Young (n = 4) and mature adults (n = 4) of the sea lamprey, Petromyzon marinus L., were used for the in situ hybridization experiments. Ten larvae were used for RNA extraction. Young adults and larvae were collected from the River Ulla (Galicia, Northwest Spain) and mature upstream migrating adults were obtained from a local commercial supplier. All experiments were approved by the Bioethics Committee of the University of Santiago de Compostela and the Consellería do Medio Rural e do Mar of the Xunta de Galicia (JLPV/IId) and conformed to the European Union and Spanish regulations for the care and handling of animals in research.

Larvae were anesthetized by immersion in 0.1% ethyl 3-aminobenzoate methanesulfonate salt (MS-222; Sigma, St. Louis, MO, USA) and the brain and spinal cord were dissected out under sterile conditions. Total RNA was isolated from these tissues using the TriPure reagent (Roche, Mannhein, Germany). The first-strand cDNA synthesis reaction from total RNA was catalyzed with Superscript III reverse transcriptase (Invitrogen, Waltham, MA, USA) using random primers (hexamers; Invitrogen). For polymerase chain reaction (PCR) cloning, specific oligonucleotide primers, 5′-TGGCACTGGCCCTGAACAAG-3′ forward and 5′- GTTGAGGTTGGGCTGCGAGT-3′ reverse; and 5′- GACAAATCTTGCTCGACGCC-3′ forward and 5′- AAACGTTGCTGAGGACACCA-3′ reverse, were designed based on the sea lamprey GABA_B1 and GABA_B2 sequences, respectively, annotated in the sea lamprey genome and deposited in the Ensembl database (Smith et al., 2013¹). The amplified fragments were cloned into pGEM-T vectors (Promega, Madison, WI, USA) and sequenced by GATC Biotech (Cologne, Germany) using Sanger sequencing. Sequence analysis and comparison was done through Basic Local Alignment Search Tool (BLAST) on NCBI and in the SMART website (Schultz et al., 1998; Letunic et al., 2015²).

For this analysis, GABA_B1 (protein GenBank ID: Canis lupus familiaris: XP_005640226.1; Drosophila melanogaster: AAK13420.1; Danio rerio: CAP09588.1; Bos taurus: AAI46242.1; Homo sapiens: AAH50532.2; Mus musculus: AAH54735.1; Rattus norvegicus: CAE84069.1; Columba livia: XP_013227121.1; Xenopus laevis: ADQ43745.1; Pan troglodytes: XP_009449053.1) and GABA_B2 (protein GenBank ID: Canis lupus familiaris: XP_538749.2; Danio rerio: NP_001137515.1; Drosophila melanogaster: AAK13421.1; Columba livia: XP_013227167.1; Homo sapiens: AAH35071.2; Mus musculus: NP_001074610.1; Rattus norvegicus: EDL98850.1; Xenopus laevis: ADQ43746.1; Pan troglodytes: XP_009455264.1; Bos taurus: XP_002689780.1) sequences of representative species were aligned with the sea lamprey GABA_B1 and GABA_B2 partial deduced protein sequences deposited in the Ensembl database. All sequences were aligned using CLUSTALW and a phylogenetic tree was constructed using the neighbor-joining method with Poisson corrected distances on amino acids and bootstrap analysis (1000 replications) using MEGA 6 software (Tamura et al., 2013). An alternative phylogenetic tree was constructed with the minimum-evolution method with Poisson corrected distances on amino acids and bootstrap analysis (1000 replications) using also the MEGA 6 software.

Templates for in vitro transcription were prepared by PCR amplification as follows. Two 459 base pairs (bp) and 446 bp fragments corresponding to GABA_B1 and GABA_B2 sequences, respectively, were obtained using the primers mentioned above. In this case, the reverse primers include the sequence of the universal T7 promoter (TAAGCTTTAATACGACTCACTATAGGGAGA). For the generation of sense probes, the sequence of the T7 promoter was included in the forward primers. The identity of the amplified fragments was confirmed by direct sequencing. Digoxigenin (DIG)-labeled riboprobes were synthesized using the amplified fragments as templates and following standard protocols using T7 polymerase (Roche Diagnostics, Germany).

In situ hybridization experiments were performed as previously described for riboprobes against the serotonin 1a receptor (5-ht1a; Cornide-Petronio et al., 2013). Briefly, the brain/rostral spinal cord of young and mature adults were fixed by immersion for 12 h in 4% paraformaldehyde. Then, they were cryoprotected with sucrose 30% and sectioned in a cryostat in the transverse plane. Two parallel series of sections were obtained. The sections of each series were incubated with the GABA_B1 or GABA_B2 DIG-labeled probes, respectively, at 70°C overnight in hybridization mix and treated with RNAse A (Invitrogen) in the posthybridization washes. Then, the sections were incubated with a sheep anti-DIG antibody conjugated to alkaline phosphatase (1:2000; Roche) overnight. Staining was conducted in BM Purple (Roche) at 37°C until the signal was clearly visible. Finally, the sections were mounted in Mowiol^® (Calbiochem; Temecula, CA, USA) and photographed with an Olympus photomicroscope (AX-70; Provis) equipped with a color digital camera (Olympus DP70; Tokyo, Japan). Images were slightly adjusted for brightness and contrast with Adobe Photoshop CS4 to compose the plates. No staining was observed when sense probes were used.

The cloned GABA_B1 and GABA_B2 cDNAs sequences corresponded to the nucleotides 302–760 for GABA_B1 (Figure 1A), and 729–1162 for GABA_B2 (Figure 1A′) of the partial cDNA sequences of these receptors annotated in sea lamprey genome of the Ensembl database. The cloned sequences had a 100% similarity with the putative GABA_B1 and GABA_B2 sequences of the Ensemble database, respectively. The cloned sequences were deposited in the GenBank database (GABA_B1: KX655780; GABA_B2: KX655781). Due to a high level of similarity between the cloned sequences and those annotated in the Ensembl database, we used the partial, but longer, GABA_B1 and GABA_B2 protein sequences (Figures 1B,B′, respectively) deduced from the sea lamprey genome sequences ENSPMAG00000006844 (corresponds to GABA_B1), which is located in the scaffold GL479777, and ENSPMAG00000004383 (corresponds to GABA_B2), which is located in the scaffold GL478877, for the subsequent analyses. Both partial deduced protein sequences contained the “ligand-binding domain of GABA_B receptors” and the “seven transmembrane sweet-taste receptor of 3 GCPR” as revealed by the SMART online tool (Figures 1C,C′).

Percentages of similarity between the sea lamprey GABA_B1 and GABA_B2 partial deduced protein sequences and those of other species are shown in Table 1. We observed a percentage of similarity of 68% to 73% and 66% to 69% between the sea lamprey partial GABA_B1 and GABA_B2 deduced protein sequences, respectively, and the corresponding deduced protein sequences of other vertebrate species. As expected, the percentage of similarity of the partial sea lamprey GABA_B1 and GABA_B2 deduced amino acids sequences with those of the protostome invertebrate D. melanogaster is low (less than 50%). The similarity between the sea lamprey partial GABA_B1 and GABA_B2 deduced protein sequences is only 33%.

The GABA_B1 and GABA_B2 deduced protein sequences annotated in the Ensembl database were aligned with GABA_B1 and GABA_B2 deduced protein sequences of representative vertebrate and invertebrate species and this alignment was used to construct a phylogenetic tree using the neighbor-joining and the minimum-evolution methods. All vertebrate GABA_B1 and GABA_B2 sequences, including the sea lamprey sequences, were clustered together in their respective groups in the resulting trees and the sea lamprey sequences were found to emerge as an outgroup to the corresponding gnathostome sequences (Figures 2A,B).

We observed a widespread expression of the GABA_B1 and GABA_B2 transcripts in neuronal populations in the brain of the adult sea lamprey. No clear and obvious differences were observed in the expression of these transcripts between young and mature adult sea lampreys. In situ hybridization signal appeared as a dotted labeling in the tissue sections.

No clear differences were noted in the expression of the transcripts of the two GABA_B subunits (Figure 3). Accordingly, only 1 set of schematic drawings of transverse sections taken from a young post-metamorphic adult showing the expression pattern of both transcripts is presented in Figure 4. Photomicrographs of representative transverse sections of the brain are shown in Figures 3, 5, 6. These photomicrographs were taken from two young postmetamorphic and three upstreaming migrating adults.

In the telencephalon, the olfactory bulbs showed numerous dots of positive GABA_B1 and GABA_B2 expression (Figures 4A, 5A). Most of the expression was observed in cells of the mitral and inner cellular layers. Some expression was also observed in cells localized between the glomeruli. In the medial pallium, GABA_B1 and GABA_B2 expression was very scarce in cells of the periventricular layer and no expression was observed in migrated cells (Figures 4C, 5B). We observed expression of both transcripts in neurons surrounding the dorsomedial neuropil (Figure 4C). The lateral pallium showed more GABA_B1 and GABA_B2 expression in the cells of the periventricular layer than in migrated cells (Figures 4B,C, 5C). In the subhippocampal lobe, most of the GABA_B1 and GABA_B2 expression was observed in the cells of the periventricular layer (Figure 4C). In the subpallium, the cells of the septum showed GABA_B1 and GABA_B2 positive dots (Figures 4B, 5D). In the striatum, GABA_B1 and GABA_B2 expression was observed mainly in migrated cells, with less expression being observed in the periventricular layer (Figures 3A,A′, 4C). In contrast, the preoptic nucleus showed wide GABA_B1 and GABA_B2 expression in periventricular cells, mainly in the intermediate layer of the preoptic nucleus (Figures 3A,A′, 4C).

The hypothalamus showed wide expression of the GABA_B1 and GABA_B2 transcripts in cells of the periventricular layers, while the migrated cells showed very scarce expression (Figure 5E). Many GABA_B1 and GABA_B2 positive dots were observed in the nucleus of the postoptic commissure, the paracommissural preoptic nucleus, the dorsal and ventral hypothalamus and the mammillary nucleus (Figures 4D–F, 5E). The pattern of expression observed in the hypothalamus was also observed in the prethalamus and thalamus (Figures 4D–F, 5F,G). GABA_B1 and GABA_B2 expression was also observed in the pineal and parapineal organs (Figure 4C). The habenula showed GABA_B1 and GABA_B2 expression, with some differences between the left and right sides (Figures 3B,B′, 4D,E). In the right habenula, which is bigger than the left one, GABA_B1 and GABA_B2 positive dots were more numerous than in the left habenula, although positive dots in cells of the left habenula were larger. The pretectum showed expression of both transcripts in cells of the periventricular layer and scarce expression in migrated neurons (Figure 4F). In the paratubercular nucleus, very few cells showed expression of the transcripts (Figure 4F). In contrast, a large number of GABA_B1 and GABA_B2 positive dots were observed in the nucleus of the medial longitudinal fascicle including the first and second Müller cells (Figures 4G, 5I).

In the mesencephalon, the optic tectum showed GABA_B1 and GABA_B2 positive dots with most of them distributed throughout the cells of the stratum griseum periventriculare, but also in cells of the stratum “album” et griseum central (Figures 3C,C′, 4H). The torus semicircularis showed numerous positive dots. The positive dots in cells of the torus were larger than those in the optic tectum (Figure 4H). The migrated neurons of both areas showed very little expression of the transcripts. There was strong GABA_B1 and GABA_B2 expression in the tegmentum and the mesencephalic reticular area (including the third Müller cell) and the motor neurons of the oculomotor nuclei showed a large amount of GABA_B1 and GABA_B2 expression in their somas (Figure 4H).

In the alar plate of the rhombencephalon, the dorsal isthmic gray showed expression of both transcripts (Figures 4I, 6A). In the octavolateral nuclei, there was GABA_B1 and GABA_B2 expression, mainly in cells of the periventricular layer (Figures 4J,K, 6B,C). The octavomotor nuclei showed GABA_B1 and GABA_B2 positive dots (Figures 4K, 6C). Numerous GABA_B1 and GABA_B2 positive dots were observed in cells of the periventricular layer of the solitary tract (Figures 4L, 6D) and dorsal column nuclei. Dorsal cells of the caudal rhombencephalon (primary medullary and spinal nucleus of the trigeminus) also showed GABA_B1 and GABA_B2 expression (Figures 6B,C). In the basal plate of the rhombencephalon, the soma of the cells of the visceromotor nuclei (V, VII, IX and X) showed many GABA_B1 and GABA_B2 positive dots (Figures 4J–L, 6E). In addition, wide GABA_B1 and GABA_B2 expression was observed in all rhombencephalic reticular nuclei (isthmic, trigeminal, middle and posterior reticular nuclei), including the giant Müller and Mauthner cells (Figures 4I–L, 6F).

In the rostral spinal cord of the adult sea lamprey, most neuronal types showed GABA_B1 and GABA_B2 expression (Figures 4M, 6G,H). Dorsal cells showed large amounts of GABA_B1 and GABA_B2 positive dots in their soma (Figures 4M, 6G). We also observed some expression of the transcripts in the edge cells (Figure 4M). GABA_B1 and GABA_B2 expression was observed in interneurons of different sizes located in the dorsal and lateral regions (Figures 4M, 6G). Expression of both transcripts was observed in the somas of motor neurons (Figure 6G). Cerebrospinal fluid-contacting (CSF-c) cells showed numerous GABA_B1 and GABA_B2 positive dots, even in their apical dendrites (Figure 6H).

The glial cells of the brain of the sea lamprey are almost exclusively ependymal cells, although astrocytes could be identified in the optic nerve. No expression of any of the two transcripts was observed in brain ependymocytes (examples in Figures 5B,C) or in the subcommissural organ (Figures 4F, 5H). The astrocytes of the optic nerve, which are easily identified because of the absence of neurons in the optic nerve, did not show expression of the transcripts (Figures 3D,D′).

Here, we report for the first time the identification and characterization of the GABA_B1 and GABA_B2 cDNA sequences and the distribution of the GABA_B transcripts in the CNS of the adult sea lamprey.

In our BLAST searches of the sea lamprey genome database, we found 1 GABA_B1 gene and 1 GABA_B2 gene, whose sequences were confirmed after PCR amplification and cloning of the respective cDNAs. The presence of only a single gene of each GABA_B subunit is also the case in mammals, birds and amphibians. Two paralogous copies of the GABA_B1 subunit gene have been found in zebrafish (Klee et al., 2012), which is probably due to the additional whole-genome duplication that occurred in the actinopterygian lineage (Taylor et al., 2001). In the phylogenetic trees, the GABA_B1 and GABA_B2 partial amino acid sequences of the sea lamprey were located at the base of the vertebrate branches clustering the GABA_B1 and GABA_B2 sequences, appearing as sister members, respectively, of the GABA_B1 and GABA_B2 sequences of gnathostomes. Each of the subunits is grouped in an independent cluster (Figure 2). The location of the sea lamprey GABA_B1 and GABA_B2 sequences at the base of the vertebrate branches of the phylogenetic tree and its higher similarity with the vertebrate sequences than with those of Drosophila are in agreement with the phylogenetic position of lampreys and confirmed the GABA_B identity of the sea lamprey sequences identified in our study.

It is currently accepted that a functional GABA_B receptor consists of a heterodimer of GABA_B1 and GABA_B2 subunits (Jones et al., 1998; Kaupmann et al., 1998; White et al., 1998; Bettler et al., 2004). GABA_B1 binds to GABA, while GABA_B2 is needed to transmit the signal, because G-protein coupling is mediated via GABA_B2 (Margeta-Mitrovic et al., 2000; Calver et al., 2001; Galvez et al., 2001; Pagano et al., 2001; Geng et al., 2013). Because of this, it seems clear that the co-expression of GABA_B1 and GABA_B2 mRNAs is necessary to form a functional GABA_B receptor. In our study, we observed an overlapping expression of the GABA_B1 and GABA_B2 mRNAs in all brain regions and in the spinal cord of the adult sea lamprey when using consecutive brain sections. Co-localization of both subunits in the same single cells of lampreys is clear in the giant individually identifiable reticulospinal neurons like the Mauthner and Müller neurons. This is also the case in D. melanogaster, where no differences were observed between the expressions of GABA_B subunits in in situ hybridization assays (Mezler et al., 2001). In zebrafish, a recent study using qPCR methods has shown that the two b1 and the b2 subunits are all expressed in the same brain regions, with the b1b and b2 being more represented than the b1a in some regions and in the brain as a whole (Cocco et al., 2016). Kuner et al. (1999) observed by analyzing serial rat brain sections, that GABA_B1 and GABA_B2 transcripts are widely expressed and that they also show considerable overlap in most regions of the brain, although in some regions the expression of the GABA_B1 transcript was enriched. This suggests that a wide and highly overlapping expression of the GABA_B1 and GABA_B2 subunits is an ancestral and conserved character of vertebrates and invertebrates.

In the mouse brain, an association between the GABA_B2 subunit and M2 muscarinic receptors has been shown which appears to enhance muscarinic signaling (Boyer et al., 2009). In lampreys, pharmacological treatments combined with electrophysiological studies have shown a role for muscarinic receptors in the modulation of the trigeminal-reticular pathway (Le Ray et al., 2004) and in the activation of reticulospinal neurons (Smetana et al., 2007). Moreover, the presence of cells immunoreactive for muscarinic receptors has been shown in the region of the posterior rhombencephalic reticular nucleus of lampreys (Smetana et al., 2007), a region that shows the expression of the GABA_B2 transcript (present results). Whether the modulation of muscarinic receptors by the GABA_B2 subunit also occurs in lampreys needs further investigation. This would show us whether this is an ancestral characteristic of vertebrates.

All brain regions and the spinal cord showed a broad expression of both GABA_B transcripts in the adult sea lamprey, which is in concordance with previous reports in invertebrates (e.g., D. melanogaster (Mezler et al., 2001), cockroaches (Blankenburg et al., 2015) or spiders (Panek et al., 2003)) and in jawed vertebrates (e.g., humans (Calver et al., 2000; Berthele et al., 2001), non-human primates (Muñoz et al., 1998; Nürnberger and Schöniger, 2001), rats (Bowery et al., 1987; Bischoff et al., 1999; Clark et al., 2000), birds (Veenman et al., 1994), frogs (Kaeser et al., 2011) and zebrafish (Tabor et al., 2008; Cocco et al., 2016)). Positive in situ signal in sea lamprey brain sections had a granular appearance probably due to low expression of these mRNAs in each single cell of the sea lamprey. Previous studies looking at the expression of different neurotransmitter receptors in lampreys have shown that the in situ hybridization signals appeared as a dotted labeling in sections of the CNS: serotonin receptor 1A (Cornide-Petronio et al., 2013, 2014), dopamine receptor D2 (Robertson et al., 2012; Fernández-López et al., 2015), and dopamine receptor D4 (Pérez-Fernández et al., 2016); suggesting that low expression levels are a common feature of different metabotropic neurotransmitter receptors.

The broad expression of these transcripts in the CNS suggests that this receptor is extensively used in the modulation of brain circuits in lampreys. The expression of both GABA_B transcripts in non-GABAergic cells, which can be identified by their size and/or location, such as the mitral cells of the olfactory bulbs, the giant reticulospinal neurons, the spinal motoneurons, the motoneurons of the visceromotor rhombencephalic nuclei, the primary sensory cells of the rhombencephalon and spinal cord or the edge cells of the spinal cord indicate that the GABA_B receptor plays a role in the modulation of the activity of non-GABAergic cells in the sea lamprey brain. The expression of GABA_B transcripts observed here is in agreement with the GABAergic modulation of spinal motoneurons and interneurons mediated by the GABA_B receptor as reported in previous pharmacological and electrophysiological studies (Alford and Grillner, 1991; Alford et al., 1991; Matsushima et al., 1993; Schmitt et al., 2004). Edge cells in the spinal cord are richly innervated by GABAergic fibers (Fernández-López et al., 2012) and recent work has shown that they are modulated by GABA (Svensson et al., 2013). Our results suggest that GABA could act trough the GABA_B receptor in the edge cells. The present results also support the idea of the role of the GABA_B receptor in the modulation of the lamprey respiratory network (Bongianni et al., 2006; Cinelli et al., 2014). Also, in agreement with our expression results, a modulatory action of GABA onto the pathway from the lateral columns to reticulospinal neurons has been suggested to be mediated by the GABA_B receptor (Vinay et al., 1998). These reticulospinal inputs are also regulated by peptidergic transmitters (Parker, 2000). Our study extends the number of neuronal populations known to be modulated through GABA_B signaling and opens the opportunity to conduct functional studies on the role of GABA and the GABA_B receptor in other circuits.

The broad expression of the GABA_B transcripts together with the broad distribution of GABAergic cells in the brain and spinal cord (Meléndez-Ferro et al., 2001; Robertson et al., 2007) suggests that the sea lamprey GABA_B receptor could modulate the activity of GABAergic cells. The expression of the GABA_B transcripts in periglomerular cells of the olfactory bulbs, which are mainly GABAergic (Meléndez-Ferro et al., 2001), or in the CSF-c cells of the spinal cord, which are mainly GABAergic as well (Rodicio et al., 2008; Villar-Cerviño et al., 2008; Fernández-López et al., 2012; Jalalvand et al., 2014), could potentially support this hypothesis. Results from pharmacological and electrophysiological studies have shown that inhibitory premotor interneurons respond to the application of GABA_B agonists (Alford and Grillner, 1991; Alford et al., 1991; Matsushima et al., 1993). Unfortunately, the different requirements of fixative for in situ hybridization and GABA immunohistochemistry precluded us from providing a definitive demonstration of the presence of GABA_B receptors in GABAergic cells of the sea lamprey. The present results show expression of the GABA_B transcripts in the dendrites of CSF-c cells of the spinal cord. Currently, some functions have been proposed for CSF-c neurons in controlling the composition of the CSF and releasing substances into the ventricular system (Vígh et al., 2004; Jalalvand et al., 2016). The GABA_B receptor in these cells could play a role for detecting GABA in the CSF. The detection of serotonin from the CSF has been also proposed for the 5-HT1A receptor due its expression in CSF-c dendrites (Cornide-Petronio et al., 2013). As stated above, the GABA_B receptor could modulate CSF-c cells acting as an autoreceptor and/or as a heteroreceptor trough synapses from other GABAergic cells.

Our results show a lack of expression of the GABA_B transcripts in ependymocytes along the lamprey brain. The ependymal cells are the main glial type present in the brain of the sea lamprey, while astrocytes are only associated to some nervous tracts. However, the spinal cord shows both types of glial cells (Retzius, 1893). In contrast to jawed vertebrates, oligodendrocytes are not present in lampreys. The glial cells of lampreys do not display immunoreactivity to glial fibrillary acid protein (GFAP), but they express cytokeratins (Merrick et al., 1995). Few works have been done to study the expression of GABA_B1 and GABA_B2 mRNAs and/or GABA_B1 and GABA_B2 subunits in the ependymal layer of vertebrates. A pharmacological study of Corns et al. (2013) reported that only the GABA_A, and not GABA_B, receptor mediates the GABAergic responses in mammalian ependymal cells surrounding the central canal. More studies in other groups of vertebrates are necessary to determine the evolution of this character due to the large evolution distance between mammals and lampreys. In addition, the cells of the subcommissural organ, a special type of ependymal cells, also lack the GABA_B receptor both in lampreys, and in other vertebrates like teleosts, frogs and mammals (Jiménez et al., 2000; Saha et al., 2000; Nürnberger and Schöniger, 2001). Studies in these vertebrates have shown that the GABAergic responses in the subcommissural organ are also mediated by GABA_A receptors.

No expression of the GABA_B transcripts was observed in astrocytes of the optic nerve although previous studies have reported the presence of GABA_B subunits in astrocytes, of rodents (Charles et al., 2003; Oka et al., 2006; Luyt et al., 2007; Beenhakker and Huguenard, 2010), and the expression of GABA_B1 and GABA_B2 mRNAs in human astrocytes (Lee et al., 2011). Again, more studies are necessary in other groups of vertebrates to determine whether the lack of GABA_B expression in astrocytes is the ancestral condition or a derived character.