Assaying Homodimers of NF-κB in Live Single Cells

Abstract

NF-κB is a family of heterodimers and homodimers which are generated from subunits encoded by five genes. The predominant classical dimer RelA:p50 is presumed to operate as "NF-κB" in many contexts. However, there are several other dimer species which exist and may even be more functionally relevant in specific cell types. Accurate characterization of stimulus-specific and tissue-specific dimer repertoires is fundamentally important for understanding the downstream gene regulation by NF-κB proteins. In vitro assays such as immunoprecipitation have been widely used to analyze subunit composition, but these methods do not provide information about dimerization status within the natural intracellular environment of intact live cells. Here we apply a live single cell microscopy technique termed Number and Brightness to examine dimers translocating to the nucleus in fibroblasts after pro-inflammatory stimulation. This quantitative assay suggests that RelA:RelA homodimers are more prevalent than might be expected. We also found that the relative proportion of RelA:RelA homodimers can be perturbed by small molecule inhibitors known to disrupt the NF-κB pathway. Our findings show that Number and Brightness is a useful method for investigating NF-κB dimer species in live cells. This approach may help identify the relevant targets in pathophysiological contexts where the dimer specificity of NF-κB intervention is desired.

Keywords: NF-κB; RelA; dimerization; microscopy; number and brightness; oligomerization; transcription factor.

Figure 1

N&B approach to discern dimerization status of RelA. (A) There are five homologous proteins of the NF-κB TF family. All contain a Rel homology domain (RHD) that promotes dimerization with other RHD-containing proteins as well as DNA binding. RelA, RelB, and c-Rel each also contain a transcription-activation domain (TAD) that enables them to activate transcription. Cleavage of p105 and p100 produces p50 and p52, respectively. Additional domains include: LZ, leucine zipper; GRR, glycine-rich region; ANK, ankyrin-repeat domain; DD, death domain. (B) 15 potential NF-κB dimers exist based on RHD interactions. (C) The N&B assay measures the oligomer state of a protein by determining its molecular brightness (ε) within a region of interest (ROI) in a cell. A fluorescent protein's molecular brightness is determined by calculating the fluctuations (variance) in mean fluorescence intensity (<I>) that are caused by the movement of protein oligomers (monomers, dimers, trimers, k-mers) within every pixel (confocal volume) of the ROI over time (confocal imaging acquisition). The ratio of the variance to the mean fluorescence intensity of the pixels is equal to the protein's brightness (ε) + 1. Because immobile proteins do not produce such movement-based fluctuations, their molecular brightness is equal to 0. (D) For quantifying a protein's brightness within an ROI, in our case the nucleus, the brightness values of each pixel comprising the ROI are extracted from a stack of N&B images (see Methods). The brightness values are then fitted to a Gaussian distribution to determine the protein's overall brightness within the ROI.

Figure 2

RelA exhibits substantial homodimer levels in mouse fibroblasts. (A) Schematic of possible mEGFP-RelA (green ellipse) interactions with mEGFP-RelA or other non-mEGFP-tagged NF-κB proteins (white ellipse). (B) Representative confocal micrographs of nuclei in 3T3 fibroblasts transiently-expressing monomer control, transiently-expressing mEGFP-RelA, or stably-expressing mEGFP-RelA under different treatment conditions. Image intensity scale was adjusted for optimal viewing. Scale bar: 5 μm. (C) Quantification of nuclear mEGFP-RelA brightness (ε) values relative to the monomer control in transfected cells (B) treated with 10 ng/ml TNFα, 100 ng/ml LPS, or 100 nM Dex. Data was obtained from at least two independent experiments performed on different days. Whiskers are drawn down to the 10th percentile and up to the 90th percentile. Number of nuclei and median values of each sample are presented below each boxplot.

Figure 3

RelA homodimers in primary fibroblasts. (A) Representative confocal micrographs of nuclei in wild-type (WT) and p50/c-Rel double-KO primary MAFs transiently-expressing monomer control or mEGFP-RelA. The latter were treated with LPS (100 ng/mL). Image intensity scale was adjusted for optimal viewing. Scale bar: 5 μm. (B) Nuclear RelA brightness (ε) values relative to the monomer control in transfected primary fibroblasts (A) treated with 100 ng/ml LPS. Data was obtained from at least two independent experiments performed on different days. Whiskers are drawn down to the 10th percentile and up to the 90th. Number of nuclei and median values of each sample are presented below each boxplot.

Figure 4

Perturbation of RelA homodimers by pharmacological agents. (A) Representative confocal micrographs of nuclei in transiently- and stably-transfected 3T3 fibroblasts pre-incubated with WFA or TSA before LPS treatment (100 ng/mL). Control and LPS images are the same as those in Figure 2B. Image intensity scale was adjusted for optimal viewing. Scale bar: 5 μm. (B) Brightness (ε) values of nuclear RelA in fibroblasts under different conditions. Control and LPS samples are the same as those in Figure 2C. Data was obtained from at least two independent experiments performed on different days. Whiskers are drawn down to the 10th percentile and up to the 90th. Number of nuclei and median values of each sample are presented below each boxplot. Unpaired two-tailed Student's t-test (**p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001) was performed for statistical comparisons.