Send to printer »

Tutorials : Feb 1, 2009 (Vol. 29, No. 3)

Accelerating Protease Drug Development

Use of the REPLi FRET Peptide Library Facilitates Identification of Substrates
  • NIcholas Ede

Proteases are involved in the regulation of a wide variety of essential physiological processes, and their dysregulation has been implicated in a number of disorders including cardiovascular disease, rheumatoid arthritis, Alzheimer’s disease, and cancer. Hence, proteases and their substrates are increasingly viewed as valuable drug targets in disease treatment. Given the large number of proteases in the human genome (>560), however, the task of characterizing the biological function in vivo of all proteases, including the identification of all protease substrates, is daunting.  

Information relating to the specific residue requirements spanning the protease cleavage site in peptide substrates is invaluable in assisting the development of specific inhibitors and in identifying possible in vivo protein substrates. In addition, specificity information can be used to design highly sensitive and specific synthetic fluorogenic substrates and these in turn enable screening of compounds to identify small molecule inhibitors.

As a result, there is an urgent need for more rapid techniques of substrate discovery on a system-wide basis, with the capability of identifying new protease substrates directly from cells, tissues, and body fluids, and quantitating differences in substrate processing of low-abundant proteins as disease markers.

Protease Substrate Identification

Various methods have been employed to characterize protease specificity, generally consisting of the screening of chemically or biologically produced peptide libraries.  The early mixture-based peptide libraries provided both general applicability and speed, with digestion of the library mixtures followed by N-terminal sequencing giving specificity information C-terminal to the scissile bond (P' sites; Figure 1).

Such methods, however, preclude the continuous monitoring of proteolytic activity and do not provide information on the protease specificity N-terminal to the scissile bond (P sites). Libraries generated using solid-phase synthesis where peptides are tagged with a fluorogenic or chromogenic group that fluoresces/absorbs light after cleavage, allows the proteolytic activity to be monitored in real time. The lack of information on the primed subsite specificity (P' sites) hinders its application to endoproteases that recognize residues C-terminal to the scissile bond.

To overcome issues such as feasibility and deconvolution, associated with large peptide libraries, Mimotopes has developed and launched, in collaboration with GlaxoSmithKline and the University of Leeds a generic fluorescence resonance energy transfer (FRET) rapid endopeptidase profiling library (REPLi) as a tool for rapidly identifying protease substrates.

To keep the number of peptides relatively small while still representing the residue requirements for the largest number of proteases, similar amino acids are paired within a tripeptide core giving rise to a relatively small library of 3,375 peptides divided into 512 distinct pools, each containing only eight peptides. The variable central core is flanked with multiple Gly residues and an additional two Lys residues are added at the C-terminus to confer adequate solubility to peptides bearing hydrophobic variable sequences (Figure 1). 

Potentially problematic amino acids that have been removed from the selection set include Cys (potential for introducing disulphide bonds), His (not generally observed within substrates at sites of protease cleavage), Met (due to both its hydrophobicity and bulky nature being shared by Leu and Ile and its propensity to being oxidized), and Trp (interference with fluorescence signal due to abs.l280nm/em.l320nm). Gly is also omitted since functionalized amino acids furnish comparatively more information and because cleavage around flanking Gly residues can be detected.

The remaining 15 amino acids are grouped in matching pairs (Ala + Val, Arg + Lys, Asp + Glu, Asn + Gln, Leu + Ile, Ser + Thr, Phe + Tyr), while Pro is left as a single residue to allow access to any potential conformational information. The judicious choice of amino acid partners ensures that maximum SAR information can be derived from an initial result. This matching-pair design simplifies the deconvolution steps. The complete peptide library is synthesized using Mimotopes’ PepSets SynPhase Lantern technology and is provided for direct usage in 96-well plate format (Figure 2).

Library Validation

To validate the library, scientists at GSK and the University of Leeds profiled the REPLi library against representative members from each of the four mechanistic protease classes—trypsin, pepsin, matrix metalloprotease (MMP)-12 (macrophage elastase), MMP-13 (collagenase-3), calpain-1 (µ-calpain) and calpain-2 (m-calpain). 

The substrate specificities of calpains-1 and -2 are almost identical and few peptide substrates for either enzyme have been reported. Thus, it was of particular interest that the REPLi returned several substrates that were preferentially cleaved by either one or other enzyme. The variable core sequence of the peptide pool that was specifically cleaved by calpain-1 was -Pro-[Ile/Leu]-[Phe/Tyr]-, while that which was specifically cleaved by calpain-2 contained -[Ile/Leu]-[Asn/Gln]-[Phe/Tyr]- in the variable core. 

The peptide pool comprising the eight peptides with the core motif [Asn/Gln]-[Ile/Leu]-[Phe/Tyr] that returned the highest signal:background ratio with calpain-2 was subjected to full LC-MS characterization (Figure 3).

This type of analysis enables the identification of the cleavage site within the peptide (MS measurement) and aids the determination of the preferred substrate sequence within the pool (LC data). To identify the actual peptide sequences in the pool cleaved by calpain-2, all eight peptides were synthesized as discrete candidates and monitored in real-time assays. These allowed the concomitant determination of the nature of the optimal residues for each of the three variable positions and also provided kinetic data for each substrate (Figure 3).

As predicted, the kinetic characterization of single peptides was in complete agreement with LC-MS data obtained for the pool and showed that the activity was confined to specific sequences. Utilizing this information obtained from the REPLi, a highly sensitive and efficient fluorogenic substrate for calpain-2 was developed from a minimum number of experimental iterations. 

The thorough validation of this peptide library with representatives from each of the four mechanistic protease classes indicates that the REPLi will be useful for the rapid identification of substrates for multiple proteases and has potential applications in high-throughput screening.