Eur. J. Biochem. 206,691 -696 (1992) (c; FEBS 1992

C-terminal sequencing of protein A novel partial acid hydrolysis and analysis by mass spectrometry

Akira TSUGITA, Keiji TAKAMOTO, Masaharu KAMO and Hiromoto IWADATE Research Institute for Biosciences, Science University of Tokyo, Yamazaki, Japan

(Received January 23, 1992) - EJB 92 0088

Peptides or proteins were hydrolyzed by vapors of 90% pentafluoropropionic acid or hepta- fluorobutyric acid at 90°C for various time periods. The hydrolyzate mixtures analyzed by both fast- atom-bombardment and electrospray ionization mass spectrometry showed a series of C-terminal successive degradation molecular ions. The degradation reaction may be due to the selective formation of an oxazolone ring at the C-terminal amino acid, followed by hydrolytic removal of the C-terminal amino acid. The major side reactions were cleavages of the peptide bonds at the C side of the internal aspartic acid residue and the N side of serine residue.

Efficient and automatic methods of sequencing micro quantities of proteins are the most demanded techniques in biotechnology. Automated Edman degradation is a widely accepted N-terminal sequencing method. However, chemical C-terminal sequencing (Inglis et al., 1991) has not yet been well established nor widely accepted.

Previously, we reported that the addition of trifluoroacetic acid to hydrochloric acid, both in liquid (Tsugita and Scheffer, 1982) and vapour phases (Tsugita et al., 1987), increased the efliciency of hydrolysis of peptides, suggesting a type of cleav- age other than the conventional hydrochloric acid hydrolysis of peptide bonds. Our work on partial acid hydrolysis using high concentrations of strong organic acids have indicated the cleavage specificity of peptide bonds to be different from those of either Sanger's high-concentration-acid method (Sanger and Thompson, 1953), or the conventional dilute-acid hy- drolysis (Inglis, 1983). The present report summarizes the results of a novel high-concentration organic acid hydrolysis of peptides which essentially allows C-terminal protein sequencing and C-terminal sequencing from specific cleavage points (i.e. at the C side of aspartic acid and/or the N side of serine). Analysis of these data was directly carried out by mass spectroscopy (MS) of the hydrolyzate. Possible mechanisms of the C-terminal successive degradation and the specific cleavages are preliminarily discussed.

MATERIALS AND METHODS

Trifluoroacetic acid and F,C,COOH were the products of Sigma Chemical CO. (USA); F7C3COOH and dithio- theritol were obtained from Wako Pure Chemical Industry (Osaka, Japan). 4-Vinylpyridine was obtained from Tokyo Kasei Kogyo Co. (Tokyo, Japan). A peptide, LWMRFA, and human des-Aspl-angiotensin I, RVYIHPFHL, were pur- chased from Serva Feinbiochem (Heidelberg, FRG). Human parathyroid hormone, fragment 69 - 84, EADKADVNVLT-

Correspondence to A. Tsugita, Research Institute for Bioscicnces,

Abbreviations. FAB, fast-atom-bombardment ; ESI, electrospray Sciencc University of Tokyo, Yamazaki, Noda 278 Japan

ionization; m/z, molecular ion; MS, mass spectrometry.

KAKSE, and Xenopus luevis magainin 1, GIGKFLHSAGK- FGKAFVGEIMKS, were the synthesized products of the Peptide Institute Inc. (Minoh, Japan). Human glucagon was obtained from Sigma Chemical Co. (USA) and hen egg white lysozyme was from Seikagaku Kogyo Co. (Tokyo, Japan).

Acid degradation of peptide or protein

A small glass tube (6 mm x 40 mm), containing dried peptide or protein, was placed in a large glass tube (13 mm x 100 mm), which contained 100 pI aqueous organic acid of a given concentration, and flame sealed under vacuum. The tube was placed in an oil bath at 90°C for various time periods. After the reaction, the tube was opened and the small tube was placed in a vacuum desiccator to remove traces of acid. Dithiotheritol (1 - 5%, mass/vol.) was added to the hydrolysis acid in the large tube to prevent oxidation or oxi- dative modification. Alternatively, the small sample tube was placed in a middle-sized glass tube (10 mm x 30 mm) contain- ing 50 mg solid dithiotheritol, and the tubes were placed in the large tube containing the acid and sealed under vacuum.

Analysis by MS

The dried hydrolyzate was disolved in 5% acetic acid and mixed with glycerol and thioglycerol(1: 1, by vol.) as matrix and put on the target. Fast-atom-bombardment mass spec- troscopy (FAB-MS) was carried out with a HX-110 mass spectrometer (JEOL, Japan), equipped with a DA 5000 data system, employing an accelerating voltage of 10 kV and xenon as an ionizing gas.

Electrospray ionization (ES1)-MS was carried out with a SX-102 mass spectrometer with an electrospray ion source (JEOL, Prototype, USA) at a flow rate of 0.5 pl/min, em- ploying an accelerating voltage of 10 kV and nitrogen as a drying gas. The sample was dissolved in 2% acetic acid/50%0 aqueous methanol.

Pyridylethylation

The protein in a small tube (6 mm x 40 mm) was placed in a large tube (13 mm x 100 mm) which contained a mixture

692

BE-

68-

48-

20-

1

a , . > 40-

m a,

1-6 .- c . 1-5 - .

cc 20:

1

1-5

40 3 26-

J 881

- \

1-6

1-3 1-1

1

1-6

I

"1

\ I \ 1-3 14 1-2

1 1 1-5

>

I 1-6

m/z(MH+)

Fig. 1. FAB-MS of the hydrolyzates of a hexapeptide, LWMRFA. The peptide (200 pmol) was hydrolyzed with 90% (by vol.) of various organic acids containing 5% dithiotheritol in the vapor phase at 90°C. (a) Before hydrolysis; (b) trifluoroacetic acid for 4 h; (c) F5CzCOOH for 4 h; (d) F7C3COOH for 4 h; (e) F5C2COOH for 24 h; (0 F7C3COOH for 24 h.

of 100 pl pyridine, 20 p1 4-vinylpyridine, 20 p1 tributyl- phosphine and 100 pl water. The large tube was flame sealed under vacuum and incubated at 60°C for 2 h. After reaction, a small tube was taken out and dried in a vacuum desiccator for 1 h. 5 pl water was added to a small tube and dried in the vacuum desiccator for 1 h to further remove trace of reagents (Yamada et al., 1991).

RESULTS AND DISCUSSION A novel organic acid successive degradation

Trifluoroacetic acid, F5C2COOH and F7C3COOH were chosen as test organic acids because of their volatility and low pK,. To avoid contamination from these acids, the vapor phase of the acids were used for the hydrolysis reaction as described in Materials and Methods.

We have noticed different modes of acid hydrolysis with high and low concentrations of organic acids in FAB-MS molecular ions. Low acid concentrations resulted in non- specific cleavages, but high acid concentrations gave C-ter- minal successive degradation including specific peptide-bond cleavages.

When high concentrations of organic acids were used, oxidation and oxidative modifications of side chains in the peptide were observed. Addition of a reducing reagent, dithiotheritol, was found effective to avoid the oxidative reac- tions of methionine, tryptophan and tyrosine residues.

Fig. 1 shows a comparison of the molecular ions produced by trifluoroacetic acid, FJ2COOH and F7C3COOH hy- drolysis of a hexapeptide LWMRFA in FAB-MS. In Fig. lb, trifluoroacetic acid at 90°C for 4 h shows several modified molecular ions, indicated by higher-mass molecular ions than that of the original peptide residues 1-6, (m/z of 823). The

693

I 00

90

80

70

a, 60 2

2

m -0 5 50 a, .- > 40 m a,

c - oc

30

20

I0

0

I (1-22) +4

300 350 400 450 500 550 600 650 700 750 800 850 900 m/Z =(M+nH?/n

-. _. 1 5 10 15 ZU 23 Gly-Ile-Gly-Lys-Phe-Leu-His-Ser-Ala-Gly-Ly~-Phe-Gly-Lys-Ala-Phe-Val-Gly-Qlu-Ile-Met-LyE-Ser I

(1-7) +2 I

+2 I

I

I t I (8-18)

t i (1-23)*y I I (1-22)+3 I I (1-21)+3 I I (1-20)+3

I I (1-18) +3 I I (1-17) +3 I I (1-16) +3 I I (1-15)+3

+3 I I (1-19)+3

I I (8-23) +3 t I (8-22) '3 t I (8-21) +3

I i (1-23)-* I (1-22) +4 +4 I

Fig. 2. EN-MS of the hydrolysate of a peptide, GIGKFLHSAGKFGKAFVGEIMKS. The peptide (125 pmol) was hydrolyzed with 90% F5C2COOH in the vapor phase in the presence of solid dithiotheritol(50 mg) for 2 h. The dried hydrolyzate was disolved in 25 pl2% acetic acid in 50% aqueous methanol solution.

successive C-terminal degradation ions were observed as 1 - 5 (m/z of 752), 1-4 (m/z of 605), 1-3 (m/z of 449) from the original peptide in addition to the modified molecules. Nonspecific cleavages were also observed, such as residues 4-6,3-5 or 3-6.

On the contrary, both FSC2COOH (Fig. lc) and F&COOH (Fig. Id) hydrolysis for 4 h at 90°C showed more clear-cut C-terminal degradation molecular ions. Fig. 1 e and

f show the successive C-terminal degradation molecular ions for both F,C,COOH and F7C3COOH after 24 h hydrolysis, respectively. No preference was shown between F5C2COOH and F7C3COOH at 4 h hydrolysis, but F5C2COOH gave more clear C-terminal degradation ions than F,C,COOH for the longer 24 h reaction time, leading to the use of F5CzCOOH for the following experiments. Of the various acid concentrations tested (50 - 98%) for several peptides, around 90% were

694

I Y Y -

100

80

6 0

40

20

z 1-9

\ 1-9

40-

1-8

m/z (MH+) Fig. 3. FAB-MS of the hydrolyzates of a nonapeptide, RVYIHPFHL. The peptide (ZOO pmol) was hydrolyzed with 90% F7C3COOH containing 5 % dithiotheritol in the vapor phase at 90°C. (a) Before hydrolysis; (b) F7C3COOH for 4 h ; (c) F7C3COOH for 24 h. The conditions for MS are same as in Fig. 1 .

found out to be optimal for this specific C-terminal degradati- on.

Fig.2 shows an ESI-MS of the 90% F5CzCOOH hy- drolyzate of a peptide, magainin 1, consisting of 23 amino acids. The multi-charged degradation molecular ions of both the intact peptide and a fragment specifically cleaved at the eighth serine residue, (8 - 23) were observed. The + 5 molec- ular ion was the maximal chargeable ion corresponding to the N-terminal amino group and four internal lysine residues in the peptide. The absence of further degraded ions in the series of f 4 and + 5 molecular ions was possibly due to the charge at Lys22. The longest successive C-terminal degradation ions, of eight residues, were detected in the + 3 molecular-ion series.

It should be noted that the amino acid analysis of the hydrolyzate mixtures detected the amino acids which corre- spond to the degradation shown in the MS data and further more, the data were able to distinguish the amino acids of identical masses such as isoleucine and leucine and lysine and glutamine (data not shown).

Susceptible and resistant peptide bonds

The peptide bonds at both sides of aspartic acid have been known to be labile in dilute acid (Inglis, 1983). The bond at the N side of serine was found to be labile for a concentrated hydrochloric acid hydrolysis (Sanger and Thompson, 1953).

69 5

16-29

600 800 1000 1200 1400 1600 1E a0 1 5 10 15 20 25 29 m/z H S Q Q T P T S D Y S K Y L D S R R A Q D F V Q W L U I T

1-9 (MHt979) I I 16-29 (Mw1753) 1-8 (Mw865) I I 16-28 (MH+1652) 1-7 (Mp778) 1 I 16-27 ( M P 1538) - 1-6 (Mp677) 16-21 (Mp733) - 1-5 (Mv530) 16-20 (MP677)

H 16-18 (Mw530) - 16-19 (MH'618)

120-126 1-14 1-15

d @la0 ' ' ' l 0 b 0 ' ' ' 1200 ' ' ' 1400 ' ' ' . 1600 ' l a 0 0 2000 2 a0

mlz 1 5 10 15 20 25 K V P G R C E L A A A M K R H G L D N Y R G ~ S L G ~ . . . .

I I 1-18 (MH'2109) I I 1-17 (MH+1993) 1 I 1-16 (MH+188O) I i 1-15 (MH:1823) I I 1-14 (MH 1686)

110 la o 129 .... I G U N A W V A W R I R C K G T D V Q A W I R G C R L I 1120-129 (MW1308) I I 120-128 (Mp1194) - 120-126 ( M P 830)

120-127 (M'l1038)

Fig. 4. FAB-MS of the hydrolyzates of proteins. (a) Human glucagon (100 pmol) was hydrolyzed with 90% F,C,COOH containing 5% dithiotheritol in the vapor phase at 90°C for 4 h. The cleavage and molecular ions are schematically illustrated. The conditions of MS were same as in Fig. 1. (b) Pyridylethylated hen egg white lysozyme (100 pmol) was hydrolyzed with 90% FsCzCOOH in the vapor phase and with solid dithiotheritol at 90°C for 4 h. The conditions of MS were same as in Fig. I.

696

Under the present conditions, the peptide bond only at the C side of the internal aspartic acid and the peptide bond at the N side of serine were selectively cleaved. In the experiments performed on various test peptides, the N side of threonine and both sides of glycine were observed to be occasionally cleaved (data not shown).

The peptide bond of Pro-Xaa is partially resistant to the present successive degradation. A nonapeptide, RVYIHP- FHL, was hydrolyzed for 4 h with 90% F7C3COOH at 90°C. Analysis of the hydrolyzate indicated resistance to hydrolysis at the peptide bond between residues 6 and 7. Prolonged hydrolysis showed further successive degradation ac- companying a small extent of nonspecific cleavage (Fig. 3 b and c).

It should be noted that under the present conditions little cleavage of the amide groups of acidic amino acids, except for the C-terminal a-amide bond, was observed.

Reaction mechanism

The successive degradation is due to the formation of an oxazolone five-membered ring at the C-terminal amino acid (Matsuo et al., 1965, Inglis et al., 1991), followed by hydrolysis of the C-terminal amino acid in the oxazolone. This mecha- nism was indicated by the observations of oxazolone molec- ular ions (- 18 from the respective peptide) in the MS, as shown in Fig. 4. The two internal cleavage reactions are at the C side of aspartic acid and the N side of serine. The internal aspartic acid forms another type five-membered ring, followed by hydrolysis of the succinylimine group (Inglis, 1983). Serine undergoes an N+O peptidyl shift for the cleavage of the peptide bond (Sanger and Thompson, 1953). The slow down of the prolyl peptide-bond cleavage may be explained by a steric hindrance for the formation of oxazolone ring. The establishment of the above reaction mechanism needs further experiments.

Application to protein sequencing Application of this method to proteins is still preliminary.

Glucagon, a peptide consisting of 29 amino acids (200 pmol), was hydrolyzed with 90% F,C&OOH at 90°C for 1 h. The result gave three series of C-terminal successive degradation molecular ions; one ion series is from the C-terminal peptide fragments, 16-29, 16-28 and 16-27, where the cleaved Deatide bond between residues 15 and 16 is Asa-Ser and the

other is that of the N-terminal peptide fragments of residues, 1 - 9, 1 - 8, 1 - 7, 1 - 6 and 1 - 5, where the cleaved peptide bond between residues 9 and 10 is Asp-Tyr. Further hydrolysis gave an additional ion series of peptides 16 - 21,16 - 20,16 - 19 and 16- 18 (Fig. 4a), where the cleaved peptide bond be- tween residues 21 and 22 is Asp-Phe.

The FAB-MS molecular ions of the hydrolyzates of lysozyme are shown in Fig.4b. The protein was pyridyl- ethylated as described in Materials and Methods in a vapor phase (Yamada et al., 1991) to protect the cysteine residues. Two series of molecular ions were observed, the C-terminal sequence of residues, 120- 129, 120- 128, 120- 127 and 120- 126, and the C-terminal degradation of an N-terminal fragment composed of residues, 1 - 18, 1 - 17, 1 - 16, 1 - 15 and 1 - 14. The direct amino acid analysis of the hydrolyzate mixtures of both proteins were carried out and the data con- firmed the above MS observations.

Analysis by FAB-MS has limitations, such as molecular size or selectivity of ionization of peptides. The present specific peptide cleavage reactions for aspartic acid and serine may favor this type of FAB-MS. ESI-MS is also useful to analyze the present hydrolyzate. Programs to calculate the successive degradation molecular ions have been developed and will become available for handling the complexed multi-charged molecular ions for ESI-MS. In conclusion, the present method of partial hydrolysis, including C-terminal successive degra- dation, may permit useful C-terminal analysis of medium- sized peptides.

REFERENCES 1. Inglis, A. S. (1983) in Methods Eiriymol. 91,324-332. 2. Inglis, A. S., Moritz, R. L., Begg, G. S., Reid, G. E., Simpson, R.

J., Graffunder, H., Matschull, L. & Wittmann-Liebold (1991) in Methods in protein sequence analysis (Joernvall, H., Hoeoeg, J.- 0. & Gustavsson, A.-M., eds) pp. 23 - 34, Birkhaeuser Verlag, Basel.

3. Matsuo, H., Fujimoto, Y. & Tatsuno T. (1965) Tetrahedron Lett.

4. Sanger, F. & Thompson, E. 0. P. (1953) Biochem. J. 53,353-366. 5. Tsugita, A. & Schemer, J . J . (1982) Eur. 1. Biochem. 124, 585-

6. Tsugita, A., Uchida, T., Mewes, H. W. & Ataka, T. (1987) J .

I . Yamada, H., Moriya, H. & Tsugita, A. (1991) Anal. Biochem. 198,

39,3465 - 3461.

588.

Biochem. 102,243 - 246.