|Home | About | Journals | Submit | Contact Us | Français|
Activity-based probes (ABPs) are reactive small molecules that covalently bind to active enzymes. When tagged with a fluorophore, ABPs serve as powerful tools to investigate enzymatic activity across a wide variety of applications. In this article, we will provide detailed protocols for using fluorescent ABPs to biochemically characterize the activity of proteases in vitro. Furthermore, we will describe how these probes can be applied to image protease activity in live animals and tissues along with subsequent analysis by histology, flow cytometry, and SDS-PAGE.
Proteases are a large family of enzymes that facilitate the degradation and turnover of proteins within the cell. In addition, they function as tight regulators of numerous signaling pathways responsible for processes such as apoptosis, cell migration, blood coagulation, and antigen presentation, among many others. Proteases also play important roles in the pathogenesis of many diseases, including cancer, atherosclerosis, arthritis, stroke, Alzheimer’s Disease, and Huntington’s Disease.
Proteases function by hydrolyzing peptide bonds of protein substrates and are classified according to their mechanism of catalysis. Cysteine, serine, and threonine proteases use an amino acid to nucleophilically attack the amide bond. Aspartic, glutamic, and metallo proteases use their catalytic amino acid to deprotonate a water molecule for substrate hydrolysis. (For further mechanistic information, see review by Deu et al, 2012) Proteolysis is tightly regulated in a temporal and spatial manner through several mechanisms. The enzymes are synthesized as inactive zymogens that become activated through structural rearrangements mediated by changes in pH, dimerization, or cleavage by upstream proteases (Deu et al., 2012; Edgington et al., 2011; Serim et al., 2012). Once activated, proteases may also be inhibited by endogenous proteins within the cell. As a result, measurements of total protein expression do not often reflect levels of protease activity. To study the function of proteases in normal cellular processes and disease states, tools to directly assess their activity are necessary.
One class of tools that has been developed to meet this need is activity-based probes (ABPs). ABPs are small molecules that bind specifically and irreversibly to active proteases (Deu et al., 2012; Edgington et al., 2011; Serim et al., 2012). Fluorescent ABPs contain a dye that allows for detection and are thus amenable to a wide variety of applications. Probe labeling can be assessed biochemically by SDS-PAGE analysis or by fluorescent imaging in live animals or excised tissues. Furthermore, ABPs can be used to study the localization of active proteases in complex tissues by microscopy as well as in single cells by flow cytometry. In this article, we will describe detailed protocols to study protease activity using ABPs both in vitro and in vivo. Basic Protocol 1 describes a procedure for labeling cell lysates with ABPs while Support Protocol 1 gives instruction on how to perform fluorescent SDS-PAGE to visualize labeled proteases. Basic Protocol 2 describes how to label proteases in live cells. Basic Protocol 3 describes a method to identify ABP targets using an immunoprecipitation assay. Furthermore, we describe the application of ABPs to image protease activity in live mice (Basic Protocol 4) and whole tissues (Basic Protocol 5). We then provide protocols for further ex vivo analysis of ABP-labeled tissues biochemically (Basic Protocol 6), histologically (Basic Protocol 7), and by flow cytometry (Basic Protocol 8).
As animal models designed to recapitulate human disease continue to improve, non-invasive imaging techniques will continue to become more powerful. The ability to monitor an enzyme’s activity in its natural environment can lend significant insight into its disease-promoting functions and aid in the development of therapeutics. Moreover, many proteases have been identified as biomarkers of disease, and functional imaging with activity-based probes has great diagnostic and prognostic potential.
Of all the classes of proteases, fluorescent ABPs have been developed most broadly for cysteine proteases, including caspases (Edgington et al., 2009; Edgington et al., 2012; Li et al., 2012), cathepsins (Blum et al., 2005; Blum et al., 2007; Gocheva et al., 2010; Joyce et al., 2004; Verdoes et al., 2012), and legumain (Lee and Bogyo, 2010). Probes for the threonine-dependent catalytic subunits of the proteasome also exist (Verdoes et al., 2006), as well for serine proteases (Liu et al., 1999). When planning experiments with ABPs, choose a probe for the protease of interest that has been well characterized. Data regarding dosing/potency, cell permeability, in vivo circulation, off-targets, etc. will be critical to the success of these experiments. If this is a new probe or such data is not available, be prepared to characterize it. The protocols below are written for testing a new probe, so some steps may be skipped if the appropriate characterization data is already available.
Proteases are involved in broad processes and many different biological models may be used to study their function. In some models, proteases are constitutively active while in others they are induced over time by outside stimuli. We could not possibly begin to address each model in this article. Therefore, as an example, we will describe protocols and advice for detecting caspase activity during apoptosis, a highly regulated form of programmed cell death. In general, apoptosis models involve induction of caspase activity through administration of drugs or other stressors. The protocols presented may be easily adapted to detect other proteases simply by changing stimulation methods and buffer conditions. In any case, the more information that is known regarding the timing and localization of protease activation in the preferred model, the better the chances of obtaining success in these protocols, particularly with in vivo imaging. Be prepared for several rounds of optimization, especially if this is a new model.
This protocol describes the most basic assay for activity-based labeling of caspases in a complex proteome. Labeling cell lysates will provide a snapshot of the caspase activity at the time the cells are harvested. The timing and degree of caspase activation will depend on the method used to stimulate apoptosis. If this information is not readily available in the literature for the cell line used, a time course and/or dose curve may be performed to determine optimal conditions for caspase activation. There are several advantages of labeling lysates over intact cells. 1) Less probe is required, which may be beneficial if the ABP is of limited supply. This is especially helpful during the early optimization steps of characterizing a new drug/stressor/cell line, etc. 2) Labeling lysates does not require that the ABP be cell permeable; therefore, it is more likely that the total pool of active caspases will be labeled rather than a subset. 3) This assay detects caspases at neutral pH, making it less likely to detect common off-targets of caspase ABPs such as legumain or cathepsins, which are active in the acidic conditions of the lysosome.
This support protocol describes how to visualize the proteases that were labeled in Basic Protocols 1, 2, 3, and 6. SDS-PAGE is a widely used technique; however, exact parameters vary by lab. Outlined here are tips that have worked well to detect caspase labeling, but other conditions may work as well.
Typhoon flatbed laser scanner or comparable device that detects wavelengths appropriate for the probe’s fluorophore
Once the optimal conditions for stimulating caspase activity have been determined using lysates, the next step is to label live, intact cells. This protocol depends on the ability of the ABP to freely penetrate cells and enter the cytoplasm (in the case of caspases). If the probe cannot reach its target, it will not label it. Furthermore, one of the major caveats of caspase probes is their varied propensity to label the lysosomal proteases (cathepsins and/or legumain) at acidic pH due to similarities in structure and mechanism of catalysis (Edgington et al., 2009; Edgington et al., 2012). If the probe enters the cell by endocytosis or macropinocytosis, it may pass through the lysosome before reaching the cytosol, labeling these off-targets in addition to caspases. This must be kept in mind when analyzing data. For these reasons, labeling live cells will often give you a more accurate prediction of the labeling profile that will be obtained in vivo.
Once ABP-labeled species have been visualized by SDS-PAGE, it may be necessary to confirm their identity. This can usually be achieved by pre-treating samples with selective inhibitors to compete for probe labeling; however, in the case of caspases, truly selective inhibitors do not exist. They tend to react broadly with members of the caspase family as well as lysosomal cysteine proteases like legumain and cathepsins. For this reason, immunoprecipitation may be the ideal way to validate probe-labeled species. If enough material remains, the sample used for SDS-PAGE analysis may be used directly in the immunoprecipitation assay.
Once the activity-based probe has been characterized in vitro, it can then be tested in vivo. Protocol details will largely depend on the mouse model and probes used, but there are certain guidelines to follow that will help increase the success rate of non-invasive imaging experiments. Rather than going through a stepwise protocol, we will outline these guidelines here. A scheme of a typical imaging experiment is outlined in Figure 1.
This model has been optimized for imaging caspase activity using ABPs (Adams, et al., 2008; Edgington, et al., 2009 and 2012). A DR5 agonist antibody is used to induce apoptosis in subcutaneous COLO205 tumors implanted in nude mice. The caspase LE22 (Edgington et al., 2012) is then administered to visualize caspase activity.
Upon completion of a non-invasive imaging experiment, it is useful to follow up with ex vivo imaging of the tissues of interest. Not only does this produce an additional, quantifiable piece of data, it can also aid in trouble-shooting of not-so-successful imaging experiments. It is often much easier to see contrast when tissues are isolated from the numerous sources of confounding background signal found in vivo. This is especially true for deeper tissues which may be difficult to image non-invasively.
One of the great advantages of activity-based probes is their tractability. Because of their covalent nature, fluorescent signal produced in vivo can be correlated with biochemical data demonstrating the actual labeling of protease targets. The tissues are homogenized and analyzed by SDS-PAGE allowing direct quantification of protease levels. In a successful imaging experiment, the levels should correlate with the fluorescent signal in the tissues. Biochemical analysis also reveals any labeling of off-target proteases that may occur in vivo.
Activity-based probes may also be used to examine the distribution of protease activity at a microscopic level. For this protocol, frozen sections are required. Fluorescent probe signal cannot be detected after paraffin embedding.
Activity-based probes are also highly amenable to flow cytometry assays. This application can be useful for determining the percentage of cells that possess protease activity within a tissue and can also correlate protease activity with cell type-specific markers.
In the last decade, the field of activity-based proteomics has expanded immensely and it continues to do so. The concept of activity-based probes has been a major breakthrough in studying the function of proteases in both normal physiology and disease(Blum, 2008; Deu et al., 2012; Edgington et al., 2011; Paulick and Bogyo, 2008; Serim et al., 2012). ABPs typically consist of three major components. The most critical moiety is a reactive electrophile that confers selectivity to a particular class of proteases. This element, often called a warhead, is susceptible to nucleophilic attack by the catalytic residue of the active enzyme, resulting in covalent, irreversible modification. An overview of electrophiles used for different proteases can be found in Serim et al., 2012. The second component is a peptide recognition element designed to confer selectivity towards a particular protease. This element is often peptidic in nature and is designed based on the natural protein substrate. The sequence can be engineered to increase potency towards the protease of interest and also to decrease affinity towards unwanted off-targets. Lastly, an ABP contains a tag that allows detection. For in vivo imaging applications, the tag is usually a fluorophore; however, ABPs with radiolabels have recently been reported for use with PET imaging modalities (Ren et al., 2011).
Some ABPs also contain a fourth component: a quenching group that absorbs photons emitted by the fluorophore until it is released by proteolytic attack of the warhead (Blum et al., 2005; Blum et al., 2007; Verdoes et al., 2012). Quenched probes are intrinsically dark until they bind the active protease; therefore, fluorescent signal specifically indicates activity.
One of the major advantages of ABPs is that they covalently and irreversibly modify the active protease (Edgington et al., 2011). This means that the fluorescent tag remains at the active site of the enzyme and does not diffuse away, allowing for precise studies of the localization of activity through imaging. Moreover, the modification survives the denaturing conditions of SDS-PAGE, which allows for a direct biochemical readout of protease binding. The fluorescent signal detected by imaging can directly be assigned to the proteases to which the probe binds. This is especially critical for identifying off-targets of probes for which absolute specificity has not been achieved.
While the covalent nature of activity-based probes is overall beneficial, one caveat is that probe binding also irreversibly inhibits enzyme activity (Edgington et al., 2011). Since proteases are often involved in complex signaling pathways, their inhibition may lead to disruption of biological processes. For most in vivo imaging experiments, the administered probe dose is so low that it only inhibits a small percentage of the total protease pool, thus it is unlikely to drastically alter the biology of the system. The use of ABPs is sufficient to obtain a snapshot of protease activity at a given time point; however, care should taken in interpreting results of experiments that monitor activity over time.
Another widely used class of probes for proteases is substrate-based probes (Edgington et al., 2011). Several adaptations of substrate-based probes exist, but the general concept is similar across all types. Like quenched ABPs, their spectral properties change upon proteolytic cleavage. Unlike APBs, however, they lack a reactive electrophile, therefore no covalent modification occurs. This means that enzyme activity is retained after substrate cleavage, which allows for signal amplification. On the other hand, the fluorescent signal is able to diffuse away from the site of cleavage, making localization studies less precise. For a more thorough discussion of the pros and cons of activity- and substrate-based probes for use in in vivo imaging studies, refer to (Blum et al., 2009; Edgington et al., 2011).
Perhaps the most critical parameter for a successful in vivo imaging experiment is timing. In order to detect protease activity, the ABP must make contact with the protease at the time that it is active. As outlined in the imaging section, achieving this may take several rounds of optimization. Information is needed regarding when the proteases are active and for how long. If the window of activity is short (as in some apoptosis models, for example, when the cells die and are quickly removed by macrophages), the timing of probe delivery will be even more important. In turn, the length of time it takes the probe to reach the target tissue is another critical factor. If using a non-quenched probe, it must also be determined how long it will take for the free probe to clear from the tissue. If little signal is observed in vivo, take time to follow up with ex vivo imaging and biochemical analysis. If no bands are observed by gel, the timing may need to be adjusted.
Another critical parameter is signal strength. In general, fairly robust protease activity is required to detect by non-invasive imaging methods. This is particularly true if the tissues are deep and the light must travel further and through other tissues. Models where caspases become activated in just a few cells, for example, are not good candidates for whole animal imaging experiments. If the precise location of these cells is known, it may be possible to see fluorescence in single cells, either by microscopy or flow cytometry. If protease activity is low, it may or may not be detected by SDS-PAGE. Typically, if the signal is strong enough to be detected by western blot (by cleaved caspase-3 staining, for example), then the threshold should be high enough for detection of ABP labeling.
Pharmacodynamic properties of the probe are also important factors that govern the effectiveness of imaging experiments. Ideally, the probe will circulate well and promptly reach the tissue of interest. Some probes are cleared by the kidney or liver before they ever reach the target tissue, and some are unstable in vivo and degrade immediately upon injection. Often, these properties are impossible to predict and require more thorough PK/PD studies to understand. Sometimes delivery or targeting strategies can be used to direct the probe to the tissue of interest (Mikhaylov et al., 2011, for example). If the probe works well in in vitro assays, it may be possible to glean some data by labeling tissue homogenates ex vivo.
Sometimes background fluorescence is so high that it obscures the true signal. Using higher wavelength, near-infrared probes, shaving the mice, and feeding non-fluorescent chow are all good practices to minimize autofluorescence. As previously mentioned, it is always a good idea to use a ‘no-probe’ mouse in parallel as a control to distinguish other areas of autofluorescence. Some probes become sequestered in the liver, gallbladder, kidneys and bladder as they clear from the body, leading to high background fluorescence in these tissues that may bleed into other areas. Imaging tissues within the peritoneum is challenging for this reason. Quenched probes are vastly preferred if available. Additionally, probes that lack absolute specificity for the protease of interest will also contribute additional background due to labeling of off-target proteases. Following up imaging studies with biochemical analysis is a must do.
These protocols should yield a reproducible and quantifiable readout of the levels of protease activity across numerous fluorescent-based applications both in vitro and in vivo, including the ability to visualize such activity by optical imaging of a live animal. Cellular localization of protease activity can also be determined using flow cytometry and histology protocols. Additionally, targets of ABPs can be confirmed using biochemical and immunoprecipitation assays.
Figure 2 provides data from representative experiments using a fluorescent ABP for caspases, LE22 (Edgington, et al., 2012). LE22 contains a Cy5 fluorophore, a Val-Glu-Ile-Asp peptide sequence, and an acyloxmethyl ketone warhead (Figure 2A). Figure 2B demonstrates the application of LE22 to label caspases in intact cells. COLO205 human colorectal cancer cells were treated with a DR-5 agonist antibody for 3.5 hours to induce caspase activity, followed by incubation with increasing amounts of LE22 for 30 minutes. Cells were then harvested according to the procedure described in Basic Protocol 2, and analyzed by SDS-PAGE (Supporting Protocol 1). Saturation of labeling occurs at 0.5 µM, indicating that this would be an optimal concentration to use for future in vitro experiments.
Multiple species are labeled by LE22 including various maturation forms of caspase-3, −6, and 7. To confirm the identity of the labeled bands, two approaches were used. The first was to pretreat the cells with a caspase-3-/7 selective inhibitor, AB13, prior to the addition of LE22. The inhibitor blocked labeling of caspase-3/−7, and the remaining bands corresponded to caspase-6. To confirm this with more certainty, an immuprecipitation experiment was performed according to Basic Protocol 3 (Figure 2C). Indeed, all of the labeled species could be assigned to caspase-3, −6, or −7, with one exception. There is one band that appears in the lane in which no probe was added (0 µM, Figure 2B). This indicates and autofluorescent protein, and highlights the importance of including a ‘no-probe’ control.
Figure 2D shows data obtained using the procedure described in the representative example of Basic Protocol 4. Nude mice bearing COLO205 human colorectal xenograft tumors were treated with anti-DR5 to induce tumor-specific caspase activity/apoptosis. At the peak of caspase activation (~11 hours in this case), LE22 was administered, and mice were imaged by IVIS one hour later. Tumors on the mice treated with anti-DR5 exhibited significantly increased accumulation of fluorescence compared to the vehicle-treated control mice. This strongly suggests that the increase in fluorescence was caspase-dependent. When the tumors were removed and imaged ex vivo according to Basic Protocol 5, the same trend in fluorescence was observed (Figure 2E). To confirm the in vivo targets of LE22, a biochemical analysis was also performed according to Basic Protocol 6 (Figure 2E). Indeed, tumors treated with anti-DR5 showed a sharp increase in the labeling of caspase-3, −6, and −7 compared to the vehicle-treated control tumors, which corresponds to the in vivo and ex vivo imaging data. Moreover, the biochemical analysis reveals that, in addition to caspases, LE22 labels low levels of the off-target protease, legumain. This suggests that some of the fluorescence detected by imaging is due to legumain rather than caspases. Caution should therefore be taken in interpreting the data. In this case, legumain levels do not change upon treatment with anti-DR5, therefore the fold-change in fluorescence is entirely due to caspase labeling. LE22 would also be amenable to histology and flow cytometry as described in Basic Protocols 7 and 8; however, this data is not shown. A similar probe, AB50, was used for these applications and data can be found in Edgington et al., 2009.
Sample preparation and probe labeling can be performed in <3 hours from the time of harvest. This does not include time to induce protease activity, as this varies widely by model. SDS-PAGE analysis also takes ~3 hours. The length of animal imaging experiments varies widely by model and the probe used and can range from 2–24 hours (sometimes more). Ex vivo imaging takes 15 minutes, or longer if more tissues are imaged. Tissue preparation for SDS-PAGE can be performed in 4 hours or less. Preparing tissues for histology takes 1.5 days, and FACS preparation will take ~ 5 hours.
This work was supported by NIH grant R01 EB005011.