Elsevier

Methods in Enzymology

Volume 501, 2011, Pages 237-273
Methods in Enzymology

Chapter twelve - Predicting Serpin/Protease Interactions

https://doi.org/10.1016/B978-0-12-385950-1.00012-2Get rights and content

Abstract

Proteases are tightly regulated by specific inhibitors, such as serpins, which are able to undergo considerable and irreversible conformational changes in order to trap their targets. There has been a considerable effort to investigate serpin structure and functions in the past few decades; however, the specific interactions between proteases and serpins remain elusive. In this chapter, we describe detailed experimental protocols to determine and characterize the extended substrate specificity of proteases based on a substrate phage display technique. We also describe how to employ a bioinformatics system to analyze the substrate specificity data obtained from this technique and predict the potential inhibitory serpin partners of a protease (in this case, the immune protease, granzyme B) in a step-by-step manner. The method described here could also be applied to other proteases for more generalized substrate specificity analysis and substrate discovery.

Introduction

The Serpin superfamily is structurally characterized by a well-conserved fold consisting of 9 α-helices and 3 β-sheets as well as a 20–25 amino acid region named the reactive center loop (RCL, sometimes also referred to as reactive site loop) (Fig. 12.1A). A large body of biochemical and structural studies (with over a hundred serpin structures solved) has shown that this metastable fold is under considerable strain in the native state. This makes serpins sensitive to mutations promoting misfolding and resulting in serpin deficiency. The native, inhibitory complexes, cleaved, latent, and polymeric forms have been characterized to atomic resolution by protein crystallography techniques and differ primarily by the structure of the RCL (Whisstock et al., 1998). We will here mostly focus on the native, cleaved, and inhibitory states of inhibitory serpins (see Law et al., 2006, Silverman et al., 2001, Silverman et al., 2004, Whisstock and Bottomley, 2006, Whisstock et al., 2010 for a full review).

Inhibitory serpins (Serine protease inhibitors) undergo a remarkable conformational change upon protease contact, switching from the native to the inhibitory complex (inhibitory pathway). The rearrangement forms an absolute requirement for the inhibitory function, where the serpin acts as a pseudosubstrate for the target protease. In the native state, the RCL adopts a disordered, solvent-exposed conformation accessible to the target protease. Situated on top of the serpin scaffold, it offers as bait a pseudosubstrate for the protease (Carrell et al., 1991), which initially docks onto the RCL (Michaelis complex) (Fig. 12.1B). The RCL contains the P1–P1′ scissile bond cleaved by the target protease. Upon its cleavage, a reversible acyl-enzyme intermediate is formed, while the inhibitory serpin rapidly undergoes the stressed-to-relaxed transition: the RCL inserts into β-sheet A and adopts a β-strand conformation (Huntington et al., 2000). The strain under which the serpin scaffold lies is thus lifted allowing the relocation of the protease to the opposite pole of the serpin, concomitant to the insertion of the amino-terminal part of the RCL into the β-sheet. This results in improved thermostability of the final serpin–protease complex, where the protease is covalently trapped by the serpin and partially disordered (Fig. 12.1C). Its active site is critically distorted, which prevents further hydrolysis of the intermediate and release of the protease. With the irreversible cleavage of the RCL, serpins achieve efficient removal of the target protease and are thus named “suicide” or “single use” inhibitors.

The RCL is key to serpin inhibitory specificity and extends from P15 to P3′ in most serpins. It is the most variable part of the molecule and bears the P1 residue, which exerts great influence over the target protease specificity. Mutation of this residue can alter the serpin inhibitory spectrum and promote disease. When the serpin α-1-antitrypsin is affected by the naturally occurring P1 Met to Arg mutation, it no longer inhibits elastase but targets coagulation cascade proteases, leading to hemophilia (Owen et al., 1983).

While the first crystal structure of a Michaelis complex revealed residues P4–P3′ are implicated in direct protease/RCL contact (Ye et al., 2001; Fig. 12.1B, boxed insert), more recent structures have highlighted the importance of residues outside the P4–P3′ positions in binding the target protease (Li et al., 2004, Li et al., 2008). Serpin exosites extending to the prime-side (P′-side) of the RCL can enhance protease interaction and seem to be a feature of serpins with more than one target (Whisstock et al., 2010). In some other serpin/protease interactions, exosite contacts also play critical roles to improve the poor ability of target proteases to recognize certain RCL sequences and allow interactions with multiple target proteases (Börner and Ragg, 2008, Huntington, 2006, Johnson et al., 2006, Li et al., 2008, Whisstock et al., 2010).

Macromolecular cofactors can also modulate serpin activity. Binding to specific sites of the protein, they can affect RCL flexibility and accessibility, thus altering serpin inhibitory function. In the case of antithrombin, the RCL is partially inserted into β-sheet A and is thus inefficient at binding target proteases. High-affinity binding of heparin provokes exposure of the RCL to the solvent and the protease, enabling antithrombin to efficiently target proteases (Li et al., 2004), while concomitantly an exosite for factor Xa is exposed on the body of the serpin (Huntington, 2006). In another example, the active conformation of plasminogen activator inhibitor 1 is maintained by interaction with vitronectin, which slows the rate of insertion of the uncleaved RCL into β-sheet A. Spontaneous RCL insertion leads to the irreversible formation of the inactive latent conformation (Keijer et al., 1991).

The kinetic nature of the inhibitory mechanism can lead serpins to follow the substrate pathway, where the serpin fails to capture the protease and is rendered unable to complete inhibition (Fig. 12.1D). If the RCL resembles the conformation of the protease's natural substrate too closely, cleavage of the intermediate will proceed before the protease can be relocated and trapped. Hence the substrate pathway leads to the insertion of the RCL in the cleaved, inactive state, while the active protease is released. The kinetic parameters that define the tendency to follow the inhibitory or the substrate pathways (kass, association constant; SI, stoichiometry of inhibition) can also be modulated by the addition of cofactors such as glycosaminoglycans to the serpin and/or protease (Olson et al., 1992, Pike et al., 2005).

Protease-controlled site-specific proteolysis is one of the most important posttranslational modifications. The key to understanding the physiological role of a protease is to identify the repertoire of its natural substrate(s) or to identify its substrate specificity (Song et al., 2011). The specificity of proteases vary, primarily depending on their active sites, which display selectivity ranging from preferences for a number of specific amino acids at defined positions to more generic proteases with limited discrimination at one position. In addition to the primary amino acid sequence of the substrate, specificity is also influenced by the three-dimensional conformation of the substrate (secondary and tertiary structures). In particular, proteases preferentially cleave substrates within extended loop regions, while residues that are buried within the interior of the protein substrate are clearly inaccessible to the protease active site. Finally, cleavage is regulated by the temporal and physical colocation of the protease and substrate. In particular, some proteases are sequestered within specific compartments, with limited access to proteins, while others are able to cleave multiple substrates in different physiological compartments (Song et al., 2011).

This chapter attempts to provide instruction in the analysis of the extended substrate specificity of proteases with a view to using the technique to mine for potential substrates. This technique has been previously used to examine the substrate specificities of a number of proteases, including the immune protease granzyme B (Kaiserman et al., 2006a, Kaiserman et al., 2006b). A potent cytotoxin, granzyme B is the most widely studied member of the granzyme family of serine proteases. Granzymes are expressed by cytotoxic lymphocytes, which transfer the contents of specialized secretory lysosomes (including granzyme B) into the cytoplasm of target cells, whereupon granzyme B cleaves several specific substrates (including Bid and caspase-3), inducing the targeted cell to undergo apoptosis (for review, see Trapani and Sutton, 2003).

Human granzyme B has a murine counterpart which, upon superficial examination, appears to be very similar. They share significant sequence homology at the amino acid level and both hydrolyze substrates after an acidic residue (mainly aspartic acid) (Odake et al., 1991). Indeed, many studies have assumed these two molecules to be functionally identical, and thus they are often used interchangeably in experimental systems. However, murine granzyme B cleaves the primary proapoptotic substrate Bid with a much lower efficiency than human granzyme B (Kaiserman and Bird, 2005, Kaiserman et al., 2006a, Kaiserman et al., 2006b). This observation, along with data demonstrating differences in the cytotoxic potential between human and murine granzyme B, suggests that (at least on a functional level) these two molecules are not as similar as previously believed.

An examination of both human and mouse granzyme B molecules using the substrate phage display, and some of the bioinformatics techniques described in this chapter, reveals differences in the extended substrate specificities at two important subsites (Kaiserman et al., 2006a, Kaiserman et al., 2006b) which could account for the inability of mouse granzyme B to efficiently cleave Bid. While these differences do not completely explain the comparatively reduced cytotoxicity of mouse granzyme B, it is convincing evidence that the origin of any granzyme B molecule should be matched with other reagents when designing in vitro or in vivo experimental systems in which the specific function of granzyme B may influence the outcome.

This chapter is intended to serve as an aid to researchers who are interested in employing phage display technology to characterize the extended substrate specificity of a protease (granzyme B is used as an example in this work), building the substrate specificity models using the PoPS (prediction of protease specificity) program (Boyd et al., 2005) and further applying these models to predict specific serpin inhibitors containing potential substrate sequences in their RCL regions. We first outline the basic principles and requirements of phage display substrate technology. Next, we describe the methodology of substrate phage library design, preparation, biopanning, and analysis. Then we discuss bioinformatics approaches to analyze and determine the substrate specificity of granzyme B based on the data generated from the phage display selection. Finally, we illustrate how to build the specificity models and make predictions for novel putative substrates of granzyme B.

Section snippets

Phage Display Methods

In the following sections, we will describe how to generate and probe a substrate phage library to characterize the substrate specificity of a protease, how to use the data generated from this system to build the substrate specificity model, and how to make a prediction of candidate serpins with the optimal substrate specificity. We will examine a data set generated for the human immune protease, granzyme B, as an example to illustrate our methodology. An overview of the steps followed is

Bioinformatics approaches for predicting protease substrates

Although proteases represent around 2% of the gene products across all the genomes, the complete repertoire of their target substrates and inhibitors remain to be fully characterized. In this context, bioinformatic modeling approaches that are capable of predicting putative cleavage sites of proteases and can provide experimentally testable information are invaluable tools. In the past few decades, a number of bioinformatics approaches have been developed to facilitate the in silico discovery

Concluding Remarks and Perspective

In-depth understanding of the substrate specificity of proteases is essential for obtaining deep insights into their function and is a prerequisite for the development of specific inhibitors to control their activities. Identification of the cognate serpin inhibitors of proteases is essential for the development of targeted therapies of protease-controlled pathogenesis. Substrate phage display is a powerful and invaluable tool for protease substrate specificity profiling and can aid in

Acknowledgments

We thank the members of the Whisstock and Bird laboratories for support and discussion. This work was supported by grants from the National Health and Medical Research Council of Australia (NHMRC), the Australian Research Council (ARC), and the Chinese Academy of Sciences (CAS). J. S. is an NHMRC Peter Doherty Fellow and is also supported by the Hundred Talents Program of CAS. J. C. W. is an ARC Federation Fellow and an honorary NHMRC Principal Research Fellow.

References (60)

  • I. Schechter et al.

    On the size of the active site in proteases. I. Papain

    Biochem. Biophys. Res. Commun.

    (1967)
  • G.A. Silverman et al.

    The serpins are an expanding superfamily of structurally similar but functionally diverse proteins. Evolution, mechanism of inhibition, novel functions, and a revised nomenclature

    J. Biol. Chem.

    (2001)
  • J.A. Trapani et al.

    Granzyme B: Pro-apoptotic, antiviral and antitumor functions

    Curr. Opin. Immunol.

    (2003)
  • P. Van Damme et al.

    Analysis of protein processing by N-terminal proteomics reveals novel species-specific substrate determinants of granzyme B orthologs

    Mol. Cell. Proteomics

    (2009)
  • J. Verspurten et al.

    SitePredicting the cleavage of proteinase substrates

    Trends Biochem. Sci.

    (2009)
  • J.C. Whisstock et al.

    Molecular gymnastics: Serpin structure, folding and misfolding

    Curr. Opin. Struct. Biol.

    (2006)
  • J. Whisstock et al.

    An atlas of serpin conformations

    Trends Biochem. Sci.

    (1998)
  • A. Zhou et al.

    The serpin inhibitory mechanism is critically dependent on the length of the reactive center loop

    J. Biol. Chem.

    (2001)
  • D.T. Barkan et al.

    Prediction of protease substrates using sequence and structure features

    Bioinformatics

    (2010)
  • C. Benarafa et al.

    The ovalbumin serpins revisited: Perspective from the chicken genome of clade B serpin evolution in vertebrates

    Proc. Natl. Acad. Sci. USA

    (2005)
  • H.M. Berman et al.

    The protein data bank

    Nucleic Acids Res.

    (2000)
  • S.E. Boyd et al.

    PoPS: A computational tool for modeling and predicting protease specificity

    J. Bioinform. Comput. Biol.

    (2005)
  • S.E. Boyd et al.

    Cooperative effects in the substrate specificity of the complement protease C1s

    Biol. Chem.

    (2009)
  • R.W. Carrell et al.

    Mobile reactive centre of serpins and the control of thrombosis

    Nature

    (1991)
  • C.T. Chen et al.

    Protease substrate site predictors derived from machine learning on multilevel substrate phage display data

    Bioinformatics

    (2008)
  • E.R. Cole et al.

    Multiple specificity of thrombin for synthetic substrates

    Nature

    (1967)
  • B.G. Condron et al.

    Frameshifting in gene 10 of bacteriophage T7

    J. Bacteriol.

    (1991)
  • S.R. Coughlin

    Thrombin signaling and protease-activated receptors

    Nature

    (2000)
  • S.P. Cullen et al.

    Human and murine granzyme B exhibit divergent substrate preferences

    J. Cell Biol.

    (2007)
  • S.E. Cwirla et al.

    Peptides on phage: A vast library of peptides for identifying ligands

    Proc. Natl. Acad. Sci. USA

    (1990)
  • Cited by (9)

    • Preparation and crystallization of a novel oral bacterial serine protease inhibitor Tannerpin

      2021, Journal of Shanghai Jiaotong University (Medical Science)
    View all citing articles on Scopus
    1

    These authors contributed equally to this work.

    View full text