Systematic Affinity Purification Coupled to Mass Spectrometry Identified p62 as Part of the Cannabinoid Receptor CB2 Interactome

Abstract

The endocannabinoid system (ECS) consists particularly of cannabinoid receptors 1 and 2 (CB1 and CB2), their endogenous ligands, and enzymes that synthesize and degrade their ligands. It acts in a variety of organs and disease states ranging from cancer progression over neuropathic pain to neurodegeneration. Protein components engaged in the signaling, trafficking, and homeostasis machinery of the G-protein coupled CB2, are however largely unknown. It is therefore important to identify further interaction partners to better understand CB2 receptor functions in physiology and pathophysiology. For this purpose, we used an affinity purification and mass spectrometry-based proteomics approach of Strep-HA-CB2 receptor in HEK293 cells. After subtraction of background interactions and protein frequency library assessment we could identify 83 proteins that were classified by the identification of minimally 2 unique peptides as highly probable interactors. A functional protein association network analysis obtained an interaction network with a significant enrichment of proteins functionally involved in protein metabolic process, in endoplasmic reticulum, response to stress but also in lipid metabolism and membrane organization. The network especially contains proteins involved in biosynthesis and trafficking like calnexin, Sec61A, tubulin chains TUBA1C and TUBB2B, TMED2, and TMED10. Six proteins that were only expressed in stable CB2 expressing cells were DHC24, DHRS7, GGT7, HECD3, KIAA2013, and PLS1. To exemplify the validity of our approach, we chose a candidate having a relatively low number of edges in the network to increase the likelihood of a direct protein interaction with CB2 and focused on the scaffold/phagosomal protein p62/SQSTM1. Indeed, we independently confirmed the interaction by co-immunoprecipitation and immunocytochemical colocalization studies. 3D reconstruction of confocal images furthermore showed CB2 localization in close proximity to p62 positive vesicles at the cell membrane. In summary, we provide a comprehensive repository of the CB2 interactome in HEK293 cells identified by a systematic unbiased approach, which can be used in future experiments to decipher the signaling and trafficking complex of this cannabinoid receptor. Future studies will have to analyze the exact mechanism of the p62-CB2 interaction as well as its putative role in disease pathophysiology.

Keywords: AP-MS; GPCR; SQSTM1; endocannabinoid system; interactome; protein–protein interaction.

FIGURE 1

(A) Flow chart of CaptiVate screening procedure. The bait CB2 was amplified and sub-cloned into the expression vector pN-TGSH (Dualsystems). The bait sequence was verified by sequencing. For constitutive expression, the bait was transfected into HEK293 Flp-In cells and selected for stable integration. Bait expression and solubility was monitored by immunoblotting using anti-HA antibodies. The expression level was compared to an internal standard and judged suitable for affinity purification and mass spectrometric analysis. (B) Background filtering of CB2 affinity purification: histogram of enrichment factors (CB2/control) of identifications to determine threshold of enrichment for specific interactors. Red area represents Gaussian distribution of unspecific interactors, yellow area is in the border and in the green area are the specific interactors (threshold as 10% of maximum peak height).

FIGURE 2

Functional Annotations were used to visualize protein–protein interaction network of 83 identified CB2 interactors by STRING. Illustrated interaction network was determined by text mining, experiments or databases with a minimum required interaction score of 0.400. In this view, the nodes represent proteins and the color saturation of the edges represents the confidence score of data support. CNR2 represents CB2 and SQSTM1 represents p62.

FIGURE 3

(A) Protein sequence of p62 with protein domains underlaid in light blue boxes (Phox and Bem1p domain (PB1), Zinc finger-type (ZZ), TNF receptor-associated factor 6 (TRAF6), LC3-interacting region (LIR), protein sequence rich in proline (P), glutamic acid (E), serine (S), and threonine (T) (PEST), UBA. The six peptides identified after tryptic digestion by mass spectrometry are highlighted in red and blue. (B) Verification of the interaction of CB2 receptors with p62 by co-immunoprecipitation (Co-IP) of transiently overexpressed flag tagged CB2 receptors (FL-CB2) with endogenously expressed p62 in HEK293 cells. The precipitated p62 protein intensity directly correlated with the amount of transfected CB2-receptor plasmid (either 1 or 2 μg) that was not influenced by 10 min stimulation with the CB2 agonist HU-210. Representative result of n > 3.

FIGURE 4

(A–G) Confocal light microscopy of HEK293 cells transiently expressing CB2 receptors. (A) Immunohistochemical staining of CB2 (green), p62 (red), and DAPI (blue). The membrane was stained by Texas-red conjugated wheat-germ-agglutinin (white) and DAPI revealed the localization of the nucleus. (B–G) 3D reconstruction modeling of confocal images using a segmentation algorithm to distinguish single signal from different channels. (D–G) Higher zoom in of 3D reconstructed cell showing a strong overlap of CB2 positive areas with the membrane marker (white). P62 was found in vesicular like structures that intermingled with membranal CB2 positive areas (yellow arrows).

FIGURE 5

(A) Schematic drawing of CB2 receptor protein with extracellular N-terminus and three extracellular loops (e1, e2, e3), seven transmembrane domains, and three intracellularly localized loops (i1, i2, i3) and the C-terminus. The red marked intracellular regions were deleted in CB2 expression plasmids and used for following Co-IP studies. (B) Representative Western blot results of CB2 precipitations using Flag antibody with lysates of HEK293 cells transiently transfected with p62-WT (1 μg) and Flag-CB2 constructs with the respective deletions (ΔC-CB2, Δ i1-CB2, Δ i2-CB2, Δ i3-CB2). Considering lower expression levels of deletion constructs compared to WT-full length Flag-CB2, the co-precipitated p62 protein levels were similar in all approaches. The IP figure was grouped from cropped parts of the same gels/blots indicated by the border line. This experiment has been performed at least two times independently.

FIGURE 6

(A) Schematic illustration of Flag-p62 deletion constructs for ZZ, PEST, and UBA motifs that were used for following co-immunoprecipitation experiments to identify interaction region. (B) Stably expressing HA-CB2 HEK293 cells were transiently transfected with different Flag-p62-deletion constructs, respectively. Protein extracts were used for precipitation of Flag-p62 using Flag antibody. The expression levels in input samples for CB2 and p62 were comparable in all constructs. Immunoprecipitation result of Flag-p62 ΔZZ revealed a significantly reduced coprecipitation of CB2 indicating that this domain in p62 is involved in the interaction with CB2. Representative results of three independent experiments were presented.