Nucleophile labeling of cysteine and serine protease substrates.

Dipeptides containing fluorescein or biotin have been incorporated into proteolytic substrate cleavage products of bovine serum albumin generated by human cathepsin S or neutrophil elastase and into a fragment of the 31-kDa interleukin 1β precursor by human interleukin 1β-converting enzyme. Incorporation of the nucleophile is blocked by prior inhibition of the enzymes, and is not seen when proteolysis occurs in the absence of label, and the protease is then inhibited before the addition of label. Labeling is dependent on the pH, the time of reaction, and the concentrations of the nucleophile and substrate. Labeling of proteins can be readily detected by SDS-polyacrylamide gel electrophoresis. The pattern of elastase-labeled bovine serum albumin bands differs among P1′ Phe, Ala, and Gly, suggesting that nucleophilic attack on acyl enzyme intermediates derived from a large protein may differ from attack on small intermediates. The only observed labeled fragment catalyzed by interleukin 1β-converting enzyme is fragment 28-116 from the interleukin 1β precursor, suggesting that the cleavage between residues 27 and 28 is at least as efficient as between residues 116 and 117. This labeling method does not require organic solvent or nonphysiological pH values and thus may be useful for the discovery of novel protease substrates in cells or other in vivo systems or for diagnostic applications.

receptors from cell surfaces (7), or the processing of cytokines such as interleukin 1␤ (8). The identification of the physiological substrates of a protease may be important for a complete understanding of the function of the protease.
The identification of small peptide substrates or protease subsite mapping has been approached by a variety of techniques. More recently, phage selection techniques have been used to screen for small peptide substrates of proteases (9,10). The identification of proteins cleaved by proteases is more difficult. This has been approached by direct examination of an isolated substrate before and after proteolysis or by the application of analytical methods such as two-dimensional gels to more complex mixtures to follow changes in specific proteins spots due to proteolysis (see Ref. 11 or 12 for examples).
The detection of protein substrates is rendered more difficult if the substrates are unknown. A sensitive method of detection that allows direct visualization of substrates without the background of other proteins may be useful for a rapid assessment of the rate and extent of proteolysis in complex mixtures, as a diagnostic method, or for substrate identification.
Peptide nucleophiles have been used previously with proteases for peptide synthesis, which is optimal when done in high levels of organic solvent (13). Other uses include the study of the aminolysis of acyl enzyme intermediates to examine structural features important in the S 1 Ј site of human leukocyte elastase (14) and to map the S 1 Ј subsite specificity of trypsin and chymotrypsin (15).
In this paper, we examine the feasibility of the use of peptide nucleophiles for the direct visualization of protease substrates, as shown in Scheme 1.
In this scheme, the acyl enzyme intermediate (in the case of a serine protease) or thioacyl enzyme intermediate (in the case of a cysteine protease) undergoes competing reactions with 55 M water (hydrolysis to give product 2 (Reaction 3)) or with ϳ1 mM labeled nucleophile (in this case aminolysis to give a labeled product 2 (Reaction 4)). Hydrolysis is expected to predominate by the principle of mass action. Much lower levels of labeled substrate may not be problematic for detection and identification if the detection methodology is adequately sensitive.
Such a process would be most useful if achieved at physio-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18

EXPERIMENTAL PROCEDURES
Materials-Electrophoretically pure BSA, leupeptin, and E-64 were obtained from Sigma. Human neutrophil elastase was obtained from Calbiochem (La Jolla, CA), and human cat S was expressed and purified as described. 2 ICE was a generous gift of Dr. Nancy Thornberry (Merck). pIL-1␤ was obtained from Cistron, Inc. (Pinebrook, NJ), and a mouse monoclonal antibody binding mature human IL-1␤ was from R&D Systems Inc. (Minneapolis, MN). Biotin-NHS ester was from Sigma). Avidin-horseradish peroxidase and the LC-RP Supersignal system were from Pierce. Cascade Blue hydrazide and lissamine rhodamine B sulfonylhydrazine were from Molecular Probes (Eugene, OR). All other chemicals were of reagent grade or better. Prepoured 4 -20% gradient SDS-gels and molecular mass standards, which contained 6and 17-kDa proteins labeled with fluorophores for visibility on fluorescence-detected gels, were obtained from Novex (San Diego, CA).
Dipeptides coupled to biotin were synthesized as in the representative examples below. The Z-protected dipeptide hydrazide Z-Ala-Pro-NH-NH 2 was synthesized by reaction of Z-Ala-OH with Pro-OBzl in tetrahydrofuran with coupling by isobutylchloroformate in the presence of N-methylmorpholine, hydrolysis of the dipeptide benzyl ester by sodium hydroxide, and coupling to Boc-hydrazide in the presence of N-methylmorpholine in tetrahydrofuran with isobutylchloroformate. After removal of the Boc group with 4 M HCl in dioxane, the resulting Z-Ala-Pro-hydrazide was coupled with biotin-NHS ester (Sigma) in dimethylformamide to yield Z-Ala-Pro-NH-NH 2 -aminocaproyl-biotin. Unreacted biotin-NHS ester was quenched overnight with 0.6 M ethanolamine, and the dipeptide was deprotected with 30% hydrobromic acid in acetic acid. Dipeptide fluorophores were synthesized by a similar procedure, except that Z-dipeptide-hydrazide was reacted with fluorescein isothiocyanate to eventually yield dipeptide-NH-NH-(CϭS)-NHfluorescein. The dipeptide-biotin or dipeptide-fluorophore was then purified by C 18 reversed phase chromatography. The structures were checked by proton NMR.
Labeling of Cathepsin S and Neutrophil Elastase Products-A reaction mixture containing 20 g of BSA, 1.2 g of cat S, and 1 mM (except as noted) Ala-Ala-biotin or Ala-Ala-fluorescein was incubated at pH 8.0 in 0.1 M borate buffer (except as noted) for variable times at 37°C for elastase and 20 -23°C for cat S. Inhibitors were added as noted. The reaction was terminated by addition of SDS-PAGE reducing sample buffer to a final concentration of 0.1 M dithiothreitol or 2.5% (v/v) ␤-mercaptoethanol, 1.2% SDS, 0.03 M Tris, pH 6.8, and heated for 5 min at 100°C. The dependence of labeling on the concentration of the nucleophile was examined by drying aliquots of labeled dipeptides (dissolved in dimethylformamide) in the Eppendorf tubes used for the reaction preceding the addition of other reagents. This was done to avoid any effect of organic solvent on the reaction.
Elastase was mixed with BSA with 1 mM concentrations of different dipeptide-fluorescein nucleophiles with or without phenylmethanesulfonyl fluoride, taken from an ethanol stock immediately before addition to the reaction. Exact conditions are listed with the figure. The reaction was terminated by addition of SDS-PAGE sample buffer.
Labeling of ICE Proteolysis Products-IL-1␤ precursor (3.6 pmol, 0.18 M) was incubated with 1 mM Ala-Ala-biotin in pH 8.0, 0.1 M borate buffer at 37°C for 30 -120 min. The labeling reaction was followed by SDS-PAGE on a 16% Tricine reducing gel, blotted onto Protran nitrocellulose paper (Schleicher and Schuell, Keene, NH) with a Bio-Rad Semi-dry Transfer Cell electroblot apparatus at 150 mA/gel for 40 min, blocked by 5% (w/v) fat-free powdered milk containing 0.1% Tween 20 in phosphate-buffered saline at room temperature for 1 h, rinsed in water for 10 min, incubated for 1 h with avidin-horseradish peroxidase conjugate (Pierce) diluted 1:10,000 with 1% BSA ϩ 0.02% (v/v) Tween 20 ϩ 0.01% thimerosal in phosphate-buffered saline (pH 7.4), washed eight times for 5 min each with 0.1% Tween 20 in phosphate-buffered saline, and incubated with luminol-based LC-RP Supersignal solution (Pierce) for 1 min. The blot was then wrapped in Saran wrap and exposed for 30 min to Biomax film (Eastman Kodak Co.).
SDS-PAGE was performed as described (17) using gels obtained from Novex. Visualization of incorporated fluorophores was with a CCD camera and software (the FACE imaging system, version 2.30) from Glyko, Inc. (Novato, CA). This system monitors fluorescent bands with excitation at 360 nm and emission at 420 nm and also records a digital image of Coomassie Blue-stained bands using a bright field background from a fluorescent plastic sheet inserted behind the gel to be imaged. This system was also used to integrate the intensity of the gel bands. It is not optimized for fluorescein-labeled dipeptides, which have excitation and emission maxima of 490 and 515 nm, but gave adequate detection of such labeled bands. Gels used for imaging were stained with Coomassie Blue R-250 immediately after imaging so that labeling could be directly compared to total proteolysis.
Escherichia coli proteins were prepared by sonication of pelleted E. coli and retention of the supernatant from centrifugation of the resulting lysate. Fig. 1 shows a comparison of labeling of BSA proteolysis products with cat S by Ala-Ala-fluorescein and Ala-Ala-biotin. These dipeptide nucleophiles were used since cat S prefers Ala in the P 1 Ј position of small peptide substrates over 11 other tested residues (18) and since for acyl transfer the most efficient dipeptide observed by Schuster et al. (19) for papain was Ala-Ala-NH 2 . In Fig. 1A, the fluorescein-labeled bands (left gel) are compared to the Coomassie-stained gel (right gel). Lane 1 shows that no bands are observed when BSA is incubated with label alone for 60 min in the absence of protease. In lane 2, BSA was incubated with cat S for 30 min, IAA was added to 1 mM and reacted for 30 min to block the enzyme, and then label was added for 30 min. While the proteolysis is apparent in the right gel by Coomassie staining, no fluorescent bands were observed (left gel). In lanes 3 and 4, cat S was first blocked with 1 mM IAA and 29 M leuptin, respectively, for 30 min, and BSA was added and incubated for 30 min, followed by addition of label for a further 30 min. No fluorescent bands were observed. Lanes 5-9 show a 1-60-min time course of labeling when BSA, cat S, and Ala-Ala-fluorescein are present. Multiple bands below BSA are labeled. Comparison of the Coomassie-stained gel (Fig. 1A, right gel) shows no evidence for proteolysis in the absence of cat S (lane 1) or in the presence of inhibitors (lanes 3 and 4), and proteolysis where expected (lanes 2 and 5-9).

Labeling with the Cysteine Protease Cathepsin S-
In Fig. 1B, the labeling nucleophile was Ala-Ala-biotin, which was detected after blotting by avidin-horseradish peroxidase. In lane 2, 1 mM Ala-Ala-biotin was added to BSA. A small band is detected at the position of unproteolyzed BSA. In lane 3, BSA was incubated with cat S for 30 min, 1 mM of the inhibitor E-64 was added for 1 min, and Ala-Ala-biotin was then added for an additional 30 min. The same faint band at the molecular mass of intact BSA was observed. In lane 4, BSA, cat S and Ala-Ala-biotin were incubated for 30 min. A smear of bands at and below intact BSA are seen, down to an apparent molecular mass just above 30 kDa.The Coomassie-stained gel is shown on the right of terminal fragment of the substrate, this should be enhanced as the pH increases above the pK a of the NH 2 terminus of the label. It may also reflect the pH dependence of catalysis by cat S, which has a bell-shaped pH rate profile with apparent pK a s of 4.5 and 7.8 (21). Fig. 2 shows the pH dependence of BSA product labeling by cat S. Lanes 1, 3,5,7,9,11, and 13 contain samples from the reaction of BSA, cat S, and 1 mM Ala-Alafluorescein at pH 6.0, 7.0, 7.0, 8.0, 8.0, 9.0 and 10.0 using different buffers as detailed in the figure legend. Labeling is barely visible at pH 6, increases at pH 7, is maximal at pH 8, is visible but lower at pH 9, and does not occur at pH 10. Control reactions at the same pH, with cat S reacting with BSA for 1 h, followed by inhibition with 1 mM IAA for 1 h, and then by addition of Ala-Ala-fluorescein for 10 min, failed to show labeling. The Coomassie-stained gel (Fig. 2, right) reflects maximal proteolysis at pH 7-8, less at pH 6 and 9, and little at pH 10.
Labeling is less effective in a nucleophilic buffer such as Tris (data not shown).
Due to the limited proteolysis of BSA, a discrete number of bands are visible. If nucleophile labeling is 100% efficient, one might expect that one-half of the bands are labeled. In Fig. 2, above the free label at the bottom of the left gel, about six fluorescent bands are seen. This is one-half the number seen by Coomassie staining (right gel). Fig. 3 shows the nucleophile concentration dependence of the labeling of BSA proteolysis products with cat S and Ala-Alafluorescein. In this experiment, the label, dissolved in dimethylformamide, was first dried in the Eppendorf tubes used for the reaction so that increasing levels of dimethylformamide did not affect the enzyme. In lanes 1-10 of the left gel, labeling increased from 50 M to ϳ1.6 mM nucleophile and then decreased, with little labeling evident at 13 or 26 mM dipeptide. Fitting the concentration dependence of the integrated intensity of the 42-and 28-kDa bands in Fig. 3 (left gel) to a single hyperbola (not shown) suggests a K 1/2 for labeling of 0.29 and 0.37 mM, respectively. The corresponding Coomassie-stained Labeling by cat S was attempted with different dipeptide nucleophiles to examine the relative importance of different P 1 Ј residues for the labeling of protein substrates significantly larger than small peptides often used for such studies. Menard et al. (18) have reported that cat S prefers P 1 Ј Ala over Phe by about 5-fold, as measured by the value of k cat /K m , for the substrate dansyl-Phe-Arg-X-Trp-Ala. Fig. 4 (left gel) shows the results of labeling with (1 mM each) Phe-Ala-fluorescein ( lanes  1-3), Ala-Ala-fluorescein (lanes 4 -6), Ala-Pro-fluorescein (lanes 7-9) and Gly-Gly-fluorescein (lanes 10 -12). In each experiment, the first lane contains cat S and BSA, and the second lane contains cat S and BSA reacted for 30 min, blocked by 40 M E-64 for 1 min, followed by the addition of the dipeptide nucleophile. The third lane contains cat S, BSA, and label without inhibitor. The dipeptide nucleophile with P 1 Ј Phe gives more intense labeling than P 1 Ј Ala, which is better than P 1 Ј Gly. Ala-Pro-fluorescein gives only a single faint band.
To examine labeling in a complex mixture of proteins, cat S was added to E. coli proteins. some of the gel bands. The time course observed by Coomassie staining would also be less obvious than that seen by fluorescence, and the pattern of fluorescent gel bands might differ from the Coomassie staining pattern. Comparison of the fluorescein-labeled gel before (Fig. 5, left gel) and after Coomassie staining (Fig. 5, right gel) is consistent with cat S substrate selectivity.
Labeling with Human Neutrophil Elastase-Labeling of BSA proteolysis products was also examined using human neutrophil elastase. Studies on the aminolysis of acyl enzymes by peptide nucleophiles have suggested that side chain hydrophobicity of the P 1 Ј residue of the nucleophile may be more important than specific structural features and that there may be no important binding interactions beyond the P 1 Ј residue (14). Thus the dipeptide nucleophile Phe-Ala-fluorescein was used for a number of preliminary labeling experiments. Fig. 6 shows substrate labeling catalyzed by elastase with a number of different dipeptide nucleophiles. Lanes 1-3 show labeling of 20 g of BSA by 0.1 or 1 g of elastase. After digestion of BSA by elastase was finished, the enzyme was blocked by 1 mM phenylmethanesulfonyl fluoride for 20 min, and different dipeptide-fluorescein labels were subsequently added. No labeled bands were seen (lanes 3, 6, 9, and 12). Higher molecular mass bands were seen with 0.1 g of enzyme (lanes 2, 5, 8, and 11), and lower bands seen with 1 g of elastase (lanes 1, 4, 7, and 10). In lanes 1-3, 4 -6, 7-9, and 10 -12, this experiment was repeated with 1 mM Phe-Ala-, Ala-Ala-, Ala-Pro-, and Gly-Gly-fluorescein, giving defined bands but a slightly lower amount of label incorporation. In lanes 7-9, 1 mM Ala-Pro-fluorescein was used, giving fewer discrete bands. The dye appears to be smeared toward the bottom of the gel in each lane. The presence of Pro in the P 2 Ј position thus appears to decrease labeling. In lanes 10 -12, the use of Gly-Gly-fluorescein results in significant incorporation into several bands. In all cases, bands were not labeled when elastase was exposed to phenylmethanesulfonyl fluoride before a dipeptide nucleophile was added, the bands appear weaker than for labeling by cat S. Fig. 6 (right gel) confirms that use of a higher level of elastase results in fewer higher molecular mass bands and more lower molecular mass bands. This is consistent with increased proteolytic degradation of the larger fragments with increased amounts of elastase.
Labeling of Interleukin 1␤ Precursor by ICE-ICE is a cysteine protease that requires a P 1 Asp for proteolysis, and it can cleave before P 1 Ј Ala or Gly in the IL-1␤ precursor (20). Due to its narrow substrate specificity, ICE was examined for labeling of one natural substrate, the IL-1␤ precursor protein. It is cleaved between Asp 116 and Ala 117 and between Asp 27 and Gly 28 (8). Fig. 7 shows the results from a labeling experiment, in which  , lanes 2-4). In lane 1 of the top gel, pIL-1␤ and mature IL-1␤ were added as standards without ICE. Lanes 2-4 show a labeling time course when ICE was added to pIL-1␤ with detection by biotin-NHS ester (bottom gel). Five bands (bands A, BЈ, C, D, and E in the bottom gel) of apparent molecular masses, based on molecular mass standards, of 31.5, 24, 17.5, 12, and 9.3 kDa, were labeled. The bands at apparent molecular masses of 24 and 9.3 kDa (bands BЈ and E) appear to be contaminants of pIL-1␤, since they are seen when biotin-NHS ester is mixed with precursor alone (data not shown). On the Western blot (top gel) in lanes 2-4, the precursor (band A) is processed to mature IL-1␤ (band C) and an intermediate (band B) with an apparent molecular mass of 29.6 kDa.
Besides the precursor protein (band A, top gel) and mature IL-1␤ (band C), possible species include fragment 28 -269 (27.7 kDa), fragment 1-116 (13.4 kDa), fragment 28 -116 (10.3 kDa), and fragment 1-27 (3.08 kDa). Addition of Ala-Ala-biotin to ICE and pIL-1␤ in lanes 7-9 (bottom gel) results in labeling of band D. Inhibition of ICE with IAA after 60 or 120 min of reaction, followed by addition of Ala-Ala-biotin to 1 mM, blocks labeling of this band (bottom gel, lanes 5 and 6). Individual bands were cut from the blots and submitted for NH 2 -terminal sequencing. Band C had an NH 2 -terminal sequence of ( )PV( )( )LN, consistent with the predicted NH 2 -terminal sequence of mature IL-1␤ of APVRSLN. Band D had an NH 2terminal sequence of ( )PKQM, consistent with the predicted sequence of the 10-kDa precursor fragment 28 -116 of GPKQMK.

DISCUSSION
In this paper, we have examined a method to label protein proteolysis products of serine and cysteine proteases. The labeling does not require addition of organic solvent or nonphysiological pH to give detectable signals in the test systems used and is at least as sensitive as Western blotting. Labeling does not appear to represent binding to the proteolytic products, since inhibiting the enzyme after digestion of BSA, E. coli proteins, or pIL-1␤ prior to addition of the dipeptide nucleo-phile results in few or no labeled bands relative to those seen in the absence of inhibitor. Inhibiting the protease before the reaction, or omitting the protease, results in little or no labeling. Weak binding would also not be expected since free dipeptides are separated from the larger labeled peptides by SDS-PAGE preceded by boiling the substrate and products in SDS sample buffer for 5 min. Labeling catalyzed by cat S appears to be efficient under the conditions examined, resulting in labeling of about one-half of the proteolytic products seen. Labeling of products catalyzed by neutrophil elastase appears to be more selective. Labeling is seen for a variety of different dipeptide nucleophile structures and has also been observed with cat S by direct addition of hydrazide-containing fluorophoric dyes lacking an attached dipeptide, such as Cascade Blue hydrazide or lissamine rhodamine B sulfonylhydrazine (data not shown). This will allow considerable flexibility in designing the detection signal, which for example could also include radiolabeled nucleophiles.
The effects on labeling of a number of parameters that might affect attack of dipeptide nucleophiles on an acyl enzyme intermediate have been examined. Labeling of proteolysis products of cat S is dependent on the pH of the reaction, increasing as pH is increased from 6 to 8 in different nonnucleophilic buffers and diminishing as pH increases from 9 to 10. The use of a nucleophilic buffer such as Tris decreases the labeling (data not shown). This pH dependence may reflect multiple molecular events, including the pK a of the nucleophile and the pH dependence of the protease. It does give information on the allowable pH range for application of this labeling with similar proteases, and it suggests that labeling in cells at pH 7.4 may be possible.
For cat S, labeling of BSA with dipeptide nucleophiles depends on the P 1 Ј residue, with a P 1 Ј Phe most preferred among those tested. This labeling preference may be a function of variables other than k cat /K m , such as the local structure of the large protein substrate (BSA or large fragments thereof) forming the acyl enzyme intermediate. This is not necessarily inconsistent with the observations of Menard et al. (18) that for the small substrate dansyl-Phe-Arg-X-Trp-Ala, cat S prefers P 1 Ј Ala over Phe by ϳ5-fold. Labeling might also be expected to be a function of the P 2 Ј residue, which for cat S in substrates is preferably a smaller amino acid (21); thus, we have used ala- nine as the P 2 Ј residue in most of these experiments. The experimentally defined P 1 Ј (and P 2 Ј) preferences suggest that occupancy of both sites may be important for efficient labeling. This is consistent with data for aminolysis of a tripeptide acyl-papain complex from Schuster et al. (19), who found that nucleophile occupancy of the P 1 Ј-P 3 Ј sites affected the rate of this reaction. An additional factor affecting the efficiency of labeling by different dipeptide nucleophiles may involve changes in the accessibility of water to the acyl enzyme intermediate, while the nucleophile is bound at the S 1 Ј binding site (16).
The P 1 Ј substrate specificity of human neutrophil elastase has been reported to be more dependent on hydrophobicity than on specific structural features of the side chain (14), and the P 2 Ј position was reported to have little effect. We have observed labeling of BSA products and somewhat different labeled product band patterns, with Phe-Ala, Ala-Ala-, and Gly-Gly-fluorescein. Some bands are labeled best with Gly-Glyfluorescein, which has the least hydrophobic P 1 Ј residue. This suggests that for elastase, as well as for cat S, dipeptide nucleophile attack on acyl enzyme intermediates formed with fragments of large substrates may behave differently regarding substrate P 1 Ј preferences than with smaller substrates. Inserting Pro in the P 2 Ј position significantly diminishes labeling; thus, this position can affect nucleophile attack, perhaps by weakening the binding of the dipeptide. The differential labeling seen with different P 1 Ј residues suggests that some selectivity of labeling might be attained if these preferences are known. The nucleophile concentration dependence for cat S shows increased labeling as concentrations approach 1.6 -3.2 mM and diminished labeling at higher levels. Diminished proteolysis above 1.6 mM on the Coomassie-stained gel is also observed, suggesting that the dipeptide nucleophiles may be acting to inhibit cat S at higher concentrations. This suggests also that there may be an optimal dipeptide nucleophile concentration for the labeling by each protease.
Besides labeling of proteolytic products for less specific enzymes such as elastase or cat S, this methodology also works for more specific proteases such as ICE. This reaction can be followed by the time-dependent disappearance of the precursor protein both in a Western blot and by biotin-NHS ester detection. Three potential fragments (residues 1-27, 28 -116, and 1-116) could possibly be labeled by cleavages at known ICEsusceptible sites, resulting in predicted bands at 3.1, 10.3, or 13.4 kDa. With the NH 2 -terminal sequence to aid interpretation, nucleophile labeling gives a band at 10 kDa consistent with the pIL-1␤ fragment 28 -116. The observed time course suggests that cleavage in pIL-1␤ between residues 27 and 28 may be at least as efficient than cleavage at 116 -117, since no labeled 1-116-residue peptide is seen. This should result in production of fragments 1-27 and 28 -116; the latter is detected as a biotinylated peptide. It would be interesting to examine the physiological activity (if any) of each of these fragments. Fragment 1-27 may have also been labeled but is not observable on this blot. Cleavage between residues 27 and 28 has been previously reported (8) and suggests that this method could be applied to such determinations for unknown substrates.
The two labeling methodologies demonstrated here (nucleophile labeling and amine labeling with biotin-NHS ester) are quite sensitive. Biotin-NHS ester labeling gives visible protein bands after electroblotting when low nanogram levels of molecular mass standards labeled by this method are loaded onto an SDS gel. This method should allow detection of any protein containing free amines and may be a convenient way to visualize most or all proteins in reactions subjected to nucleophile labeling or Western blotting. Nucleophile labeling of 3.6 pmol of pIL-1␤ gives a strong band at 10 kDa even after electroblotting, with about the sensitivity of detection seen with Western blots of the same sample. Combined with the lack of dependence of labeling on organic solvent and its demonstration at pH 7-8, this suggests that nucleophile labeling might be useful for the identification of natural substrates in cells or other in vivo systems.