Protease profiling using a fluorescent domino peptide cocktail

Abstract

Five hexapeptides were prepared containing, in a domino-type arrangement, all 25 possible dipeptides between (1) aromatic, (2) hydrophobic, (3) positively charged, (4) negatively charged, and (5) small and polar amino acids. The peptides were fluorescence labeled at the N-terminus with a (7-coumaryl)oxyacetyl group, allowing the selective detection of N-terminal cleavage products. The five peptides were used as a cocktail reagent in an HPLC analysis. The cocktail produced specific cleavage patterns, or fingerprints, for a variety of proteases. This domino peptide cocktail can be used as a general reagent for protease identification and functional profiling.

---

Introduction

The analysis of protein function is central to the deciphering and understanding of biological processes.¹ Enzymes, which represent a large portion of proteins, can be investigated using catalysis assays. Such assays are based on reactive probes such as specific substrates undergoing turnover,² or covalent labeling reagents.³ Enzyme assays are used in enzyme discovery, enzyme engineering, drug discovery and medical diagnostics. Assays based on single substrates are suitable for screening specific enzyme types. One can also combine several substrates to produce enzyme activity profiles or fingerprints.⁴ Such multi-substrate assays provide much richer information than assays based on single substrates. The fingerprint information can be used to distinguish between closely related enzymes on a functional basis, and can also be related to the genetic phylogeny as shown for P450 enzymes.⁵

A critical aspect of enzyme fingerprinting studies is the availability of a simple and practical method for recording the fingerprints. Enzyme-specific fingerprints can be obtained for hydrolytic enzymes on the basis of a few tens of substrates in microtiter-plate based parallel assays.^6–8 However this method consumes large quantities of enzyme and substrates and is complicated to handle, although improvements are possible using a solid support.⁹ Recently, we showed that enzyme fingerprints can be accurately recorded from enzyme reactions involving substrate cocktails, which are quantitated by HPLC.¹⁰ The method provides reproducible fingerprints in a simple and flexible single-analysis format and was demonstrated for fingerprinting closely related lipases. Herein we report the application of cocktail fingerprinting to proteases (Scheme 1).

Scheme 1 Cocktail fingerprinting of protease with N-terminal labeled hexapeptide substrates. Sequences are selected by domino design (see Fig. 1).

Proteases are ubiquitous enzymes associated with a variety of diseases, in particular cancer and viral infections.¹¹ Protease assays are used in high-thoughput screening of inhibitors and for biochemical and clinical investigations. Several non-specific protease reagents are known, e.g. fluorescence-labeled casein.¹² These reagents detect proteolytic activities but do not provide any information about the particular type of protease involved. Specific protease assays are possible with fluorogenic or chromogenic peptides, for example for HIV-protease.¹³ Such assays allow the selective identification of a single protease if only the target protease cleaves the substrate,¹⁴ and may be suitable for in vivo imaging.¹⁵ In this format however a different assay must be applied for each protease. A general approach to functional profiling of proteases has been described by several groups using combinatorial libraries of peptides,¹⁶ either as on-bead substrates,¹⁷ positional scanning libraries of substrates¹⁸ or inhibitors,¹⁹ single compounds,²⁰ or as glass-bound substrates on a chip.²¹ These methods typically involve hundreds or thousands of different peptide substrates and lead to consensus cleavage sequences for each protease studied.

Library-based profiling experiments require a large number of parallel analyses and represent a full-blown study for each enzyme. We reasoned that an assay based on only a few well-chosen peptide substrates and the analysis of the products might provide sufficient differential reactivity information to allow a functional classification of various proteases in a simple assay setup, this without necessarily covering the entire peptide sequence space exhaustively. Here we report a protease identification reagent consisting of a cocktail of five fluorescence labeled hexapeptides. All possible dipeptide cleavage sites are represented in the cocktail in a domino-type arrangement. Protease specific fingerprints are produced using a single assay by HPLC-analysis of the N-terminal fragments.

Results and discussion

Assay design

The biological function of proteases depends in large part on their substrate specificity, which usually translates into preferred cleavage sites. Cleavage sites are defined in terms of the amino acid preferences for positions P1, P2, P3 etc. and P1′, P2′, P3′ etc. describing the amino acids extending from the siscile peptide bond on the carboxyl-side (N-terminal part of the peptide substrate) and the amine-side (C-terminal part of the peptide substrate), respectively. The combinatorial protease profiling experiments cited above have shown that protease specificity usually depends on the amino acids at only one or two positions, most often P1 and P2. Well known examples include trypsin, which cleaves peptide bonds when P1 is a positively charged amino acid, and chymotrypsin, which prefers aromatic or hydrophobic amino acids at P1. In addition to sequence specificity, the proteolytic cleavage activity of any protease also strongly depends on the reaction conditions, including pH, buffer, temperature and solvent. Another important factor for protease reactivity is the kinetic availability of the substrate. Readily soluble short peptides of five amino acids or more are optimally suited for cleavage, while large folded proteins or poorly soluble peptides are cleaved more slowly or not at all.

We aimed for a cocktail reagent for protease profiling that would result in at most 20–30 different fragments, so as to be separable by HPLC. The cocktail was designed as a series of five hexapeptides carrying a fluorescent label at the N-terminus. In this manner the analysis would involve only thirty one possible N-terminal labeled fragments, including the five substrates and 26 cleavage products. The use of five different substrates instead of a single long peptide would ensure that most reactions would produce at least five different detectable fragments. In addition, the synthesis of five hexapeptides by solid-phase peptide synthesis would provide much higher overall yields than the synthesis of a single 30-mer peptide. (7-Coumaryl)oxyacetic acid was selected as the N-terminal label because coumarin-labeled substrates are generally well-suited to enzyme assays with respect to solubility and enzyme reactivity, and can be detected selectively by UV or fluorescence.²²

Domino-sequence design

The sequences of the five labeled peptides were chosen to maximize the probability of protease reactivity by providing all possible dipeptide arrangements. The 20 proteinogenic amino acids fall into five different classes, which are (1) aromatic, (2) hydrophobic, (3) positively charged, (4) negatively charged, and (5) small and polar amino acids. This classification implies that there exists 25 different types of dipeptides. These 25 dipeptide types were assembled into five hexapeptides by playing a simple domino-game with pieces representing the dipeptides. Amino acids were then introduced in place of amino acid types to obtain actual peptides (Fig. 1).

Fig. 1 Domino design of peptide cocktail. (a) Classification of proteinogenic amino acids; (b) all 25 possible types of dipeptide serve as domino pieces; (c) one possible solution of the domino game; (d) contraction of the domino solution into five hexapeptides and one possible realization with actual amino acids. Note that the domino-solution selected also realizes 16 of 25 possible 1,3-arrangements of amino acid types. The missing 1,3-arrangements are AXA, HXA, HXH, PXP, PXS, NXP, NXN, SXA, SXS.

Synthesis

The five target hexapeptides were prepared by Fmoc-type solid phase peptide synthesis (SPPS) on rink-amide or Wang resin following a standard protocol.²³ The sequences were terminated by coupling of (7-couramyl)oxyacetic acid under standard conditions. Cleavage from the resin and purification by preparative reverse-phase HPLC gave the pure peptides in good yields. All reference N-terminal cleavage fragments were also prepared by Fmoc-type SPPS on Wang resin, followed by cleavage and purification by preparative reverse-phase HPLC. Analysis conditions were optimized to obtain the best possible separation of all fluorescence-labeled substrates of the cocktail and their fragments in a single HPLC-analysis, allowing the separation of 26 from the 31 fragments. The retention time of each fragment was identified by individual elution in the gradient. The information from fingerprint chromatograms of proteolytic assays was sufficient to reconstitute the cleavage pattern in each case (Table 1).

Table 1 Cocktail fingerprinting of proteases by analysis of N-terminal labeled fragments. For each protease reaction, the percentages of total peak area at 320 nm (coumarin label) are given for the corresponding fragment, identified by its retention time. Assay conditions: 50 µM total cocktail in 20 mM bis-tris buffer, 37 °C, 1 h, then HPLC analysis: Vydac 218TP54, RP-C18, 22 × 0.4 cm, eluent 1.5 mL min⁻¹ gradient water–acetonitrile + 0.1% TFA, detection by UV at 320 nm. Duplicate assays are labeled -1

Fragments	Sequence	t R/min^a	Trypsin-1	Trypsin 2 h	Trypsin 5 h	Trypsin	Trypsin 10 min	Trypsin 30 min	Subtilisin	Subtilisin-1	Prot-K	Prot-K-1	Thermolysin	Thermolysin-1	Prot-C 10 h	Prot-C 1 h	Chymotrypsin	Papain	Pepsin
a * is the (7-coumaryl)oxyacetyl label and the capital letters are standard amino acid codes.
1P5	*K	6.86									0.4%		1.5%		0.7%
1P2	*KDES	6.89								2.0%
1P3	*KDE	7.23
1P4	*KD	7.63									2.2%	1.0%			0.8%
3P5	*E	9.15											17%	11%
P6	*	11.4													10%
1P	*KDESYR	11.7	20%	20%	20%	20%	20%	20%	17%	18%	19%	22%	23%	22%	23%	22%	9%	22%	22%
2P5	*A	12.2													2.0%
4P2	*YARK	13.7	0.2%	0.1%		0.3%	0.2%	0.2%			4.2%	6.0%	0.4%	0.1%		0.7%
1P1	*KDESY	14.1							4.8%	3.5%							14%
4P3	*YAR	15.3	22%	23%	22%	19%	23%	23%	14%	19%	11%	7.0%				0.3%	3.0%
2P	*AVPERS	17.5	19%	19%	18%	19%	19%	19%	20%	19%	14%	18%	18%	18%	3.0%	16%	18%	25%	18%
4P4	*YA	18		0%	0.6%	0.8%			4%	0.3%					6.0%	13%	10%
2P1	*AVPER	18.1									6.0%	2.4%	3.9%		3.0%	2.9%			0.6%
4P5	*Y	18.6											14%	22%	3.3%		7.2%	2.1%
2P2	*AVPE	19.9		0.2%	0.5%		0.2%	0.2%	1.0%	0.3%					8.0%	2.0%
2P4	*AV	20.2										0.9%
5P4	*LK	22.5	13%	16%	18%	23%	5%	10%	6.8%	5.0%	0.5%	2.0%			8.0%	2.1%	0.8%	3.7%
2P3	*AVP	23.4													0.5%	0%
4P1	*YARKL	24.1
3P	*EFVGSD	25.1	20%	20%	20%	18%	20%	20%	4.6%		0.4%	0.6%			19%	21%	19%	22%	18%
3P1	*EFVGS	26.6							14%	19%	21%	23%
3P4	*EF	28.7																	0.6%
3P2	*EFVG	28.7							0.3%		0.4%		5.2%	6.0%				1.5%
5P3	*LKY	30.1									3.2%	1.1%	17%	18%	6.5%	17%	1.2%		23%
5P5	*L	31													6.2%	0.4%
3P3	*EFV	32.8												2.4%
4P	*YARKLF	35	0.5%	0.2%			0.5%	0.5%				3.0%				0.6%		7.0%	18%
5P1	*LKYFD	39
5P2	*LKYF	41.1							14%	11%	16%	0.7%				2.2%	13%
5P	*LKYFDI	43	5.0%	2.2%	1.1%		12%	8.0%		4.0%	2.2%	13%				0.4%	5.0%	16%
Conversion			35%	40%	41%	43%	29%	34%	59%	60%	65%	44%	59%	60%	55%	40%	49%	7.0%	25%

---

Cocktail fingerprinting

The fluorescent label allowed the detection of submicromolar concentrations of substrates and products. The cocktail was formulated as a mixture containing 10 µM of each substrate under assay conditions, corresponding to 50 µM total concentration of peptide substrates. A series of proteases were incubated at their optimal pH with the domino peptide cocktail, and the products were analyzed by HPLC. Detection at 320 nm showed only the starting peptide and the N-terminal cleavage products. The percentage conversion of each substrate to the products was quantified by comparing peak areas. The data was reported as a percentage of each substrate or N-terminal fragment observed for each enzyme reaction (Table 1).

Data analysis

Statistical analysis was carried out using product-only data. The protease fingerprints were expressed in terms of product selectivity by scaling the sum of all products formed to 100% for each reaction. The product percentages were then used as variables for hierachical cluster analysis of the different protease fingerprints using Ward's method on the basis of squared Euclidean distances.²⁴ A color-coded representation of the product fingerprints is shown in Table 2.

Table 2 Color-coded protease fingerprints from domino-peptide cocktail. The relative distribution of products is shown by coloring the P1 position of cleavage of each detected N-terminal coumarin-labeled fragment relative to the most abundant fragment detected (shown in black) using the indicated scale. Unreacted substrates are not color coded (complete data in Table 1). Proteases are ordered according to the hierarchical clustering shown at right (Ward's clustering using squared Euclidean distances with product percentages as variables. The variables were not normalized). * is the (7-coumaryl)oxyacetyl label and the capital letters are standard amino acid codes. C-termini of the peptide substrates are carboxamide (peptide 1, 3 ,5) or carboxylate (peptide 2, 4)

---

It must first be noted that all proteases tested produced significant cleavage products with the cocktail, suggesting that the domino-sequence design helped to provide a well-differentiated set of peptides for protease activity analysis. There were strong reactivity differences between the different peptides composing the cocktail. Thus, peptide 5 with two aromatic residues in its center was cleaved by all proteases and often delivered the strongest product. By contrast, peptides 1 and 2 were cleaved only rarely. However even the less reactive peptides were essential in providing cleavage specificity information. For example, Proteinase K and subtilisin differ mostly by their weak cleavage activity on peptide 1 and 2, while their activity on peptides 3, 4 and 5 is very similar.

Pepsin showed the most selective reaction, and hydrolyzed almost exclusively peptide 5 at the central peptide bond between two aromatic residues. Trypsin cleaved peptides 4 and 5 at the postively charged amino acid as expected. However there was no reaction of trypsin with peptides 1 and 2 despite the fact that these also carry a positively charged amino acid. This may be due to the fact that these tryptic sites are close to the C- or N-terminus in peptides 1 and 2. Papain hydrolyzed peptide 5 at the same site as trypsin, but cleaved peptides 3 and 4 at different sites corresponding to small and aromatic residues at P1, respectively. Thermolysin also cleaved peptides 3, 4 and 5, with the strongest cleavage occurring in peptide 5 between the two central aromatic amino acids. Protease C, a metallo-protease secreted by Erwinia chrysanthemi belonging to the serralysin family,^25,26 produced a fingerprint similar to thermolysin. In this case, extending the reaction time up to 10 h resulted in further cleavage of the peptides, including a significant cleavage of the N-terminal coumarin label. Chymotrypsin, which is known to cleave preferentially at aromatic residues, produced the expected cleavage products with peptide 1 and 5. However the phenylalanine–valine linkage in peptide 3 remained untouched. The trace of tryptic cleavage observed in peptide 4 with the chymotrypsin sample probably reveals a trypsin contamination. Subtilisin and proteinase K produced very similar cleavage patterns comprising three main cleavage products from peptides 3, 4 and 5.

The cocktail reagent delivers specific cleavage patterns for each protease in a single assay. The statistical analysis of these patterns was carried out considering product selectivity only. Product selectivity analysis circumvents the problem of assigning activity units to each sample tested.^7e More importantly the selectivity fingerprint of an enzyme reaction at two different time points or using different amounts of the same enzyme should give the same selectivity fingerprints. While this principle is applicable in cocktails where each product is formed from a different substrate,¹⁰ the situation of the peptide cocktail is different because one substrate forms several products, and the product distribution may evolve in the course of the reaction. This is exemplified in the case of protease C, which was measured at either 1 h or 10 h incubation time. The product pattern at 10 h shows further degradation of the cleavage products, including the formation of the free (7-coumaryl)oxyacetic acid label. Hierarchical clustering however indicates a strong similarity between the pair of measurements with protease C relative to the other proteases tested. Analysis of the product pattern at different time points was also investigated in the case of trypsin. In this case variation of incubation time did not produce significant changes, as shown by cluster analysis. The cleavage product patterns observed were generally reproducible, as exemplified for the case of trypsin, subtilisin, proteinase K, thermolysin, and protease C. The duplicate measurements were repeated at 3 month intervals with the same cocktail reagent, which was completely stable upon storage. Most importantly, variations were smaller within repeated measurements of the same protease than between different proteases.

Statistical analysis based on the relative percentage of products gives the strongest statistical weight to the largest peaks of product formed, including their actual integration value. Clustering was also investigated using a logarithmic scale (P_i = 2 + log₁₀(0.01% + percent product i)). Logarithmic weighting dampens small differences between large peaks and gives more statistical weight to trace amounts of small peaks. In this manner the dataset is interpreted in terms of a product pattern rather than a quantitative product distribution. Clustering in this logarithmic scale gave comparable results to the relative percentage analysis presented above, with the exception of the assignment of the protease C analyses at 1 h and 10 h in two distant clusters due to the presence of additional peaks in the later sample. This suggest that product pattern analysis is not suitable for data produced by our cocktail reagent.

Although the domino-cocktail experiment provides specific cleavage patterns for all proteases, these patterns may not be directly interpretable in terms of protease selectivity. Indeed the assignment of selectivity to a particular dipeptide arrangement as realized by the domino-sequence is not possible without knowing which amino acid positions are important for cleavage. Selectivity is probably determined in most cases by the amino acids P1 and P2 relative to the cleavage site. Interestingly, the domino-cocktail used here also realizes 16 of the 25 possible 1,3-arrangements of amino acid pairs, such that the reagent also surveys a number of P1–P3 combinations (Fig. 1). It must also be noted that peptide bonds in the middle of the hexapeptides are probably favored by most proteases, and that cleavages occurring near the C-terminus may be masked by subsequent cleavages since only the N-terminal fragments are detected.

The key technical difficulty in the fingerprinting assay is the separation and identification of products by HPLC. A possible method refinement would consist in confirming the identity of each peak by LC-MS in each assay. The advantage of the (7-coumaryl)oxyacetyl label is that products are detectable selectively and with high sensitivity at 320 nm, either by UV or by fluorescence, using a standard HPLC-instrument. We have used this assay to test crude cellular extracts and microbial cultures successfully without noticing any interference. The cocktail protease assay is well suited for testing protease activities in a variety of situations.

Conclusion

Fingerprinting of protease activities was realized using five fluorescence-labeled hexapeptides. Peptide sequences were selected using a domino-type arrangement of all 25 possible dipeptides between different amino acid types (aromatic, hydrophobic, positive, negative, small and polar). The domino sequence selection produced a cocktail reactive with all proteases tested. The reagent should be generally useful to produce a specific pattern, or fingerprint, for any given protease. This fingerprint may then be used for the identification of this protease, or for functional comparisons with other proteases. Further experiments are underway to assess whether the reactivity information obtained by cocktail fingerprinting is sufficient to classify all proteases by their reactivity.

Experimental section

(2-Oxo-2H-chromen-7-yloxy)-acetic acid (*)

A solution of umbelliferone (3.24 g, 20 mmol), ethyl bromacetate (4.4 g, 26 mmol), K₂CO₃ (8.3 g, 60 mmol) and [18-crown-6] (10 mg, cat.) in DMF (15 mL) was stirred at 60 °C for overnight. After completion of the reaction (TLC), aqueous workup (AcOEt/water, brine) and evaporation of the organic phase, the residue was dissolved in acetone (20 mL) and water (50 mL) and treated with NaOH (1.3 g, 33 mmol) at 50 °C overnight. The reaction was diluted with water (300 mL) and neutralized by dropwise addition of 3N HCl to pH <2 under stirring, leading to precipitation of the product, which was filtered and dried to yield (2-oxo-2H-chromen-7-yloxy)-acetic acid (4.2 g, 19 mmol, 95%) as colorless solid, m.p. 216–218 °C. ¹H NMR (300 MHz, DMSO): δ = 7.96 (d, 1H, J = 9.42 Hz), 7.61 (t, 1 H, J = 6.57 Hz), 6.93 (dd, 2H, J = 7.14, 2.64 Hz), 6.28 (d, 1H, J = 9.39 Hz), 4.82 (s, 2 H); ¹³C NMR (75 MHz, DMSO): δ = 169.6, 160.8, 160.2, 155.1, 144.2, 129.5, 112.8, 112.7, 112.5, 101.4, 64.8.

Peptide 1

Starting with 100 mg of Rink amide resin (0.061 mmol), Fmoc-type SPPS followed by cleavage and purification by preparative RP-HPLC gave *KDESYR-NH₂ (48 mg, 0.049 mmol, 64%) as colorless foamy solid. ESI⁺-MS: calc. for C₄₄H₆₀N₁₁O₁₆ [M + H]⁺: 998.4219, found 998.4240.

Peptide 2

Starting with 100 mg of Fmoc-Ser_(t-bu)-Wang resin (0.063 mmol), Fmoc-type SPPS followed by cleavage and purification by preparative RP-HPLC gave *AVPERS-OH (33 mg, 0.039 mmol, 56%) as colorless foamy solid. ESI⁺-MS: calc. for C₃₈H₅₄N₉O₁₄ [M + H]⁺: 860.3790, found 860.3779.

Peptide 3

Starting with 100 mg of Rink amide resin (0.061 mmol), Fmoc-type SPPS followed by cleavage and purification by preparative RP-HPLC gave *EFVGSD-NH₂ (17.5 mg, 0.020 mmol, 34%) as colorless foamy solid. ESI⁺-MS: calc. for C₃₉H₄₇N₇O₁₅ [M + H]⁺: 854.3208, found 854.3215.

Peptide 4

Starting with 100 mg of Fmoc-Phe-Wang resin (0.10 mmol), Fmoc-type SPPS followed by cleavage and purification by preparative RP-HPLC gave *YARKLF-OH (48 mg, 0.048 mmol, 48%) as colorless foamy solid. ESI⁺-MS: calc. for C₅₀H₆₇N₁₀O₁₂ [M + H]⁺: 999.4939, found 999.4964.

Peptide 5

Starting with 100 mg of Rink amide resin (0.061 mmol), Fmoc-type SPPS followed by cleavage and purification by preparative RP-HPLC gave *LKYFDI-NH₂ (21 mg, 0.021 mmol, 31%) as colorless foamy solid. ESI⁺-MS: calc. for C₅₁H₆₇N₈O₁₃ [M + H]⁺: 999.4827, found 999.4796.

Enzymes

The following commercial enzymes were used: trypsin from pig pancreas 1645 U mg⁻¹ (Fluka 82495); subtilisin from Bacillus licheniformis 10.5 U mg⁻¹ (Fluka 85968); α-chymotrypsin from bovine pancreas 74.6 U mg⁻¹ (Fluka 27270); proteinase K from Tritirachium album 299 U mg⁻¹ (Fluka 82495); thermolysin from Bacillus thermoproteolyticas rokko 50–100 U mg⁻¹ (SIGMA p-1512); papain from papaya latex 14 U mg⁻¹ (SIGMA p-4762); pepsin from porcine stomach mocusa 4500 U mg⁻¹ (SIGMA p-6887). Purified protease C from Erwinia chrysanthemi was provided through a collaboration.^25b

Assay conditions

Peptides 1–5 were conditionned as 2 mM stock solutions in 1 ∶ 1 water–acetonitrile by dissolving 1–2 mg solid into the appropriate volume. The cocktail was prepared my mixing 5 µL of each peptide stock solution and 5 µL of a 2 mM stock solution of the internal standard 4-aminomethyl-umbelliferone in 920 µL 20 mM aq. bis-tris buffer adjusted in advance to the following pH using 1 N HCl: trypsin: pH 8, subtilisin: pH 6.5, chymotrypsin: pH 8, pepsin: pH 4, papain: pH 7.5, protease C: pH 7, proteinase K: pH 9, thermolysin: pH 7. The assay was started by adding 5 µL of a freshly prepared 1 mg mL⁻¹ stock solution of the protease in water to 95 µL of the cocktail. The assay concentration under these conditions was 10 µM for each of the five peptide substrates and 50 µg mL⁻¹ for the protease. After 1 h at 37 °C, the reaction was analyzed by RP-C18 HPLC using a Vydac 218TP54 column, 0.4 × 22 cm, detection by UV at 320 nm. Eluents: A = 0.1% CF₃CO₂H in water, B = 1 ∶ 1 acetonitrile–water. Flow rate: 1.5 mL min⁻¹. Linear gradient: t = 0: A/B = 80/20, t = 30 min: A/B = 50/50, t = 40 min: A/B = 30/70. The column was then washed for 10 min at 100% B and reequilibrated to the initial conditions for 10 min.

Acknowledgements

This work was supported by the University of Berne, the Swiss National Science Foundation, and Protéus SA, Nîmes, France. The authors thank Prof. U. Baumann for providing a sample of protease C and for helpful discussions.

References

1. L. M. Steinmetz and R. W. Davis, Nat. Rev. Genet., 2004, 5, 190–201.
2. (a) Enzyme Assays: A Practical Approach, ed. Robert Eisenthal and Michael Danson, Oxford University Press, Oxford, UK, 2002; (b) S. Gul, S. K. Sreedharan and K. Brocklehurst, Enzyme Assays: Essential Data, John Wiley & Sons, 1998; (c) M. T. Reetz, Angew. Chem. Int. Ed., 2001, 40, 284–310; (d) D. Wahler and J.-L. Reymond, Curr. Opin. Chem. Biol., 2001, 5, 152–158; (e) D. Wahler and J.-L. Reymond, Curr. Opin. Biotechnol., 2001, 12, 535–544; (f) J.-P. Goddard and J.-L. Reymond, Trends Biotechnol., 2004, 22, 363–370; (g) J.-P. Goddard and J.-L. Reymond, Curr. Opin. Biotechnol., 2004, 15, 314–322.
3. (a) Y. Liu, M. P. Patricelli and B. F. Cravatt, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 14694–14699; (b) H. Scholze, H. Stutz, F. Paltauf and A. Hermetter, Anal. Biochem., 1999, 276, 72–80; (c) B. F. Cravatt and E. J. Sorensen, Curr. Opin. Chem. Biol., 2000, 4, 663–668; (d) D. Kidd, Y. Liu and B. F. Cravatt, Biochemistry, 2001, 40, 4005–4015; (e) G. C. Adam, E. J. Sorensen and B. F. Cravatt, Mol. Cell. Proteomics, 2002, 1, 781–790; (f) D. A. Jeffery and M. Bogyo, Curr. Opin. Biotechnol., 2003, 14, 87; (g) G. Y. J. Chen, M. Uttamchandani, Q. Zhu, G. Wang and S. Q. Yao, ChemBioChem, 2003, 336–339; (h) R. Birner-Grunberger, H. Scholze, K. Faber and A. Hermetter, Biotechnol. Bioeng., 2004, 85, 147–154; (i) A. E. Speers and B. Cravatt, Chem. Biol., 2004, 11, 535–546; (j) Q. Zhu, A. Girish, S. Chattopadhaya and S. Q. Yao, Chem. Commun., 2004, 1512–1513.
4. Review: J.-L. Reymond and D. Wahler, ChemBioChem, 2002, 3, 701–708.
5. V. Abecassis, D. Pompon and G. Truan, Nucleic Acids Res., 2000, 28, e88.
6. A. M. F. Liu, N. A. Somers, R. J. Kazlauskas, T. S. Brush, F. Zocher, M. M. Enzelberger, U. T. Bornscheuer, G. P. Horsman, A. Mezzetti, C. Schmidt-Dannert and R. D. Schmid, Tetrahedron Asymmetry, 2001, 12, 545–556.
7. (a) D. Wahler, F. Badalassi, P. Crotti and J.-L. Reymond, Angew. Chem. Int. Ed., 2001, 40, 4457–4460; (b) D. Wahler, F. Badalassi, P. Crotti and J.-L. Reymond, Chem. Eur. J., 2002, 8, 3211–3228; (c) J.-L. Reymond, D. Wahler, F. Badalasssi and H. K. Nguyen, Patent, WO0192563, 2001-12-07; (d) J. Grognux and J.-L. Reymond, ChemBioChem, 2004, 5, 826–831; (e) J. Grognux, D. Wahler, E. Nyfeler and J.-L. Reymond, Tetrahedron Asymmetry, 2004, 15, 2981–2989.
8. (a) D. Wahler and J.-L. Reymond, Angew. Chem. Int. Ed. Engl., 2002, 41, 1229–1232; (b) D. Wahler, O. Boujard, F. Lefèvre and J.-L. Reymond, Tetrahedron, 2004, 60, 703–710.
9. P. Babiak and J.-L. Reymond, Anal. Chem., 2005, 77, 373–377.
10. J.-P. Goddard and J.-L. Reymond, J. Am. Chem. Soc., 2004, 126, 11116–11117.
11. (a) A. Barrett, N. Rawlings and J. Woessner, Handbook of proteolytic enzymes, Elsevier Academic Press, London, 2nd edn., 2004; (b) X. S. Puente, L. M. Sanchez, C. M. Overall and C. Lopez-Otin, Nat. Rev. Genet., 2003, 4, 544–558.
12. S. S. Twining, Anal. Biochem., 1984, 143, 30–34.
13. (a) E. D. Matayoshi, G. T. Wang, G. A. Krafft and J. Erickson, Science, 1990, 247, 954–958; (b) E. M. V. Toth and G. R. Marshall, Int. J. Peptide Protein Res., 1990, 36, 544–550; (c) F. Badalassi, H.-K. Nguyen, P. Crotti and J.-L. Reymond, Helv. Chim. Acta, 2002, 85, 3090–3098.
14. N. Marmé, J.-P. Knemeyer, J. Wolfrum and M. Sauer, Angew. Chem. Int. Ed., 2004, 43, 3798–3801.
15. Review: A. Baruch, D. A. Jeffery and M. Bogyo, Trend. Cell. Biol., 2004, 14, 29–35.
16. Review: D. J. Maly, L. Huang and J. A. Ellman, ChemBioChem, 2002, 3, 16–37.
17. (a) K. Breddam and M. Meldal, Eur. J. Biochem., 1992, 206, 103–107; (b) M. Meldal, I. Svendsen, K. Breddam and F. I. Auzanneau, Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 3314–3318; (c) P. M. St. Hilaire, M. Willert, M. A. Juliano, L. Juliano and M. Meldal, J. Comb. Chem., 1999, 1, 509–523.
18. (a) N. A. Thornberry, T. A. Rano, E. P. Peterson, D. M. Rasper, T. Timkey, M. Garcia-Calvo, V. M. Houtzager, P. A. Nordstrom, S. Roy, J. P. Vaillancourt, K. T. Chapman and D. W. Nicholson, J. Biol. Chem., 1997, 272, 17907–17911; (b) J. H. Harris, B. J. Brackes, F. Leonetti, S. Mahrus, J. A. Ellmann and C. S. Craik, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 7754–7759; (c) J. L. Harris, P. B. Alper, J. Li, M. Reichsteiner and B. J. Backes, Chem. Biol., 2001, 8, 1131–1141; (d) T. A. Rano, T. Timkey, E. P. Peterson, J. Rotonda, D. W. Nicholson, J. W. Becker, K. T. Chapman and N. A. Thornberry, Chem. Biol., 1997, 4, 149–155; (e) C. Pinilla, J. R. Appel, P. Blanc and R. A. Houghten, Biotechniques, 1992, 13, 901–905; (f) A. F. Kisselev, M. Garcia-Calvo, H. S. Overkleeft, E. Peterson, M. W. Pennington, H. L. Ploegh, N. A. Thornberry and A. L. Goldberg, J. Biol. Chem., 2003, 278, 35869–35877.
19. (a) M. Bogyo, S. Verhelst, V. Bellingard-Dubouchaud, S. Toba and D. Greenbaum, Chem. Biol., 2000, 7, 27–38; (b) D. Greenbaum, K. F. Medzihradszky, A. Burlingame and M. Bogyo, Chem. Biol., 2000, 7, 569–581; (c) T. Nazif and M. Bogyo, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 2967–2972.
20. J. E. Sheppeck, II, H. Kar, L. Gosink, J. B. Wheatley, E. Gjerstad, S. M. Loftus, A. R. Zubiria and J. W. Janc, Bioorg. Med. Chem. Lett., 2000, 10, 2639–2642.
21. (a) C. M. Salisbury, D. J. Maly and J. A. Ellman, J. Am. Chem. Soc., 2002, 124, 14868–14870; (b) Q. Zhu, M. Uttamchandani, D. Li, M. L. Lesaicherre and S. Q. Yao, Org. Lett., 2003, 5, 1257–1260.
22. J.-L. Reymond, Food Technol. Biotechnol., 2004, 42, 265–269.
23. D. Lagnoux, E. Delort, C. Douat-Casassus, A. Esposito and J.-L. Reymond, Chem. Eur. J., 2004, 10, 1215–1226.
24. Introduction to Multivariate Analysis, ed. C. Chatfield and A. J. Collins, Chapman and Hall Ltd., London, 1983.
25. (a) P. Delepelaire and C. Wandersman, J. Biol. Chem., 1989, 264, 9083–9089; (b) H. Thomas and U. Baumann, J. Mol. Biol., 2001, 314, 187–193.
26. W. Bode, F. Grams, P. Reinemer, F. X. Gomis-ruth, U. Baumann, D. B. Mckay and W. Stocker, Adv. Exp. Med. Biol., 1996, 389, 1–11.

---

Footnote

† Electronic supplementary information (ESI) available: data for reference fragments, MS-traces of cocktail peptides and HPLC traces of cocktail references and protease fingerprints. See http://www.rsc.org/suppdata/mb/b4/b419446b/.

---