Tissue- and cell-specific expression of a splice variant in the II-III cytoplasmic loop of Cacna1b.

Presynaptic CaV 2.2 (N-type) channels are fundamental for transmitter release across the nervous system. The gene encoding CaV 2.2 channels, Cacna1b, contains alternatively spliced exons that result in functionally distinct splice variants (e18a, e24a, e31a, and 37a/37b). Alternative splicing of the cassette exon 18a generates two mRNA transcripts (+e18a-Cacna1b and ∆e18a-Cacna1b). In this study, using novel mouse genetic models and in situ hybridization (BaseScope™), we confirmed that +e18a-Cacna1b splice variants are expressed in monoaminergic regions of the midbrain. We expanded these studies and identified +e18a-Cacna1b mRNA in deep cerebellar cells and spinal cord motor neurons. Furthermore, we determined that +e18a-Cacna1b is enriched in cholecystokinin-expressing interneurons. Our results provide key information to understand cell-specific functions of CaV 2.2 channels.

cornu ammonis 1

cornu ammonis 3

cannabinoid receptor 1

cholecystokinin

cannabinoid receptor 1 gene

deep cerebellar nuclei

dorsal raphe nuclei

granular cell layer

glutamate decarboxylase 2

glyceraldehyde 3‐phosphate dehydrogenase

hilus

locus coeruleus

molecular layer

pyramidal cell layer

pyramidal neurons

stratum pyramidale

stratum radiatum

substantia nigra

substantia nigra pars compacta

ventral tegmental area

Presynaptic CaV2.2 (N‐type) channels are key mediators of transmitter release across the nervous system. The calcium that enters through CaV2.2 channels in response to action potentials triggers transmitter release 1, 2. Cacna1b is a multi‐exon gene that encodes the CaVα1 pore‐forming subunit of CaV2.2 channels. The Cacna1b pre‐mRNA contains more than 40 exons that are spliced during mRNA maturation 3. Most Cacna1b exons are constitutively included in the final Cacna1b mRNA; however, some are selectively included through alternative splicing (e18a, e24a, e31a, and 37a/37b) 4. Alternative splicing of the Cacna1b pre‐mRNA originates splice variants that are translated into CaV2.2 channels with distinct biophysical properties and pharmacological profiles 5, 6. CaV2.2 splice variants with differences in activation, inactivation, and current density 7, 8, 9, as well as in their response to neurotransmitters and drugs such as GABA and morphine, have been previously characterized 10, 11. Thus, alternative splicing of the Cacna1b pre‐mRNA is thought to provide functional diversification to cells that utilize CaV2.2 channels to release neurotransmitter.

Among the alternative spliced exons in the Cacna1b, pre‐mRNA is the cassette exon 18a (e18a) 12. Alternative splicing of e18a generates two splice variants, +e18a‐Cacna1b (e18a is included) and ∆e18a‐Cacna1b (e18a is skipped) (Fig. 1A). E18a encodes 21 amino acids within the ‘synprint’ region of the II‐III cytoplasmic loop of CaV2.2 (LII‐III), an essential region for interaction of CaV2.2 channels with presynaptic proteins 13, 14, 15. Functionally, +e18a‐CaV2.2 channels are more resistant to cumulative inactivation induced by repetitive stimulation and closed‐state inactivation than ∆e18a‐CaV2.2 channels 9. Furthermore, +e18a‐CaV2.2 channels exhibit larger calcium currents relative to ∆e18a‐CaV2.2 channels in both mammalian expression systems and neurons 16. Despite all of this information, very little is known about the functional role of +18a‐Cacna1b and ∆e18a‐Cacna1b splice variants in the nervous system.

Figure 1

Targeted deletion of e18a in the Cacna1b gene. (A) Schematic of splicing patterns for Cacna1b (CaV2.2) pre‐mRNA in WT and ∆e18a‐only mice. In WT mice, e18a is spliced to generate two mRNA transcripts, +e18a‐ and ∆e18a‐Cacna1b, which in turn are translated into two different CaV2.2 channels, +e18a‐ and ∆e18a‐CaV2.2. The Cacna1b gene was modified to remove e18a and its flanking intronic regions to generate e∆18a‐only mice, which generate only the ∆e18a‐CaV2.2 splice variant. (B) Top, Schematic of Cacna1b pre‐mRNA. Arrows indicate the approximate location of RT‐PCR primers flanking e18a and qPCR primers in constitutive exons 45 and 46. Black box shows the approximate location of qPCR probe spanning exon junction 45–46. Bottom left, RT‐PCR from whole‐brain samples of WT and ∆e18a‐only mice. Bottom right, comparison of whole‐brain Cacna1b mRNA levels between WT and ∆e18a‐only mice. Data are shown as mean (filled symbols) ± SEM and individual values for biological replicates (empty symbols).

To understand the functional role of +e18a‐Cacna1b and ∆e18a‐Cacna1b splice variants, it is necessary to determine their tissue and cell‐type expression. Previous studies have shed light on this. In the adult nervous system, the abundance of +e18a‐Cacna1b mRNA differs among tissues. Higher levels of +e18a‐Cacna1b mRNA are observed in dorsal root ganglia, superior cervical ganglia, and spinal cord relative to whole brain 6, 16. Interestingly, the relative abundance between +e18a‐Cacna1b and ∆e18a‐Cacna1b mRNAs also differs among brain subregions. Less than 10% of Cacna1b splice variants contain e18a in whole cerebral cortex and hippocampus, whereas 30–60% of Cacna1b splice variants contain e18a in the thalamus, cerebellum, hypothalamus, and midbrain 12. Within the midbrain, ~ 80% of Cacna1b splice variants contain e18a in monoaminergic regions including the ventral tegmental area (VTA), substantia nigra (SN), dorsal raphe nuclei (DRN), and locus coeruleus (LC) 17.

The expression of +e18a‐Cacna1b is also cell‐specific. +e18a‐Cacna1b mRNA is enriched in tyrosine hydroxylase‐expressing cells of the SN pars compacta (SNc) and VTA 17. +e18a‐Cacna1b has also been identified in magnocellular neurosecretory cells of hypothalamus and capsaicin‐responsive neurons of dorsal root ganglia 7, 18. The molecular mechanisms underlying the tissue‐ and cell‐specific expression of +e18a‐Cacna1b and ∆e18a‐Cacna1b mRNAs are beginning to be elucidated. The RNA‐binding protein, Rbfox2, is a splicing factor that represses e18a by binding an intronic region upstream of this exon 16, 19. Rbfox2 expression and activity depend on the cell type; thus, this would help to explain the cell‐specific expression of +e18a‐Cacna1b and ∆e18a‐Cacna1b 20, 21.

In this study, we utilized a novel version of in situ hybridization (ISH) (BaseScope™) and genetic mouse models, to precisely determine the expression of +e18a‐Cacna1b in the central nervous system. We generated a mouse model that lacks expression of +e18a‐Cacna1b, thereby expresses only the ∆e18a‐Cacna1b splice variant (∆e18a‐only). This mouse was used to control for probe specificity in our BaseScope™ experiments. With these new approaches and tools, we confirmed that +e18a‐Cacna1b mRNA is expressed in monoaminergic regions (SN, VTA, DRN, and LC). Furthermore, we identified new cell populations that express +e18a‐Cacna1b mRNA such as cells of the deep cerebellar nuclei (DCN) and spinal cord motor neurons. Finally, using fluorescence‐activated cell sorting (FACS) of genetically identified cell populations coupled to RT‐PCR, we found that +e18a‐Cacna1b mRNA is more abundant in cholecystokinin‐expressing interneurons (CCK+INs) relative to Ca2+/calmodulin‐dependent protein kinase IIα‐expressing projection neurons [CaMKIIα+ pyramidal neurons (PNs)].

Materials and methods

Housing conditions

A combination of adult males or females was used in all of our experiments, and no association with sex was found in the amount of e18a in Cacna1b in whole brain. Mice were housed with food and water ad libitum in temperature‐controlled rooms with a 12‐h light/dark cycle. All experimental procedures followed the guidelines of the Institutional Animal Care Committee of the University of New Hampshire.

Mouse lines

C57BL/6 wild‐type mice were used in our experiments. Mice lacking e18a or ∆e18a‐only (Cacna1b) were back‐crossed in C57BL/6 (Charles River) background for 6–8 generations, and for details on how this mouse line was generated, see Ref. 16. To label CCK+INs with tdTomato (tdT), we performed intersectional genetic labeling as previously reported 22, 23. We utilized CCK‐Cre (Cck/J, Jax: 012706) in C57BL/6 background, Dlx5/6‐Flpe (Tg(mI56i‐flpe)39Fsh/J, Jax: 010815) in FVB/NJ background, and Ai65‐D (B6;129S‐Gt(ROSA)26Sort, Jax: 021875) in C57BL/6;I129 background mice. Using these three mouse lines, we generated a triple transgenic line, CCK‐Cre; Dlx5/6‐Flpe; Ai65‐D (named CCK;Dlx5/6;tdT) as follows: First, CCK‐Cre mice were crossed with Dlx5/6‐Flpe two times to produce a dual transgenic mouse (homozygous for CCK‐Cre and heterozygous for Dlx5/6‐Flpe). Next, this dual transgenic mouse line was crossed with homozygous Ai65‐D mice. From the resulting offspring, we selected only heterozygous mice for the three alleles. To induce the expression of tdT in PNs, we crossed CaMKIIα‐Cre mice (B6.Cg‐Tg(Camk2a‐cre)T29‐1Stl/J, Jax: 005359) in a mixed C57BL6 background with Ai14 mice (B6.Cg‐Gt(ROSA)26Sor, Jax: 007914) in C57BL6. The resulting dual transgenic mouse line CaMKIIα;tdT was heterozygous for both alleles.

Genotyping

Conventional toe biopsy was performed on P7‐P9 pups. Genomic DNA was extracted using Phire Animal Tissue Direct Kit II (Thermo Fisher Scientific, Waltham, MA, USA, F140WH) according to the manufacturer’s instructions. Next, PCR was performed with AmpliTaq Gold® 360 Master Mix (Thermo Fisher Scientific) using the following conditions: a hot start of 95 °C for 10 min, followed by 35 cycles (95 °C, 30 s; 60 °C, 30 s; and 72 °C, 1 min), and final step of 72 °C for 7 min. Primers and expected products are shown in Table 1. Primers were added together to genotype each mouse line.

Table 1

Primers to perform genotyping of mouse lines and expected products.

Mouse line	Primers	Expected products
e18a‐null	F‐MT: CCATGTCTGCTGCCAATATCT F‐WT: GCAGCAGAGGTTCTGTTTGC R‐C: CTCGGTGCTTTCTGTCCTGTCC	Hom: 395 bp Het: 395 and 880 bp WT: 880 bp
CCK‐Cre	F‐WT1: GGGAGGCAGATAGGATCACA F‐MT1: TGGTTTGTCCAAACTCAT CAA R: GAGGGGTCGTATGTGTGGTT	Hom: 180 bp Het: 180 bp and 468 bp WT: 468 bp
Dlx5/6‐Flpe	F‐T1: CAGAATTGATCCTGGGGAGCT ACG R‐T1: CCAGGACCTTAGGTGGTGTTT TAC F‐C: CAAATGTTGCTTGTCTGGTG R‐C: GTCAGTCGAGTGCACAGT TT	Transgene: 406 bp PCR positive control: 200 bp PCR conditions do not differentiate between heterozygous and homozygous mice
Ai14 and Ai65‐D	F‐WT2: AAGGGAGCTGCAGTGGAG TA R‐WT2: CCGAAAATCTGTGGGAAG TC R‐MT1: GGCATTAAAGCAGCGTAT CC F‐MT2: CTGTTCCTGTACGGCATGG	Hom: 196 Het: 196 and 297 WT: 297
CaMKIIα‐Cre	F‐T1: GTT CTC CGT TTG CAC TCA GG R‐T1: CAG GTT CTT GCG AAC CTC AT F‐C: AGT GGC CTC TTC CAG AAA TG R‐C: TGC GAC TGT GTC TGA TTT CC	Transgene: ~ 500 bp PCR positive control: 521 bp

All primers are reported in 5’ to 3’ direction.

In situ hybridization (BaseScope™)

Mice were deeply anesthetized with Euthasol (Virbac, Centurion, South Africa, 200‐071). Next, mice were transcardially perfused with 1× PBS for 10 min and subsequently with 10% neutral buffered formalin (NBF, ~ 4% formaldehyde) fixative solution (Sigma, St. Louis, MO, USA, HT501128) for 10 min, and then again with 1× PBS for 10 min. Both PBS and NBF were kept on ice during the perfusion. Brain and spinal cord were immediately removed and postfixated in 10% NBF at 4 °C for 24 h. After washing with 1× PBS, tissue was sequentially dehydrated in 15% and 30% sucrose : PBS solutions for at least 18 h in each sucrose concentration, or until tissue sank to the bottom of the tube. Next, tissue was cryopreserved in optimal cutting temperature compound, or OCT (Fisher, 4585), with isopentane prechilled in dry ice. Twelve micrometer cryosections from brain and spinal cord were collected (Shandon, ThermoFisher Scientific, 77200222) and placed in prechilled 15‐mm Netwell™ inserts (Corning, NY, USA, 3478). Sections were allowed to free‐float in 1× PBS and mounted on positively charged microscope slides (VWR, 48311‐703) using a paintbrush. After air‐drying for 10–20 min, sections were incubated at 60 °C in a drying oven for 30 min to aid tissue adhesion to the slides. Before ISH, sections were postfixed in 10% NBF at 4 °C for 15 min and then dehydrated with 50%, 70%, and two rounds of 100% ethanol for 5 min in sequential steps. Slides were air‐dried for an additional 5 min before incubating with RNAscope® hydrogen peroxide solution for 10 min (ACD, Newark, CA, USA, 322381). Next, sections were washed with Milli‐Q water and transferred to RNAscope® Target Retrieval solution, preheated to 99 °C for 15 min (ACD, 322000). Sections were briefly washed with Milli‐Q water, transferred to 100% ethanol for 3 min, and then placed in drying oven at 60 °C for 30 min. Sections were isolated with a hydrophobic barrier pen (ACD, 310018) and air‐dried overnight at room temperature. Following incubation with RNAscope® Protease III for 30 min (ACD, 322381) at 40 °C in the ACD HybEZ™ Hybridization System (ACD, 310010), sections were exposed to a probe spanning e18 and e18a (BaseScope™ Mm‐Cacna1b‐e18e18a) (ACD, 701151) for 2 h at 40 °C in the ACD HybEZ™ Hybridization System. BaseScope™ Detection Reagents AMP 0–AMP 6 and FastRed (ACD, 322910) were applied according to the manufacturer’s instructions and washed using RNAscope® Wash Buffer (ACD, 310091). To visualize nuclei, sections were counterstained in Gill’s Hematoxylin I for 2 min at RT (American Master Tech, Lodi, CA, USA, HXGHE). Next, sections were washed with tap water three times, briefly transferred to 0.02% ammonia water, and washed again with tap water. Finally, sections were dried at 60 °C for 15 min and mounted using VectaMount™ mounting medium (Vector Laboratories, Burlingame, CA, USA, H‐5000). Phase‐contrast images were acquired using an Olympus IX‐81 microscope.

Fluorescence‐activated cell sorting

Adult CCK;Dlx5/6;tdT and CaMKIα;tdT mice were deeply anesthetized with isoflurane. After decapitation, brains were quickly removed and placed on a petri dish with Earl’s balanced salt solution (EBSS) (Sigma, E3024) containing 21 U·mL−1 of papain. Rapid dissection (< 45 s) of cerebral cortex and hippocampus was performed. Then, tissue was dissociated using a modified version of Worthington Papain Dissociation System® (Worthington Biochemical Corporation, Lakewood, NJ, USA, LK003150). After incubating with papain for 45 min at 37 °C on a rocking platform, tissue was triturated with three sequential diameter fire‐polished glass pipettes. Next, cell suspensions were centrifuged at 300  for 5 min. After discarding supernatants, pellets were resuspended in 3 mL of EBSS containing 0.1% of ovomucoid protease inhibitor and 0.1% BSA (Worthington, LK003182) to quench papain. Cell suspension was centrifuged at 270  for 6 min and resuspended in EBSS (3 mL). To isolate tdT‐expressing cells, we performed FACS in a Sony SH800 flow cytometer using a 561 nm laser to excite and a 570‐ to 630‐nm filter for event selection. At least 300 000 events were collected directly into TRIzolTM LS Reagent (Thermo Fisher Scientific, 10296028). Collection was performed keeping 1 : 3 (v/v) sorted cell suspension: TRIzolTM LS ratio. Cell suspension was kept on ice throughout the sorting session.

RT‐PCR and RT‐qPCR

Total RNA from tissue was extracted with RNeasy Mini Kit columns (Qiagen, Hilden, Germany, 74134) according to the manufacturer’s instructions. Total RNA from sorted cells was extracted using TRIzol LS and isopropanol precipitation with the addition of 30 μg of GlycoBlue® Coprecipitant (Thermo Fisher Scientific, AM9516) to facilitate visualization of RNA pellet. 1 μg (tissue) or 300 ng (sorted cells) of total RNA was primed with oligo‐dT and reverse‐transcribed with Superscript IV First‐Strand Synthesis System (Thermo Fisher Scientific, 18091050) according to the manufacturer’s instructions. To quantify the relative amount of e18a, PCR was performed using AmpliTaq Gold® 360 Master Mix (Thermo Fisher Scientific, 4398881) with primers flanking e18a (F: 5′GGCCATTGCTGTGGACAACCTT and R: 5′CGCAGGTTCTGGAGCCTTAGCT) with the following conditions: hot start at 95 °C for 10 min, 28 cycles (95 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min), and a final step of 72 °C for 7 min. These sets of primers quantify +e18a‐ and ∆e18a‐Cacna1b splice variants simultaneously. PCR products were run in 3% agarose gel stained with ethidium bromide, and densitometric analysis was performed using imagej 24. This quantification method has been validated before 16. To confirm band identity, the two bands were cloned and sequenced. To quantify total mRNA levels for Cacna1b, glutamate decarboxylase‐2 (Gad‐2) and cannabinoid receptor 1 (Cnr1), we performed TaqMan® real‐time PCR assays (Thermo Fisher Scientific) with the following probes: Cacna1b, Mm01333678_m1; Gad‐2, Mm00484623_m1; Cnr1, Mm01212171_s1; and glyceraldehyde 3‐phosphate dehydrogenase (Gapdh), Mm99999915_g1, which was used as constitutive control. First‐strand cDNA was diluted 1 : 5, and 4 μL of this dilution was used in a 20 μL qPCR containing Taq polymerase master mix (Applied Biosystems, Foster City, CA, USA, 4369016) and the predesigned probes mentioned above. RT‐qPCRs were run on an ABI 7500 Fast Real‐Time PCR System (Applied Biosystems) with the following conditions: 1 cycle 95 °C for 10 min, 4 °C ycles (95 °C for 15 s and 60 °C for 1 min). Each sample from at least five different mice per genotype (biological replicates) was run in triplicate (technical replicates). C values were determined by 7500 Software v2.3 (Applied Biosystems). Relative quantification of gene expression was performed with ∆∆‐C method 25.

Statistical analysis

Two‐tailed unpaired Student’s t‐test was performed in Excel (Microsoft).

Results and Discussion

Validation of a mouse model to detect +e18a‐Cacna1b mRNA in the central nervous system

To determine the localization of +e18a‐Cacna1b mRNAs in the central nervous system (CNS), we performed a modality of ISH (BaseScope™). We used a mouse line with targeted deletion of e18a (∆e18a‐only) to control for probe specificity. In this mouse model, DNA sequences between e18 and e19 (e18‐e18a intron, e18a, and e18a‐e19 intron) were removed using homologous recombination (Fig. 1A, 16). To confirm deletion of e18a sequence, we performed RT‐PCR in whole‐brain samples from WT and ∆e18a‐only mice. We utilized primers flanking e18a to quantify both +18a‐ and ∆e18a‐Cacna1b transcripts. Amplicons for each splice variant were resolved on gel electrophoresis based on size (Fig. 1B). In WT whole‐brain samples, we observed two bands, ~ 290 and ~ 230 bp, corresponding to +18a‐ and ∆18a‐Cacna1b transcripts, respectively (Fig. 1B, bottom left panel, lanes 1–3). As expected, the upper band was absent in samples from ∆e18a‐only mice (Fig. 1B, bottom left panel, lanes 4–6). Next, we tested whether the targeted deletion of e18a alters the overall Cacna1b mRNA levels. We quantified total Cacna1b mRNA using RT‐qPCR with a probe that spans the splice junction between two constitutive exons, e45 and e46 (Fig. 1B, top panel). We found no significant differences in the total amount of Cacna1b mRNA between WT and ∆e18a‐only mice (Fold change relative to control ± SEM: WT = 1 ± 0.12, n = 7; e18a‐null = 0.89 ± 0.18, n = 7. P = 0.61. Two‐tailed, unpaired, Student’s t‐test. Fig. 1B, right lower panel). Our results show that the e18a sequence was successfully eliminated from the Cacna1b gene and that this deletion does not alter the total Cacna1b mRNA levels in whole brain. Therefore, this model is ideal to control for the specificity of probes directed to e18a, thereby allowing the localization of +e18a‐Cacna1b splice variants in CNS tissue.

+e18a‐Cacna1b mRNA is expressed in substantia nigra and ventral tegmental area

In the CNS, expression of Cacna1b is restricted to neurons; however, the cell‐specific expression of the Cacna1b splice variants has not been fully determined 4. Previous studies using microdissections and RT‐PCR showed that +e18a‐Cacna1b mRNA is abundantly expressed in SN and VTA 17. Furthermore, +e18a‐Cacna1b mRNAs colocalize with tyrosine hydroxylase mRNA in both of these brain areas 17, suggesting that +e18a‐Cacna1b mRNA is enriched in dopaminergic neurons. Here, we confirmed these findings using BaseScopeTM with a probe designed against the e18a sequence. Briefly, ‘Z’ probes containing a short complementary region bind to the e18a sequence. This binding leads to the assembly of a signal amplification system, thereby allowing the detection of short RNA sequences (Fig. 2A). In our experiments, we used brain sections from ∆e18a‐only mice to control for probe specificity. Red signal indicates the presence of +e18a‐Cacna1b splice variants, and blue indicates counterstaining with hematoxylin (Hem) (Fig. 2A). Each dot detects a single mRNA molecule. We performed BaseScopeTM in midbrain sections of WT and ∆e18a‐only mice. We compared our BaseScopeTM images (Fig. 2B, top panels) to conventional ISH images for the dopaminergic marker, Slc6a3 dopamine transporter, from SN and VTA found in the Allen Mouse Brain Atlas (Fig. 2B, bottom panels) 26. We observed that staining for e18a follows a pattern similar to Slc6a3 in the midbrain. No signal for e18a was detected on sections from ∆e18a‐only mice (Fig. 2B, bottom panels). These results confirm previous studies showing that e18a is expressed in SN and VTA regions. It is important to note that we observed +e18a‐Cacna1b expression in other cells near VTA and SNc; the identity of these cells is currently unknown. Some cells expressing +e18a‐Cacna1b were also observed in SN pars reticulata.

Figure 2

Localization of e18a‐Cacna1b in dopaminergic midbrain areas using BaseScopeTM. (A) Schematic of workflow for BaseScope™. Complementary Z probes were designed to target the junction e18 and e18a. These two independent Z probes bind to the Cacna1b mRNA in tandem, thereby allowing the assembly of an amplification complex. Subsequent signal amplification leads to the development of red coloration. Sections were counterstained with Hem. (B) Top panels, BaseScope™ images of VTA and SN from WT mouse brain (left). Insets of VTA and SN (middle and right, respectively). Blue indicates all nuclei stained with Hem. Red dots indicate the presence of +e18a‐Cacna1b mRNA molecules. Middle panels, BaseScopeTM images from areas similar to top panels in ∆e18a‐only mice. Bottom panels, ISH images for Slc6a3 in SN and VTA from brain‐map.org (left). Black squares represent sections within SN and VTA that were magnified for clarity (insets 1–6). Images credit: Allen Institute. Scale bar = 400 μm.

To our knowledge, antibodies specific for +e18a‐CaV2.2 channels are unavailable. This would help to convincingly show that +e18a‐CaV2.2 channels are present in dopaminergic neurons. However, our use of BaseScopeTM with adequate negative controls sheds light on the localization of e18a in the midbrain. Previous results have shown that dopamine release from VTA and SNc heavily relies on CaV2.2 channels 27, 28. Our studies and mouse models combined with previous functional studies of +e18a‐ and ∆e18a‐CaV2.2 channels in mammalian systems and neurons will enable to propose further studies to unveil a potential role of e18a splicing on dopamine release.

Expression of +e18a‐Cacna1b mRNA in dorsal raphe nuclei and locus coeruleus

DRN and LC contain neurons that release serotonin and norepinephrine, respectively. DRN is located ventral to the cerebral aqueduct (AQ) (Fig. 3A, top right panel). Prior studies showed that +e18a‐Cacna1b mRNA is expressed in DRN and LC 17. We next performed BaseScope™ in these brain areas. To guide our analysis, we compared images for conventional ISH staining for the serotoninergic marker, Slc6a4 (serotonin transporter), from the Allen Mouse Brain Atlas (Fig. 3A, top left panel 26) to our BaseScopeTM staining. Signal for e18a was observed ventral to AQ and followed a pattern similar to Slc6a4, suggesting that +e18a‐Cacna1b splice variants are expressed in DRN (Fig. 3A, middle and bottom left panels). Sections of ∆e18a‐only mice in a similar area show little to no signal for e18a (Fig. 3A, middle and bottom right panels). Our results show that e18a is present in the DRN; however, further studies are needed to determine the cell populations within the DRN that express +e18a‐Cacna1b mRNA. CaV2.2 channels are involved in the release of serotonin 27, and CaV2.2‐null mice show enhanced aggression that has been linked to the control of serotonin neurons excitability 29. Our results open the possibility that splicing e18a is linked to the activity of the serotonin system.

Figure 3

Localization of e18a‐Cacna1b in the DRN and LC. (A) Left top panel, ISH images for Slc6a4 in DRN. AQ. Right top panel, Nissl staining image from brain‐map.org showing the approximate location of DRN relative to AQ. Image credit: Allen Institute. Left middle panel, BaseScope™ images from DRN on a section of a WT mouse. Right middle, BaseScope™ images of DRN sections from ∆e18a‐only mice. (B) Top panel, ISH images for Slc6a2 in LC. V4 = fourth ventricle. CENT2 = central lobule of the cerebellum 2. Middle panel, BaseScope™ images from WT mice in low and high magnification. Note the presence of e18a lateral to the V4. Red dots indicate the presence of +e18a‐Cacna1b mRNA. Blue denotes nuclei stained with Hem. Insets are shown and numbered for clarification. Scale bar = 400 μm.

To determine whether +e18a‐Cacna1b mRNA is expressed in LC, we stained sections containing the fourth ventricle (V4) and the central lobule of the cerebellum II (CENT2) (Fig. 3B, top panel). We found that the signal for +e18a‐Cacna1b mRNA is located lateral to V4 (Fig. 3B, middle and bottom panels). This signal is similar to the pattern of expression for Slc6a2 (norepinephrine transporter, Net) observed in ISH images from the Allen Mouse Brain Atlas 26 (Fig. 3B, top panel). Our results suggest that e18a is present in LC, as previously reported 17.

Distribution of +e18a‐Cacna1b mRNA in cerebellum

In cerebellum, CaV2.2 channels control the release of neurotransmitter from climbing fibers and parallel fibers synapsing onto Purkinje cells 30, 31, 32. CaV2.2 channels are also critical for the intrinsic firing of DCN neurons by coupling to calcium‐dependent potassium channels 33. In cerebellum, ~ 20% of Cacna1b splice variants contain e18a 16. However, the distribution of +e18a‐Cacna1b splice variant in the cerebellum is unknown. Using the well‐defined anatomy of cerebellum as landmark, we determined the expression of e18a in this area. In cerebellar cortex, little e18a signal was observed in the Purkinje cell layer, as well as the molecular and granular layers (m.l. and g.l.) (Fig. 4A, left panel, inset 1). However, in the DCN, we observed several cell bodies stained for e18a (Fig. 4A, left panel, inset 2 and 3). Very little signal was detected in cerebellar sections from ∆e18a‐only mice (Fig. 4B, right panel and insets 4–6). Our results provide a framework to test whether splicing of e18a influences the firing properties of neuronal populations present in DCN.

Figure 4

Localization of +e18a‐Cacna1b in cerebellum. (A) BaseScopeTM images from cerebellum of WT. Insets 1, 2, and 3 indicate magnified areas. Inset 1 represents cerebellar cortex. Layers of the cerebellar cortex are shown: m.l., p.c.l., and g.l. Inset 2 and 3 represent deep cerebellar areas. Red dots indicate the presence of +e18a‐Cacna1b mRNAs. Blue denotes nuclei stained with Hem. (B), BaseScopeTM images from cerebellar areas similar to (A) from ∆e18a‐only mice. Inset 4 represent cerebellar cortex and insets 2 and 3 the deep cerebellar areas. Scale bar = 400 μm.

Expression of +e18a‐Cacna1b in spinal cord

CaV2.2 channels have been previously reported in spinal cord at nociceptive afferents, interneurons, and motor neurons 34. In spinal cord, we have previously shown that e18a‐containing splice variants represent ~ 55% of the Cacna1b pool of transcripts 12. We next determined the pattern of expression of +e18a‐Cacna1b in spinal cord. We found very little signal for e18a in laminae I‐III (Fig. 5A, insets 1 and 2). In contrast, large cell bodies in ventral areas (lamina IX) of spinal cord contained > 6 red dots, indicating the presence of e18a (Fig. 5A, inset 2). Note the absence of e18a signal in spinal cord sections of ∆e18a‐only mice (Fig. 5B, insets 3 and 4). It is well known that dorsal laminae of spinal cord primarily contain neurons that receive sensory inputs from dorsal root ganglia, whereas ventral areas contain circuits that control motor neuron activity. Given that motor neurons in lamina IX have larger cell bodies relative to interneurons 35, 36, our results suggest that e18a is expressed in motor neurons. However, further studies are needed to determine whether +e18a‐Cacna1b mRNA is more abundant in motor neurons relative to interneurons of spinal cord. Previous studies have shown that some neuromuscular junctions rely on CaV2.2 channels to release acetylcholine (phrenic nerve‐diaphragm), whereas others utilize almost exclusively CaV2.1 (sciatic nerve‐tibialis muscle) 37, 38. The cell‐specific expression of +e18a‐Cacna1b mRNA could provide an explanation for this neuromuscular junction‐specific role of CaV2.2 channels.

Figure 5

Distribution of +e18a‐Cacna1b in spinal cord. (A) BaseScopeTM images of spinal cord from WT mice. Inset 1 represents sensory areas containing laminae I‐IV. Inset 2 represents motor areas containing laminae VIII and IX. Red dots are localized in large nuclei in motor areas. Blue denotes counterstaining with Hem. (B) BaseScopeTM images of similar regions as described in A, but in e∆18a‐only mice. Insets 3 and 4 represent sensory and motor areas of the spinal cord, respectively. Scale bar = 200 μm.

Expression of +e18a‐Cacna1b splice variants in hippocampus

The functional role of CaV2.2 channels in controlling neurotransmission has been extensively described in hippocampal synapses. Interestingly, the distribution of Cacna1b splice variants in the hippocampus is unknown. Interneurons reside in stratum radiatum (s.r.), whereas PNs are arranged along stratum pyramidale (s.p.) of both cornu ammonis 1 (CA1) and cornu ammonis 3 (CA3) regions. Using these anatomical landmarks, we determined the expression of e18a in CA1 and CA3 regions. Signal for +e18a‐Cacna1b mRNA was detected in nuclei of s.r. and s.p. Interestingly, only a subpopulation of s.r. nuclei showed strong signal for e18a in both regions (Fig. 6A,B, left panels). In dentate gyrus (DG), interneurons are located in the hilus (h.), whereas granular cells (GC) are localized in the granular cell layer (g.c.l). Here, we observed signal for e18a in several nuclei located in h. and g.c.l. (Fig. 6C). In sections from ∆e18a‐only mice, signal for e18a was absent throughout all regions of hippocampus analyzed (Fig. 6A–C, right panels). We also compared the percentage of cells that show more than one single dot in the DG of hippocampal sections from WT and ∆18a‐only mice; less than one dot per cell is likely to be background 39. We found that ~ 30% of cells in DG from WT contained more than a single dot, compared to ~ 5% the ones from ∆18a‐only mice (% cells with more than one dot ± SEM. WT = 29.69 ± 1.23, n = 5; D18a‐only = 4.81 ± 0.17, n = 4, P < 0.00001, Student’s t‐test). These results further support the specificity of the e18a‐Cacna1b probe. Based on anatomical landmarks for hippocampus, our results suggest that e18a is broadly distributed among PN, GC, and a subpopulation of interneurons in the hippocampus.

Figure 6

Localization of e18a‐Cacna1b in hippocampus. (A) Left, BaseScopeTM images of CA1 region of ventral hippocampus in WT mice. Note the presence of red dots in both s.r. and s.p.. Right, similar to left, but in sections of ∆e18a‐only mice. (B) Left, BaseScopeTM images of CA3 region of ventral hippocampus in WT mice. Right, similar to left but in CA3 region of sections from ∆e18a‐only mice. (C) Left, representative images of DG of WT mice stained for e18a. DG layers are shown molecular layer (m.l.), h., and g.c.l. Right, similar to left for ∆e18a‐only mice. In all images, red dots are indicative of the presence of e18a, and blue denotes counterstaining of nuclei with Hem. Insets are numbered for clarity. Scale bar = 400 μm.

+e18a‐Cacna1b splice variants are enriched in cholecystokinin‐expressing interneurons

CaV2.2 channels, together with CaV2.1 and CaV2.3, control transmitter release of excitatory terminals 40. CaV2.2 channels also couple to calcium‐dependent potassium channels in PNs, thereby controlling neuronal firing 41. Furthermore, CCK+INs in s.r. of CA1 and CA3, and h. of DG heavily rely on CaV2.2 channels to release GABA 42. To determine the cell‐specific pattern of expression for +e18a‐Cacna1b mRNA, we compared the relative amounts of +e18a‐ and ∆e18a‐Cacna1b mRNAs in CCK+INs and PNs by combining genetic labeling, FACS, and RT‐PCR. To label CCK+IN, we used genetic intersectional labeling with Cre and FLPe recombinases, which resulted in the expression of tdT in CCK+INs (Fig. 7A and see Methods). To identify PNs, mice expressing Cre recombinase from the CaMKIIα promoter were crossed with mice containing tdT with an upstream floxed STOP codon (CaMKIIα;tdT) 43. Next, we performed FACS on dissociated tissue from cortex and hippocampus of CCK;Dlx5/6;tdT and CaMKIIα;tdT mice. Total RNA was extracted and reverse‐transcribed to quantify the relative amounts of +e18a‐ and ∆e18a‐Cacna1b mRNA. We found that Cacna1b splice variants that contain e18a are more abundant in CCK+INs relative to CaMKIIα+ PNs (% e18a relative to total Cacna1b mRNA, mean ± SEM: CCK+INs = 19.5 ± 2.5, n = 7; CaMKIIα+PNs = 2.5 ± 1.7, n = 8. n represents the number of mice. P = 0.006, Student’s t‐test. Fig. 7B). To validate the RNA extracted from sorted cell populations, we quantified Gad‐2 mRNA in both CCK+INs and CaMKIIα+PNs. As expected, CCK+INs express higher levels of Gad‐2 mRNA relative to CaMKIIα+PNs (% fold change, mean ± SEM. CCK+INs = 6.9 ± 1.57, n = 7; CaMKIIα+PNs = 1 ± 0.41, n = 5. P = 0.02, Student’s t‐test. Fig. 7C, left panel). Several groups have reported that CCK+INs are enriched with cannabinoid receptor 1 (Cnr1) mRNA 44, 45, 46; therefore, to further validate our cell sorting, we compared the levels of Cnr1 mRNA between CCK+INs and CaMKIIα+PNs. We found that CCK+INs express significantly higher levels of Cnr1 than CaMKIIα+PNs (% fold change, mean ± SEM. CCK+IN = 102 ± 23.0, n = 7; CaMKIIα+PNs = 1.0 ± 0.30, n = 8, P = 0.0002, Student’s t‐test. Fig. 7C, right panel). Taken together, our results strongly suggest that +e18a‐Cacna1b pre‐mRNA is expressed at higher levels in CCK+INs relative to CaMKIIα+PNs.

Figure 7

+e18a‐Cacna1b mRNA is enriched in CCK+INs. (A) Intersectional genetic labeling of CCK+INs. Cre recombinase is expressed from the CCK promoter active in CCK interneurons and projection neurons, whereas Flpe recombinase is expressed from the Dlx5/6 promoter active in forebrain GABAergic interneurons. The tdT allele present in Ai65‐D mouse contains two STOP codons (white hexagons) flanked by loxP and FRT sites. In a mouse with these three alleles, the combined action of Cre‐loxP and Flpe‐FRT systems removes both STOP codons allowing expression of tdT in GABAergic CCK+INs, but not in PNs. (B) Left panel, representative gel of RT‐PCR in RNA purified from sorted CCK+INs and CaMKIIα+PNs. Right panel, comparison of +e18a‐Cacna1b relative to total Cacna1b mRNA in CCK+IN and CaMKIIα+PN. (C) Validation of sorted cells. Total amounts for glutamate decarboxylase 2 (Gad2) and Cnr1 mRNA were compared between CCK+INs and CaMKIIα+PNs using RT‐qPCR. Data are shown as mean (filled symbols) ± SEM and the averages of individual biological replicates (empty symbols).

With our current results, we are unable to discard the possibility that e18a‐Cacna1b is also present in other types of GABAergic interneurons in cortex and hippocampus. However, we focus on CCK+INs because 100% of GABA release relies on CaV2.2 channels; therefore, the functional significance of e18a‐Cacna1b has high potential 42. Furthermore, this type of interneurons are linked to mood disorders 47, 48, express relatively high levels of the cannabinoid receptor protein (CB1R) 49, and contribute to the behavioral tolerance of tetrahydrocannabinol 50. Given that CB1R agonists downregulate CaV2.2 channels to inhibit GABA release in CCK+INs, our findings suggest an anatomical link between e18a splicing in Cacna1b and the effects of CB1R agonists of transmitter release. CCK+INs in hippocampus show robust asynchronous release and spontaneous release that relies on CaV2.2 channels 42. Due to the slower inactivation rates and positive shift in voltage‐dependent inactivation of +18a‐CaV2.2 channels relative to ∆18a‐CaV2.2 channels, it is possible that inclusion of e18a in this neuronal type enhances asynchronous and spontaneous release.

Conflict of interest

The authors declare no conflict of interest.

Author contributions

AB performed all ISH experiments. BL performed cell‐sorting experiments. MA assisted in cell sorting and ISH. All authors wrote the manuscript. AA designed the study.