ANALYTICAL
BIOCHEMISTRY
Selective
l%i,224-229
(1983)
isolation of Tryptophan-Containing by Hydrophobicity Modulation
Peptides
TATSURU SASAGAWA, KOITI TITANI,
AND KENNETH A. WALSH Department of Biochemistry and Howard Hughes Medical Institute Laboratory. University of Washington, Seattle, Washington 98195 Received April 4, 1983
Tryptophan-containing peptides am selectivelyisolated from complex digests by taking advantage of changes in hydrophobicity and chromatographic mobility induced by reaction with onitrophenylsulfenyl chloride. The peptides are first located in crude fractions by monitoring the fluorescence during high-performance liquid chromatography and then chemically modified to facilitate their separation from contaminants during subsequent rechromatography. KEY WORDS: HPLC; nitrophenylsulfenyl chloride; tryptophan peptide.
Tryptophanand methionine-containing peptides are of particular importance in the sequence analyses of proteins, whether directly by amino acid- or indirectly by DNA-sequencing procedures. Direct analysis often demands isolation from enzymatic digests of small tryptophan- or methionine-containing peptides required to overlap fragments separately derived by chemical cleavage at these residues (l-3). For the indirect approach, the design of oligonucleotide probes requires knowledge of at least a pentapeptide sequence, preferably including one or more methionyl or tryptophanyl residues, which alone have unique codons (4). We have recently reported a procedure for the selective isolation of methionyl peptides by hydrophobicity modulation (5). For the isolation of tryptophanyl peptides, two methods have been so far reported. The method of Rubinstein et al. (6) is based on the specific and reversible binding of tryptophanyl residues to an insoluble polymer, whereas that of Wilchek and Miron (7) is based on the specific atlinity of 2,4dinitrophenylsulfenylated tryptophanyl residues to antidinitrophenyl antibody columns. The procedure proposed herein is similar in principle to that previously designed for methionyl peptides (5), which involved (i) the 0003-2697183 $3.00 Copyright 0 1983 by Academic Press. Inc. All rights of reproduction in any form resewed.
initial separation of a digest of [ “C]S-methyl protein into relatively crude fractions by HPLC, (ii) treatment of the radiolabeled fractions with a residue-specific demethylating agent, and (iii) rechromatography under identical conditions. As pointed out earlier (5), these procedures are analogous to the “diagonal paper electrophoresis” methods reviewed by Hartley (8). However, our procedures are based on changes in hydrophobicity and HPLC mobility, whereas the other procedures were based on changes in charge and electrophoretic mobility. In our case, the protein is first digested (e.g., with trypsin) and subjected to preliminary reverse-phase HPLC. Tryptophan-containing peptides are located in crude fractions on the basis of their characteristic fluorescence at 348 nm (9). Each such fraction is then treated with o-nitrophenylsulfenyl chloride (10) which specifically modifies tryptophanyl residues, selectively increasing the hydrophobicity of tryptophan-containing peptides, but not of their contaminants, during rechromatography. MATERIALS
AND METHODS
Glycyl-L-tryptophan was purchased from Schwarz/Mann. Proteins and reagents were purchased from the indicated sources as fol224
TRYPTOPHANYL
lows: hen lysozyme and horse cytochrome c (Sigma), bovine TPCK trypsin (Worthington), o-nitrophenylsulfenyl chloride and trifluoroacetic acid (Pierce), acetonitrile (Burdick and Jackson), and acetic acid (J. T. Baker). Lysozyme was reduced and S-carboxymethylated according to the method of Crestfield et al. (11). Tryptophan-containing peptides were onitrophenylsulfenylated at the 2-position of their indole rings according to the method of Scoffone et al. (10). Each tryptophan-containing peptide (30 nmol) in 200 ~1 of glacial acetic acid was treated with 100 nmol of o-nitrophenylsulfenyl chloride in the dark at room temperature for 18 h. Peptides were separated by reverse-phase HPLC using a Varian liquid chromatograph, Model 5000, and a Waters PBondapak Cl 8 column. The mobile phase was 0.1% trifluoroacetic acid and the modifier was acetonitrile containing 0.07% trifluoroacetic acid ( 12). The concentration of acetonitrile was increased linearly (2%/min) during 30 min at a flow rate of 2 ml/min. Elution was monitored by absorbance at 2 10 nm and by fluorescence at 348 nm (with excitation at 287 nm, 1 nmol is easily detected) using a PerkinElmer MDF-44A fluorescence spectrophotometer. Amino acid analyses were performed with a Dionex amino acid analyzer (Model D-500). Automated sequence analysis was performed with a Beckman sequencer as described previously ( 13). NPS’-tryptophan was prepared from 1 mg of L-tryptophan by the method of Scoffone et al. (lo), and separated from excess reagent by HPLC as above. The phenylthiohydantoin was prepared by treating 500 c(g (in I ml of 10% pyridine) with 4 ~1 of phenylisothiocyanate at 40°C for 30 min. The phenylthiocarbamyl derivative was extracted in 1 ml of benzene. After the solvent was removed with a stream of Nz, the residue was treated with 1 ml of 1 N HCl for 20 min at 85°C. The product was separated by HPLC as above and lyophilized.
’ Abbreviation
used: NPS, o-nitrophenylsulfenyl.
PEPTIDE
225
ISOLATION
This twicederivatized product (the phenylthiohydantoin of NPS-tryptophan) was identified by mass spectrometry using direct-insertion probe analysis in a VG 7070H instrument (VG Analytical, Manchester) operated in the electron impact-ionization mode. A major fragment at m/z 283 corresponded to the 2-(o-nitrophenylsulfenyl)-3-methylene indole ion. In addition, in contrast to the simple phenylthiohydantoin of tryptophan (where molecular ions were observed at 3 19 and 32 1), the phenylthiohydantoin of NPS-tryptophan displayed molecular ions larger by 153 (the additional mass of o-nitrophenylsulfenylation). RESULTS
After 5 h of treatment of Gly-Trp (30 nmol in 200 rl of acetic acid) with 100 nmol of onitrophenylsulfenyl chloride, an aliquot of the reaction mixture was injected into the HPLC system (Fig. 1). The Gly-Trp (peak 3) was replaced by a new peak 4, corresponding to NPS-Gly-Trp, with a later retention time in accord with the expected increase in hydrophobicity. This decrease in mobility appeared
0
10 TIME
20 (MI
N)
FIG. 1. Mobility of Gly-Trp and its NPS- form. 30 nmol of Gly-Trp was treated with 100 nmol of o-nitrophenylsulfenyl chloride for 5 h at room temperature and then subjected to HPLC separation. Peak 1 is acetic acid, peak 2 is an internal standard (Gly-Phe), peaks 3 and 4 are Gly-Trp and its NPS- form, respectively, and peak 5 is reagent.
226
SASAGAWA,
TITANI,
to be sufficient to serve as the basis for a tryptophan-specific separation scheme. To test this procedure, two proteins, cytochrome c (30 nmol) and S-carboxymethylated lysozyme (60 nmol), were separately digested with trypsin (100:2 (w/w) in 0.1 M NH4HC03 for 6 h at 37°C) and fractionated by HPLC (Figs. 2A and 3). Seven fluorescent fractions were collected and lyophilized. These fractions were then each treated with o-nitrophenylsulfenyl chloride, selectively increasing the hydrophobicity and thus the retention time of each component tryptophanyl peptide. Subsequent chromatography yielded in each case a tryptophan-containing peptide. For example, in Fig. 2 the single tryptophan residue of horse cytochrome c was found in peptides corresponding to the sequences GITWK (peak I-2) and GITW (peak X1-2). In each case contaminants with the original retention times (peaks I- 1 and II- 1) were separated and excess reagent eluted at a much later retention time. The compositions (Table 1) indicated that the tryptic peptide surrounding tryptophan 59 of cytochrome c was selectively isolated, although chymotryptic contamination of trypsin yielded an additional peptide lacking the terminal lysine.
AND WALSH
I
0.2 AU
-NV’
FIG. 3. Separation on a pBondapak Cl 8 column of tryptophan-containing peptides from 60 nmol of a tryptic digest of S-carboxymethylated lysozyme.
Analogous experiments with S-carboxymethylated lysozyme gave similar results. The six tryptophanyl residues were distributed among five separate peak fractions (III-VII, Fig. 3). Fraction VI yielded, after o-nitrophenylsulfenylation and rechromatography
A I
TIM 0-W)
C QZAU
II.
II-1 &-jl 20
; 0
10
20
TIME ;RIN) FIG. 2. Isolation of tryptophan-containing peptides from cytochrome c. (A) 30 nmol of tryptic digest of cytochrome c was fractionated on a pBondapak Cl8 column. (B,C) Fluorescent fractions I and II were collected and subjected to o-nitrophenylsulfenylation (5 h, using 300 nmol of NPS chloride) and rechromatographed under identical conditions. Peak I-2 and II-2 were identified by composition as GITWK and GITW, respectively (Table 1). The original peak positions are indicated by arrows.
TRYPTOPHANYL
PEPTIDE TABLE
227
ISOLATION
1
AMINOACIDCOMFQSIT~ON OFFTJRIFIEDPEFTIDES Peptides Amino acids Cmc (C) Asx (B) Thr (T) k 6) ax (Z) GUY (G) Ala (A) Val (V) Met(M) Ile (I) Lea (L) TY~ W) LYS W) Arg CR) Trp (WI Residues Yield (W)
I-2
II-2
III- 1
0.6 (I) 1.8 (2) I.1 (1)
III-2
0.4 (I) I .8 (2)
1.0 (I)
1.0 (I,
1.0 (I)
0.8 (I)
0.9 (I)
1.0 (I)
IV-2
IV-I
1.1 (I)
4.0 2.0 1.9 I.1 2.1
(4) (2) (2) (1) (2)
1.9 (2) 0.9 (I) 1.0 (I)
ND (1) 56-60 51
ND (1) 56-59 31
1.0 (I)
ND 62-68 23
(2)
0.6 (1) 1.1 (1) 1.0 (I)
1.0 (I) 1.0 (I)
1.0 I.1 1.0 I.1
1.0 1.0 1.0 1.0
(I) (1) (I) (I)
0.8 (I)
(I) (I) (I) (I)
0.8 (1)
VI- I
V-I
0.5 (I) 1.0 (I)
2.x (3,
1.0 (1)
I.1 (I,
1.0 (I)
2.1 2.0 1.3 0.4 0.4
2.0 (2) 1.8 (2) 0.8 (I)
2.0 2.0 1.4 0.7 0.4
(2) (2) (2) (I) (I)
0.9 (I) 1.1 (I)
1.0 (1)
0.6 (I) 1.0 (I)
ND (0)
ND(l)
ND(l)
ND
46-6 91
115-125 35
117-125 64
97-l 23
I
VII-I
2.9 (3)
I.1 (I)
1.0 (I) 1.0 (1) ND (2) 62-68 30
IV-3
6-4 12
I.0 (I, 1.0 (I) 0.9 (I)
ND(l) 22-23 86
I.1 (I) ND (2) W-112 39
Nore. Residues/molecule by amino acid analysis or (in parentheses) from the sequence (16, 17). The purification profiles are illustrated in Figs. 2-4. Peptides l-2 and II-2 are isolated from a tryptic digest of cytochrome c. and peptides III-1 through VII-I are from tryptic digest of S-carboxymethylated lysozyme.
(Fig. 4E), a single tryptophanyl peptide (Table 1) from an uncontaminated pool. Fraction VII also yielded a single peptide which happened to coelute with excess reagent and to contain two tryptophanyl residues. Fraction V contained a peptide (V- 1) overlapping VII1 but resulting from tryptic cleavage between the lysyl residues in a Lys-Lys sequence rather than after the second lysine. Fraction IV, after o-nitrophenylsulfenylation and rechromatography (Fig, 4C), yielded two tryptophanyl peptides (IV-2 and IV-3), leaving a contaminant peptide IV- 1, lacking tryptophan (residues 46-61) at the original chromatographic elution position. Amino acid analyses identified the two tryptophanyl peptides as residues 115- 125 and 117-125 where the longer included an uncleaved S-carboxymethyl-CysLys at the amino terminus of the shorter peptide. Fraction III was the most complex. Treatment with reagent and rechromatography yielded two new peptides (III-1 and III2 in Fig. 4A) with identical amino acid compositions after acid hydrolysis. Further reaction of fraction III-1 (for 24 h) with o-nitro-
(2) (2) (21 (II (I )
a
phenylsulfenyl chloride resulted in an increase of fraction III-2 at the expense of fraction III1 (Fig. 4B). Since each of these peptides must contain two tryptophanyl residues (WWCNDGR), it appears that the original modification of peptide III-1 was simply incomplete. Peptide III-2 was subjected to Edman degradation and aliquots of the products were compared on HPLC (Fig. 5) with the phenylthiohydantoins of tryptophan and NPStryptophan (track 8) and with other standards (not shown). The phenylthiohydantoin of NPS-tryptophan was observed as the product in both cycles 1 and 2, in accord with the known sequence of residues 62-68 in lysozyme. DISCUSSION
The described method offers a rapid and a simple procedure for selective isolation of tryptophan-containing peptides from complex mixtures. The method is analogous to that used for isolation of methionine-containing
228
SASAGAWA,
i 0
TITANI,
AND WALSH
_II1 10
20
F 0
TIME(MIN)
4. Rechromatography of the fluorescent Ii-actions from Fig. 3 after reaction with o-nitrophenylsulfenyl chloride (200 nmol). Peaks were identified as follows: III-1 and 111-2,WWCNDGR (in (A) the modification of both tryptophan residues was incomplete, whereas in (B) all of peptide III- 1 was converted to 111-2);IV1, NTDGSTDYGILQINSR (lacks tryptophan and had original mobility); IV-2, CKGTDVQAWIR, IV-3, GTDVQAWIR; V-l, KIVSDGDGMNAWVAWR; VI- I, GYSLGNWVCAAK, VII- I, IVSDGDGMNAWVAWR. The original peak positions are indicated by arrows. RG.
peptides and takes advantage of changes in polarity induced by residue-specific modification. Several tryptophan-specific modification methods have been proposed by various laboratories. We tested two other tryptophanspecific reagents, 2-hydroxy-5-nitrobenzyl bromide (14) and formic acid-HCl ( 15), but neither gave satisfactory results in our system. 2-Hydroxy-5nitrobenzyl bromide yielded more than two products, whereas formic acidHCl caused some nonspecific cleavage of peptide bonds. Thus, o-nitrophenylsulfenyl chloride became our reagent of choice. The proposed procedure has several advantages over previously reported approaches (6,7). The primary difference is that in the present method, individual tryptophan-con-
taming peptides are separated from each other. As shown in Fig. 4 and Table 1, the six tryp tophan residues of lysozyme were distributed among seven different peptides in a tryptic digest. Incomplete cleavage of a Lys-Lys bond and an S-carboxymethyl-Cys-Lys-Gly sequence added to the expected number of pep tides. It should be noted that the two tryptophanyl peptides (IV-2 and IV-3) which eluted in the same fraction during the first chromatography were separable during recbromatography after modification as shown in Fig. 4C. Incomplete modification of the sterically hindered TrpTrp sequence accounted for the apparent redundancy of peptides III-1 and 111-2. Since NPS-chloride can react with cysteine
TRYPTOPHANYL
PEPTIDE
ISOLATION
229
ACKNOWLEDGMENTS The authors are grateful to Mr. Roger D. Wade for performing the amino acid analyses. This work was supported in part by NIH Grants GM 15731 and HL 29595. Note added in proof: Our attention has been drawn to related work by F. M. Veronese, A. Fontana, and E. Boccu (Acta vitaminol. enzymol. (Milan) 29, 243-247. 1975) where a carboxy-substituted NPS-reagent was used as the basis for a diagonal electrophoretic method of isolation of a tryptophan derivative.
E c 0 a?
REFERENCES
0
10
20 TIME(M1
jo
N)
FIG. 5 Identification of phenylthiohydantoins released by Edman degradation of peptide 111-2. HPLC tracings 1-7 correspond to the number of cycle of the Edman degradation. Tracing 8 is a mixture of the phenylthiohydantoins of Trp and NPS-Trp. The HPLC conditions were as described for peptide separations under Materials and Methods. These data verify the sequence WWCNDGR.
1. Gross, E. in Methods in Enzymology (Hirs, C. H. W., ed.), Vol. 11, pp. 238-255, Academic Press, New York. 2. Omenn, G. S., Fontana, A., and Anfinsen, C. B. (1970) J. Bioi. Chem. 245, 1895-1902. 3. Mahoney, W. C., and Hem&son, M. A. (1979) Biochemistry 18, 3810-3814. 4. Suggs, S. V., Wallace, R. B., Hirose, T.. Kawashima, E. H., and Itakura, K. (I 981) Proc. Natl. Acad. Sci. USA 78,66 13-66 17. 5. Sasagawa, T., Titani, K., and Walsh, K. ( 1983) Anal. Biochem. 128, 371-376. 6. Rubinstein, M., Shechter, Y., and Patchomik, A. (1976) Biorhem. Biophys. Rex Commun. 70,12571263.
(18) as well as tryptophan residues, it is important to prevent cysteine side reactions by prior alkylation as was done with lysozyme in the present study. The Scarboxymethyl cysteine in peptide III-2 was obtained as the normal phylthiohydantoin (Fig. 5, track 3) in spite of the NPS-chloride treatment during the isolation of the peptide. Our experience with lysozyme is in accord with the claims of Scoffone et al. ( 10) that only tryptophan residues are reactive with NPS-chloride under their experimental conditions. A combination of the present method for tryptophan-containing peptides and the method previously described for methionine-containing peptides (5) should facilitate sequence analyses of proteins, either directly by amino acid- or indirectly by DNA-sequencing procedures,
7. Wilchek, M., and Miron, T. (1972) B&him. Bioph,vs. Acta 278, l-7. 8. Hartley, B. S. (1970) Biochem. J. 119, 805-822. 9. Udenfiiend, S. (1962) Fluorescence Assay in Biology and Medicine, p. 129, Academic Press, New York. 10. SCOffOne,E., Fontana, A., and Rocchi. R. (1968) Bi+ chemistry 7,97 l-979. 1 I. Crestfield, A. M., Moore, S., and Stein, W. H. ( 1963) 12. Dunlop III, C. E., Gentleman, S., and Lowney, I. (1978) J. Chromatog. 160, 191-198. 13. Sasagawa, T., Ericsson, L. H., Walsh, K. A., Schreiber, W. E., Fischer, E. H., and Titani, K. ( 1982) Biochemistry 21, 2565-2569. 14. Horton, H. R. and Koshland, D. E.. Jr. (1965) J. Amer. Chem. Sot. 87, 1126-1132. 15. Previero, A., Antonia, M., Previero, C., and C&adore, J. C. (1967) Biochim. Biophys. Acta 147,453-46 1. 16. Margoliash, E., Smith, E. L., Kriel, G., and Tuppy. H. (1961) Nature (London) 192, 1125-I 127. 17. Canfield, R. (1963) J. Biol. Chem. 238, 2698-2707. 18. Fontana, A., Scoffone, E., and Benassi, C. A. (1968) Biochemistry 7, 980-986.