Evolution of dopamine receptors: phylogenetic evidence suggests a later origin of the DRD_2l and DRD_4rs dopamine receptor gene lineages

DNA data and phylogenetic analyses

We used bioinformatic procedures to retrieve dopamine receptor genes in species of all major groups of vertebrates. Our sampling included mammals, birds, reptiles, amphibians, coelacanths, teleost fish, holostean fish, cartilaginous fish and cyclostomes (Table S1). We identified genomic pieces containing dopamine receptor genes in the Ensembl database using BLASTN with default settings (Maximum number of hits to report =100; Maximum E-value for reported alignments =10; Word size for seeding alignments =11; Match/Mismatch scores =1,  − 3; Gap penalties: opening =5, Extension =2) or NCBI database (refseq_genomes, htgs, and wgs) using tbalstn (Altschul et al., 1990) with default settings (Max target sequences =100; Expect threshold =10; word size =6; Max matches in a query range =0; Matrix = BLOSUM62; Gap cost: Existence =11, Extension =1; Compositional adjustments: Conditional compositional score matrix adjustment) (human and chicken sequences were used as a starting point). Conserved synteny, based on previous literature (Yamamoto et al., 2013; Yamamoto et al., 2015), was also used as a criterion to define the genomic region containing dopamine receptor genes. Once identified, genomic pieces were extracted including the 5′ and 3′ flanking genes. After extraction, we curated the existing annotation by comparing known exon sequences to genomic pieces using the program Blast2seq with default parameters (Max target sequences =100; Expect threshold =10; word size =28; Max matches in a query range =0; Match/Mismatch scores: 1,  − 2; Gap costs = Linear) (Tatusova & Madden, 1999). Putatively functional genes were characterized by an open intact reading frame with the canonical exon/intron structure typical of vertebrate dopamine receptors. Sequences derived from shorter records based on genomic DNA or cDNA were also included in order to attain a broad and balanced taxonomic coverage. We also included sequences of the α₂-adrenoreceptors (ADRA2A, ADRA2B, ADRA2C, ADRA2D), and β-adrenoreceptors (ADRB1, ADRB2 and ADRB3) (Table S1). Our final dataset contained 396 sequences. Amino acid sequences were aligned using the FFT-NS-i strategy from MAFFT v.7 (Katoh & Standley, 2013). We used the proposed model tool of IQ-Tree (Trifinopoulos et al., 2016) to select the best-fitting model of amino acid substitution (JTT + R9). We performed a maximum likelihood analysis to obtain the best tree using the program IQ-Tree (Trifinopoulos et al., 2016); support for the nodes was assessed with 1,000 bootstrap pseudoreplicates using the ultrafast routine. Phylogenetic analyses were performed 20 times in order to better explore the tree space. Human ADRA1A, ADRA1B, and ADRA1D sequences were used as outgroups.

Assessments of conserved synteny

We examined genes found upstream and downstream of the dopamine receptor genes of representative vertebrate species. We used the estimates of orthology and paralogy derived from the EnsemblCompara database (Herrero et al., 2016); these estimates are obtained from an automated pipeline that considers both synteny and phylogeny to generate orthology mappings. These predictions were visualized using the program Genomicus v90.01 (Louis et al., 2015). Our assessments were performed in humans (Homo sapiens), chicken (Gallus gallus), spotted gar (Lepisosteus oculatus) and elephant shark (Callorhinchus milii). In the case of the elephant shark, flanking genes were examined using the entrez gene database from the National Center for Biotechnology Information (NCBI) (Maglott et al., 2011).

Molecular structure and graphics

Molecular visualization and analyses of the human DRD₄ protein structure were performed with the UCSF Chimera package (Pettersen et al., 2004) using the 1.96 Å resolution structural file PDB ID: 5WIV (Wang et al., 2017). Molecular dynamics simulation of site-directed mutagenesis was performed using the Chimera structure editing tool and choosing the Dunbrack rotamer library (Dunbrack, 2002) to visualize the probability of a particular amino acidic conformation. The rotamer displaying the highest probability was selected: Y rotamer : 72.6% probability; I rotamer 79% probability. Sequences were aligned using Vector NTI Express (Thermo Fisher) using default parameters (Display setup: identity value = 1, Similarity value = 0.5, Weak similarities value = 0.2. Showing weak similarities = checked. Multiple alignment options: slow, Protein weight matrix = GONNET, gap open penalty = 15, gap extension penalty = 6.66, percentage of identity for delay = 30. Protein gap parameters: hydrophilic residues = GPSNDQEKR, Gap separation distance = 4, residue specific penalties = checked, hydrophilic penalties = checked). Human protein sequences DRD₂: NP_000786.1 and DRD₄: NP_000788 were used as reference for the numbering and alignment.

Overview of the evolution of dopamine receptors

In this work we performed an evolutionary study of dopamine receptors in representative species of all major groups of vertebrates. We combined gene phylogenies and synteny analyses with the main goal of understanding the duplicative history of the DRD₁ class of dopamine receptors and the time of origin of the DRD_2l and DRD_4rs gene lineages. Our phylogenetic tree recovered the monophyly of the two groups of dopamine receptors (Fig. 1). In the first clade we recovered the sister group relationship between the DRD₁ class of receptors and a clade containing β-adrenoreceptors (Fig. 1); in the second clade, the DRD₂ receptors were recovered sister to the α₂-adrenoreceptors (Fig. 1). This phylogenetic arrangement is in agreement with previous results (Yamamoto et al., 2013; Spielman, Kumar & Wilke, 2015; Céspedes et al., 2017; Zavala et al., 2017) and reflects the fact that the ability to bind dopamine was acquired twice during the evolutionary history of biogenic amine receptors (Callier et al., 2003; Yamamoto et al., 2015).

Phylogenetic relationships among the DRD₁ class of dopamine receptors

According to our phylogenetic analyses, the monophyly of the DRD₁ class of dopamine receptors, as well as the monophyly of each paralog (DRD₁, DRD₅, DRD_1C and DRD_1E), were recovered with strong support (Fig. 1). In all cases synteny analyses provided further support for the identity of the four DRD₁ clades recovered in our phylogenetic tree (Fig. 2). Phylogenetic relationships among the different DRD₁ lineages were well resolved (Fig. 1). We recovered the sister group relationship between the DRD_1C and DRD_1E dopamine receptors (Fig. 1), and this clade was recovered sister to a cyclostome sequence (Fig. 1). The DRD₁ clade was recovered sister to the aforementioned clade, and the group containing DRD₅ sequences was recovered sister to all other DRD₁ paralogs (Fig. 1). Although in the literature there are studies reporting dopamine receptor phylogenies (Callier et al., 2003; Le Crom et al., 2004; Yamamoto et al., 2013; Yamamoto et al., 2015; Haug-Baltzell et al., 2015), they are not directly comparable as the taxonomic and/or family membership sampling differ. Beyond this point, phylogenetic relationships among the DRD₁ class of dopamine receptors seem to still be a matter of debate. In some cases DRD₁ has been recovered sister to DRD₅, a clade that in turn is recovered sister to DRD_1C; in these studies DRD_1E is recovered sister to all other DRD₁ (Callier et al., 2003; Le Crom et al., 2003; Yamamoto et al., 2013). In other studies the clade containing DRD₁ sequences has been recovered sister to DRD_1C, and this group is sister to DRD₅ (Le Crom et al., 2004). A case in which the monophyly of DRD_1E is not recovered has also been reported (Haug-Baltzell et al., 2015). Finally, there is also a case in which the DRD₁ class of receptors has been recovered as two different clades, one that includes DRD₁ and DRD₅ and another grouping DRD_1C and DRD_1E (Yamamoto et al., 2015). Thus, our results propose a new phylogenetic hypothesis regarding the evolution of the DRD₁ class of dopamine receptors (Fig. 1). Overall, we believe that our hypothesis is well supported based on a taxonomic sampling that covered all main groups of vertebrates, as well as, the phylogenetic context of the monoamine receptors (Spielman, Kumar & Wilke, 2015).

Phylogenetic relationships among the DRD₂ class of dopamine receptors

We recovered the monophyly of the DRD₂ class of dopamine receptors with strong support (Fig. 1). The monophyly of all paralogs of this class of receptors are also well supported, defining clear orthology and paralogy (Fig. 1). Synteny analyses provide further support for the evolutionary identity of all DRD₂ dopamine receptors (Fig. 3). In our phylogenetic tree DRD₂ was recovered sister to DRD₃ with strong support (Fig. 1), whereas DRD₄ was sister to the DRD₂/DRD₃ clade (Fig. 1). In contrast to the lack of phylogenetic resolution among the DRD₁ class of dopamine receptors, phylogenetic relationships among the DRD₂ class of receptors seem to be well resolved as all studies, including ours, show the same topology ((DRD₂,DRD₃),DRD₄) (Callier et al., 2003; Le Crom et al., 2003; Haug-Baltzell et al., 2015; Spielman, Kumar & Wilke, 2015; Yamamoto et al., 2015).

Phylogenetic evidence for the origin of the DRD_2l gene lineage in the ancestor of gnathostomes

In agreement with Yamamoto et al. (2015), our phylogenetic analyses also suggest the presence of an extra dopamine receptor gene lineage that is related to DRD₂ gene (Boehmier et al., 2004; Boehmler et al., 2007) (Fig. 4). Although our results agree with Yamamoto et al. (2015) regarding the presence of a new dopamine receptor gene lineage, our results suggest a different time of origin.

According to our results, we recovered a strongly supported clade containing the DRD_2l sequences of teleost fish, holostean fish, and coelacanths (Fig. 4) sister to the clade containing DRD₂ sequences of gnathostomes (Fig. 4). This tree topology suggests that in the ancestor of gnathostomes, between 615 and 473 mya, the DRD₂ gene underwent a duplication event that gave rise to an extra DRD₂ gene copy—the DRD_2l—that was independently lost in the ancestor of tetrapods and cartilaginous fish (Fig. 5). In support of this scenario, our phylogenetic tree recovered a cyclostome sequence sister to the DRD₂/DRD_2l clade (Fig. 4). The pattern of gene conservation found up, and downstream of DRD₂ and DRD_2l genes, provides further support for the presence of two DRD₂ dopamine receptor gene lineages (Fig. 3). For example, in the spotted gar (Lepisosteus oculatus), a species that possesses both DRD₂ gene copies, DRD₂ and DRD_2l are found in different chromosomal locations. The identity of their genomic locations is defined by the presence of upstream and downstream flanking genes all across gnathostome vertebrates. Thus, the upstream genes ANKK1 and TTC12 and the downstream genes TMPRSS, ZW10, USP28 and HTR3B define the genomic location of the DRD₂ gene lineage, whereas the upstream gene XRCC1 and downstream genes ETHE1, PHLDB3 and IRQQ1 define the genomic location of the DRD_2l gene lineage (Fig. 3). Importantly, this pattern of conservation is also found in species that lost the DRD_2l gene from their genomes (Fig. 3).

The evolutionary hypothesis proposed here is different from that proposed by Yamamoto et al. (2015) in which the clade containing DRD_2l sequences was recovered sister to a clade containing DRD₂ sequences of vertebrates. Thus, according to their phylogeny the duplication event that gave rise to the DRD_2l gene would have occurred in the ancestor of vertebrates, between 676 and 615 mya, even though they claim that the origin of this gene occurred after the Osteichthyes-Chondrichthyes divergence, between 473 and 435 mya (Yamamoto et al., 2015). Beyond this discrepancy, both evolutionary scenarios proposed in the study of Yamamoto et al. (2015) are different from ours.

An amino acid alignment of both DRD₂ gene lineages revealed that in the case of the spotted gar (Lepisosteus oculatus) and the coelacanth (Latimeria chalumnae) the distance, defined as the percentage of amino acid residues that are different between two sequences, between DRD₂ and DRD_2l receptors is approximately 30% whereas it is approximately 45% in zebrafish (Danio rerio). These estimates are in agreement with previous reports (Boehmier et al., 2004). Additionally, the human DRD₂ amino acid sequence was aligned to the zebrafish, coelacanth and spotted gar DRD_2l sequence to infer functionally significant changes (Fig. 6). The binding sites for dopamine and DRD₂ agonists and antagonists are conserved among these species. However, the adjacent hydrophobic pocket, which confers ligand specificity to DRD₂ is not conserved (Fig. 6). While in humans, coelacanths and spotted gar the second amino acid of the third transmembrane domain (TM3) is phenylalanine (F), it is leucine (L) in zebrafish. This change from an aromatic to an aliphatic amino acid could change the zebrafish DRD_2l ligand specificity and therefore its function. The site that confers specificity to the human G protein subunit G_αi (Senogles et al., 2004) (Fig. 6; orange asterisks) is not conserved among species. The side chain size, shape and polarity changes observed could potentially influence the receptor/G protein coupling specificity, suggesting important evolutionary differences.

Phylogenetic evidence for the origin of DRD_4rs gene lineage in the ancestor of gnathostomes

Also in agreement with Yamamoto et al. (2015) our phylogenetic reconstruction identified an extra dopamine receptor gene lineage that is related to the DRD₄ gene (Figs. 1 and 7). According to our phylogenetic tree, a strongly supported clade that contains dopamine receptors of bony fish and coelacanths was recovered sister to the DRD₄ clade of gnathostomes (Fig. 7). Similarly to the DRD_2l gene lineage, this topology suggests that the DRD₄ gene underwent a duplication event in the ancestor of gnathostomes, between 615 and 473 mya, giving rise to an extra copy of the DRD₄ gene. During the radiation of the group, one of the copies (DRD₄) was retained in all main groups of gnathostomes, whereas the other was only retained in bony fish and coelacanths (Fig. 5). In agreement with this hypothesis, our phylogenetic reconstruction recovered a lamprey sequence sister to the DRD₄/DRD_4rs clade (Fig. 7). Synteny analyses provide further support to our phylogenetic tree, as the genomic locations that harbor both DRD₄ gene lineages are different (Fig. 3). Thus, there are four upstream genes (SCT, CDHR5, IRF7 and PHRF1) and four genes downstream (DEAF1, TMEM80, EPS8L2 and TALDO1) that define the identity of the DRD₄ genomic location (Fig. 3). Similarly, there are upstream genes (KCP, CDHR5, IRF5 and TNOP3) and downstream genes (ATP6V1F) of the DRD_4rs gene that define the identity of its genomic location (Fig. 3). Similar to that found for the DRD₂ genes, our evolutionary hypothesis regarding the origin of the DRD_4rs gene lineage is different from the scenario proposed by Yamamoto et al. (2015). According to their results, the clade containing DRD_4rs sequences was recovered sister to a clade containing DRD₄ sequences of vertebrates. Therefore, their phylogenetic tree suggests that the evolutionary origin of the DRD4rs gene lineage would be in the ancestor of vertebrates, between 676 and 615 mya, as a product of two rounds of whole genome duplication (Yamamoto et al., 2015). Thus, both studies suggest different evolutionary scenarios regarding the time of origin of the DRD_4rs gene lineage.

The distance between the DRD₄ and DRD_4rs gene lineages was found to be higher compared to that estimated for the DRD₂ gene lineages. In the case of the spotted gar (Lepisosteus oculatus) and the coelacanth (Latimeria chalumnae) the distance, defined as the percentage of amino acid residues that are different between two sequences, was approximately 45% whereas in zebrafish (Danio rerio) it was approximately 49%. The human DRD₄ amino acid sequence was aligned to the zebrafish, coelacanth and spotted gar DRD_4rs sequence (Fig. 8). The binding sites for dopamine and DRD₄ agonists and antagonists are conserved among species. Interestingly, two sites in the hydrophobic pocket of the dopamine receptor differ. The first site is located in the selectivity region of DRD₄, where a change from tyrosine (Y) to phenylalanine (F) occurs at position 91 (F91) of the human receptor sequence (Fig. 8; green asterisk). At the second site (Fig. 8; green asterisk) in position 193 of the human DRD₄, the isoleucine (I) in the corresponding spotted gar sequence is changed to valine (V) in the other species (V193). To understand the potential effects that these changes might have on DRD₄ function we used the recently uncovered crystal structure of the human DRD₄ sequence coupled to the antipsychotic drug nemonapride (Wang et al., 2017). All amino acids within 4 Å of the active site are conserved (Figs. 9A and 9B, red amino acids; 9C, red dots) with the exception of the two discussed above. First, F91 located in the recently characterized extended binding pocket, which is a region poorly conserved among dopamine receptors and is key for receptor class specificity (Wang et al., 2017). Second, V193 located in the classic orthosteric-binding pocket known to modulate agonist responses (Lane et al., 2013) (Fig. 9B green amino acids, Fig. 9C, green dots). When we performed molecular dynamics simulation of site-directed mutagenesis to convert the human sequence to the amino acids present in the spotted gar sequence (Fig. 9A, green amino acids), we found that the most likely conformation of Y91 modifies the shape and the ionic properties of the extended binding pocket. Specifically, the polar hydroxyl group oriented along the surface of the pocket would favor interactions with more hydrophilic ligands. Given that the human sequence contains the nonpolar F91 ring, these results could suggest an important evolutionary change in ligand specificity and receptor function. Simulated mutagenesis of V193I, which is located towards the periphery of the orthosteric-binding pocket, slightly modified the shape of the binding pocket; however the nonpolar nature of the amino acid was maintained. Taken together, both substitutions in the human sequence caused the dopamine binding site to be more hydrophobic with less protruding amino acidic side chains, suggesting a structural/functional evolutionary refinement.

Duplicative history and ancestral gene repertoires

To understand the duplicative history of dopamine receptors, including the definition of ancestral repertoires, it is necessary to reconcile the evolutionary history of the gene lineages with the sister group relationships among the species involved. According to our results, the presence of differentiated dopamine receptors in vertebrates (Fig. 5) allowed us to infer that at some point of time the vertebrate ancestor possessed two dopamine receptors, one of each class (Fig. 10). After the two rounds of whole genome duplications (WGD) that occurred in the ancestor of the group (Garcia-Fernàndez & Holland, 1994; Dehal & Boore, 2005) each ancestral gene (DRD_1anc and DRD_2anc) gave rise to four genes in each class of dopamine receptors (Fig. 10). In support of this hypothesis, the DRD₁ and DRD₂ classes of dopamine receptors appear in the repository of genes that originated and were retained after the WGDs occurred in the ancestor of vertebrates (Singh, Arora & Isambert, 2015). The fact that non-vertebrate chordates possess just one DRD₁ (Kamesh, Aradhyam & Manoj, 2008; Burman et al., 2009) and that the four chromosomal locations where the DRD₁ class of receptors are located in humans derive from a single linkage group in the chordate ancestor (Putnam et al., 2008) provide support to our hypothesis. Overall, three out of the four DRD₁ originated as a product of the WGDs were retained in the genome of the vertebrate ancestor (DRD₁, DRD₅ and DRD_1C∕E; Fig. 10). After that, in the gnathostome ancestor the DRD_1C∕E gene underwent a duplication event that gave rise to the actual DRD_1C and DRD_1E genes (Fig. 10). In support of this, we recovered a cyclostome sequence sister to the clade containing the DRD_1C and DRD_1E genes. Thus, the gnathostome ancestor that existed between 615 and 473 mya had a repertoire of four DRD₁ genes: DRD₁, DRD₅, DRD_1C and DRD1_E (Fig. 10). In teleost fish, a group that experienced an extra round of whole genome duplication (Meyer & Van de Peer, 2005; Kasahara et al., 2007; Sato & Nishida, 2010; Glasauer & Neuhauss, 2014), all DRD₁ doubled in number, however, three out of the four gene lineages retained duplicated copies (Fig. 5) (Yamamoto et al., 2013; Yamamoto et al., 2015).

Similarly to the DRD₁ class of receptor, the vertebrate specific WGDs originated four DRD₂ genes, three of which were maintained in the genome of extant species (DRD_2∕2l, DRD₃ and DRD_4∕4rs; Figs. 5 and 10). In the ancestor of gnathostomes the DRD_2∕2l gene underwent a duplication event that gave rise to the actual DRD₂ and DRD_2rs genes (DRD_2l; Figs. 4 and 10). In this case both genes followed different evolutionary trajectories. On one hand DRD₂ was retained in the genome of all of the main groups of vertebrates (Fig. 5) whereas DRD_2l was only retained in coelacanths and bony fish (Fig. 5) (Yamamoto et al., 2015). Similarly, the DRD_4∕4rs gene also underwent a duplication event that gave rise to the actual DRD₄ and DRD_4rs genes (DRD_4rs; Figs. 7 and 10). This case is similar to that found for the DRD₂ gene, as one of the copies (DRD₄) was retained in the genome of all of the main groups of vertebrates, while the other was independently lost in tetrapods and cartilaginous fish (Fig. 6). Consequently, the ancestor of gnathostome vertebrates possessed a repertoire of five DRD₂ class of dopamine receptors: DRD₂, DRD_2l, DRD₃, DRD₄ and DRD_4rs (Fig. 10). As a consequence of the teleost-specific genome duplication (Meyer & Van de Peer, 2005; Kasahara et al., 2007; Sato & Nishida, 2010; Glasauer & Neuhauss, 2014), teleost fish doubled their number of DRD₂ receptors, however extant species retained duplicated copies in just two gene lineages (Fig. 5) (Yamamoto et al., 2015).