|Home | About | Journals | Submit | Contact Us | Français|
Protein thiol modifications occur under both physiological and pathological conditions and have been shown to contribute to changes in protein structure, function, and redox signaling. The majority of protein thiol modifications occur on cysteine residues that have a low pKa; these nucleophilic proteins comprise the “reactive thiol proteome.” The most reactive members of this proteome are typically low abundance proteins. Therefore, sensitive and quantitative methods are needed to detect and measure thiol modifications in biological samples. To accomplish this, we have standardized the usage of biotinylated and fluorophore-labeled alkylating agents, such as biotinylated iodoacetamide (BIAM), N-ethylmaleimide (Bt-NEM), and BODIPY-labeled IAM and NEM, for use in one- and two-dimensional proteomic strategies. Purified fractions of cytochrome c and glyceraldehyde-3-phosphate dehydrogenase were conjugated to a known amount of biotin or BODIPY fluorophore to create an external standard that can be run on standard SDS-PAGE gels, which allows for the quantification of protein thiols from biological samples by Western blotting or fluorescence imaging. A detailed protocol is provided for using thiol-reactive probes and making external standards for visualizing and measuring protein thiol modifications in biological samples.
Cysteinyl protein thiols play crucial roles in enzyme catalysis, protein structure, maintenance of the cellular redox potential, and cell signaling . The properties that make cysteine ideal for these redox-based reactions, however, also make them exceptionally vulnerable to oxidation by reactive oxygen or nitrogen species (ROS/RNS) or to modification by environmental or endogenous electrophiles [1, 2]. The average pKa of the cysteine residue thiol is ~8.5, which at cytosolic physiological pH, is less likely to react with ROS/RNS or electrophiles. However, many proteins have domains which result in a substantial lowering of the pKa thiol group, such that they are predominantly in the reactive thiolate anion form at physiological pH. It is also important to note that the differences in the local intracellular environment (e.g., pH and hydrophobicity) will also impact protein thiol reactivity. For example, the intra-mitochondrial pH is typically more alkaline than the cytosol which will likely affect the composition of thiol reactive proteins [3, 4]. These proteins collectively make up the “reactive thiol proteome.”
Several methods have been used to detect and measure thiol modifications. These include direct detection techniques for individual modifications (e.g., protein modifications induced by nitric oxide, glutathione, or electrophilic lipids) and strategies for the detection and quantification of overall cellular thiol modification [1, 2]. Detection of specific thiol modifications utilizes either antibody-based detection approaches or chemical approaches that facilitate tagging of the modified thiol group, whereas detection of the reactive thiol proteome is predominantly based on tagging methodologies that employ thiol-reactive probes. Both of these approaches have advantages and disadvantages. The obvious advantage of probing for individual modifications is that specific modifications can be monitored and associated with pathological or physiological mechanisms; however, the drawback with such an approach is that other modifications to the thiol proteome which may occur simultaneously may be overlooked. In this respect, the development of external standards is particularly important since it allows for the quantification of protein thiol modifications. This information can be valuable in assessing the biological impact of a modification. For example, if only one protein molecule in 100 molecules is modified, then it is unlikely that this modification will have a significant biological impact. Tagging the unmodified or reduced thiol pool allows for a broad view of the redox state of the cell or tissue and can be used to identify oxidatively modified proteins. However, this approach does not distinguish thiols which are specifically S-nitrosated, S-glutathiolated, oxidized, or modified by electrophilic lipids. Several excellent reviews and articles on the detection of specific modifications and approaches for tagging the unmodified protein thiol pool are available [1, 2, 5–10]. The purpose of this article is to provide a detailed protocol for the detection and quantification of the reactive thiol proteome using biotin- and fluorescence-based proteomic approaches.
Alkylation of free thiols and detection of alkylated proteins is an effective strategy for evaluating the reactive thiol proteome. N-ethylmaleimide (NEM), iodoacetamide (IAM), and iodoacetic acid (IAA) are commonly used for protein thiol alkylation. Additionally, radiolabeled, biotin-conjugated, and fluorophore-labeled forms of these and similar compounds are commercially available. Nevertheless, the decision of which compound to use experimentally should be based on the suitability of the thiol alkylation chemistry and the detection method employed. Moreover, the usage of IAA may or may not be desirable for proteomic strategies (depending on the experimental question being addressed) because it carries with it a negative charge that may shift the isoelectric point of labeled proteins.
The underlying chemistries of sulfhydryl modification by thiol alkylating agents are distinct and confer differences in their reactions with proteins. IAM and IAA yield carbamidomethylated and carboxymethylated cysteines, respectively, by bimolecular nucleophilic substitution (SN2) reactions [11, 12]. The lone pair of electrons in the deprotonated thiol (thiolate anion; S−) act as the nucleophile and attack the electron deficient electrophilic center of IAM/IAA expelling iodine anion as the leaving group (Fig. 1A). This reaction is second order, with the rate of reaction depending on the nucleophile concentration (S−), the concentration of the substrate itself (IAA/IAM), and the pH and proticity of the solvent. The reaction of NEM with thiols is based on a Michael-type addition reaction , where the thiolate anion attacks the electrophilic center of the C=C bond of the maleimide group to form a thioether bond between the thiol and the maleimide (Fig. 1B). The reaction of NEM with thiols is faster than IAM or IAA and less dependent on pH [11, 12]. However, NEM may be less specific than iodo derivatives; at alkaline pH, NEM also reacts with the side chains of lysine and histidine . The comparative effectiveness of protein thiol alkylation between NEM, IAM, and IAA was demonstrated in a study by Rogers et al. .
Biotin-based tagging techniques have been used to monitor intracellular thiol status of proteins after exposure to ROS/RNS or electrophilic compounds [10, 13–17]. To assess protein thiol modification in general, thiols can be labeled directly with biotin-tagged reagents such as biotinylated IAM (BIAM) or biotinylated NEM (Bt-NEM), and the biotin signal can subsequently be measured by standard immunoblotting-type protocols using streptavidin-conjugated horseradish peroxidase (HRP). In this case, the loss of the biotin signal is proportional to the degree of thiol modification. To quantify protein thiol content, we have developed biotinylated standard proteins containing known amounts of biotin per mole of protein .
The major advantages of this biotin-thiol tagging technique include: 1) the extremely high affinity of avidin and streptavidin for biotin (Kd ≈ 10−15 M) [18, 19]; 2) the binding of streptavidin, which unlike an antibody, is not readily affected by flanking residues at the site of protein modification; 3) affinity resins are available for purification of the biotinylated proteins; and 4) the biotin tag can be easily and accurately quantified using biotinylated standards. Nevertheless, there are some drawbacks to using biotin as a tag which must also be considered. For example, in contrast to antibody-based methods, the high affinity of streptavidin for biotin results in an association which is practically (though not formally) irreversible. The technical implications are that Western blots using enzyme-conjugated avidin/streptavidin as a means of detection cannot be stripped after development. In addition, some proteins such as carboxylases have biotin covalently attached as an enzymatic cofactor, and these proteins can give false positive results that must be accounted for during data interpretation and quantification. Lastly, the conditions for transfer of the proteins from the SDS-PAGE gel to PVDF or nitrocellulose membranes must be standardized to ensure that all of the protein is transferred. In many cases, proteins are not thoroughly transferred from the gel, and proteins of different molecular weights tend to transfer at different rates. This could lead to inaccurate estimates of the amount of biotin incorporated based on the initial protein loaded for SDS-PAGE.
Fluorophore-tagged alkylating agents are also an option for use in the assessment of the reactive thiol proteome [9, 20]. A major advantage to using fluorophore-tagged alkylating agents is that the transfer step can be omitted since Western blotting is not required. After alkylation and separation by one-dimensional (1D) SDS-PAGE or by 2D proteomics techniques, the level of alkylation can be measured in-gel using detection techniques based on the emission/excitation characteristics of the particular fluorophore. False positives that occur with biotin tagging techniques (due to endogenous biotin-containing enzymes) are also circumvented using fluorescence-based tagging techniques. Generally, fluorescent-based approaches such as these are highly sensitive and have a large dynamic range. However, one disadvantage is that some fluorophores are light sensitive, and consequently, much of the work must be done under low light conditions. Also, unless an antibody exists to the fluorophore itself, pull-down assays cannot be performed using fluorophore-labeled compounds. In this methods article, we also describe a BODIPY-labeled standard that can be used for quantifying the reactive thiol proteome.
We have standardized the conjugation of biotin to cytochrome c for quantifying biotin in Western blotting procedures . This conjugation reaction utilizes an amine-reactive succinimidyl ester that will react with lysine residues to form a stable amide bond.
The cyt c concentration is most accurately measured using the spectral properties of the covalently bound heme prosthetic group in its reduced state. For other proteins, we use the Lowry protein assay method to determine protein concentration . The amount of biotin (moles) for a given amount of protein (moles) can be calculated from the concentrations of biotin (determined below by the HABA assay) and protein. We have also confirmed these calculations with detailed mass spectrometric analysis .
Biotin incorporation is determined using a colorimetric HABA dye displacement assay. In this assay, the colored dye HABA reversibly binds avidin and is displaced by biotin, resulting in a decrease in the absorbance of HABA at 500 nm .
We have observed that the linear range of this assay, where an increase in biotin is represented by a proportional decrease in absorbance, is somewhat narrow . Therefore, it is necessary to repeat measurements using varying amounts of biotinylated protein, usually 1–20 μg. At this point, the absorbance at 500 nm may be plotted as a function of the protein amount added (in μg). In the linear range of the plot, quantitate the biotin incorporation in solution according to the extinction coefficient 34,000 M−1cm−1. In the case of colored proteins, it may be necessary to correct for the absorbance of the protein itself by preparing a solution of protein in PBS which does not contain HABA-avidin. This absorbance should be subtracted from that of the HABA-avidin solution containing the protein.
We have also standardized the conjugation and usage of BODIPY for the quantification of protein thiols using BODIPY-IAM. Similar to biotinylated cyt c, the BODIPY conjugation reaction utilizes an amine-reactive succinimidyl ester that will react with lysine residues to form a stable amide bond.
Note: Bromophenol blue can quench fluorescence of BODIPY. It is advised that dyes are omitted from the SDS sample loading buffer in this protocol. A 5× bromophenol blue-free loading buffer can be prepared using 62.5 mM Tris, pH 6.8, containing 8% SDS and 30% glycerol.
The BIAM- or BODIPY-labeled alkylating agents can be used at the time of cell lysis or tissue homogenization to evaluate the thiol redox state. This gives a “snap-shot” of thiol status at one point in time. Below is a general protocol for snap-shot thiol labeling:
As shown in Fig. 4A–C, BIAM will alkylate fewer thiols than Bt-NEM under these conditions. This is due to the more reactive nature of maleimide compared to iodoacetamide. If probing the “reactive thiol proteome” it may be advisable to use BIAM. This will ensure that only the most reactive cysteines will be labeled, which could give more insight into the protein targets affected during conditions such as oxidative stress. The buffers used in these protocols are also compatible with 2D electrophoresis. As shown in Fig. 4D and E, BIAM-labeled proteins resolve well on IEF-SDS-PAGE gels.
Note: Although fluorescent, chemiluminescent, and chemifluorescent techniques work well, the quality of the results will be dictated by the imaging system available and the level of alkylation achieved during the reaction. For the best possible results, the protocol should be optimized in each laboratory and with each imaging system. In our experience, the best resolution is achieved using BODIPY-labeled alkylating agents on a Typhoon imager; biotin-labeled alkylating agents also work well in chemifluorescent systems. Chemiluminescent systems work well, but may have less resolution than the above two methods. This is shown in Fig. 5, where identical samples were derivatized with BODIPY-IAM or BIAM; the images show greater resolution of samples derivatized with BODIPY-IAM and imaged by in-gel fluorescence compared to BIAM-derivatized gels imaged by chemiluminescent methods.
We have found that BODIPY-IAM is also useful for labeling thiols in situ in cell culture. This would be more desirable when compared to “snap-shot” labeling if the experimenter wishes to understand how redox status changes temporally under a given set of conditions. An example of how this could be used in cell culture is shown in Fig. 6. Mesangial cells were treated with 0 or 1 mM diamide for 10 min. Cells were then treated with 50 μM BODIPY-IAM (added directly into the media) for 30 min, followed by cell lysis in buffer containing 1 mM DTT. As described in section E.8, excess alkylating agents such as NEM can be used instead of DTT to prevent further alkylation reactions from occurring after cell lysis. After cell harvest, proteins in the lysates were resolved by SDS-PAGE and imaged in-gel using a Typhoon Trio imager. As shown in Fig. 6A–C, cells treated with diamide were labeled less extensively (i.e., −5.1 ± 0.9 nmoles BODIPY/mg protein) than untreated cells, which is likely due to increased formation of protein-mixed disulfides. Since the ratio of reaction of iodoacetamide with thiols is 1:1, this suggests that the diamide treatment led to the oxidation of ~5 nmoles thiol/mg protein. When separated by 2D electrophoresis, diamide was shown to oxidize a large number of proteins (Fig. 6D and E), which could be identified by mass spectrometry.
A key aspect to accurate quantitation of the biotin tag using Western blot analysis is the use of high resolution digital imaging techniques. The use of film is rarely optimal because the narrow linear range of chemiluminescence using film can result in image saturation. The dynamic range of digital camera imagers and fluorescent imagers (e.g., Typhoon imagers) have a wider dynamic range and are therefore recommended for Western blotting applications using biotin tags and/or in-gel fluorescent imaging.
Biotinylated protein samples are separated by 1D- or 2D-SDS-PAGE. To measure biotin content in experimental samples, biotinylated protein standards should be included on the gel. For 1D gels, a standard curve may be constructed by varying amounts of one biotinylated protein in a few lanes. For 2D gels, the biotinylated standard protein(s) can be run in a lane adjacent to the first dimension gel strip. Duplicate gels can be run so that proteins from one gel can be stained with Sypro Ruby or comparable protein stains and proteins from the other gel can be transferred to nitrocellulose or PVDF and blotted for biotin detection. To quantify moles biotin per mole protein, non-biotinylated protein can be mixed with biotinylated protein, so that the standard can be used to calibrate both stained gels and Western blots .
To blot proteins, transfer to nitrocellulose or PVDF membrane at 100 V for 2 h with cooling or at 25 V overnight at 4°C. If using a Typhoon imager in later steps (see below), a low fluorescence PVDF membrane such as Hybond-LFP should be used. Block nonspecific binding sites with 5% milk in TBS-T for 1 h. Wash membranes thoroughly to remove all milk. (Note: some milk blotting formulations contain biotin. Since this could interfere with streptavidin binding to biotinylated proteins, milk should not be included after the blocking step). Incubate blots with streptavidin-HRP (1:10,000 dilution in 10 ml TBS-T) for 1 h. Wash membranes 3 times for 10 min each. Add chemiluminescent substrate evenly to the blot, ensuring coverage of the entire surface. For chemifluorescent imaging of biotin incorporation, use ECL Plus substrate and incubate on the membrane for 5 min under dim-lighting conditions or in the dark.
If using fluorophore-tagged alkylating agents (e.g., BODIPY-IAM or –NEM), the image may be acquired directly from the gel. However, it is very important to run the dye front completely off of the gel prior to imaging, because the unreacted alkylating agent can interfere with the image and result in poor image quality in the lower half of the gel. Prior to imaging, wash the outside of the glass plates containing the gel with 70% ethanol, followed by distilled water. It should be noted that standard glass plates often have intrinsic fluorescent properties that could affect image quality. In our experience, conventional glass plates do not show any background fluorescence on a Typhoon imager at the excitation and emission settings described below for the BODIPY FL fluorophore. However, if fluorophores other than BODIPY FL or multiple fluorophores with different spectral properties will be used, it is advisable to use low fluorescence glass plates. Scan using the proper wavelength and emission filter settings; the excitation maximum for BODIPY FL is 505 nm and the emission maximum is at 513 nm. Refer to the respective instrument manual for guidelines on emission and excitation settings.
For Western blotting applications, acquire a series of images using a CCD camera imager (AlphaInnotech), a fluorescent imager (Typhoon; Amersham Biosciences), or similar instrumentation. For chemiluminescent imaging, the substrate should be left on the membrane during imaging. Imaging can be performed using a “movie” function, which integrates serial exposures. The result is a “movie strip” containing images of increasing intensity. Images should be saved as TIFF format files, which are used for subsequent analyses. Images containing saturated pixels should not be used for quantitation purposes.
For chemifluorescent imaging, the membranes should be dried prior to imaging or placed in distilled water and imaged while wet. In our experience using a Typhoon Variable mode imager, less background is observed when the membrane is dried completely before imaging. The signal is stable for > 48 h after development using ECL Plus detection reagents. After exposure to ECL Plus, transfer the membrane to a stacked layer of Kimwipes and gently blot the membrane dry. Afterwards, place the membrane on a piece of filter paper and allow it to dry for at least 20 min prior to imaging. Failure to allow the blot to dry completely will result in uneven background and image artifacts. Lay the blot face-down on the Typhoon platen and overlay with a piece of non-fluorescent plastic such as 3M Dual Purpose transparency film. Scan image in the fluorescence acquisition mode using the 520 BP 40 emission filter and the blue (488 nm) laser (note: if using a Typhoon Trio imager, use the ECL + excitation setting). Set the PMT voltage to 400–450 V initially. Do not press the sample; the plastic transparency film will adequately press the membrane against the platen. Set the pixel size to 500–1000 μm initially. Adjust the PMT voltage to obtain an image that is intense but not saturated. Take the final image using a pixel size of 100 μm.
Acquire Sypro Ruby images using UV light if using a CCD camera imaging system or with Sypro Ruby filter settings and the appropriate laser conditions if using a Typhoon imager.
The amount of biotin is quantitated by determining the density (in arbitrary units; AU) of the selected area with AlphaEaseFC software if using a CCD camera or the with ImageQuant TL software if using the Typhoon imager. Several alternative densitometry programs exist and can be used in a manner similar to those described here. For 1D gels, the densities of each lane containing experimental samples and biotinylated standard proteins are determined. Note that background subtraction settings are particularly important here, and, in our experience, the “minimum profile” setting is optimum for assessment of overall biotinylation or BODIPY signal from 1D membranes or gels.
This protocol describes the use of alkylating agents for the detection of reactive protein thiols. Depending on the intent of the experiment or project, other types of probes may also be used to interrogate the thiol proteome. For instance, disulfide-forming probes or probes similar to those used in protein spin-labeling studies such as biotinylated glutathione, methyl methanethiosulfonate (MMTS), or N-(6-(biotinamido)hexyl)-3′-(2′-pyridyldithio)-propionamide (biotin HPDP) may be used alone or in concert to detect specific protein modifications [1, 2, 5–8, 24]. The advantage to using these probes is that particular reductants or derivatization reagents (e.g., ascorbate  or dimedone ) allow for detection and discrimination of specific reversible post-translational protein modifications such as S-nitrosated or sulfenic acid-modified proteins. The protocols described here, however, focus mainly on identifying the overall proteome that is modified. It should be noted that these protocols could easily be adapted and used jointly with other fluorophore-labeled probes and protocols to give further insight into and to quantify potentially important protein thiol modifications. For example, it is possible to extend this protocol to not only detect the overall thiol proteome that is modified but also to discern the proportion of those proteins that are S-nitrosated, -glutathiolated, or -oxidized.
The choice of alkylating reagent depends on the experimental intent and the imaging facilities available. IAM will alkylate many reduced protein thiols at high pH; alternatively, NEM can be used to alkylate for this purpose at neutral pH or below. To label thiols that are most highly reactive (i.e., those that have a low pKa), IAM should be used at neutral pH or below. These thiols, which make up the “reactive thiol proteome,” are of interest in most protein modification studies because of their potential involvement in redox signaling or their role in pathology due to modification during excessive oxidant stress. The caveat to using iodo derivatives is incomplete alkylation of the reactive thiol proteome. In studies by Rogers et al., NEM was shown to more fully alkylate thiols than eightfold higher concentrations of IAM or IAA . Hence, the amount of thiols modified by IAM or IAA per mg protein will likely be an underestimation of the actual reactive thiol proteome. Because of this, measurement of thiols using this protocol is not absolute. Nevertheless, if all samples are treated equally, the measurements are quantitative and exceed more semi-quantitative measurements based only on assessing arbitrary or relative fluorescence units. The external standards used in this protocol are the primary advancement over previous methods. In our experience, the fluorescence standards have a much wider linear dynamic range and can be detected in femtomole quantities.
This study was supported by the UAB-UCSD O’Brien Core Center (NIH/NIDDK 1P30 DK 079337) and an American Heart Association (Scientist Development Grant to A. L.).
Conflicts of Interest