AU760221B2 - Method for modifying and identifying functional sites in proteins - Google Patents

Abstract

The invention relates to a method for identifying one or more functional sites in proteins. Said method is characterized by the following steps: a) contacting the target protein with a binding partner (BP tag) linked with a laser-activatable marker (tag), resulting in a complex consisting of target protein and BP tag, b) irradiating the complex consisting of target protein and BP tag with laser light, resulting in the formation of free radicals which selectively modify the bound target protein at the binding site, and c) identifying the selectively modified region of the protein by a combination of protein cleavage and mass spectrometry. The invention further relates to a device for carrying out the inventive method.

Description

I.

XERION PHARMACEUTICALS GMBH X29166PCAU BO/HW Method for modifying and identifying functional sites in proteins The invention relates to a method for the specific modification and identification of functional sites in a target protein. A correlation can be made between the specific amino acid(s) and their function in the protein or their function as ligand binding sites through the loss/gain of function of the protein and subsequent determination of the modified amino acids (epitope mapping). The method is based on combined use of chromophore-assisted laser inactivation (CALI) and mass spectrometry or protein sequencing. The invention also relates to an apparatus for carrying out this method.

The techiques of recombinant DNA and of protein preparation make deeper understanding of the relations between protein structure and protein function easier.

Methods of random mutagenesis, specific genetic selection and screening methods make it possible rapidly to gain information at the protein level, but have the disadvantage that particular functional domains of the protein are very tolerant to amino acid substitutions, that some mutagenesis reagents preferentially attack particular DNA sequences, and possible mutations are limited. The information provided by these methods is thus limited.

The alternative of site-specific mutagenesis methods provides a more direct and more specific way of identifying amino acids which are important for the biological function, such as, for example, the alanine scanning mutagenesis method. In this method there is systematic mutation of each amino acid to alanine, and the function is determined. Each loss of function is -2related to the specific amino acid. This technique can be applied to proteins without the relevant threedimensional structure and without unknown functional domains. In addition, this entails in each case modification of only one amino acid, which is why specific functions involving a plurality of amino acids are not detected. In addition, the method becomes more complicated as the molecular size increases. This is a disadvantage for analysing the function of gene products in connection with diseases because these are often large proteins containing multiple structural domains.

Furthermore, other recombinant DNA techniques have been used for epitope mapping and for mapping ligand binding sites, in which there is expression of random antigen fragments with mutations introduced at the DNA level (cf. J. Immunol. Meth., 141, 245-252, 1991, J. Immunol. Meth., 180, 53-61, 1995, Virology, 198, 346-349, 1994). Recently, the phage display of random peptide libraries have been used for epitope mapping Immunol., 153, 724-729, 1994, PNAS, 93, 1997-2001, 1996) Another variant of epitope mapping of proteins is the rapid automatic synthesis of selected peptides Endocrinol., 145, 169-174, 1995) and the use of combinatorial peptide libraries (FEBS Letters, 352, 167-170, 1994). Although these methods are reliable for linear epitopes, they have been unsuccessful with nonlinear or discontinuous epitopes. The variant of using overlapping peptides was also unsatisfactory with discontinuous epitopes. This entails screening individual peptides or peptide mixures for binding to the ligand by means of ELISA, with free and bound peptides competing for the ligand (for example antibody). However, this method is elaborate and timeconsuming.

3 Another approach used a combination of protein modification and mass spectrometry (Anal. Biochem., 196, 120-125, 1991). This entails the ligand being bound to the receptor, and the complex being modified with acetic anhydride, resulting in acetylation of the lysine residues. The proteolytic cleavage mixtures of the two proteins are analysed by mass spectrometry and compared with the corresponding fragments from the untreated complex. The modified lysines can easily be detected, these lysines not being protected in the complex, and the unmodified lysines forming part of the interaction between ligand and receptor. This technique is thus confined to interactions involving lysine residues. Another variant is based on differential proteolysis (Protein Science, 4, 1088-1099, 1995), with sites sensitive to proteolysis becoming proteaseresistant after complex formation.

There has also been a description of identification by mass spectrometry of a proteolytic cleavage of free peptide antigen comparing with the pattern from the peptide antigen bound to an antibody.

The identification took place by 252 Cf plasma desorption mass spectrometry (PNAS 87, 9848-9852, 1990). There has furthermore been a description of a combination of immunoprecipitation and matrix-assisted laser desorption mass spectrometry (MALDI-MS) (PNAS, 93, 4020-4024, 1996). This entails an antigenic protein being cleaved into smaller fragments and precipitated with an antibody of interest. The immunoprecipitated peptides are identified by MALDI-MS, and the antibodybinding region is determined. In this method there was also separation of proteolytically cleaved peptides by affinity capillary electrophoresis and identification by electrospray mass spectrometry (ACE-MS) (Anal.

Chem., 69, 3008-3014, 1997). Injection of the peptide mixture is followed by injection of the antibody.

Peptides which bind to the antibody are trapped and 4 therefore do not migrate. The bound peptide is investigated by the subtraction screening method in order to determine the epitope residue on the peptide.

However, this technique is confined to linear epitopes and cannot be applied to discontinuous epitopes.

The present invention is therefore based on the object of overcoming the problems of the prior art mentioned. The intention is to provide a method with which any functional sites in any proteins can be identified. It is intended preferably to be able to identify sites involved in a ligand interaction, and epitopes. In particular, the method according to the invention should be applicable to nonlinear and discontinuous epitopes and without knowledge of the three-dimensional structure of a protein. It is also intended according to the invention for determination of the protein function to be possible without inactivating the molecule. It is additionally intended that the method be simple to use, quickly carried out and automatable. It is further intended to provide an apparatus for carrying out the method according to the invention simply.

This object is achieved by a method for identifying one or more functional sites in a protein, which is characterized in that a) the target protein is contacted with a binding partner linked to a laser-activatable marker (tag), (BP-tag), to form a complex of target protein and BPtag, b) the complex of target protein and BP-tag is irradiated with laser light to generate free radicals which selectively alter the bound target protein at the binding site, and c) the selectively altered region of the protein is identified by a combination of protein cleavage and mass spectrometry, and by an apparatus for carrying out the method.

5 It has been found, surprisingly, that combination of the CALI technique with mass spectrometry permits reliable and rapid identification of functional sites on proteins. According to the invention, a target protein is modified by CALI to inactivate it. Subsequently, the modified region of the protein is determined by mass spectrometry. Tandem MS and/or de novo sequencing are preferably used. This makes it possible in a simple way to make a correlation between the amino acid and its biological function.

This correlation between structure and function allows information to be gained for example about the correlation between a particular metabolic activity and the corresponding site in the molecule, about proteins with pathological alterations, cancer-promoting proteins etc. It is thus also prossible specifically to inactivate unwanted (for example pathological) proteins.

The target protein is initially contacted with a binding partner under conditions such that complex formation takes place, but the proteins are not denatured. Ideally, the conditions correspond to the physiological conditions of the cellular environment of the proteins. The binding partner is linked to a laseractivatable marker. The laser-activatable marker can be any marker suitable for covalent or noncovalent linkage to a binding partner, i.e. an amino acid sequence, and can be activated so that it is able to generate free radicals. The marker is preferably activated with laser light, but activation is also possible by peroxidases (hydrogen peroxide/horseradish peroxidase system).

Examples of such markers are AMCA-S, AMCA, BODIPY and variants thereof, Cascade Blue, Cl-NERF, dansyl, dialkylaminocoumarin, 4',5'-dichloro-2',7'-dimethoxyfluorescein, DM-NERF, eosin, eosin F3S, erythrosin, hydroxycoumarin, Isosulfan Blue, lissamine rhodamine B, malachite green, methoxycoumarin, naphthofluorescein, 6 NBD, Oregon Green 488, 500, 514, PyMPO, pyrene, Rhodamine 6G, Rhodamine Green, Rhodamine Red, Rhodol Green, 2',4',5',7'-tetrabromosulphonefluorescein, tetramethylrhodamine, Texas Red or X-rhodamine. The marker is preferably malachite green isothiocyanate, fluorescein isothiocyanate or 4',5'-bis(1,3,2dithioarsolan-2-yl)fluorescein. The irradiation takes place with laser light of a wavelength which is absorbed by the particular chromophore.

The binding partner can be any binding partner for the appropriate protein. It is preferably scFv, Fab, a diabody, an immunoglobulin-like molecule, a peptide, RNA, DNA, PNA or a small organic molecule.

The binding partner is preferably derived from a combinatorial library. This can be any combinatorial library, for example protein library, peptide library, cDNA library, mRNA library, library with organic molecules, scFv library with immunoglobulin superfamily, protein display library etc. The following can be presented in the libraries: all types of proteins, for example structural proteins, enzymes, receptors, ligands, all types of peptides including modifications, DNAs, RNAs, combinations of DNAs and RNAs, hybrids of peptides and RNA or DNA, all types of organic molecules, for example steroids, alkaloids, natural substances, synthetic substances etc. The presentation can take place in various ways, for example as phage display system (for example filamentous phages such as M13, fl, fd etc., lambda phage display, viral display etc), presentation on bacterial surfaces, ribosomes etc.

Using the CALI technique, target proteins are directly and specifically inactivated (cf. PNAS, 5454-5458, 1988; Trends in Cell Biology, 6, 442-445, 1996). CALI can be used to convert binding reagents such as antibodies or other ligands into functionblocking molecules. This entails a binding partner (BP) 7 being linked for example to the dye malachite green (MG) On irradiation with laser light of a wavelength which is not significantly absorbed by the cellular components, this dye generates free radicals. For the determination, the BP linked to MG (=BP-MG) is incubated with the protein sample of interest. The region to be inactivated is selected and irradiated with a laser beam at 620 nm. This light is absorbed by MG to produce short-lived free radicals which selectively inactivate the proteins bound to the BP-MG in a radius of 15 A through irreversible chemical modification. This system can be used for in vitro and in vivo assays and for intra- and extracellular target molecules.

The protein inactivated in this way is then cleaved. It is possible to use for this purpose a protease which cleaves specifically after a residue, for example Lys-C, Glu-C, Asp-N. Examples thereof are trypsin, chymotrypsin, papain etc. Chemical cleavage is also possible, for example with cyanogen bromide (specific for Met), 3-bromo-3-methyl-2-(2nitrophenylmercapto)-3H-indole (BNPS-skatole; specific for Trp), 2-nitro-5-thiocyanatobenzoic acid (specific for Cys) and Fe-EDTA.

The cleavage mixture is fractionated by electrophoresis. It is then possible to evaluate the modified amino acids by mass spectrometry and comparison with the untreated target protein.

Very recent developments in mass spectrometry have led to rapid identification of proteins (PNAS USA, 93, 14440-14445, 1996). As soon as the target protein has been inactivated by CALI it is possible to identify the modified amino acids of the inactivated protein which are presumably responsible for the inactivation by tandem mass spectrometry and, where appropriate, de novo sequencing (Rapid Commun. Mass Spectrom., 11, 8 1015-1024, 1997; Rapid Commun. Mass Spectrom., 11, 1067-1075, 1997).

The identity of the proteins complexed according to the invention with a binding partner is established by a combination of protein cleavage and mass spectrometry, where appropriate de novo sequencing.

This entails a protein which has been treated in this way and is provided with a marker initially being separated from the target protein by an electrophoresis or by a chromatography. The isolated and modified protein is then cleaved either chemically or proteolytically by the methods described above. This can take place either in the gel (that is to say by direct elution of the target protein from the gel after the separation, followed by subsequent protein separation) or in solution. Methods for cleavage in the gel are known and described, for example, in Advanced Methods in Biological Mass Spectrometry, EMBL- Laboratory, Heidelberg, 1997 or in Shevchenko, et al., Anal. Chem. 68:850-858, 1996.

The MALDI analysis is carried out in a manner known per se.

It is necessary for the nanoelectrospray analysis (nanoES) to extract the tryptic peptides from the pieces of gel. To do this, the pieces of gel are washed successively with ammonium bicarbonate, acetonitrile, dilute formic acid and again with acetonitrile. The supernatants are combined and dried in a vacuum centrifuge. The sample is dissolved in strength formic acid, rapidly diluted with water and then desalted.

The analysis by mass spectrometry can be carried out in various ways known per se, for example using an ionization source such as electrospray (Chapman, et al., Methods in Molecular biology, 61, JR Chapman editor, Humana Press Inv. Totowa NJ, 9 USA, 1996) including nanoelectrospray (Wilm. M. and Mann, Anal. Chem. 68, 1-8, 1996) and matrixassisted laser desorption and ionization (MALDI) (Siuzdak, G. Mass Spectrometry for Biotechnology, Academic Press Inc. 1996) or using a combination of mass analysers such as triple, quadrupole, time of flight, magnetic sector, Fourier transformation ion cyclotron resonance and quadrupole ion capture.

If the peptides in the cleavage mixture are insufficient for unambiguous establishment of the identity of the investigated site in the protein, further sequence information can be obtained by further fragmentation in the mass spectrometer such as, for example, by decomposition downstream of the source in MALDI-TOF, MS/MS (tandem mass spectrometry), MS n the proteins can additionally be identified by de novo sequencing.

Since in CALI there is photochemical modification of certain amino acids of the target protein, the MS can be used in two stages to identify which amino acids have been modified in what way. The method can be carried out with low and high resolution.

At low resolution it is firstly possible to determine by peptide mass mapping (Anal. Chem., 69, 4741-4750, 1997; Biochem. Soc. Transactions, 24, 893- 896, 1996; Anal. Chem., 69, 1706-1714, 1997) which segments of the protein have been modified. In the second stage, at higher resolution, tandem mass spectrometry and/or de novo sequencing on selected peptides can be used to determine the site of the modifications. It is possible with a mass spectrometer configuration able to resolve masses from 0 up to 0.03 dalton (Rapid Commun. Mass Spectrom., 11, 1015- 1024) (qQTOF and other mass analysers) to determine which specific amino acids have been modified by CALI; it is even possible to determine the type of modification from the incremental mass gain or loss.

10 The free radicals generated by CALI lead to a modification of the oxidation-sensitive amino acid side chains such as His, Met, Cys and Trp. There may also be modification of other amino acid side chains. This results in significant changes in mass, because oxygen has been added on. Comparison of the peptide fragments after the cleavage between CALI-treated fragments and untreated sample leads to an approximate localization of the modified amino acids. These amino acids are identified more accurately by de novo sequencing.

It is unnecessary to know the exact nature of the modification. The method according to the invention permits the modification to be recognized from the difference between the treated and untreated samples.

The advantage of this method is that it can be used to elucidate discontinuous epitopes even if the threedimensional structure of the protein is unknown.

The modifications induced by CALI can also be elucidated by using other methods such as parent ion scans Mass Spectrom, 32, 94-98, 1997), which increases the speed of analysis. This detects only the peptides altered by CALI.

If insufficient information is obtainable because of limited peptide fragmentation or doubtful assignment of mass, it is possible to carry out a de novo mass sequencing (Rapid Commun. Mass Spectrom., 11, 1015-1024, 1997). In this method, the peptides are labelled with 180. This is done by carrying out the tryptic cleavage in the gel with a cleavage buffer which contains 50% (vol/vol) H 2 18 0 purified by microdistillation. A QqTOF mass spectrometer is preferably used. Clear results are possible by reading the doublet owing to the 1:1 160/180 ratio. The mass difference of the doublet is an indicator of the charge state of the specific peptide. Using different charge states it is possible to read off up to 15 amino acids in a sequence for a given peptide. Comparison of the 11 18 0-labelled spectra of the untreated and CALI-modified proteins provides unambiguous information about the modified amino acids.

Figure 1 shows the scheme for identifying functional sites in proteins.

The invention also relates to an apparatus for carrying out the method according to the invention.

This is shown in Figure 2. This apparatus is an automated system of integrated independent units/parts which can be used for identifying functional sites in proteins. Part A relates to the preparation of the LBPtag. An automated LBP screening machine reveals specific LBPs directed against specific target molecules/ligands, while the chromophore synthesis appratus produces chromophores of choice. The selected LBPs and the synthesized chromophores are chemically linked in an LBP-chromophore coupling apparatus, resulting in the LBP-tag. This LBP-tag is transferred into a loading apparatus which transfers the LBP-tab into predetermined cavities coated with the target molecule/ligand in the assay platform.

A transfer robot then moves the assay platform into the laser system in order to initiate the second part B. The samples are irradiated with the laser at the required wavelength in order to induce a modification by free radicals. The irradiated samples are then transferred by a sample transfer robot into an LBT-tag/ligand separating apparatus in order to isolate the ligand. The ligand is then cleaved in a protein cleavage apparatus. The peptide fragments are then analysed with a mass spectrometer in order to detect changes in mass, or in order to carry out a direct sequencing. Data from the mass spectra are then used for the analysis in the data base, which finally leads to identification of the amino acids or peptide fragments. All parts of the apparatus are connected to a central computer system for control and analysis.

12 The invention is explained in detail by means of the following examples.

Example 1 Epitope mapping of P-galactosidase by CALI-MS IgG antibodies against P-galactosidase were purchased from Sigma or Cappell and labelled with malachite green isothiocyanate or fluorescein isothiocyanate as described in PNAS 85, 5454-5458, 1988 or in PNAS, 95, 4293-4298, 1998. This entailed malachite green isothiocyanate or fluorescein isothiocyanate (from Molecular Probes, Eugene, OR) being added from a stock solution (20 mg/ml or 2 mg/ml) in DMSO stepwise until the concentration was 120 pg/ml to the antibodies in a concentration of 600 pg/ml in 500 ml of NaHCO 3 (pH After incubation with stirring at room temperature for one hour or incubation in ice for four hours, the solution was desalted on a column in 150 mM NaCl/50 mM Na 3

PO

4 pH 7.3 (for malachite green it is necessary to centrifuge the precipitate before changing the buffer) in order to separate the labelled protein from free marker.

The laser arrangement and the irradiation with the laser beam for carrying out CALI are essentially as described in Methods Cell Biol., 44, 715-732, 1994 or in PNAS, 95, 4293-4298, 1998. 20 pl of P-galactosidasecontaining sample (10 pg/ml) and dye-labelled antibodies against P-galactosidase (200 pg/ml) were placed in an ELISA plate (Nunc). The entire volume of the well was irradiated with a laser beam with different durations. The activity of these samples was measured as described in Meth. Enzymol., 5, 212-219, 1962. After the CALI, each sample was placed on a 1D PAGE gel in order to separate CALI-modified Pgalactosidase from the antibody.

For the MS, both the CALI-modified P-galactosidase and the untreated P-galactosidase were 13 cleaved in the gel with trypsin (Anal. Chem., 68, 850- 858, 1996; Protein Structure, A Practical Approach, 2 nd edition, editor Creighton, Oxford University Press: Oxford, UK, 1997, pp. 29-57). An aliquot of 0.3- 0.5 ul of cleaved protein was used for the MALDI analysis, and the remainder was used for MS/MS and/or de novo sequencing by electrospray MS.

To prepare samples for MALDI for the lowresolution CALI-MS, 0.3 to 0.5 pl of cleaved protein was added as a drop to a 0.3 pl drop of 5-10% formic acid on a vapor-deposited matrix/nitrocellulose surface on the MALDI sample loading apparatus (cf. Anal. Chem., 69, 4741-4750, 1997; Anal. Chem., 66, 3281-3287, 1994; J. Am. Soc. Mass Spectrom., 5, 955-958, 1994; Rapid Commun. Mass Spectrom., 10, 1371-1378, 1996; Anal.

Chem., 68, 850-858, 1996; Org. Mass Spectrom., 27, 156- 158, 1992). The MALDI analyses of the untreated and CALI-modified P-galactosidase were carried out in direct succession in order to ensure identical instrumental conditions for the two samples. The MALDI peptide mass mapping (Anal. Chem., 69, 4741-4750, 1997; Biochem. Soc. Transactions, 24, 893-896, 1996; Anal.

Chem., 69, 1706-1714, 1997) was previously carried out on the untreated P-galactosidase in order to identify peaks in the MALDI mass spectrum corresponding to P-galactosidase part-sequences. Peaks not belonging to P-galactosidase are, for example, matrix and trypsin autolysis peaks. It is additionally possible on the basis of a list of theoretical tryptic peptides to establish the actual extent of the tryptic cleavage.

Analysis of the MALDI mass spectrum of the untreated -p-galactosidase was followed by analysis of the spectrum of the CALI-modified sample, and differences in the spectra indicated the regions of the protein which were modified by CALI and are therefore associated with a function.

14 Tandem MS and de novo partial sequencing of selected peptides based on the previously obtained MALDI results were used for the high-resolution CALI-MS to determine the actual amino acid(s) which has/have been modified.

Based on the crystal structure of P-galactosidase, the amino acids responsible for substrate binding and catalysis were identified in this way. It emerged that the amino acids Met, His and Trp are involved in the catalysis. The masses of these amino acids result in peaks of increased mass, compared with the unmodified amino acids, in the mass spectrum.

A de novo sequencing confirmed the presence of the modified amino acids.

It is also possible to determine the type of modification through the incremental mass loss or gain.

As previously, the untreated and the CALI-modified P-galactosidase underwent successive MALDI analysis in order to ensure identical instrumental conditions for the two samples. The tandem MS or the MS/MS were carried out with a nanoelectrospray source as described, for example, in Nature, 379, 464-469, 1996 or in Biochem. Soc. Transactions, 24, 893-896, 1996.

The result obtained concerning the amino acids involved in the catalysis was the same as previously.

The beta-galactosidase activity was followed as a function of time after CALI. The activity was measured as assay units and is shown in relation to the activity of untreated beta-galactosidase.

Time (min) active beta-galactosidase CALI CALI 0 100% 100% 1 100% 100% 3 75% 100% 22% 100% 5% 100% 15 100% The P-galactosidase inactivated by CALI after minutes was used for further processing. The tryptic cleavage led to the fragments listed below in the sequence of decreasing mass as detected by MALDI. Pos.

refers to the number of the position of trypsin cleavage in the enzyme. The stated ranges correspond to the number of the amino acid of the enzyme starting from the N terminus. The underlined fragments showed a change in the molecular mass and correspond to those modified after the CALI treatment. The amino acids His, Trp and Cys in these fragments were modified, as was confirmed by resequencing, and are consistent with structural investigations from which it is evident that the modified amino acids are responsible for the enzyme activity.

Pos from to mol. wt.

43 523 523 146.2 775 775 146.2 68 855 855 146.2 2 15 15 174.2 57 756 756 174.2 37 448 449 289.3 811 812 332.4 35 441 443 374.4 26 335 337 402.4 940 943 435.5 69 856 858 438.5 28 354 357 444.5:His357 (new peptide mass observed 460.5) 39 475 477 445.6 22 290 293 487.6 42 519 522 532.6 36 444 447 532.7 12 180 184 545.5 46 559 562 562.6 -19 253 256 565.6 21 62 33 67 58 13 23 61 7 73 17 30 mass 34 77 64 6 4 51 76 41 24 11 49 71 3 63 59 78 18 1 284 783 39 206 427 849 757 185 294 775 54 553 911 232 382 observed 433 613 159 954 802 45 28 623 944 507 301 1015 168 601 883 16 788 762 963 239 1 289 787 44 211 432 854 761 191 300 782 60 558 918 238 389 978 440 622 167 962 810 53 38 631 953 518 311 1024 179 612 895 27 801 774 974 252 14 16 637.6 665.7 704.7 709.9 715.7 735.9 749.8 801.1 812.9 840.0 860.9 870.0 896.0 899.9 962.0: Cvs389 (new peptide 989.1 1064.2 1067.1 1083.2 1099.2 1100.2 1252.4 1265.4 1299.4 1341.5 1361.5 1367.6 1394.6 1400.6 1414.6 1428.5 1457.5 1496.7 1507.6 1547.8 1577.9 17 52 632 646 1742.9 72 896 910 1757.9 27 338 353 1776.2 14 192 205 1787.9 47 563 578 1891.2: Trp2568 (new peptide mass observed 1907.2) 9 142 158 1949.2 31 390 405 2005.2: His39l (new peptide mass observed 2021.2) 16 212 231 2265.5 32 406 426 2408.7: His4l8 (new peptide mass observed 2424.7) 312 334 2416.7 48 579 600 2447.4 70 859 882 2466.7 74 919 939 2500.8 54 679 700 2517.9 701 722 2522.8 38 450 474 2744.9 20 257 283 2848.1 29 358 381 2867.3: His359;His36O (new peptide mass observed =2883.3, 2899.3) 478 506 3071.3: Met5O2 (new peptide mass observed 3087.3, 3103.3) 44 524 552 3132.6 53 647 678 3424.9 56 723 755 3524.0 66 813 848 3776.2 79 975 1014 4325.7: Trp999 (new peptide mass observed 4341.7) 8 61 141 9178.1 It was shown that free radicals lead to oxidation. Thus, oxidation of the above amino acids led to a mass gain of 16 da for each modified amino acid, except for Met with a gain of up 32 da, depending on the oxidation state. The modified amino acids are 18 located near or in the active site of the enzyme, which explains the loss of function. This is consistent with independent studies which show the role of the amino acid Met502 in catalysis (Arch Biochem Biophys 283:342- 350, 1990; J. Biol. Chem. 265:5512-5518, 1990).

Figures 3 and 4 depict part of the MALDI spectrum of the tryptic beta-galactosidase peptides.

The peptide 478-506 with a molecular weight of 3071.3 is depicted. Met502 in this peptide is modified by CALI so that masses are obtained at 3087.3 da (sulphoxide) and/or 3103.3 da (sulphone). Only the alteration to the sulphone is shown for the illustration. All the expected peaks including trypsin autolysis peaks and matrix peaks have been omitted.

The MALDI spectrum shows only the peaks which represent beta-galactosidase peptides in addition to the peptide 478-506. The masses from the MALDI-MS spectra are monoisotopic molecular masses with the addition of one hydrogen atom The following table indicates the average and monoisotopic peptide ions MH+ in the spectrum: Peptide position Average MH+ Monoisotopic MH+ 257-283 2848.1 2847.42 358-381 2867.3 2866.38 478-506 3071.3 3070.43 478-506* 3103.3 3102.41 524-552 3132.6 3131.58 647-678 3424.9 3423.74 *CALI-modified peptide The MS/MS or de novo sequencing spectrum shows only the peaks of the peptide 478-506 (MH+ 3070.43 da). MS/MS was performed on the doubly charged peak MH2+ at m/z 1535.215 as parent ion. The monoisotopic MH+ peak is observed as in a MALDI spectrum. The following table gives a list of peptides 19 from the tryptic peptide 478-506. Tryptic peptides normally lead to a number of Y fragment ions which are C-terminal ions resulting from the cleavage of the amino acid bond between the carbonyl carbon and the amide nitrogen. Leucine and isoleucine have identical masses and cannot be distinguished. The Y ion on the C-terminal side of proline gives a weak signal or may be completely absent, but the following Y ion which contains the proline itself gives a prominent signal.

Monoisotopic MH+ peaks in the de novo sequencing of the tryptic peptide 478-506 from betagalactosidase (only partial list) Yn ion Sequence of the MH+ (before MH+ (after ion CALI) CALI) 1 R 175.12 175.12 2 AR 246.16 246.16 3 YAR 409.22 409.22 4 MYAR 540.26 572.25 PMYAR 637.31 669.30 6 CPMYAR 740.32 772.31 7 ICPMYAR 853.41 885.40 8 IICPMYAR 966.49 998.48 Example 2 Determination of the binding site of an RNA aptamer The binding site of an RNA aptamer on the UlsnRNP-A protein was investigated using CALI and mass spectrometery. RNA aptamers have a consensus sequence for binding to the RNA hairpin structure complexed with U1A. The 3D structure thereof has been determined.

Aptamers specific for the Ul-snRNP-A protein were labelled with fluorescein in a manner known per se. Then 20 pl of sample containing 10 pg/ml protein and a dye-labelled aptamer (200 pg/ml) were placed in the well of an ELISA plate. A further 20 ul of sample containing the U1A protein (10 pg/ml) and an aptamer 20 (200 pg/ml) without dye was placed in another well.

Irradiation was carried out with laser light of a wavelength of 488 nm.

Both the CALI-treated and the untreated sample was mixed with SDS sample buffer. Electrophoresis was carried out on a 12% SDS polyacrylamide gel. The gels were stained and the bands corresponding to the UlA proteins were cleaved with trypsin as described above.

A 0.3-0.5 p1 aliquot of the cleaved protein was used for the MALDI analysis, and the remainder was kept for the MS/MS analysis or the de novo sequencing with nanoES/MS.

To prepare for the MALDI analysis, 0.3-0.5 p1 of cleaved protein was added as a drop to a 0.3 p1 drop of 5-10% strength formic acid on a vapor-deposited matrix/nitrocellulose surface on the MALDI target (cf., for example, Anal. Chem., 1997, 69, 4741-4750). The MALDI analysis was then immediately carried out for both samples. The spectra obtained in this way for the treated and untreated samples were compared.

Differences in the spectra indicate the regions where the protein has been modified by CALI and thus define residues connected with the binding region. It is additionally possible to use a list of theoretical tryptic peptides for determining the actual extent of the tryptic cleavage.

The result already obtained on the basis of the X-ray structure concerning the binding of RNA ligands was confirmed (Nature 1994, 372:432-438) A high-resolution CALI-MS was then carried out.

This comprises a tandem MS and a partial de novo sequencing. This makes it possible to determine the amino acids which have been modified, and the nature and extent of the modification. The tandem MS (MS/MS) was carried out as described in NATURE, 1996, 379, 464- 469 and Biochem. Soc. Transactions, 1996, 24, 893-896.

21 The tryptic cleavage of the N-terminal U1A fragment led to the fragments listed below in the sequence of decreasing mass as detected by MALDI. Pos.

refers to the number of the position of trypsin cleavage in the enzyme. The stated ranges correspond to the number of the amino acid of the enzyme starting from the N terminus. The underlined fragments showed a change in the molecular mass and correspond to those modified after the CALI treatment. The amino acids His and Met in these fragments were modified, as was confirmed by resequencing, and are consistent with structural investigations from which it is evident that the modified amino acids are responsible for the RNA binding.

The amino acid sequence of N-terminal U1A fragment is known to be as follows: 1

MAVPETRPNHTIYINNLNEKIKKDELKKSLYAIFSQFGQILDILVSRSLKMRGQA

FVIFKEVSSATNALRSMQGFPFYDKPMRIQYAKTDSDIIAKMK

98 Tryptic cleavage of the N-terminal U1A fragment afforded the following fragments with the corresponding molecular weights: Pos from -to 3 23 23 28 28 2 21 21 14 97 98 8 51 52 321.4, 337.4) 7 48 50 4 24 27 12 84 88 13 89 96 mol. wt.

146.2 146.2 259.3 277.4 305.4: Met51 (new poetide mass observed 346.4 503.5 621.7 861.9 22 9 53 60 909.1 61 70 1047.1 11 71 83 1603.9: Met72;Met82 (new peptide mass observed 1619.9, 1635.9, 1651.9, 1667.9) 6 29 47 2170.5 1 1 20 2354.7: HislO (new peptide mass observed 2370.7) It was shown that free radicals lead to the oxidation. Thus, oxidation of the above amino acids led to a mass gain of 16 da for each modified amino acid, except for Met with a gain of up to 32 da depending on the oxidation state. The modified amino acids are located near the RNA binding site of the protein, which is thus consistent with the structural studies which show the binding site of the RNA.

Figures 5 and 6 depict part of the MALDI spectrum of the tryptic peptides of the N-terminal U1A fragment. The peptide 71-83 with a molecular weight of 1603.9 is depicted. Met72 and Met82 in this peptide are modified by CALI to result in masses at 1667.9 da (two sulphones). Only the alteration to the sulphone is shown for the illustration. Other expected peaks including trypsin autolysis peaks and matrix peaks have been omitted.

The average and monoisotopic peptide ions MH+ in the spectrum are indicated in the table below: Peptide position Average MH+ Monoisotopic MH+ 89-96 862.96 862.96 53-60 910.10 909.52 61-70 1048.14 1047.54 71-83 1604.89 1603.74 71-83* 1668.89 1667.72 29-47 2171.55 2170.19 -23 *CALI-modified peptide The MS/MS or de novo sequencing spectrum shows only the peaks of the peptide 71-83 (MH+ 1603.74 da).

MS/MS was performed on the doubly charged peak MH2+ at m/z 801.87 as parent ion. The monoisotopic MH+ peak is observed as in a MALDI spectrum. The following table gives a list of peptides from the tryptic peptide 71- 83. Tryptic peptides normally lead to a number of Y fragment ions which are C-terminal ions resulting from the cleavage of the amino acid bond between the carbonyl carbon and the amide nitrogen. Leucine and isoleucine have identical masses and cannot be distinguished. The Y ion on the C-terminal side of proline gives a weak signal or may be completely absent, but the following Y ion which contains the proline itself gives a prominent signal.

Monoisotopic MH+ peaks in the de novo sequencing of the tryptic peptide 71-83 of the N-terminal U1A fragment (only partial list) Yn ion Sequence of the ion 1 R 2 MR 3 PMR 4 KPMR

DKPMR

6 YDKPMR 7 FYDKPMR 8 PFYDKPMR 9 FPFYDKPMR

GFPFYDKPMR

11 QGFPFYDKPMR MH+ (before

CALI)

175.12 306.16 403.21 531.31 646.33 809.40 956.47 1053.52 1200.59 1257.61 1385.67 MH+ (after

CALI)

175.12 338.15 435.20 563.30 678.32 841.39 988.46 1085.51 1232.58 1289.60 1417.66 24 12 MQGFPFYDKPMR 1516.71 1580.69 13 SMQGFPFYDKPMR 1603.74 1667.72