E Lederer’s group at the Institute for Chemistry of Natural Substances in Gif sur Yvette, France, which studied natural peptidolipids among other compounds. Mickey Barber, who at that time worked at AEI in Manchester, got involved m the From Methods In Molecular B/o/ogy, Vol 61 Protern and Pepbde Analysrs by Mass Spectrometry E&ted by J R Chapman Humana Press Inc , Totowa, NJ 2 Roepstorff work of E. Lederer. The French group had isolated a peptldohpld called fortultine, which was analyzed by Mickey on the MS9 mass spectrometer Fortultine appeared to be a very fortuitous compound. It was N-terminally blocked with a mixture of fatty acids, was naturally permethylated, and con- tamed an esterlfied C-termmus Mickey Barber obtained a perfect El spectrum of this 1359-Da peptide and was able to interpret the spectrum (4). 1 believe that this, at that time, was the largest natural compound ever analyzed by MS. The achievement contains many of the elements now considered to be only possible with the most contemporary MS techniques. Thus, heterogeneity both m the sequence and the secondary modlficatlons was determined and the frag- ment ions, always present in EI spectra, allowed sequencmg. The realization of the effect of iV-methylatlon on the volatlhty of the peptldohprd resulted m development of the permethylatlon procedure for peptlde analysis by MS (5) For me, personally, the fortuitme paper has also been very fortmtous. Shortly after Its publication, I became involved m peptlde synthesis by the Merrlfield method, which, unfortunately, did not always yield the expected product. Modi- fications, which researchers had no practical method to analyze, were frequent. The fortultine paper inspired me to investigate the possible use of MS for analy- sis of these modlficatlons. Smce then, MS has been my mam tool m protein studies. A few years later, I first met Mickey during a visit to AEI m Manches- ter. I was very fascinated by his lively engagement m the subject, and also realized over a pmt of beer m a nearby pub that MS was not his only scientific interest. He was deeply involved m surface science and had been also very active in the development of equipment for photo-electron spectroscopy, as well as in exploring the possibihties of electron spectroscopy for chemical analysis (ESCA). Shortly after we first met, he moved from AEI to a lecture- ship at the University of Manchester Institute of Science and Technology, where he became a full professor m 1985. In the same year, he was elected Fellow of the Royal Society. In the 197Os, MS of peptldes progressed slowly and, although Its potential was demonstrated by a number of applications to structure elucidation of modl- fied peptides, the field was stagnating by the end of the decade. Two new ion- ization methods, chemical lontzation and field desorptlon, appeared during that period. They created new hopes for improvements m peptlde analysis by MS, but unfortunately, they did not result m a real breakthrough In that period, I had no real contact with Mickey Barber. It is my impression that he, mtuitlvely or consciously, was realizing that the opening of new possl- blhtles for mass spectrometric analysis had to come from surface science. Any- way, among other subjects, he started to investigate surface analysis by secondary-ion mass spectrometry (SIMS). This led to his discovery of a new technique for desorptlon and ionization of mvolatile organic compounds. He Mass Spectrometry 3 termed the technique FAB because, instead of the primary ions used m SIMS, he used a beam of 3-l 0 keV argon atoms to effect desorption and iomzation. The choice of atoms instead of ions was mainly determined by a desire to avoid surface charging phenomena, which could disturb ion focusing m the sector mstrument used. It was later realized that primary ions work just as well as atoms, and the fast atom gun is now frequently replaced by a cesmm gun creat- mg 20-30 keV primary cesmm ions. The concept of SIMS of orgamc solids was not new. Benninghoven et al. (6), at the University of Munster in Germany, had already, some years earlier, demonstrated mass spectra of organic solids, mcludmg amino acids, by SIMS. The spectra, however, were only transient because the surface was quickly destroyed by the high flux of primary ions. As a matter of fact, the real discov- ery by Mickey Barber was the use of a liquid matrix and the technique IS now often termed hquid secondary-ion mass spectrometry (LSIMS) when a pri- mary beam of cesium ions is used Instead of fast atoms. The trick of using a matrix was that the matrix surface was contmuously replenished with sample so that secondary ions could be produced continuously over a long period of time. This feature also made the technique directly compatible with scanning mass spectrometers. In fact, an important reason for the immediate success of FAB was that it could readily be mstalled on existmg sector field and quadru- pole mass spectrometers. In my own laboratory, for example, we installed FAB in 1982 simply by replacing the standard solids inlet probe with a simple rod and by placing an ion gun m the place of the GC mlet on a Varian MAT 3 1 IA double-focusing sector Instrument. To my knowledge, Mickey was the first to introduce the use of a matrix m MS and, as is described later in this chapter, the use of a matrix is essential in all mass spectrometric techniques used for analy- sis of peptides and protems. Another technique that allowed the analysis of large, mvolatile organic mol- ecules was plasma-desorption mass spectrometry (PDMS) developed as early as 1974 by Torgerson et al. (7). The technique had been shown to be capable of the analysis of large underivatized peptides (8), and, soon after, of proteins, by the demonstration of the first mass spectrum of insulin in 1982 (9) and of a number of snake toxms with molecular masses up to 13 kDa (ZO). Instrumen- tation for PDMS became commercially available a few years later, and the real breakthrough for this techmque came 2 years later with the simulta- neous discoveries of the advantages of using nitrocellulose as support (2 2) or reduced glutathione as matrix (12). Shortly after the publication of the PD- spectrum of insulin, FAB mass spectra of insulin were also published (13,14), and in the followmg years, mass spectra of proteins as large as 25 kDa were published using both techniques, concomitantly with the gradual acceptance of the potential of mass spectrometrtc analysts by a number of protein chemists. 4 Roepstorff In that period, Mickey Barber visited my laboratory for a period during which he “played” with our plasma-desorption mass spectrometers to get a personal feeling of the potential of this technique compared to FABMS. Mickey’s enthusiastic engagement made his stay a great mspiration for me as well as for my students, and we had many long discusstons about the status and future perspectives of the field It seemed to us that both techmques had fundamental hmitations that would prevent then full acceptance among protem chemists. The major limitations were that it was drffcult to extend the mass range beyond 25 kDa, and that this range could only be attamed for a few ideally behaved protems. The sensitivity, which was m the low- to mid-picomole range, was also not as good as desired, and finally, mixture analysis was SubJect to consid- erable selectivity owmg to suppression effects. In 1988, two new mass spectrometrtc techniques, which dramatically extended the potential of MS for protein analysis, were published, and it was soon appreciated that they were able to overcome most of the hmrtations men- tioned. At the Amerrcan Society for Mass Spectrometry (ASMS) conference m June m San Francisco, John Fenn from Yale University gave a lecture on the apphcatton of a new ionization technique, termed electrospray ionization (ESI), for protem analysts (15) Those of us who attended the lecture walked away with the feeling that we had witnessed a real breakthrough for the mass spec- trometrtc analysis of large btomolecules. A few months later at the Interna- tional Mass Spectrometry Conference m Bordeaux, France, Franz Hillenkamp gave a lecture descrtbmg another new tomzatton technique, matrix-assisted laser desorption/iomzation (MALDI) (16). In this lecture, he showed molecu- lar ions of proteins up to 117 kDa using time-of-flight analysis. MALDI seemed at least as promising for protein analysis as ESI. A few years later, commercial mstruments were available for both tech- niques, and the questron was really which of the two techniques would be the future method of choice m the protein laboratory. In fact, at present, I consider the two techniques to be highly complementary. They have both dramatically improved the perspectives for the application of MS in protein chemistry to such an extent that a protein chemistry laboratory without access to these two techniques or at least one of them cannot be considered up to date. A common feature of both techniques is that, if the idea of a matrix IS considered in its widest sense, both can be considered to be matrix-dependent Just as FAB and PD. MALDI IS, as indicated by its name, a matrix-dependent method. How- ever, ESI, in spite of an entirely different ionization mechanism, can be consid- ered to be matrix-dependent, since a prerequisite is that the analyte IS dissolved m an appropriate solvent prior to ion formation m the electrospray process, Mickey, unfortunately, died in May 1991, and therefore did not get the chance to see how his dream about the role of MS m protein studies and hts Mass Spectrometry 5 concept of usmg a matrrx have made then triumphal progress durmg the past 5 years. We m the scienttfic commumty have been deprived of the posstbihty to obtain his interpretation of which elements are common to the four tech- niques that have created the progress in MS applied to protein chemistry In the absence of hts mterpretation, I will try to use the way of thinking and argumg I have experienced m my dtscusstons with Mickey to outline what I consider to be the main function of the matrix. This 1s independent of whether tt 1s the nitrocellulose support in PD, the liquid matrix in FAB, the solid matrix m MALDI, or the solvent m ESI. A common feature 1s that the matrix/support is present in a large molar excess relative to the analyte. This indicates that a prime purpose of the matrix is to isolate the protein molecules and prevent aggregation Next, the matrix must create a platform that can be removed, leaving the single analyte mol- ecules free in the gas phase. This 1s effected by different means m the four techniques. In PDMS, the mtrocellulose support most likely decomposes on high-energy impact, so that the analyte molecules are left free and pushed off the surface by the resultmg pressure wave. In FAB, the analyte molecules are sputtered from the liqutd matrix surface, maybe still partly solvated in microdroplets of ltquid matrix, followed by desolvatton by multiple colhsions just above the matrix surface. In MALDI, the solid matrix absorbs most of the laser energy, and decomposes or evaporates leaving the analyte molecules free m the expanding matrix plume. Finally, in ESI, the mtcrodroplets created in the electrospray process by combined evaporation and coulombic explosions are subdivided until each droplet contains only one or a few analyte molecules, which, on final desolvatton, leaves the analyte molecules free. The last step is that the analyte molecules must be ionized. Several different ionization mecha- nisms are wtthout doubt operattve m the different techniques and also within a single technique. In PD, FAB, and MALDI, chemical iomzatton is most likely to be the dominant tomzatton mechanism, although preformed ions, as well as other mechanisms, may also play a role. The iomzatton mechamsm in ES1 is still controversial, and tt ts outside the scope of this introductory chapter to enter thrs debate. In summary, the matrix serves to isolate single analyte molecules, to create a removable platform from which the analyte molecules can be brought mto the gas phase, and to create a medium that can tomze the analyte molecules. To be able to create ions of the proteins ts, however, not suffictent to make the techniques usable in the protein laboratory. They must be compatible with the procedures generally used m protein chemistry m terms of senstttvtty and acceptance of solvents, detergents, and buffers, They must be able to handle impure samples and complex mixtures. Last, but not least, the information gained must be of sufficient value to justify the effort and cost needed to obtain 6 it. The numerous apphcations of MS to protein studies published during the last decade and the rapid acceptance of MS m the protein community clearly show that these conditions are now fulfilled. The followmg chapters m thts volume describe these mass spectrometric techniques in detail and demonstrate a wide variety of applications: The use of MS for protein identificatton m com- bination with high-resolutton separation techntques, such as 2D-PAGE, its compatibthty with buffers and detergents, and its use m combinatron wtth HPLC are illustrated. Several methods for the sequencing of pepttdes, determi- nation of disulfide bonds, and different methods for the localization and struc- ture determmatlon of secondary modificattons, including glycosylation, are described. Even examples of tasks considered very dtfficult, such as the analy- sis of very hydrophobic protems (e.g , membrane proteins), the highly specific quantitation of btologrcally active pepttdes, and studies of noncovalent mter- actions between protems or between protems and low-mol-wt hgands, are now wtthm reach. In my mind, there IS no doubt that MS will be an essential tech- nique m all protein studies m the future. References 1, Barber, M , Bordoh, R. S., Sedgwick, R. D., and Tyler, A. N (1981) Fast atom bombardment of solids (FAB) a new ion source for mass spectrometry J Chem Sot Chem Commun 1981,325-327 2 Shemyakm, M. M , Ovchinnikov, Yu. A., Kiryushkin, A A., Vinogradova, E I , Miroshmkov, A I., Alakhov, Yu. B., et al. (1966) Mass spectrometric determina- tion of the ammo acid sequence of peptides Nature 211,361-366 3 Biemann, K and Vetter, W (1960) Separation of peptide derivatives by gas chro- matography combmed with mass spectrometrtc determmation of the ammo acid sequence. Blochem Blophys Res Commun 3,578-584 4 Barber, M., Jolles, P., Vilkas, E , and Lederer, E. (1965) Determmation of ammo acid sequence in oligopeptides by mass spectrometry. I The structure of fortuitme, an acyl-nonapeptide methyl ester Biochem Bzophys Res Commun l&469-473 5 Das, B C , Gero, S D , and Lederer, E. (1967) N-methylation of N-acyl ohgopeptides Blochem Bzophys Res Commun 29,2 1 l-2 15 6. Benninghoven, A., Jaspers, D , and Srchtermann, W. (1976) Secondary-ion emts- sion of ammo acids. Appl. Phys 11,35-39. 7 Torgerson, D F., Skowronski, R P , and Macfarlane, R D (1974) New approach to the mass spectrometry of non-volatile compounds. Btochem Bluphys. Res Commun 60,616-621 8 Macfarlane, R. D and Torgerson, D F. (1976) Californium-252 plasma desorp- tion mass spectrometry Sczence 191,92&925. 9. Hakansson, P , Kamensky, I , Sundqvist, B., Fohlman, J., Peterson, P., McNeal, C. J., and Macfarlane, R. D (1982) 127-I plasma desorption mass spectrometry of insulm J Am Chem Sot 104,2948,2949 Mass Spectrometry 7 10 11 12. 13 14 15 16 Kamensky, I., Hakansson, P , KJellberg, J., Sundqvist, B , Fohlman, J., and Petterson, P. (1983) The observation of quasi molecular ions from a tiger snake venom component (MW 13309) usmg 252-Cf plasma desorption mass spectrom- etry FEBS Lett 155, 113-l 16 Jonsson, G. P , Hedm, A B , Hbkansson, P L , Sundqvist, B U R , Save, G S , Nielsen, P F , et al ( 1986) Plasma desorption mass spectrometry of peptides and proteins absorbed on mtrocellulose Anal Chem 58, 1084-1087 Alai, M , Demirev, P , Fenselau, C , and Cotter, R J (1986) Glutathione as matrix for plasma desorption mass spectrometry of large peptides Anal Chem 58, 1303-1307 Dell, A and Morris H R (1982) Fast atom bombardment-high field magnetic mass spectrometry of 6000 Dalton polypeptides Blochem Blophys Res Commun 106, 14561461 Barber, M , Bordoh, R S , Elhot, G J , Sedgwick, R D., Tyler, A. N., and Green, B N. (1982) Fast atom bombardment mass spectrometry of bovine msulm and other large peptides J Chem Sot Chem Commun 1982,93&938 Meng, C. K , Mann, M , and Fenn, J B (1988) Electrospray Ionization of Some Polypeptides and Small Proteins Proceedings of the 36th ASMS Conference on Mass Spectrometry and Allied TOPICS, San Francisco, CA, June 5-10, pp 77 1,772 Hillenkamp, F (1989) Laser desorption mass spectrometry’ mechanisms, tech- niques and apphcattons, m Advances in Mass Spectrometry, vol 11 (Longevialle, P , ed ), Heyden and Sons, London, pp 354-362 [...]... (1988) Mass Spectrometry /Mass Spectrometry Technzques and Applications of Tandem Mass Spectrometry VCH, New York 16 Hogeland, K E , Jr and Demzer, M L (1994) Mass spectrometrlc studies on the N-linked oligosaccharides of baculovnus-expressed mouse mterleukm-3 Bzol Mass Spectrom 23, 2 18-224 3 Charged Derivatives for Peptide Sequencing Using a Magnetic Sector Instrument Joseph Zaia 1 Introduction Peptides... chromatography mass spectrometry Part II Techniques and mechanisms of thermospray Mass Spectrom Rev 9,63 l&669 9 Dawson, P H (1976) Quadrupole Mass Spectrometry and Its Applrcatlons Elsevler, New York 10 Duckworth, E , Barber, R C., and Venkatasubramaman, V S (1988) Mass Spectroscopy, 2nd ed Cambridge University Press, Cambridge, UK 11 Cotter, R J (1992) Time-of-flight mass spectrometry for the structural analysis. .. a quadrupole massanalyzer of wide range of solvent polarities TS has been used for the analysis of many different solute types, including small peptides, but shows poor sensmvny with higher mol-wt analytes, and has not been used m a practical manner for the analysis of peptides or proteins 3 Mass Analyzers The function of the mass analyzer (Table 2) is to separate Ions according to then mass- to-charge... P., and Martin, S, A (1995) Delayed extraction matnxassisted time-of-flight mass spectrometry Rapid Commun Mass Spectrom 9, 1044-1050 13 March, R E and Hughes, R J (1989) Quadrupole Storage Mass Spectrometv Wiley, New York 14 Asamoto, B (ed.) (1991) FT-ICR MS Analytxal Applications of Fourier Transform Ion Cyclotron Resonance Mass Spectrometry VCH, Wemheim, Germany 15 Busch, K L., Glish, G L., and. .. Chapman, J R (1993) Practical Organic Mass Spectrometry, 2nd ed Wiley, Chichester, UK 2 Harrison, A G (1983) Chemical Ionlzatlon Mass Spectrometry CRC, Boca Raton, FL 3 Barber, M , Bordoh, R S , Elhott, G J , Sedgwlck, R D , and Tyler A N (1982) Fast-atom bombardment mass spectrometry Anal Chem 54, 645AA57A 4 Capnoh, R M (ed.) (1990) Contmuous-Flow Fast Atom Bombardment Mass Spectrometry Wiley, Chichester,... accomplished only From Methods m Molecular Bology, Vol 61 Profem and PeptIde Analysrs by Mass Spectrometry Edlted by J R Chapman Humana Press Inc , Totowa, NJ 29 Zala 30 at the expense of ion transmtssion (.5,6) By contrast, low-energy collisions, in the range of 1O-l 00 eV, are produced in triple-quadrupole (7,8) and hybrid 1e , sector-quadrupole (9,l O), mass spectrometers The ion series produced from high-energy... and 2B, respectively The effect of TMAA derivattzatton 1sto produce a complete series of a, and d, ions for both peptides, although the glytines do not produce charge-remote ions, as seen in Fig 2B by the absence of the a2 and a3 ions With the knowledge that a fixed positive charge at the N-terminus produces a,, and d, ions, useful mformation can be gamed by noting the mass difference between a,, and. .. denoted a,,,b,, c,, and d, for cleavages of the nth pepttde bond, with charge retained by the N-terminal fragment Ions produced with charge retention by the C-terminal fragment are denoted v,, w,, x,,, yn, and z, (13) Th e ions denoted d,, v,, and w, are produced by cleavage of the amino acid side chain, and such tons are found only m high-energy CID spectra (3, II) The productton of d, and w, servesis... ions are produced and that these tons are predommantly v, and w, series (3,1I) For the correspondmg peptides with an N-termmal arginine, a,, and d, series predominate Since the proton resides on the argmme side chain, the generation of product ions must occur by charge-remote fragmentation in these peptides (16) A terminal argirune residue provides for optimal sequencmg of a peptide by high-energy CID... (SYSMEHFRWGKPVG) Peptides lacking an argmme residue will generally produce a mixture of N-terminal and C-termmal tons resultmg from CID The completeness of the CID spectra and the presence of gaps m the fragmentation patterns vary greatly from peptide to peptide Generally, of course, the interpretabihty of the CID spectra of these peptides will decrease with mcreasmg mass An example is shown in Fig 2A for the peptide . 1 Mass Spectrometry in the Analysis of Peptides and Proteins, Past and Present Peter Roepstorff When the editor, John Chapman, asked me to write the introductory chapter to this volume and. to be capable of the analysis of large underivatized peptides (8), and, soon after, of proteins, by the demonstration of the first mass spectrum of insulin in 1982 (9) and of a number of snake. Polypeptides and Small Proteins Proceedings of the 36th ASMS Conference on Mass Spectrometry and Allied TOPICS, San Francisco, CA, June 5-10, pp 77 1,772 Hillenkamp, F (1989) Laser desorption mass