Targeting Proteases in Cardiovascular Diseases by Mass Spectrometry-Based Proteomics

Abstract

Proteases hydrolyze peptide bonds, thereby controlling the function of proteins and peptides on the posttranslational level. In the cardiovascular system, proteases play pivotal roles in the regulation of blood pressure, coagulation and other essential physiological processes. Accordingly, proteases are prime targets for therapeutic interventions and diagnostics. Proteases are part of complex proteolytic networks comprised of enzymes, inhibitors, activators, substrates and cleavage products. Analyzing these networks on a system-wide level is essential to understanding cardiovascular function and how dysregulation can lead to pathological conditions. Mass spectrometry-based quantitative and dynamic proteomics approaches are leading the way to enhance our knowledge of proteolytic networks such as the renin-angiotensin-system. Here, we critically review proteomics tools utilized in protease biology and provide an overview on how these methods can be used to characterize and validate protease function.

Introduction

Proteases catalyze the hydrolysis of peptide bonds, which results in the cleavage of protein and peptide chains and thereby lead to an irreversible change of protein structure. The fundamental nature of this process makes proteolysis a powerful post-translational modification that can control protein function and abundance. Digestive proteases in the gastrointestinal tract break down proteins fairly indiscriminately, while peptidases involved in cell signaling catalyze very specific cleavage reactions to regulate the abundances of bioactive peptides. Uncontrolled proteolysis could have disastrous physiological consequences, therefore a multitude of mechanisms exist to tightly regulate proteolytic processing. One of the more basic regulatory principles is substrate specificity, in which the three-dimensional structure of the protease determines which substrates are accessible to the active site. Other regulatory mechanisms include the activation of proteases from inactive precursors (i.e., zymogens) and the limitation of protease activities to specific pH ranges and compartments (e.g., lysosomal proteases). Another regulatory element is the spatial and temporal interplay of proteolytic network components: Devoid all required factors (proteases, substrates and their respective activators and inhibitors) present, reactions may not occur (Figure 1).

Figure 1.

A simplified view of the essential functional relationships between members of a proteolytic network. Activators and inhibitors can modulate the activity state of proteases. For cleavage reactions to occur, the spatial and temporal distributions of activated proteases and substrates need to overlap.

The renin-angiotensin-system can serve as an example how a proteolytic network regulates a physiological process¹: Briefly, angiotensinogen is cleaved by renin to produce angiotensin I, which in turn is cleaved by angiotensin-converting enzyme (ACE) to form the effector peptide angiotensin II. Binding of angiotensin II to the AT1 receptor mediates vasoconstriction, while binding to AT2 results in vasodilation (Figure 2). The renin-angiotensin-system tightly controls arterial blood pressure and assures the uninterrupted perfusion of vital organs with oxygen and nutrients. Recent discoveries of alternative angiotensin-processing pathways illustrate that the renin-angiotensin-system is more complex than previously thought². Dysregulation of the renin-angiotensin-system can lead to pathological conditions such as hypertension, which is a major risk factor for congestive heart failure, stroke and myocardial infarction. Due to their pivotal roles, proteolytic enzymes have emerged as prime targets for the pharmaceutical treatment of cardiovascular diseases³. The increasing awareness of the modulatory effect of the biological context mandates that protease function is analyzed on a system-wide level. Mass spectrometry (MS)-based proteomics provides a unique technological platform to capture the complexity and dynamic nature of proteolytic networks. Here, we review the current state of proteomics approaches utilized in protease biology and highlight strategies particularly suited to characterize protease activities in the context of cardiovascular disorders.

Figure 2.

Processing of angiotensin peptides by ACE, ACE2 and neprilysin as part of the rennin-angiotensin system. Renin cleaves angiotensinogen to produce angiotensin I. ACE converts angiotensin I to angiotensin II. In a second processing axis, angiotensin I is cut by ACE2 resulting in Ang [1–9], which is cleaved by either ACE or neprilysin (NEP) to produce Ang [1–7] (bracketed numbers refer to the amino acid positions within the peptide sequences). Ang [1–7] can also result from processing of angiotensin I by NEP or angiotensin II by ACE2. Binding of angiotensin II to the AT1 receptor activates vasoconstriction. In contrast, binding to the AT2 receptor mediates vasodilation, which can also be initiated by the binding of Ang [1–7] to the Mas receptor.

Mass spectrometry-based methods to unravel proteolytic networks

The renin-angiotensin-system demonstrates how MS-based proteomics can help filling the need for a detailed understanding of proteolytic networks. While renin and ACE were the original proteases of the renin-angiotensin-system, the updated view includes additional proteases such as ACE2, chymase, neprilysin and aminopeptidases (A and N) ^4–7. Likewise, additional bioactive peptides such as the angiotensin metabolites Ang III [2–8], Ang IV [3–8] and Ang [1–7] that were previously considered functionally inactive, are now included⁸. MS-based technologies played a leading role in the identification of these novel system constituents.

Historically, protease function is assessed on an individual basis by in vitro enzyme assays after biochemical purification. The advent of new analytical technologies makes it now feasible to investigate protease activity and function in complex environments^9–13. Generally, these approaches can be categorized based on whether they are designed to (i) identify constituents of the proteolytic network (proteases, substrates), (ii) screen for protease inhibitors/activators/modulators or (iii) characterize the dynamics of proteolytic processing. In Table 1, we provide an overview of the most commonly applied strategies in the field and which aspects of proteolysis they address. To choose the appropriate method, it is essential to define what aspect of protease biology needs to be addressed. Figure 1 depicts the functional relationships between members of proteolytic networks and can serve as a guide to develop research strategies: Is the protease of interest catalytically active? How does the zymogen differ from the active form of the protease? Which molecules can modulate the activity and substrate specificity of the protease? What are the endogenous substrates of the protease? What are the spatial and temporal distributions of the protease and other components of the proteolytic network? Many of these questions can be readily answered by online resources: the MEROPS database (http://merops.sanger.ac.uk) curates published information about peptidases, their substrates and inhibitors and offers indications about overall substrate specificity²⁵; PMAP (http://www.proteolysis.org) combines five databases (ProteaseDB, SubstrateDB, CutDB, ProfileDB and PathwayDB) and a computational toolkit (including cleavage site predictions) to create an integrated reasoning environment to analyze proteolytic networks²⁶.

Table 1.

Commonly applied techniques to study proteolytic reactions.

Method	Description Application	Protease	Substrate	Inhibitor	Process
General techniques for protease research
Yeast two-hybrid systems	Yeast libraries carrying cloned open reading frames (ORF)14 Sequencing of positive PCR amplification products Identification of protein interactions on a large scale Two target ORFs are analyzed per experiment	•	•	•
Phage-display	Affinity selection of clones and screening of peptide libraries9 DNA sequencing of remaining amplification products High throughput screening of protein interactions and inhibitors One target protein or peptide is analyzed per experiment	•	•	•
2D differential gel electrophoresis (2D DIGE)	Gel-based protein separation of fluorescent labeled samples15 Proteins are imaged by fluorescence and identified by MS Quantitation of proteins and posttranslational modifications Comparison of two samples with internal standards	•	•
Multidimensional protein identification technique (MudPIT)	Peptide products are separated by SCX and RP-HPLC13 Identification of peptide products by MS Comparison of an unlimited number of samples	•	•
Isotope-coded affinity tag (ICAT)	Labeling of cysteine-residues in proteins followed by digestion16 Relative quantitation of peptide products by MS Comparison of two samples		•
Isobaric tags for relative and absolute quantitation (iTRAQ)	Isobaric amine-specific tagging of peptide products17 Relative and absolute quantitation of protease activity by MS Comparison of up to eight samples		•
Stable isotope labeling with amino acids in cell culture (SILAC)	Expressed proteases/peptide products labeled by amino acids18 Identification and quantitation of peptide products by MS Comparison of an unlimited number of samples	•
Multiple reaction monitoring (MRM)	Multiple reaction monitoring of known peptide fragments by targeted MS19 Absolute quantitation of known peptides Comparison of an unlimited number of samples		•	•
Techniques specifically designed for protease research
Cellular libraries of peptide substrates (CLiPS)	Combinatorial approach to measure substrate hydrolysis11 Identification of substrates by quantitative screening of whole-cell fluorescence		•		•
Positional scanning synthetic libraries	Screening of tetra-peptide libraries by proteolysis-dependent signal intensities10 Screen for P1–P4 substrate specificities		•
Colloidal barcoding bead- based protease profiling	Screening of combinatorial libraries using polyelectrolyte-coated fluorescent silica reporter particles20 Identification of consensus proteolytic cleavage sites	•	•	•	•
Near infrared (NIR) fluorogenic reporters	NIR fluorescence signal upon cleavage of protease-sensitive peptide linkers21 In vivo imaging and quantitation of protease activities	•	•	•	•
C- and N-term enrichment of cleavage products	Negative or positive enrichment of C- or N-terminal peptide cleavage products22 Identification of proteolytic peptides and cleavage sites Comparison of two samples		•
Activity-based probes (ABPs)	Chemical probes with affinity and fluorescent tags23 Report on the structure and reactivity of enzyme active sites in cells and tissues	•		•	•
Proteinase activity labeling employing 18O- enriched water (PALeO)	18O-labeling of proteolytic peptides during hydrolysis24 Quantitation and identification of protease activity, peptide substrates and cleavage products Comparison of an unlimited number of samples	•	•	•	•

MS-based proteomics has emerged as the premier tool to identify and quantify proteins and peptides^27–29. Briefly, in a typical proteomics identification workflow, proteins are separated by one- or two-dimensional gel electrophoresis, digested by exogenous proteases (i.e., trypsin). Resulting peptide fragments are recovered, fractionated by reverse-phase chromatography and their molecular masses measured by MS. Selected peptides are fragmented in tandem MS experiments and resulting fragmentation data submitted to search engines (i.e., Mascot) that match them to fragmentation patterns predicted from protein databases³⁰. Alternatively, in a peptidomics workflow the digestion step is omitted and peptide products formed by endogenous proteases are directly isolated³¹. Peptide sequence assignments are particularly challenging in peptidomics-type experiments, however recent advances in bioinformatics are starting to approach this problem³².

How to develop a quantitative view of proteolytic processing?

While MS readily provides protein and peptide identities, MS-measurements are inherently poorly quantitative³³. Accurate quantitation of protein and peptide expression levels is a prerequisite to capture the dynamic nature of proteolysis and to gain functional insights. Quantitation by MS is achieved by either label-free or stable isotope labeling methods. Measurements of chromatographic ion intensity (e.g., peak areas) and spectral counting are the most commonly applied label-free approaches³⁴. Label-free methods provide relative quantitation for an unlimited number of samples. Stable isotope labeling strategies, in contrast, can provide relative and absolute quantitation, however the specifics of the labeling reactions can limit the number of samples interrogated. Over recent years, a broad variety of stable isotope labeling methods have been developed^34,35. Here, we will focus on the most prominent examples that have been utilized in the context of protease biology. In general, the presented techniques are capable of simultaneously identifying and quantifying novel and known proteases and their substrates. They can also be used to test the efficacy of protease inhibitors.

Isotope-coded affinity tags (ICAT) are one of the best-known techniques to measure protein abundances. ICAT reagents are comprised of three functional components: (i) a reactive group specific toward cysteinyl residues, (ii) a stable isotope label, and (iii) a biotin affinity tag¹⁶. The affinity tag allows for the selective enrichment of cysteine containing peptides, thus reducing sample complexity. The stable isotope label introduces either a light or heavy tag, which in turn allows for the comparative analysis of protein expression levels across two states. After labeling with ICAT reagents, two samples are combined, digested and cysteine-containing peptides enriched by avidin affinity chromatography. LC-MS/MS analysis yields both protein quantity and identity. The ICAT approach has been successfully used to identify novel protease substrates in complex cellular environments and to quantify protease activity³⁶. The exclusive reliance on cysteine-containing peptides limits the applicability of the ICAT approach as a general quantitation approach.

This shortcoming has been addressed by the next generation of chemical labeling strategies that tag peptide N-termini and lysine side chains using N-hydroxy-succinimide (NHS) chemistry. The TMT (tandem mass tag)³⁷ and iTRAQ (isobaric tags for relative and absolute quantitation)³⁸ strategies share as an important design feature an isobaric stable isotope moiety, which renders differentially labeled samples indistinguishable during chromatographic and mass spectrometric analyses. Only upon MS/MS fragmentation, low molecular weight reporter ions are released and their relative ion abundances can be used for quantitation. Currently, there are four and eight reporter ions available for iTRAQ^38,39 and two and six for TMT^37,40, each allowing for multiplexed analysis in single LC-MS/MS experiments. iTRAQ labeling has been successfully applied to quantify peptides generated by endogenous proteases⁴¹ and to monitor protease activity in cell cultures⁴².

SILAC (stable isotope labeling by amino acids) has been developed as a metabolic labeling method alternative to chemical labeling approaches. SILAC has been applied in cell culture systems¹⁸ and more recently in entire organisms⁴³. SILAC relies on the in vitro incorporation of essential amino acids with substituted stable isotope nuclei (e.g., Arg and Lys labeled with ¹³C, ¹⁵N). Heavy and light versions of the amino acids are incorporated into every protein in the system, which, in turn, allows for comparative expression analyses. SILAC has been used to determine which substrates are trapped by a catalytically inactivated protease⁴⁴ and more recently to quantify Granzyme B catalyzed proteolysis⁴⁵. Compared to chemical derivatization strategies, SILAC offers more efficient and reproducible labeling.

High levels of sample complexity and large cleavage products are a challenge for proteomic analyses. To reduce complexity and determine cleavage sites within large proteins, investigators developed strategies to enrich for N-terminal cleavage products that are of interest⁴⁶. In positive enrichment methods, α-amino groups that are newly formed by proteolysis are captured by chemical or enzymatic processes^{47, 48}. In depletion methods, internal peptides are removed by either altered chromatographic properties²² or by chemical means⁴⁹. Recently, a positive enrichment strategy termed C terminomics has been introduced for the until then elusive C-terminal cleavage products⁵⁰. Combining N- and C-terminal peptide data provides complementary information and increased coverage of cleavage sites. However, such enrichment steps considerably extend sample processing and analysis time.

Another strategy to address sample complexity is targeted MS. These types of experiments are typically performed on triple quadrupole instruments operating in multiple reaction monitoring (MRM) mode. MRM assays determine the abundances of peptide analytes by measuring the intensities of their characteristic MS/MS fragment ions and can be used in relative and absolute quantitation mode. Compared to other MS approaches, MRM provides higher sensitivity and specificity²⁸. MRM assays have been used to study the pharmacokinetics of enalapril, an ACE inhibitor used in the treatment of hypertension and congestive heart failure⁵¹. The MRM approach was also used to simultaneously quantify seven angiotensin II receptor antagonists in plasma of patients undergoing cardiovascular treatment⁵². Despite their broad usage in pharmacological studies, MRM assays have so far not been employed in protease research, which can partially be attributed to the laborious assay development that requires optimized LC and MS conditions for each analyte¹⁹.

How to determine the substrate specificity and activity state of proteases?

Proteases are typically low abundant molecules, which makes their functional characterization by MS-based workflows challenging. Therefore, activity-based probes (ABPs) have been devised as chemical warheads directed at the active sites of enzymes. ABP-based profiling has been used to detect and capture the subset of catalytically active proteases in complex biological systems and to give insights into the catalytic mechanism and substrate specificity of enzymes⁵³. ABPs report on the functional state of proteases by selectively labeling enzymes that share active site features (affinity, reactivity). Currently, ABPs have been developed for serine-, cysteine-, aspartyl- and metallo-proteases²³. One of the most important design features of ABPs is their ability to distinguish active proteases from inactive precursors (zymogens)⁵⁴ or proteases whose activity is modulated by other mechanisms⁵⁵. ABP-profiling thereby provides functional information that is beyond standard protein expression data. ABPs with various combinations of affinity- and fluorescent tags have been developed to study diseases in vivo^{56, 57}. In conjunction with downstream analytic technologies like MS, ABP-profiling offers the unique advantage to enrich for proteome fractions that share functional properties. ABP-profiling can generally be classified between (i) gel-based methods and (ii) LC-MS strategies^{58, 59}. The LC-MS-based platform uses biotin-tagged probes that capture intact enzymes on streptavidin beads. On-bead digestion and subsequent LC-MS analysis provide protein identification⁶⁰. In the bottom-up variation of the strategy, the enrichment step occurs on the peptide level after all proteins have been digested⁶¹. Shortcomings of ABP-profiling include the limited specificity and availability of functional probes. Also, some of the probes are toxic and cannot be used in vivo. Finally, ABPs inactivate captured proteases and therefore enzyme function cannot be further studied.

How to gain insights into the dynamics of proteolytic processing?

In contrast to the methods described above that primarily aim at the identification of protease network components, few experimental approaches exist that capture the dynamics of proteolytic processing⁶². Classically, chromo- or fluorogenic protease cleavage assays are used to measure the kinetics of individual proteolytic reactions. Fluorescence-based assays rely on the activation or dequenching of fluorophores upon proteolytic cleavage. The design of such assays requires precise knowledge of the targeted cleavage sites, so that appropriate constructs can be synthesized that specifically report substrate-to-product conversions⁶³. Introduction of fluorescent reporter functionalities may alter substrate-enzyme interaction kinetics. Also, cleavage-assays are likely to fall short when complex proteolytic reactions are being investigated, i.e. when a protease recognizes multiple, alternate cleavage sites on the same substrate molecule. MS can fill this critical technological gap by its ability to detect and identify unknown peptide metabolites on a large-scale in a virtually unbiased way. MS-based monitoring of enzymatic reactions has been described for electrospray and MALDI-ionization techniques⁶⁴. For example, in the MALDI-MES strategy (MS-assisted enzyme screening)⁶⁵ endogenous proteases are captured from crude samples and immobilized on functionalized beads. The beads are subsequently incubated with substrate solutions, whose compositions are monitored over time by MALDI-MS analyses. Enzyme immobilization and the use of separate substrate solutions minimize interferences from the biological matrix (e.g., salt, proteins) during MS-analysis. Villanueva et al. showed that even against complex biological backgrounds (i.e., serum) ex vivo incubation of endogenous proteases and substrates can yield characteristic protease activity signatures that can be used as cancer biomarkers⁶⁶. In the PALeO approach (protease activity labeling employing ¹⁸O-enriched water)²⁴, ex vivo incubations occur in the presence of ¹⁸O-enriched water. Hydrolysis of peptide bonds results in the concomitant incorporation of solvent ¹⁸O-atoms into the C-termini of nascent cleavage products, which can be readily detected by MS based on their characteristic isotope patterns⁶⁷. Interestingly, some substrate-protease interactions extend beyond the initial cleavage reaction: Cleavage products formed by serine proteases (e.g., trypsin) have been shown to rebind to the protease and reform acyl-enzyme intermediates. In the presence of H₂¹⁸O, hydrolysis of the acyl-enzyme intermediate facilitates the incorporation of a second ¹⁸O-atom into the C-terminus⁶⁸. ¹⁸O-based strategies such as PALeO can therefore provide additional information regarding aspects of protease-substrate interaction and enzymatic mechanism. The ¹⁸O-label allows positively selecting cleavage products and disregarding signals derived from the biological matrix background. Strategies such as PALeO are important steps towards mapping proteolytic networks in more global and dynamic ways. Studying complex proteolytic networks, however, remains an analytical challenge and accurate quantitative information is needed. Such efforts can be aided by providing non-degradable quantitation standards (e.g., all-d-amino acid peptides) and defined protease substrates⁶⁹. Stable isotope labeling of these peptides (e.g., by acid-catalyzed ¹⁸O-exchange) allows to track their degradation in complex mixtures⁷⁰.

The emerging field of dynamic proteomics promises to greatly facilitate the elucidation of complex proteolytic networks. Already, there is ample evidence that such approaches can shed new light into even well studied subject areas such as the renin-angiotensin-system. Using MALDI-MES, Schlüter et al. determined that renal Cathepsin G can generate angiotensin II⁷¹. Using the PALeO-assay, we demonstrated that endothelin-converting enzyme-1 (ECE-1), a member of the neprilysin protease family, can also generate angiotensin II (Figure 3) (M. Hardt, unpublished data, 2011). ECE-1-mediated peptide processing included additional cleavage reactions that could only be resolved by temporal analysis⁷². Further, the data reconfirmed that pH conditions influence the substrate specificity and activity profile of ECE-1⁷³. The existence of alternative angiotensin II biosynthesis pathways illustrates that the renin-angiotensin-system is far more complex than previously thought.

Figure 3.

PALeO, a dynamic MS-based protease assay, shows that ECE-1 converts angiotensin I to multiple biologically active metabolites including angiotensin II. The waterfall-plot shows the subsequent generation of Ang[1–9], angiotensin II and Ang[1–7] at extracellular pH (7.4). Red letters in the insert specify sites of proteolytic ¹⁸O-incorporation. Blue arrows indicate novel cleavage sites detected by PALeO, while the black arrow denotes the previously known cut site.⁶⁸.

Future perspectives

Proteomics and peptidomics technologies continue to positively impact protease research by aiding in the identification and quantitation of proteases, their substrates and inhibitors. Advances in MS-instrument technology that improve mass accuracy and sensitivity will further increase identification rates⁷⁴. Advances in MS-quantitation methods will help shift the analytical focus from static to dynamic measurements and assist in moving the field from identification- to activity-based workflows. In addition, improved quantitation techniques will enhance comparative studies that, for example, investigate the effect of perturbations (e.g., pathological conditions) on proteolytic networks.

Due to their key roles in diseases, proteases have tremendous potential as therapeutic targets and biomarkers. Taking advantage of the catalytic properties could lead to more specific and sensitive diagnostic tools compared to what is achievable with non-catalytic biomarkers. Proteases such as ACE are prime examples for how proteolytic enzymes have been targeted in drug discovery³. MS provides valuable tools to these endeavors. However, it is important to keep in mind that a single analytical platform is unlikely to explain complex and dynamic proteolytic systems by themselves. In vivo validation, such as classical overexpression and knock out experiments, are needed to confirm the biochemical roles of proteases. Novel live imaging probes are changing the validation process. Near infrared (NIR) fluorescence that penetrates through tissue offers the ability to noninvasively monitor the spatial and temporal distributions of enzyme activities²¹. NIR-probes have been used to visualize specific proteolytic cleavages in intact organisms⁷⁵ and validate in vivo targets of protease inhibitors⁷⁶. NIR-based imaging has an immense potential for revolutionizing the drug development process and clinical practice. Knowledge gained from activity-based and dynamic proteomics can accelerate the development of these exciting technologies.

Conclusion

MS-based technologies are increasingly recognized as methods of choice for studying proteolytic enzymes. Here, we provided an overview of the most prominent examples in the field and illustrated how these methods contribute to cardiovascular research. In the study of the renin-angiotensin-system, system-wide proteomics analyses were essential to discover alternative pathways generating vasodilatory and vasoconstrictive peptides. Similarly, activity-based proteomics workflows provided unprecedented details of the dynamics of proteolytic processing that assisted in the discovery of novel pharmaceutical targets. Despite these success stories, it is important to consider that proteomics approaches have shortcomings: Most quantitative MS-strategies are technically limited to small sample numbers, which renders them less suitable to clinical studies with large cohorts. Emerging strategies such as targeted MRM-assays can increase throughput but at the cost of the discovery research component. Single-handedly, none of the presented MS-technologies provides a complete picture of complex protease networks. Likewise, MS-based techniques are not directly suitable for in vivo measurements. These limitations postulate the careful integration of MS-based approaches with other experimental techniques such as live imaging. MS-based proteomics has proved its utility for elucidating protease function in biological systems. The rapidly evolving field promises to continue contributing to our understanding of the renin-angiotensin-system, blood coagulation and other cardiovascular systems that entail complex proteolytic networks.