Discovery and characterization of two novel CB1 receptor splice variants with modified N-termini in mouse

Abstract

Numerous studies have been carried out in the mouse model, investigating the role of the CB1 cannabinoid receptor. However, mouse CB1 (mCB1) receptor differs from human CB1 (hCB1) receptor in 13 amino acid residues. Two splice variants, hCB1a and hCB1b, diverging in their amino-termini, have been reported to be unique for hCB1 and, via different signaling properties, contribute to CB1 receptor physiology and pathophysiology. We hypothesized that splice variants also exist for the mCB1 receptor and have different signaling properties. On murine hippocampal cDNA, we identified two novel mCB1 receptor splice variants generated by splicing of introns with 117 bp and 186 bp in the N-terminal domain, corresponding to deletions of 39 or 62 amino acids, respectively. The mRNAs for the splice variants mCB1a and mCB1b are expressed at low levels in different brain regions. Western Blot analysis of protein extracts from stably transfected HEK293 cells indicates a strongly reduced glycosylation due to the absence of two glycosylation sites in mCB1b. On-cell Western analysis in these stable lines revealed increased internalization of mCB1a and mCB1b upon stimulation with the agonist WIN55,212-2. Results also point towards an increased affinity to SR141716 for mCB1a, as well as slightly enhanced inhibition of neurotransmission compared to mCB1. In mCB1b, agonist-induced mitogen-activated protein kinase (MAPK) phosphorylation was decreased compared to mCB1 and mCB1a. Identification of mouse CB1 receptor splice variants may help to explain differences found between human and mouse endocannabinoid systems and improve the understanding of CB1 receptor signaling and trafficking in different species.

Graphical abstract

Two novel introns in the mouse CB1 receptor were identified. Splicing leads to variants with shortened N-terminal tails, but leaves the highly-conserved N-terminal loop which contributes to the ligand-binding pocket intact. Differing from the variants identified in human and having a much lower expression, this might help to explain some of the differences seen between human and mouse endocannabinoid systems.

Introduction

The CB1 cannabinoid receptor is a member of the G protein coupled receptor family and is activated by various endogenous and exogenous ligands (Pertwee et al. 2010). The CB1 receptor plays key roles in a variety of physiological functions including fear, anxiety, reward, appetite, and pain perception, while imbalances in the endocannabinoid system can cause many pathologies (Di Marzo et al. 2015; Maccarrone et al. 2014; Lutz et al. 2015). Many preclinical studies investigating the role of CB1 receptor signaling have been carried out in mice because of the versatility and availability of many transgenic models (Lutz 2014). CB1 receptor function can be very well studied in mouse, as the receptor is well conserved between mouse and human with 97% amino acid identity and 90% nucleotide identity (Abood 2005). However, for the human CB1 (hCB1) receptor, two splice isoforms differing at the N-terminus of the receptor have been identified (Shire et al. 1995; Ryberg et al. 2005; Xiao et al. 2008) and were described to contribute to normal physiological processes (Bagher et al. 2013). Furthermore, possible implications in pathophysiology have recently been proposed (González-Mariscal et al. 2016).

The first variant, hCB1a, has an altered N-terminal amino acid sequence truncated by 61 amino acid residues with a substitution of the first 89 amino acids for a different 28-residue sequence due to a frame shift and the use of a different start codon (Shire et al. 1995). The second variant, hCB1b has an in-frame deletion of 33 amino acids also in the hCB1 receptor amino terminus (Ryberg et al. 2005). Both hCB1a and hCB1b mRNAs are found in several tissues including brain (Shire et al. 1995; Ryberg et al. 2005; Xiao et al. 2008) and hCB1a and hCB1b receptor proteins are co-expressed with hCB1 throughout the brain (Bagher et al. 2013). Whereas the membrane location and trafficking of hCB1a and hCB1b does not differ from hCB1 (Straiker et al. 2012), signaling properties of the human receptor splice variants are distinct from one another as well as from rodent CB1 receptors (Straiker et al. 2012; Ryberg et al. 2005; Xiao et al. 2008). Co-expression of hCB1 and each of the splice variants was shown to increase cell surface expression of the human cannabinoid hCB1 receptor and agonist-induced ERK phosphorylation (Bagher et al. 2013). The two human splice variants rescued depolarization induced suppression of excitation in a similar range as the rat CB1 receptor when expressed in cultured hippocampal neurons from CB1 receptor-deficient mice, whereas the hCB1 receptor signaled less robustly under identical conditions (Straiker et al. 2012).

The classical splice donor and acceptor sequences that are present in the coding sequence of the human CB1 receptor gene are not conserved in the mouse CB1 receptor gene. Thus, it was assumed that the N-terminal splicing is unique for the human CB1 receptor (Howlett et al. 2002; Xiao et al. 2008). However, in this study, we identified two splice variants for the mouse CB1 receptor (mCB1) and characterized their trafficking and signaling properties.

Materials and Methods

Animals

All experimental protocols conducted in Germany were carried out in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and approved by the Ethical Committee on animal care and use of Rhineland-Palatinate (23 177-07/A 10-1-009), Germany. Experiments conducted in the United States were approved by the Institutional Animal Care and Use Committee of Indiana University Bloomington and conform to the Guidelines of the National Institutes of Health on the Care and Use of Animals. Mice (C57BL/6N (RRID:IMSR_CRL:27) for identification of novel transcripts and CB1 knockout in CD1 (Ledent et al. 1999; RRID:MGI:4881723) background for neuronal cultures) derived from in-house breeding colonies and were housed in a temperature- and humidity-controlled room (22°C±1; 50%±1) with a 12 h-12 h light-dark cycle (lights on at 7 am) and had access to food and water ad libitum.

Identification of novel transcripts

Animals (5 males, C57BL/6N, 8–12 weeks) were sacrificed by decapitation under deep isoflurane anesthesia. Brains were isolated and snap-frozen on dry-ice. Mounted on Tissue Tek (Polysciences, Warrington, PA, USA), coronal sections were cut on a cryostat Microtome HM560 (Microm, Walldorf, Germany). Brains were trimmed until the area of interest was reached (distance from bregma according to (Paxinos and Franklin 2008)). Tissue punches were taken with sample corers (Fine Science Tools, Heidelberg, Germany) from infralimbic cortex (AP +1.94, ML ±0.00, DV +3.10; 1 mm diameter, 0.5 mm depth), caudate putamen (AP +1.10, ML ±1.80, DV +3.10; 1 mm diameter, 1 mm depth), basolateral amygdala (AP −0.82, ML ±2.80, DV +4.70; 0.8 mm diameter, 1.2 mm depth), dorsal hippocampus (AP −1.52, ML ±1.40, DV +1.75; 1 mm diameter, 1 mm depth) and cerebellum (AP +5.80, ML ±1.20, DV +2.10; 2 mm diameter, 2 mm depth) and were stored at −80°C.

Frozen tissue samples were homogenized in RLT buffer (RNeasy Mini-Kit; Qiagen, Hilden, Germany) with a Precellys 24 (Peqlab, Erlangen, Germany) at 6000 rpm for 20 s. Total RNA was isolated using the RNeasy Mini-Kit (Qiagen). Total RNA was reverse-transcribed using the High Capacity cDNA Reverse Transcription Kit with random primer hexamers (Applied Biosystems, Carlsbad, CA).

Primer pairs flanking putative introns were designed using Vector NTI software (Invitrogen, Darmstadt, Germany; forward primer: 5′-GGTTATGAAGTCGATCTTAGACGG; reverse primer: 5′-TCCCCACACTGGATGTTGT). PCR on cDNA templates was performed with Phusion High-Fidelity DNA Polymerase (New England Biolabs (NEB), Frankfurt, Germany). Conditions were 95°C 5 min, followed by 30 cycles of 95°C 30 s, 56°C 30 s, 72°C 90 s, followed by a final extension of 72°C 5 min.

PCR-amplified DNA was separated on 1.5% (w/v) agarose gels and DNA bands were extracted using the NucleoSpin Extract II Kit (Macherey Nagel, Dueren, Germany). A-overhangs were added using a Taq-Polymerase (Promega, Madison, WI). PCR products were cloned into a TOPO-TA vector according to the manufacturer’s instructions (Eurofins MWG Operon, Ebersberg, Germany). The intron and exon sequences were identified by comparing the sequences of cDNA clones with the genomic sequence of the mouse CB1 receptor gene using Nucleotide BLAST (http://blast.ncbi.nlm.nih.gov/).

Quantitative PCR using TaqMan assays

Quantification of cDNA was performed with an ABI 7300 real time PCR cycler (Applied Biosystems, Carlsbad, CA). The cDNA was amplified using commercial FAM dye-labeled TaqMan assays (Applied Biosystems) for mouse CB1 cannabinoid receptor (Cnr1; Mm00432621_s1) and mouse glucuronidase beta (Gusb; Mm00446953_m1). For mCB1a and mCB1b, custom Taqman assays were designed using Primer Express Software (Applied Biosystems): mCB1a forward primer 5′-GATACCACCTTCCGTACCATCAC, mCB1a reverse primer 5′-AATGTTGGTTGTGTCTCCTTTGATAT, mCB1a FAM-labelled reporter 5′-TCCTCTACGTGGGCTC; mCB1b forward primer 5′-GATACCACCTTCCGTACCATCAC, mCB1b reverse primer 5′-GTTGTCCTCGTTCTCCTGAATGT, mCB1b FAM-labelled reporter 5′-ACAGACCTCCTCTACGTGG. Primer pairs were pretested for specificity in a non-quantitative PCR. PCR was performed in a volume of 20 μl containing 10 μl TaqMan Gene Expression Mastermix (Applied Biosystems), 1 μl TaqMan assay and 9 μl cDNA (prediluted to approximately 3 ng/μl) with the cycle protocol 50°C 2 min, 95°C 10 min, 50 cycles of 95°C for 15 s and 60°C for 1 min. Expression levels of the quantified transcripts were normalized to that of the reference gene Gusb. To compare the expression levels of the novel splice variants mCB1a and mCB1b with that of mCB1, expression levels of mCB1a and mCB1b were normalized to the expression level of mCB1 (assay Cnr1).

Cloning of mCB1a and mCB1b

Hippocampal cDNA was used to clone mCB1a and mCB1b in a 2-step process (see Figure S1 for details, and Table S1 and S2 for primer sequences and PCR conditions, respectively). PCRs were performed with Phusion High Fidelity DNA Polymerase (NEB). PCR products from interim and final cloning steps were purified by excision of bands from 1% agarose gel. DNA was purified using the NucleoSpin Extract II Kit (Macherey Nagel). PCR products were then digested with the restriction enzymes KpnI and NotI (NEB) and purified by gel extraction as above. The vector pcDNA3 was also digested with KpnI and NotI, dephosphorylated with Antarctic phosphatase (NEB). Vector backbone and insert were then ligated with T4 DNA ligase (NEB). Ligated DNA was transfected into chemically competent DH5α E. coli, positive clones were picked and expanded (Sambrook et al. 2001). Plasmid DNA was purified with the NucleoSpin Plasmid Kit (Macherey Nagel). To verify correct cloning, the plasmids were sequenced.

A hemagglutinin (HA)-epitope tag was added to each splice variant via PCR, and these labeled constructs were then used to generate stable cell lines. Primers were designed using ApE cDNA plasmid software (http://biologylabs.utah.edu/jorgensen/wayned/ape/). The 5′ primer (5′-GCGGATCCACCATGGCATACCCATATGATGTCCCCGACTACGCGAAGTCGATCTTAGACGGCCTTG) consisted of a BamHI restriction site followed by a strong Kozak consensus sequence in frame with the HA11 epitope, and the common beginning of the three variants. The antisense primer (5′-GGCGCGGCCGCTCACAGAGCCTCGGCAGA) contained a NotI restriction site directly after the stop codon. The PCR products for the three mCB1 variants were then digested with BamHI and NotI and subcloned into both CAG and pcDNA3 vectors. The constructs were verified by sequencing.

Cell culture and transfection

HEK293 cells were grown in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum (FBS) and 100 units/ml penicillin and 100 μg/ml streptomycin (all substances from Gibco, Carlsbad, CA, USA) at 37°C in a 5% CO₂ humidified incubator and split the day before transfection. Cells were transfected with the CB1 receptor variant-containing plasmids via Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Stable cell lines of all constructs were generated by selection using Geneticin (G418; Invitrogen). G418-resistant colonies were evaluated for the CB1 surface expression by live cell immunostaining using an antibody directed towards the N-terminal extracellular HA11 epitope tag (Cat#MMS-101P, Covance, Berkeley, CA, USA: RRID:AB_291230) and a fluorescein isothiocyanate (FITC) secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA; RRID:AB_2340792). Clones expressing uniform and moderate to high levels of CB1 receptor were expanded and used for subsequent experiments. The relative levels of expression of CB1 receptor between the novel-variant and wild-type lines were not significantly different (see Fig. 3A for Western Blots including GAPDH internal standard). Three stable cell lines were generated and analyzed for each CB1 receptor variant.

Figure 3. Molecular weight and glycosylation of mCB1, mCB1a and mCB1b.

(A) Representative Western blot shows bands for HA-immunodetection of mCB1 and the splice variants mCB1a and mCB1b stably expressed in HEK293 cells in the upper panel. GAPDH bands are shown in the lower panel to show similar expression levels of the receptors between the stable cell lines. (B) Amino acid sequence of the N-terminal tails of mCB1, mCB1a and mCB1b. Putative N-linked glycosylation consensus sequences are indicated in bold letters, and putatively glycosylated asparagines are marked with arrows.

For neuronal cell culture for autaptic neurons, mouse hippocampal neurons isolated from the CA1–CA3 region were cultured on microislands as previously described (Bekkers and Stevens 1991; Furshpan et al., 1976). Neurons were obtained from CB1 knockout animals of either sex (at postnatal day 0–2, killed via rapid decapitation without anesthesia) and plated onto a feeder layer of hippocampal astrocytes that had been laid down previously (Levison et al., 1991). Cultures were grown in high-glucose (20 mM) minimum essential media containing 10% horse serum, without mitotic inhibitors and used for recordings after 8 days in culture and for no more than 3 h after removal from culture medium (Straiker et al., 2005). All electrophysiological experiments were performed exclusively on excitatory neurons.

We transfected neurons using a calcium phosphate-based method adapted from Jiang et al. (2004). Briefly, plasmids for the protein of interest and for an EYFP expression plasmid (2 μg/well) were combined with 2 M CaCl₂ and water and then gradually added to HEPES-buffered saline (HBS, 130 mM NaCl, 5.4 mM KCl, 1.8 mM MgCl₂ and 10 mM HEPES, pH 7.5); the resulting mixture was added to the serum-free neuronal media. Coverslips were incubated with this mixture for 2.5 hours while extra media was placed in a 10% CO₂ incubator to induce acidification. At the end of 2.5 h, the reaction mixture was replaced with acidified serum-free media for 20 min. After this, cells were returned to their home wells. Each data set was taken from at least two different neuronal platings.

Western blot

Stably transfected HEK293 cell lines grown to approximately 90% confluency in 6-well dishes were chilled on ice for 5 min. Following a wash with ice-cold 1x Phosphate-buffered saline PBS (137 nM NaCl, 10 mM NaH₂PO₄, 2.7 mM KCl, pH 7.4), cells were covered with 200 μl lysis buffer (100 mM Tris (pH 7.4), 150 mM NaCl, 0.5% CHAPS, 1 mM EDTA, 6 mM MgCl₂ and 100 mM PMSF) and incubated on ice for 5 minutes. Cells were then scraped and lysates were sonicated and spun down at 10,000 × g and 4°C. The supernatant was collected and protein concentration was determined by the method of Bradford (Bradford 1976). The samples were normalized to total protein, and 25 μg protein of each sample was run on a 10% Tris-glycine SDS-PAGE. The separated proteins were transferred to nitrocellulose and immunoblotting was performed using a mouse monoclonal anti-HA11 antibody and anti-GAPDH antibody (Cat# 649201, Biolegend, San Diego, CA, USA: RRID:AB_2107422). Both primary antibody was diluted 1:1000 in Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE, USA). A secondary antibody, goat anti-mouse conjugated IR680 dye (Cat# A21057, Invitrogen; RRID:AB_141436) was diluted 1:5000 in a 1:1 mixture of PBS and Odyssey blocking buffer. Western blots were scanned on an Odyssey near-IR scanner (LI-COR Biosciences), and images were processed using Photoshop CE.

Quantitative internalization assay

Agonist-induced internalization of the mCB1 receptor and its splice variants was assessed in stably transfected HEK293 cell lines as described previously (Daigle et al. 2008b). Briefly, HEK293 cells stably expressing mCB1, mCB1a or mCB1b were seeded onto poly-D-lysine coated 96-well plates and grown until ~95% confluent. Before drug treatment, cells were washed once in HEPES-buffered saline (HBS) containing 0.2 mg/ml bovine serum albumin (BSA, Sigma-Aldrich) and patted dry. Cells were then incubated in HBS/BSA containing WIN55,212-2 with/without SR141716 with concentrations and incubation times indicated in the results. For internalization in presence of antagonist, 10 nM SR141716 was applied together with WIN55,212-2 and incubated for 30 min. At the end of the incubation, wells were emptied and the plate was placed on ice. Cells were immediately fixed in 4% paraformaldehyde (PFA) for 20 min at RT. Cells were washed five times for 5 min in PBS and blocked for 60 min with LI-COR Odyssey Blocking Buffer (LI-COR Biosciences). Cells were then incubated overnight at 4°C in mouse monoclonal anti-HA11 antibody (1:150, Cat# 901503, BioLegend; RRID:AB_2565005) with gentle shaking. The following day, cells were washed five times for 5 min in Tris-buffered saline (TBS) containing 0.05% Tween-20 (TBST; Tris-Base 10 mM, 137 mM NaCl, 0.05% Tween-20, pH 7.4) and then incubated for 1 h in the dark with donkey anti-mouse IgG conjugated to IRDye 800CW (1:800; Rockland Immunochemicals, Gilbertsville, PA, USA; RRID:AB_1660938) in Odyssey Blocking Buffer. Cells were then washed four times for 5 min in TBST in darkness. After a short rinse in TBS, the immunocomplex was visualized on a LI-COR Odyssey near-IR scanner (LI-COR Biosciences, Lincoln, NE, USA).

Immunocytochemistry

Cells were plated onto coverslips coated with Poly-D lysine and grown overnight They were then washed in cold PBS followed by cold 4% PFA for 20 minutes at RT. Following five washes in PBS, cells were covered with wheat germ agglutinin (WGA488, Cat# W11261, Invitrogen) diluted 1:300 in PBS for 5 minutes at RT. Samples were washed three times with PBS and immediately blocked in blocking buffer for 45 minutes at RT. Mouse anti HA-11 antibody was added (1:150). Cells were incubated overnight at 4°C. The next day, cells were washed six times with TBST and incubated in secondary antibody (donkey anti-mouse Alexa594, Cat# 715-585-150, Jackson Immunoresearch; RRID:AB_2340854) diluted 1:500. Coverslips were then incubated for 60 minutes at RT, washed five times in PBS and three times in water. After allowing to dry, coverslips were mounted using Fluoromount-G with DAPI (Cat# 00-4959-52, Invitrogen). Images were collected on a Leica SP-5 (LMIC, Indiana University) using UV, 488 and 594 nm lasers. Images were processed using NIH ImageJ and Photoshop CS5 (Adobe Software, vs 12.0).

Quantitative measurement of MAPK phosphorylation

Stably expressing HEK293 cells were seeded on poly-D-lysine coated 96-well plates and grown until ~95% confluent. Cells were serum-starved overnight in 100 μl Dulbecco’s modified Eagle’s medium containing penicillin/streptomycin and only 0.1% FBS prior to the experiment. Before drug treatment, media containing 0.1% FBS was replaced and an additional 100 μl media with WIN55,212-2 was added to a final concentration of 100 nM. Cells were then incubated at 37°C for 5–120 min. After drug treatment, medium was removed and the cells were fixed by immediate addition of 100 μl ice-cold 4% PFA and incubated for 15 min on ice and for an additional 30 min at RT. Cell membranes were permeabilized by addition of 100 μl ice-cold methanol and incubation at −20°C for 20 min. Cells were then washed five times for 5 min with Triton wash solution (0.1% Triton X-100 in PBS or TBS) and patted dry. Cells were blocked for 1.5 h in 100 μl Odyssey Blocking Solution with gentle shaking and incubated overnight at 4°C with anti-phospho-p44/42 MAPK antibody (Thr202/Tyr204; 20G11, Cell Signaling Technologies, Danvers, MA, USA: RRID:AB_331772) diluted 1:200 in Odyssey Blocking buffer. Cells were then washed five times for 5 min in TBST, briefly rinsed in PBS and incubated for 1.5 h with donkey anti-mouse IgG conjugated to IRDye 800CW (1:800) in Odyssey Blocking Buffer protected from light. Cells were washed 5 times for 5 min in TBST and patted dry. Immunofluorescence was scanned with a LI-COR Odyssey near-IR scanner (LI-COR Biosciences, Lincoln, NE, USA). Integrated intensity values were used for analysis and were averaged and normalized to basal level activity.

Electrophysiology

When a single neuron is grown on a small island of permissive substrate, it forms synapses – or ‘autapses’ – onto itself. All electrophysiology experiments were performed on isolated autaptic neurons. Whole-cell, voltage-clamp recordings from autaptic neurons were carried out at room temperature using an Axopatch 200B amplifier (Axon Instruments, Burlingame, CA, USA). The extracellular solution contained (mM) NaCl 119, KCl 5, CaCl₂ 2, MgCl₂ 1, glucose 30 and HEPES 20. Recording pipettes of 1.8–3 MΩ were filled with solution containing (mM): potassium gluconate 121.5, KCl 17.5, NaCl 9, MgCl₂ 1, HEPES 10, EGTA 0.2, MgATP 2 and LiGTP 0.5. Access resistance was monitored and only cells with a stable access resistance were included for data analysis.

Conventional stimulus protocol

The membrane potential was held at –70 mV and excitatory postsynaptic currents (EPSCs) were evoked every 20 seconds by triggering an unclamped action current with a 1.0 ms depolarizing step. The resultant evoked waveform consisted of a brief stimulus artifact and a large downward spike representing inward sodium currents, followed by the slower EPSC. The size of the recorded EPSCs was calculated by integrating the evoked current to yield a charge value (in pC). Calculating the charge value in this manner yields an indirect measure of the amount of neurotransmitter released while minimizing the effects of cable distortion on currents generated far from the site of the recording electrode (the soma). Data were acquired at a sampling rate of 5 kHz.

DSE stimuli

After establishing a 10–20 second 0.5 Hz baseline, DSE was evoked by depolarizing to 0 mV for 50 msec, 100 msec, 300 msec, 500 msec, 1 sec, 3 sec and 10 sec, followed in each case by resumption of a 0.5 Hz stimulus protocol for 20–80+ seconds, allowing EPSCs to recover to baseline values. This approach allowed us to determine the sensitivity of the synapses to DSE induction. To allow comparison, baseline values (prior to the DSE stimulus) are normalized to one. DSE inhibition values are presented as fractions of 1, i.e. a 50% inhibition from the baseline response is 0.50 ± standard error of the mean. The x-axis of DSE depolarization-response curves are log-scale seconds of the duration of the depolarization used to elicit DSE.

Depolarization response curves are obtained to determine pharmacological properties of endogenous 2-AG signaling by depolarizing neurons for progressively longer durations (50 msec, 100 msec, 300 msec, 500 msec, 1 sec, 3 sec and 10 sec). The data are fitted with a nonlinear regression, allowing calculation of an ED₅₀, the effective dose or duration of depolarization at which a 50% inhibition is achieved. Statistical significance in these curves is based on non-overlapping 95% confidence intervals of the ED₅₀’s.

Data analysis

The results were analyzed using Prism6 Software (GraphPad, La Jolla, CA, USA) or SPSS Statistics Software version 19 (IBM, Armonk, NY, USA). Differences were considered statistically significant at p<0.05. All data are expressed as mean +/− SEM. Statistical analysis of the expression of the splice variants with shortened N-terminal were performed with one-way ANOVA. Assays for quantitative measurement of internalization and p44/p42 MAPK phosphorylation were analyzed using two-way ANOVA. Significant genotype effects were further analyzed using Bonferroni’s post-hoc analysis for multiple comparisons. Statistical analysis of depolarization response curves involved calculating the nonlinear fit of each curve and then comparing the 95% confidence interval for the curves. Non-overlapping 95% confidence intervals were taken as an indication of statistical significance.

Results

Identification

To analyze whether splice variants for the mCB1 receptor do exist, primers designed to amplify the sequence coding for the N-terminal extracellular domain of the mCB1 receptor were used in a PCR on cDNA templates from several brain tissues (forward primer binding to −4 to +20 and reverse primer binding to +203 to +284, counting the coding sequence (CDS) start as +1, Fig. 1A). Gel electrophoresis of the PCR products showed a band of approximately 300 bp corresponding to the unspliced mCB1 receptor sequence and two less abundant, smaller products for some of the template cDNAs (e.g. from hippocampus, basolateral amygdala; Fig. 1B). The 300 bp band (corresponding to mCB1), and the gel areas between 150 and 250 bp and between 100 bp and 150 bp were excised and DNA was extracted. A second PCR on the extracted DNA resulted in strong, distinct bands of approximately 300 bp (corresponding to mCB1, 306 bp), 190 bp and 120 bp separated on an agarose gel. The PCR products were purified and subcloned, and insert DNA was sequenced. Alignment of the resulting sequence from the 190 bp PCR product with the sequence of the mCB1 receptor showed 100% sequence identity up to +103, then a gap corresponding to an intron of 117 bp and again 100% sequence identity from +220 (Fig. 1C). The 120 bp PCR product had 100% sequence identity up to +87, then a gap corresponding to an intron of 186 bp and again 100% sequence identity from +274. As the intron sizes of 117 bp and 186 bp are divisible by three, and there is no alternative ATG upstream of the exon junction, it is highly likely that the same start codon for mCB1 is used for mCB1a and mCB1b. Comparison of the splice donor and acceptor sites with the conserved sequences (Lim and Burge 2001) showed that the donor site for mCB1a or mCB1b contains AG/GA or AG/TA at the 3′ end/5′ start of the intron, instead of the canonical AG/GT, respectively. Investigation of the splice acceptor sites of mCB1a and mCB1b revealed the presence of the canonical AG at the 3′ splice site with an upstream region with enrichment of pyrimidine and the presence of the conserved G at the 5′ site of the acceptor exon for both variants.

Figure 1. Detection of two novel splice variants, mCB1a and mCB1b, in cDNA derived from tissue punches of mouse brain.

(A) Schematic representation of the protein coding sequence (CDS) of the CB1 receptor. Black box: CDS, grey box: transmembrane domain (TM), arrows: primer binding sites in the sequence coding for the N-terminus, white box: location of the putative intron. (B) Detection of novel splice variants by PCR. Products from the first PCR on cDNA prepared from infralimbic cortex (IL), caudate putamen (CPu), basolateral amygdala (BLA), dorsal hippocampus (Hi) and cerebellum (Cer) were separated on a gel. The strong band at about 300 bp (mCB1), and the two weaker bands at 200 bp (a) and 130 bp (b, marked by red boxes) were excised and DNA was extracted. The bands detected in a second PCR represent mCB1, mCB1a and mCB1b. (C) Sequence alignments of mCB1 and the two novel variants mCB1a and mCB1b show the introns in the sequence coding for the N-terminal extracellular domain. Splicing of an intron of 117 bp or 186 bp at position +103 or +87 leads to different N-termini for mCB1a and mCB1b, respectively. The binding sites of the primers used to identify the novel variants are indicated in the sequence. Non-coding sequences are in lower-case letter, coding sequences are in upper-case letters, +1 indicates the translation start.

Quantification

To compare the expression levels of the novel splice variants mCB1a and mCB1b with those of mCB1, Taqman qPCR assays (primer-probe pairs) were used to selectively detect mCB1a or mCB1b. The reverse primers were designed to overlap the exon junctions in the coding sequences of mCB1a or mCB1b (Fig. 2A). To detect the overall amount of all isoforms of the mCB1 receptor, an assay amplifying a sequence in the transmembrane domain was used. Expression levels of mCB1a and mCB1b were normalized to the expression level of all isoforms. Quantification revealed that both mCB1a and mCB1b have very low expression levels, between 0.02% and 0.1% relative to the overall mCB1 receptor level in all the analyzed brain tissues of C57Bl/6N mice (Fig. 2B; n=5). The expression level of mCB1b was significantly lower than expression of mCB1a in the infralimbic cortex (p<0.001) and cerebellum (p<0.001), with no significant difference in the hippocampus (p=0.098).

Figure 2. Quantification of mCB1a and mCB1b.

(A) Schematic illustration of the amplicons of the Taqman qPCR assays used to detect mCB1a, mCB1b and mCB1. For selective amplification of mCB1a and mCB1b, the assays have exon junction-spanning reverse primers. The assay detecting all three splice variants recognizes an amplicon in the transmembrane domain. (B) Quantification of the ratio of the novel splice variants to the complete amount of mCB1 receptor in infralimbic cortex (IL), caudate putamen (CPu), basolateral amygdala (BLA), dorsal hippocampus (Hip) and cerebellum (Cer) of C57BL/6N mice (n=5). Data are mean + SEM; ***p<0.001.

Protein variants and potential differences in N-linked glycosylation

To analyze the molecular weight of mCB1a and mCB1b and compare it with the molecular weight of mCB1, Western blots of whole protein extracts from stably transfected cells were performed (Fig. 3A). The calculated molecular weights are 53.9 kDa for HA-mCB1, 49.7 kDa for HA-mCB1a and 47.0 kDa for HA-mCB1b. In the N-terminal extracellular tail, mCB1 has two putative N-linked glycosylation sites with the consensus sequence Asn(N)-X-Ser(S)/Thr(T) at amino acid positions 77 and 83 (Fig. 3B). Glycosylated mCB1 is running approximately 11 kDa higher than the non-glycosylated protein, resulting in the observed band of around 64 kDa. The splice variant mCB1a is running a little lower than mCB1, but still seems to be fully glycosylated as it is running approximately 11 kDa higher than the calculated molecular weight of 49.7 kDa. In variant mCB1b, both of the glycosylation sites are removed by splicing. The band representing mCB1b is running below 50 kDa, which indicates a strong reduction of glycosylation.

Signaling efficiencies of mCB1 and the two novel splice variants mCB1a and mCB1b

The CB1 receptor is known to rapidly internalize following binding of several agonists and receptor activation (Hsieh et al. 1999). To investigate whether the modified N-termini of mCB1a and mCB1b influence trafficking of the receptors upon activation, agonist-induced internalization was analyzed by confocal imaging (Fig. 4A). After 1 h treatment with the synthetic CB1 receptor agonist WIN55,212-2 at a concentration of 1 μM, surface receptor loss was stronger for mCB1a (Fig. 4A columns iii, iv) and mCB1b (Fig. 4A colums v, vi) than for unspliced mCB1 (Fig. 4A columns i, ii). Internalization was then quantified using on-cell Western analysis. Cells expressing mCB1, mCB1a or mCB1b were stimulated with different concentrations of WIN55,212-2 and after 1 h, the amount of receptors on the surface was quantified. Starting from a concentration of 5 nM or 10 nM, respectively, mCB1a and mCB1b showed greater internalization than the unspliced mCB1 receptor upon treatment with increasing agonist concentrations (Fig. 4B). Analysis of the time course of surface receptor loss after stimulation with 1 μM WIN55,212-2 showed that shortly after stimulation (5 min) more mCB1a is internalized compared to mCB1b (Fig. 4C). Fifteen minutes after stimulation with agonist, mCB1a shows greater internalization than unspliced mCB1. Between 30 minutes and 2 h after stimulation, both mCB1a and mCB1b showed similar internalization, which was significantly greater than that of the unspliced mCB1 receptor.

Figure 4. Different trafficking of mCB1a and mCB1b as compared with mCB1.

(A) HA-immunostaining of HEK293 cells stably expressing mCB1-HA (columns i,ii), mCB1a-HA (columns iii,iv) or mCB1b-HA (columns v,vi). Cells were treated with vehicle (i,iii,v) or stimulated with 1 μM WIN55,212-2 (ii,iv,vi) for 1 h. Following agonist treatment, mCB1 and the splice variants mCB1a and mCB1b were internalized. Surface receptor loss was stronger for mCB1a and mCB1b. Cells were counterstained with WGA (green) and DAPI (blue) to visualize cell membranes and nuclei, respectively. (B) Dose-response curve: Following 1 h exposure to increasing concentrations of WIN55,212-2, mCB1 (black) and the novel splice variants mCB1a (red) and mCB1b (blue) internalized in a concentration-dependent manner, with the internalization of the variants mCB1a and mCB1b being stronger at any agonist concentration than of the unspliced mCB1 receptor. (C) Time course: Cells expressing mCB1, mCB1a or mCB1b were stimulated with 1 μM WIN55,212-2 and the loss of surface receptors with time was quantified. After 5 min, surface mCB1a was reduced compared with mCB1 and mCB1b. At later time points, mCB1a and mCB1b were reduced on the cell surface compared with mCB1. Data are mean +/− SEM; n=12 from 4 independent experiments; ***p<0.001, **p<0.01, *p<0.5 in post-hoc test vs. mCB1; ^p<0.05 in post-hoc test of mCB1a vs. mCB1b.

Analysis of receptor trafficking to different concentrations of WIN55,212-2 was also measured in presence of CB1 inverse agonist. 10 nM SR141716 caused a shift to the right in the WIN55,212-2 dose response curve for all three mCB1 variants (Fig. 5A). The strongest EC₅₀ shift for mCB1a suggests a potentially higher affinity for SR141716 of this splice variant (Fig 5B).

Figure 5. Receptor internalization is differentially attenuated by SR141716.

(A) Surface receptor loss in HEK293 cells stably expressing mCB1 (black), mCB1a (red) or mCB1b (blue) in response to increasing concentrations of WIN55,212-2 in absence (solid squares) or presence (hollow triangles) of 10 nM SR141617. Lines show nonlinear regression fits. (B) Bar graph displaying EC₅₀ ± 95% confidence interval (CI), n=4. * indicates non-overlapping 95% CIs. See Methods for statistical details.

To analyze whether the splice variants differentially influence downstream signaling pathways, activation of the MAPK pathway was investigated since CB1 receptor activation leads to phosphorylation of p42/p44 MAPK (Bouaboula et al. 1995). Stimulation of cells expressing the mCB1 receptor with 100 nM WIN55,212-2 resulted in a transient elevation of p44/42 MAPK phosphorylation with a peak activation of 150% over baseline at 5 min (Fig. 6). Cells expressing mCB1a showed a similar response to cells expressing mCB1. Stimulation of cells expressing mCB1b led to significantly less MAPK phosphorylation at 5 min with a slightly delayed peak response of 50% over basal activation at 7.5 min.

Figure 6. Agonist stimulation of mCB1b transfected cells leads to reduced p44/42 MAPK activation.

The time course of p44/p42 phosphorylation was quantified in HEK293 cells stably expressing mCB1 (black), mCB1a (red) or mCB1b (blue) after stimulation with 100 nM WIN55,212-2. Cells expressing mCB1b showed a strongly reduced response compared with cells expressing mCB1 or mCB1a. Data are mean +/− SEM; n=9 from 3 independent experiments; ***p<0.001 in post-hoc test vs. mCB1 or mCB1a.

mCB1 and splice variant receptors signal similarly in autaptic hippocampal neurons

Depolarization of autaptic hippocampal neurons can result in a form of retrograde inhibition termed depolarization induced suppression of excitation (DSE) (Straiker and Mackie 2005). This can be quantified by a series of successively longer depolarizations (50 ms, 100 ms, 300 ms, 500 ms, 1 s, 3 s, 10 s) resulting in progressively greater inhibition of neurotransmission (Straiker et al. 2011). This produces a “depolarization-response curve” that permits the characterization of some pharmacological properties of cannabinoid signaling including the calculation of an effective-dose 50 (ED₅₀), corresponding in this case to the duration of depolarization that results in 50% of the maximal response. We have previously reported that DSE is fully rescued by transfection of rCB1 into neurons cultured from CB1 receptor deficient mice (Straiker et al. 2011). We found that transfection of mCB1 similarly rescued DSE (Fig. 7A,B; ED₅₀ (95% CI): 1.94 s (1.20–3.41)), and that mCB1 splice variants signaled similarly to mCB1 (Fig 7C,D; mCB1a ED₅₀ (95% CI): 0.68 s (0.40–1.17); mCB1b ED₅₀ (95% CI): 1.33 s (0.70–2.55)). No differences were found in the extent of DSE (maximal response) between the splice variants. A statistically significant difference in ED₅₀’s from mCB1 was seen for mCB1a but not mCB1b. However, the difference is modest and contrasts to our findings for hCB1 splice variants, where much larger differences in DSE were detected (Straiker et al. 2012).

Figure 7. mCB₁ and splice variant receptors signal similarly in autaptic hippocampal neurons.

(A) Depolarization response curve with successively longer depolarizations. mCB1 (black squares, n=7) and rCB1 (green circles, n=8) are quite similar (rCB₁ from Straiker et al. 2011). (B) Sample time course (left panel) and EPSCs (right panel) for mCB1a-transfected into an autaptic hippocampal neuron from a CB1 knockout mouse. (C) Depolarization response curve for mCB1 (black, n=7, included from panel A for reference), mCB1a (red, n=4) and mCB1b (blue, n=5) D) Bar graph displaying EC₅₀ ± 95% confidence interval (CI) show only modest, though statistically significant, difference in signaling between mCB1 and mCB1a. * indicates non-overlapping 95% CIs. See Methods for experimental and statistical details.

Discussion

In this study, we identified two novel mouse CB1 receptor splice variants. The first variant, mCB1a, results from the excision of a 117 bp intron resulting in an in-frame deletion of 39 amino acids following amino acid residue 35. The second variant, mCB1b, results from the excision of a 186 bp intron resulting in an in-frame deletion of 62 amino acids, including two putative glycosylation sites, following amino acid residue 28.

As for the human CB1 receptor splice variants, both variants of the mouse CB1 receptor are generated by alternative splicing events in which a specific sequence may either be spliced out as an intron or be retained (Matlin et al. 2005), which is the rarest mode of alternative splicing in mammals (Sammeth et al. 2008). However, N-terminal splice variants were described to be unique for the human CB1 receptor (Howlett et al. 2002; Xiao et al. 2008) as the consensus splice donor and acceptor sequences present in the CDS of the human CNR1 gene are absent from the same positions in the mouse Cnr1 gene. The introns we found in the mouse CB1 receptor gene are at different positions from the human variants generated by alternative splicing. In addition, only the 3′ end of the exons matches the splice consensus sequence (AG), whereas the 5′ end of the introns differs from the consensus sequence (consensus: GT, mCB1a: GA, mCB1b: TA). The splice acceptor sites correspond to the consensus sequence (AG/G). The transcript variants of the mouse CB1 receptor that are generated by alternative splicing are expressed at very low levels compared with the unspliced CB1 receptor mRNA. This might be explained by less efficient assembly of the splice machinery at splice sites differing from the consensus sequences because of lower affinity interactions with the splicosome components.

Expression levels of the human CB1 receptor splice variants have only been estimated by semi-quantitative RT-PCR (Shire et al. 1995; Ryberg et al. 2005; Xiao et al. 2008; Bagher et al. 2013) and to our knowledge have not yet been analyzed by qPCR. The large N-terminus of the hCB1 receptor was shown to inhibit efficient receptor translocation across the membrane of the endoplasmatic reticulum (ER), leading to large amounts of misfolded CB1 receptor that are rerouted towards proteasome degradation (Andersson et al. 2003). Shortening the N-terminus of the hCB1 receptor or the inclusion of a signal peptide for ER export greatly increases receptor stability, and both result in increased targeting to the cell surface (Andersson et al. 2003). Partial truncation of the N-terminal tail of the human CB1 receptor was detected in various cell lineages in vitro, and is due to the fast proteolytic processing of de novo synthesized receptors in the cytoplasm prior to their translocation over the ER via a mechanism independent of the proteasome (Nordström and Andersson 2006). Co-expression of the hCB1 receptor and each of the splice variants increased cell surface expression of the hCB1 receptor in HEK cells (Bagher et al. 2013), but the cellular distribution in vivo still remains to be determined. We could not investigate the relative amount of mCB1a and mCB1b and their cellular distribution in vivo as there are no isoform-specific antibodies available, but it is conceivable that shorter N-terminal tails might facilitate the surface targeting of the murine splice isoforms.

In order to characterize possible differences in protein expression and signaling efficiencies, HA-tagged mCB1, mCB1a, or mCB1b were stably expressed in HEK cells. In Western blots with whole protein extracts, mCB1 and the splice variant mCB1a had a higher molecular weight than calculated from the amino acid sequences, but variant mCB1b was running as expected for its calculated molecular weight. An elevated molecular weight was previously described for the mature, glycosylated CB1 receptor in several species (Song and Howlett 1995; Onaivi et al. 1996; Andersson et al. 2003). In the N-terminal extracellular tail, CB1 has two putative N-linked glycosylation sites with the consensus sequence Asn(N)-X-Ser(S)/Thr(T) at amino acid positions 77 and 83 that are conserved in human, rat and mouse. In the rat, both potential N-linked glycosylation sites are glycosylated, as treatment with endoglycosidase shifted the 64 kDa band to two 59 kDa and 53 kDa bands (Song and Howlett 1995). In the splice variant mCB1a, these putative glycosylation sites are still present, whereas in mCB1b, they are removed by splicing, thus explaining the lower molecular weight. The functional significance of the N-glycosylation in the N-terminal domain of the CB1 receptor is not yet clear (Basavarajappa et al. 2008). Human CB1 receptor with deletion of the first 89 amino acids (and thus, the two putative N-glycosylation sites), remained stably expressed at the cell surface, suggesting that N-linked glycosylation may not be required for transport of CB1 receptor to the plasma membrane (Andersson et al. 2003). However, in that study, receptor trafficking upon stimulation was not analyzed. In other GPCRs (e.g. β-adrenergic receptors and muscarinic receptors), mutations of N-glycosylation sites abolished glycosylation, but had no obvious effect on receptor expression and function (Dohlman et al. 1991). In recent studies, a crystal structure of the human CB1 receptor was determined (Hua et al. 2016; Shao et al. 2016). A truncated CB1 receptor lacking the first 89 residues maintained basic inverse-agonist and agonist-binding properties, and the membrane-proximal N-terminal extracellular surface of CB1 formed a critical part of the ligand binding pocket. The ligand binding pocket is capped by the N-terminal loop interactions (Met103^N-term and Ile105^N-term), which are still present in mCB1a and mCB1b.

The mCB1a and mCB1b splice variants demonstrated significant differences from mCB1 in agonist-induced internalization analyzed in stably expressing cell lines. Both novel splice variants had increased surface receptor loss in response to increasing concentrations of the agonist WIN55,212-2. Furthermore, the time course of internalization in response to 1 μM WIN55,212-2 was different, as mCB1a and mCB1b showed significantly faster internalization together with a stronger reduction of surface receptors compared with the unspliced mCB1 receptor. As the two splice variants with shortened N-termini showed very similar trafficking to each other, it is unlikely that the differences in N-linked glycosylation induced the increased surface receptor loss. Thus, either shortening of the N-terminal tail or the loss of a specific functional domain formed by amino acids missing in the splice variants leads to the observed stronger internalization. It was previously shown that phosphorylation of residues in the distal C-terminus play a role in endocytosis of the CB1 receptor (Daigle et al. 2008a), but it has not yet been studied whether the N-terminal tail is involved in the regulation of receptor trafficking.

Agonist-induced stimulation of cells expressing the mCB1 receptor resulted in a transient activation of p44/42 MAPK phosphorylation with the classical response (Dalton and Howlett 2012): a maximal response in the first 5 min followed by a rapid decline after 5–10 min. Cells expressing mCB1a showed a similar response to cells expressing mCB1. Stimulation of cells expressing mCB1b led to significantly decreased p44/42 MAPK activation at 5 min with a lower and slightly delayed peak after 7.5 min. The rapid decline after agonist-induced p44/42 MAPK activation was similar for mCB1b as compared with mCB1 and mCB1a. Previously, it was shown that the duration of p44/42 MAPK activation by the CB1 receptor is regulated by receptor desensitization (uncoupling of receptor signal transduction) and not by internalization of the receptor (Daigle et al. 2008a). The rapid decline in ERK phosphorylation after 5–10 minutes involves PKA inhibition and serine/threonine phosphatase activation (Dalton and Howlett 2012). As the rapid decline of p44/p42 MAPK phosphorylation was similar for mCB1, mCB1a and mCB1b, receptor desensitization seems to be unchanged in the splice variants. However, the peak p44/42 MAPK activation is significantly lower in cells with stable expression of mCB1b. It has been shown that the maximal response of p44/p42 MAPK activation depends on receptor-stimulated, ligand-independent transactivation of multiple receptor tyrosine kinases, requires G_i/o βγ subunit-stimulated phosphatidylinositol 3-kinase activation and Src kinase activation, and is modulated by inhibition of cAMP/PKA (Dalton and Howlett 2012). Thus, it is possible that in cells expressing the mCB1b variant, one or several of these kinases are not efficiently recruited. One might speculate that differences in N-linked glycosylation of the N-termini could account for the different maximal responses, as it was also shown that the human splice variant hCB1a, which also lacks two putative glycosylation motifs, also displayed a reduced MAPK activity after agonist-induced stimulation (Rinaldi-Carmona et al. 1996). However, in that study dose-dependent MAPK activation was measured after 10 min. At this time point, MAPK activity would be expected to decline from the maximal response. The possibility that loss of N-linked glycosylation of the N-terminus leads to reduced peak MAPK activation needs to be carefully tested by comparing the maximal p44/p42 MAPK activation in CB1 receptors with mutations in one or several of the glycosylated Asn residues in the N-terminus. Furthermore, it would be interesting to analyze the maximal response in these mutants upon inhibition of the kinases that were shown to be involved in the strength of p44/p42 MAPK activation (Dalton and Howlett 2012).

In contrast to our results in MAPK and internalization assays, cannabinoid-receptor mediated inhibition of neurotransmission in autaptic hippocampal neurons was similar between full-length receptor and either splice variant. Though statistically significant for mCB1a vs. mCB1, the differences were quite modest in comparison to our findings for hCB1 (Straiker et al. 2012). One explanation for the difference between signaling pathways is that the splice variants exhibit pathway dependence in their signaling. They might therefore profoundly alter MAPK-mediated signaling while leaving neurotransmission unaltered. This finding also contrasts with our previous findings in neurons for human splice variants (Straiker et al. 2012); in that study we found that the splice variants substantially altered DSE. It should be noted that DSE represents one of several identified forms of neuromodulation mediated by cannabinoids (reviewed in Kano et al. 2009). Hence, the splice variants may differentially alter various forms of cannabinoid-mediated neuronal plasticity.

The physiological relevance of the novel splice variants of the mouse CB1 receptor is not yet clear, and might be expected to be modest due to their low expression on mRNA level. However, the splice variants can be valuable tools to improve our understanding of CB1 receptor signaling and trafficking, in particular the role of the N-terminal domain and conformational alterations brought upon by posttranslational modifications.

List of abbreviations used in the text

Δ⁹-THC

Δ⁹-tetrahydrocannabinol

BSA

bovine serum albumin

CA

Cornu ammonis

CB1

cannabinoid receptor type 1

CDS

coding sequence

DSE

depolarization induced suppression of excitation

ED₅₀

effective-dose 50

ER

endoplasmatic reticulum

FBS

fetal bovine serum

FITC

fluorescein isothiocyanate

HA

hemagglutinin

HBS

HEPES-buffered saline

hCB1

human CB1

MAPK

mitogen-activated protein kinase

mCB1

mouse CB1

PFA

paraformaldehyde

PBS

phosphate-buffered saline

TBS

Tris-buffered saline

WIN55,212-2

R-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinyl)methyl]pyrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanonemesylate

SR141716

5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-N-1-piperidinyl-1H-pyrazole-3-carboxamide