The BS3 chemical crosslinking assay reveals reduced cell surface GABAA receptor expression in mouse brains under chronic psychosocial stress conditions.
Anxiety is a state of emotion that variably affects animal behaviors, including cognitive functions. Behavioral signs of anxiety are observed across the animal kingdom and can be recognized as either adaptive or maladaptive responses to a wide range of stress modalities. Rodents provide a proven experimental model for translational studies addressing the integrative mechanisms of anxiety at the molecular, cellular, and circuit levels. In particular, the chronic psychosocial stress paradigm elicits maladaptive responses mimicking anxiety-/depressive-like behavioral phenotypes that are analogous between humans and rodents. While previous studies show significant effects of chronic stress on neurotransmitter contents in the brain, the effect of stress on neurotransmitter receptor levels is understudied. In this article, we present an experimental method to quantitate the neuronal surface levels of neurotransmitter receptors in mice under chronic stress, especially focusing on gamma-aminobutyric acid (GABA) receptors, which are implicated in the regulation of emotion and cognition. Using the membrane-impermeable irreversible chemical crosslinker, bissulfosuccinimidyl suberate (BS3), we show that chronic stress significantly downregulates the surface availability of GABAA receptors in the prefrontal cortex. The neuronal surface levels of GABAA receptors are the rate-limiting process for GABA neurotransmission and could, therefore, be used as a molecular marker or a proxy of the degree of anxiety-/depressive-like phenotypes in experimental animal models. This crosslinking approach is applicable to a variety of receptor systems for neurotransmitters or neuromodulators expressed in any brain region and is expected to contribute to a deeper understanding of the mechanisms underlying emotion and cognition.
Neurotransmitter receptors are localized either at the neuronal plasma membrane surface or intracellularly on the endomembranes (e.g., the endosome, the endoplasmic reticulum [ER], or the trans-Golgi apparatus) and dynamically shuttle between these two compartments depending on intrinsic physiological states in neurons or in response to extrinsic neural network activities1,2. Since newly secreted neurotransmitters elicit their physiological functions primarily through the surface-localized pool of receptors, the surface receptor levels for a given neurotransmitter are one of the critical determinants of its signaling capacity within the neural circuit3.
Several methods are available to monitor surface receptor levels in cultured neurons, including the surface biotinylation assay4, the immunofluorescence assay with a specific antibody in non-permeabilized conditions5, or the use of a receptor transgene genetically fused with a pH-sensitive fluorescent optical indicator (e.g., pHluorin)6. By contrast, these approaches are either limited or impractical when assessing surface receptor levels in vivo. For example, the surface biotinylation procedure may not be practical for processing large quantities and sample numbers of in vivo brain tissues due to its relatively high price and the subsequent steps necessary for purifying the biotinylated proteins on avidin-conjugated beads. For neurons embedded in three-dimensional brain architecture, low antibody accessibility or difficulties in microscope-based quantification may pose a significant limitation for assessing the surface receptor levels in vivo. To visualize the distribution of neurotransmitter receptors in intact brains, non-invasive methods, such as positron emission tomography, could be used to measure receptor occupancy and estimate the surface receptor levels7. However, this approach critically relies on the availability of specific radio ligands, expensive equipment, and special expertise, making it less accessible for routine use by most researchers.
Here, we describe a simple, versatile method for measuring surface receptor levels in experimental animal brains ex vivo using a water-soluble, membrane-impermeable chemical crosslinker, bis(sulfosuccinimidyl)suberate (BS3)8,9. BS3 targets primary amines in the side chain of lysine residues and can covalently crosslink proteins in close vicinity to each other. When brain slices are freshly prepared from a region of interest and incubated in a buffer containing BS3, the cell surface receptors are crosslinked with neighboring proteins and, thus, transform into higher-molecular weight species, whereas the intracellular endomembrane-associated receptors remain unmodified. Therefore, the surface and intracellular receptor pools can be separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and quantitated by western blot using antibodies specific to the receptor to be studied.
Unpredictable chronic mild stress (UCMS) is a well-established experimental paradigm for inducing chronic psychosocial stress in rodents10. UCMS elicits anxiety-/depressive-like behavioral phenotypes and cognitive deficits via the modulation of an array of neurotransmitter systems, including GABA and its receptors10,11. In particular, the α5 subunit-containing GABAA receptor (α5-GABAAR) is implicated in the regulation of memory and cognitive functions12,13, suggesting the possible involvement of altered functions of this subunit in UCMS-induced cognitive deficits. In this protocol, we used the BS3 crosslinking assay to quantitate levels of surface-expressed α5-GABAAR in the prefrontal cortex of mice exposed to UCMS as compared with non-stressed control mice.
All the animal work in this protocol was completed in accordance with the Ontario Animals for Research Act (RSO 1990, Chapter A.22) and the Canadian Council on Animal Care (CCAC) and was approved by the Institutional Animal Care Committee.
1. Preparation of animals
2. Preparation of the stock solutions
3. Preparation of the workstation
4. Preparation of the working solutions and buffers
NOTE: On the morning of the assay, prepare the following solutions. This calculation is based on the necessary solutions to process two brain regions (i.e., the PFC and HPC) from 12 mice.
5. Dissection of brain tissues
NOTE: From this step on, at least two people should work together in a coordinated manner. While one person focuses on the animal dissection (steps 5.2-5.10 and step 6.3), the other person should work as a timekeeper and help coordinate the assay (step 5.1, step 6.1, step 6.2, step 6.4, and step 6.5)
6. Crosslinking reaction
7. Tissue lysis, protein preparation, and western blot
To demonstrate the feasibility of the BS3 crosslinking assay for evaluating the surface α5-GABAAR levels in the mouse PFC, we ran 10 µg each of BS3-crosslinked and non-crosslinked protein samples on SDS-PAGE and analyzed the proteins by western blot using an anti-α5-GABAAR antibody (rabbit polyclonal) (Figure 7). The non-crosslinked protein samples gave the total amount of α5-GABAAR at ~55 kDa, while the BS3-crosslinked protein samples gave a certain amount of endomembrane-associated α5-GABAAR (migrating at ~55 kDa) along with higher-molecular weight protein species representing protein complexes covalently crosslinked to α5-GABAAR. For the quantification of the surface α5-GABAAR levels, we can assess the extent of depletion of α5-GABAAR at ~55 kDa from the total pool upon crosslinking, as the subunit then shifts to higher-molecular weight positions. Practically, the surface levels of α5-GABAAR can be calculated by subtracting the amount of endomembrane-associated α5-GABAAR (at ~55 kDa in the crosslinked sample lane) from the total α5-GABAAR levels (at ~55 kDa in the non-crosslinked sample lane).
We next evaluated the effects of UCMS (at 3 weeks and 5 weeks) on the surface and total α5-GABAAR levels in the mouse PFC. In this experiment, in order to follow the time course of the crosslinking reaction, we prepared the samples at 1 h, 2 h, and 3 h after the addition of BS3. Since BS3 crosslinking reactions appeared to reach a plateau by 2 h, the data at the 1 h time point were used to plot the graph. We observed a significant and progressive reduction in surface α5-GABAAR levels in the PFC at 3 weeks and 5 weeks of UCMS, as compared with the no-stress control mice (Figure 8). The data showed negligible or no apparent changes in the total receptor levels under these experimental conditions, suggesting that chronic stress specifically impacted α5-GABAAR trafficking to the cell surface.
Figure 1: Dissection tools used in the BS3 crosslinking assay. Filter paper is moistened with ice-cold PBS and placed on the flat surface of blue ice. A Petri dish filled with PBS and stored at −20 °C 1 day prior to the assay is placed on ice. The brain matrix and razor blades are pre-chilled on ice. Scissors (one large, one small), forceps (one large, a few small ones), and a tissue punch are all cleaned prior to the assay. Please click here to view a larger version of this figure.
Figure 2: Whole mouse brain dissected out of the skull and placed in ice-cold PBS. Immediately after the whole brain was removed from the skull, it was submerged in ice-cold PBS for 10-15 s in a Petri dish on ice. This slows down brain metabolism and minimizes the protein trafficking and degradation while also helping the tissue to become harder, which makes the subsequent brain slicing easier. Please click here to view a larger version of this figure.
Figure 3: A mouse brain placed in the brain matrix. The chilled mouse brain submerged in ice-cold PBS was transferred to the brain matrix and placed with the ventral side facing up. The razor blades and the matrix were pre-chilled on ice. Please click here to view a larger version of this figure.
Figure 4: Coronal sectioning of a mouse brain in the brain matrix. The first pre-chilled razor blade was inserted through the border between the olfactory bulb and the olfactory peduncle to start sectioning the brain coronally. Please click here to view a larger version of this figure.
Figure 5: Serial coronal sections of a mouse brain. The anterior part of the mouse brain was cut coronally by inserting five razor blades serially into the brain matrix (1 mm intervals). All the inserted blades were held together, lifted off the matrix, separated from one another using forceps, and placed on the chilled flat surface with the brain slice facing up. (Top left) The olfactory bulb; (middle left) the first section; the second (bottom left) and third (top right) sections were used for sampling the PFC tissues; (bottom right) the fourth section. Please click here to view a larger version of this figure.
Figure 6: Posterior part of the brain for dissecting the hippocampus. After the anterior part of the brain was coronally cut with razor blades (left), the posterior part of the brain (middle) was taken out of the brain matrix and placed on PBS-moistened filter paper on the chilled surface (right). Please click here to view a larger version of this figure.
Figure 7: BS3 crosslinking of GABAAR on the cell surface. In the presence of BS3, α5-GABAAR on the plasma membrane surface is crosslinked (XL) with anonymous proteins close to it, such as other GABAAR subunits within the pentameric GABAAR assembly or additional neighboring proteins (X), but not with proteins (Y) far away from it. Thus, the α5-GABAAR appears as a high-molecular weight (HMW) protein species on the blot. α5-GABAAR on the endosome remains intact and migrates at the expected size of ~55 kDa. Please click here to view a larger version of this figure.
Figure 8: Surface α5-GABAAR levels affected by UCMS. Both the total and surface levels of α5-GABAAR in the PFC of mice under the no-stress and UCMS (3 weeks and 5 weeks) conditions were evaluated. To follow the time course of the crosslinking reaction, the samples were prepared at 1 h, 2 h, and 3 h after the addition of BS3. Female mice (2-3 months, N = 4/group) were used. The intact α5-GABAAR levels at 55 kDa for each condition were first normalized by the corresponding αTubulin levels and then used to calculate the total and endomembrane-associated α5-GABAAR levels (from the no-crosslinked [No Xlink] and crosslinked [Xlink] samples, respectively). Subsequently, the surface receptor levels were calculated by subtracting the endomembrane-associated amount from the total levels. Since the BS3 crosslinking reactions appeared to reach a plateau by 2 h, the data at the 1 h time point were used to plot the graph (mean ± SEM). Significant effects of chronic stress on surface α5-GABAAR levels were observed, and this effect was UCMS duration-dependent, as a progressive decrease in the surface α5-GABAAR levels was identified across the UCMS period. *p < 0.05, **p < 0.01 (Kruskal-Wallis test with Dunn's multiple comparisons). Please click here to view a larger version of this figure.
Working conc. | Stock solution | Amount of stock solution to dispense |
1.2 mM CaCl2 | 480 mM (400x)* | 100 μL |
20 mM HEPES | 1 M (50x) | 800 μL |
147 mM NaCl | 5 M (34x) | 1176.5 μL |
2.7 mM KCl | 1.08 M (400x) | 100 μL |
1 mM MgCl2 | 400 mM (400x) | 100 μL |
10 mM Glucose | 2.5 M (250x) | 160 μL |
Deionized water | 37.563 mL | |
Total | 40 mL | |
* CaCl2 stock should be freshly prepared on the day of experiment. |
Table 1: The composition of artificial cerebrospinal fluid.
Working conc. | Stock solution | Amount of stock solution to dispense |
25 mM HEPES | 1 M (40x) | 500 μL |
500 mM NaCl | 5 M (10x) | 2 mL |
2 mM EDTA | 0.5 M (250x) | 80 μL |
1 mM DTT | 1 M (1000x) | 20 μL |
0.1% NP-40 | 10% (100x) | 200 μL |
Protease inhibitor cocktail | 100x | 200 μL |
Deionized water | 17 mL | |
Total | 20 mL |
Table 2: The composition of the lysis buffer
Video 1: Isolating the PFC using a tissue punch. Please click here to download this Video.
Video 2: Mincing the PFC using forceps. Please click here to download this Video.
Video 3: Dissecting the HPC using a curved probe and forceps. Please click here to download this Video.
Video 4: Inversion-mixing a tissue sample. Please click here to download this Video.
Although the impact of chronic psychosocial stress on behaviors (i.e., emotionality and cognitive deficits) and molecular changes (i.e., reduced expression of GABAergic genes and accompanying deficits in GABAergic neurotransmission) are well-documented10, the mechanisms underlying such deficits need further investigation. In particular, given the recent study showing that chronic stress significantly affects the neuronal proteome through overload on the ER functions and, thus, elevated ER stress15, a question remains as to whether chronic stress affects GABAAR trafficking through the ER membrane and how altered trafficking or surface levels of GABAAR could be causally linked to psychopathology.
The BS3 crosslinking assay shown in this protocol will serve as a powerful approach for answering some of these questions. For example, UCMS is known to elicit a series of behavioral changes, including cognitive deficits, anxiety-like behavior, and anhedonic phenotypes, depending on the duration of UCMS10. Therefore, it would be possible to study the time course and the degree of UCMS-induced changes in surface α5-GABAAR levels by sampling mouse brains at successive time points from the start of UCMS (e.g., 1 week, 3 weeks, 5 weeks), and it would also be possible to cross-compare with the behavioral phenotypes observed at each time point. Additional GABAAR subtypes (e.g., α1, α2) and the receptor for neuromodulators known to be affected by chronic stress (e.g., the TrkB receptor for brain-derived neurotrophic factor [BDNF])10 could be tested simultaneously using the same samples. Since some of these receptor systems and subtypes are selectively implicated in specific behavioral domains (e.g., sedation, anxiety, cognition)11, it is worth correlating the behavioral changes with the type and degree of receptors affected by UCMS at each time point.
Below are the critical points to consider for several steps in the protocol. The first step is to determine the necessary and sufficient numbers of animals to use based on the experimental design and power analysis. For example, to study the effect of chronic stress on surface GABAA receptor levels, we routinely prepare one group of mice (N = 6/sex) to be exposed to UCMS for 5 weeks and another group (N = 6/sex) to be kept under no-stress conditions. This group size (N = 24 in total, including both sexes) is expected to give enough statistical power to detect a ~20% difference in receptor levels, thus allowing evaluation of both stress effects and sex effects. Notably, it is reported that chronic stress causes sex-dependent differences in behavioral and molecular outcomes10. For example, women are generally more prone to developing depressive symptoms than men. Consistently, our studies using human postmortem and rodent brains indicate higher levels of behavioral and molecular pathologies in female subjects; downregulated levels of somatostatin (SST), a molecular marker for depression, are more robust in women among depressed patients16, and increased emotionality is more robust in female mouse models replicating aspects of depressive pathology than in males15,17. Therefore, it is advised that any experimental design addressing the effects of chronic stress on psychopathological outcomes should include adequate numbers of both males and females to ensure statistical power in analyzing the data and to identify possible sex-dependent effects.
The optimal incubation time with BS3 should be determined in pilot experiments for each receptor to be studied. It is reported that receptor trafficking may occur slowly, even at low temperatures during sample incubation9. To capture accurate surface receptor levels at the time of brain dissection, it would be ideal to minimize the incubation time and choose the time point right before the crosslinking reaction reaches the plateau (30 min to 2 h at 4 °C).
We noticed several operational limitations associated with BS3 crosslinking assays. (1) First, due to the relatively short half-life (2-3 h) of BS3 caused by spontaneous hydrolysis within the physiological pH range, one needs to complete the animal dissection and crosslinking reaction within this time frame. This led us to limit the number of animals we could dissect at a time to a maximum of 12. If the experimenter plans on dissecting more than 12 animals, it is advised to divide the experimental cohort into several groups, with each containing less than 12 animals. After the assay for one group is completed, a new batch of BS3 should be freshly prepared to use in subsequent crosslinking assays for the next group. In the same vein, the number of regions of interest to be dissected from one animal should be limited. We routinely sample from two brain regions (PFC, HPC) for the crosslinking assay, and this is the maximum number of brain regions we can dissect in order to obtain consistent results with minimal variability across samples. (2) Second, the specificity of the antibodies used in western blot should be carefully evaluated. The BS3 chemical crosslinking reaction may have an unexpected effect on the antigenicity of the proteins detected by each antibody. We found that one α5-GABAAR antibody (specified in the Table of Materials) erroneously detected a strong ~50 kDa band specifically in BS3-crosslinked samples, regardless of the presence or absence of the α5-GABAAR protein in the sample; this ~50 kDa band was seen even in tissue samples from α5-GABAAR knockout mice, suggesting that this antibody began cross-reacting with irrelevant antigens accidentally generated by the BS3 crosslinking reaction. It is advised that the antibody specificity be thoroughly determined with and without BS3 crosslinking reactions, ideally using knockout tissue samples, if available, for a given protein of interest. (3) The current protocol described here does not address the cell types in which each GABAAR subtype is expressed (e.g., neurons and astrocytes). For some GABAAR subtypes (e.g., α1, α5) that are predominantly expressed in neurons18, the crosslinking assay data obtained using bulk brain tissues, as described in this protocol, should provide information reflecting the neuronal surface levels. However, for other GABAAR subtypes (e.g., α2) that are highly expressed in both neurons and astrocytes18, it is of interest to study the receptor surface levels in neurons versus astrocytes separately. To this end, conventional cell sorting (e.g., fluorescence-activated cell sorting [FACS]19) may be integrated into the BS3 crosslinking protocol; the cell dissociation step for cell sorting can be done after the BS3 quenching step, but one needs to validate that all the procedures and conditions for FACS (e.g., the buffers to use, temperature, incubation time) are compatible with those in the crosslinking assay. In addition, the FACS approach may be used only for GABAAR subtypes known to be localized to the perisomatic cellular compartment (e.g., α2) but not for the subtypes enriched in distal dendrites (e.g., α5)18, because peripheral or distal cell compartments are likely lost during the extensive cell dissociation step necessary for cell sorting. (4) Finally, we found the results to be more consistent when we calculated the surface receptor levels by subtracting the amount of endomembrane-associated receptors from the total amount of receptors rather than directly evaluating the surface levels based on the densitometry of high-molecular weight protein species. This is likely because the transfer efficiency of these high-molecular weight protein species onto the PVDF membrane is more variable than that of the original, intact protein of a smaller size (e.g., ~55 kDa for α5-GABAAR). It is, therefore, recommended to follow the method described in the results section and the legend of Figure 8 to calculate the surface receptor levels.
Apart from the effects of chronic stress on GABAAR, the BS3 crosslinking assay can also be applied to genetically modified mice or rodent models with experimental manipulations to investigate a number of neurological or neuropsychiatric conditions. This assay has been successfully used to capture cocaine-induced effects on the surface expression of glutamate receptors in the nucleus accumbens of rat brains20,21. The assay has also been used to show reduced surface expression of α5-GABAAR in the PFC of heterozygous BDNF knockout mice22. In another previous study, in the model of hepatic encephalopathy in rats, the accompanying spatial learning deficits were causally linked to altered surface expression of glutamate and GABAA receptors based on this crosslinking assay23. In summary, the BS3 chemical crosslinking assay provides a versatile tool for capturing brain region-specific and context-dependent changes in a range of receptor systems in the brain, as well as in virtually any other peripheral tissues or organs. This assay can also be conducted in parallel with other methods for evaluating the surface receptor levels, such as the electrophysiological recording of tonic inhibition (especially in the case of α5-GABAAR), cryogenic electron microscopy, and surface biotinylation, to cross-compare and validate the results.
The authors have nothing to disclose.
The authors thank the CAMH animal facility staff for caring for the animals over the study duration. This work was supported by the Canadian Institute of Health Research (CIHR Project Grant #470458 to T.T.), the Discovery Fund from the CAMH (to T.P.), the National Alliance for Research on Schizophrenia and Depression (NARSAD award #25637 to E.S.), and the Campbell Family Mental Health Research Institute (to E.S.). E.S. is the founder of Damona Pharmaceuticals, a biopharma dedicated to bringing novel GABAergic compounds to the clinic.
0.5 M EDTA, pH 8.0 | Invitrogen | 15575020 | |
1 M HEPES | Gibco | 15630080 | |
10x TBS | Bio-Rad | 1706435 | |
2.5 M (45%, w/v) Glucose | Sigma | G8769 | |
2-mercaptoethanol | Sigma | M3148 | |
4x SDS sample buffer (Laemmli) | Bio-Rad | 1610747 | |
Bis(sulfosuccinimidyl)suberate (BS3) | Pierce | A39266 | No-Weigh Format; 10 x 2 mg |
Brain matrix | Ted Pella | 15003 | For mouse, 30 g adult, coronal, 1 mm |
Calcium chloride (CaCl2) | Sigma | C4901 | |
Curved probe | Fine Science Tools | 10088-15 | Gross Anatomy Probe; angled 45 |
Deionized water | milli-Q | EQ 7000 | Ultrapure water [resistivity 18.2 MΩ·cm @ 25 °C; total organic carbon (TOC) ≤ 5 ppb] |
Dithiothreitol (DTT) | Sigma | 10197777001 | |
Filter paper (3MM) | Whatman | 3030-917 | |
Forceps (large) | Fine Science Tools | 11152-10 | Extra Fine Graefe Forceps |
Forceps (small) | Fine Science Tools | 11251-10 | Dumont #5 Forceps |
GABA-A R alpha 5 antibody | Invitrogen | PA5-31163 | Polyclonal Rabbit IgG; detect erroneous signal upon chemical crosslinking |
GABA-A R alpha 5 C-terminus antibody | R&D Systems | PPS027 | Polyclonal Rabbit IgG; cross-reacts with mouse and rat |
Glycine | Sigma | W328707 | |
Horseradish peroxidase-conjugated goat anti-rabbit IgG (H+L) | Bio-Rad | 1721019 | |
Magnesium chloride (MgCl2·6H2O) | Sigma | M2670 | |
Nonidet-P40, substitute (NP-40) | SantaCruz | 68412-54-4 | |
Potassium chloride (KCl) | Sigma | P9541 | |
Protease inhibitor cocktail | Sigma | P8340 | |
PVDF membrane | Bio-Rad | 1620177 | |
Scissors (large) | Fine Science Tools | 14007-14 | Surgical Scissors – Serrated |
Scissors (small) | Fine Science Tools | 14060-09 | Fine Scissors – Sharp |
Sodium chloride (NaCl) | Sigma | S9888 | |
Sonicator (Qsonica Sonicator Q55) | Qsonica | 15338284 | |
Table-top refregerated centrifuge | Eppendorf | 5425R | |
Tissue punch (ID 1 mm) | Ted Pella | 15110-10 | Miltex Biopsy Punch with Plunger, ID 1.0 mm, OD 1.27 mm |
Trans-Blot Turbo 5x Transfer buffer | Bio-Rad | 10026938 | |
Tube rotator (LabRoller) | Labnet | H5000 |