An adenylyl cyclase signaling pathway predicts direct dopaminergic input to vestibular hair cells

Abstract

Adenylyl cyclase (AC) signaling pathways have been identified in a model hair cell preparation from the trout saccule, for which the hair cell is the only intact cell type. The use of degenerate primers targeting cDNA sequence conserved across AC isoforms, and reverse transcription-polymerase chain reaction (RT-PCR), coupled with cloning of amplification products, indicated expression of AC9, AC7 and AC5/6, with cloning efficiencies of 11:5:2. AC9 and AC5/6 are inhibited by Ca(2+), the former in conjunction with calcineurin, and message for calcineurin has also been identified in the trout saccular hair cell layer. AC7 is independent of Ca(2+). Given the lack of detection of calcium/calmodulin-activated isoforms previously suggested to mediate AC activation in the absence of Gαs in mammalian cochlear hair cells, the issue of hair-cell Gαs mRNA expression was re-examined in the teleost vestibular hair cell model. Two full-length coding sequences were obtained for Gαs/olf in the vestibular type II-like hair cells of the trout saccule. Two messages for Gαi have also been detected in the hair cell layer, one with homology to Gαi1 and the second with homology to Gαi3 of higher vertebrates. Both Gαs/olf protein and Gαi1/Gαi3 protein were immunolocalized to stereocilia and to the base of the hair cell, the latter consistent with sites of efferent input. Although a signaling event coupling to Gαs/olf and Gαi1/Gαi3 in the stereocilia is currently unknown, signaling with Gαs/olf, Gαi3, and AC5/6 at the base of the hair cell would be consistent with transduction pathways activated by dopaminergic efferent input. mRNA for dopamine receptors D1A4 and five forms of dopamine D2 were found to be expressed in the teleost saccular hair cell layer, representing information on vestibular hair cell expression not directly available for higher vertebrates. Dopamine D1A receptor would couple to Gαolf and activation of AC5/6. Co-expression with dopamine D2 receptor, which itself couples to Gαi3 and AC5/6, will down-modulate levels of cAMP, thus fine-tuning and gradating the hair-cell response to dopamine D1A. As predicted by the trout saccular hair cell model, evidence has been obtained for the first time that hair cells of mammalian otolithic vestibular end organs (rat/mouse saccule/utricle) express dopamine D1A and D2L receptors, and each receptor co-localizes with AC5/6, with a marked presence of all three proteins in subcuticular regions of type I vestibular hair cells. A putative efferent, presynaptic source of dopamine was identified in tyrosine hydroxylase-positive nerve fibers which passed from underlying connective tissue to the sensory epithelia, ending on type I and type II vestibular hair cells and on afferent calyces.

Fig. 1. First Row: First panel shows agarose gel resolution of AC isoform PCR products obtained from trout saccular hair cell cDNA

Degenerate primers, targeting C2 domain sequence conserved across the nine AC isoforms of higher vertebrates (Table 1), directed amplification of 220-300 bp products. Second and third panels show agarose gel separation of calcineurin amplification products from trout saccular hair cell cDNA. Degenerate primers targeting calcineurin from the trout saccular hair cell cDNA yielded PCR products of predicted size, 599 bp (Calcn Δ599) (Table 1, 2b) and 593 bp (Calcn Δ593) (Table 1, 2a) for hair cell preparations. No amplification was observed for negative control in the absence of reverse transcriptase. The PCR products were cloned with pGEM T-easy vector and sequenced, with the sequence deposited into GenBank (GU073385). Last four panels show Gαs/olf messages in the trout saccular hair cell preparation. Application of degenerate primers targeting Gαs sequence for Xenopus laevis and Drosophila melanogaster (Table 1, 3a) in PCR yielded an amplification product of predicted size, 479 bp (Gαs Δ479), for trout saccular hair cell and trout brain cDNA. A 5’ RACE gene-specific primer 4 (5’racegsp4down, Table 1) directed amplification of hair cell Gαs/olf-2 cDNA (Gαs Δ1300) but not brain Gαs/olf-2 cDNA, presumably due to one nucleotide difference in the respective primer regions for hair cell and brain. The 5’ RACE gene-specific primer 3 (5’racegsp3 down, Table 1), reflecting sequence of Gas/olf-1 and not Gas/olf-2, elicited amplification of brain cDNA at ~ 2 Kb (Gαs Δ2000) but not for hair cell cDNA. Primers specific to trout brain Gas/olf-1 sequence directed amplification of the predicted 294 bp product (Gαs Δ294), with sequence confirmation for hair cell cDNA and brain cDNA. Second Row: Amplification of saccular hair cell Gαi cDNA with degenerate and specific primers (see Table 1, Gαi-1 through Gαi-4 refer to primer sets and not Gαi isoforms). First panel. Agarose gel shows a PCR product of predicted size 530 bp produced with degenerate primers (Gαi-1, Table 1) from saccular hair cells with reverse transcriptase but not in its absence (Gαi-1, Δ530). Second panel. A PCR product of predicted size, 248 bp, was obtained for hair cell cDNA with medaka-specific upstream and trout-specific downstream primers (Gαi-2, Table 1) (Gαi-2, Δ248). Third panel. A 771 bp predicted product resulted from the use of trout specific primer upstream coupled with medaka sequence downstream on hair cell cDNA (Gαi-3 primers, Table 1) (Gαi-3, Δ771). Fourth panel. A 1,016 bp PCR product was obtained with Gαi-4 primers (Table 1) (Gαi-4, Δ1016) in the presence of reverse transcriptase but not in its absence. The 771 bp and 1,016 bp products were later cloned and sequenced. Third Row: Agarose gels of dopamine D1 PCR products from the trout saccular hair cell layer cDNA. First and second panels. Amplification results for dopamine D1 receptor transcripts from RT-PCR applied to trout brain and trout saccular hair cells from the use of degenerate primers (Table 1). The predicted 690 bp PCR products (D1 Δ690, panel 1 and panel 2) not present in negative controls. Last three panels. Predicted PCR products amplified from trout saccular hair cell cDNA, 391 bp (D1 Δ391), 497 bp (D1 Δ497), and 691 bp (D1 Δ691). Fourth Row: Agarose gel separation of dopamine D2 PCR products amplified from the trout saccular hair cell cDNA. Dopamine D2 receptor primers (see Table 1) yielded predicted amplification products (D2 Δ208, D2 Δ599, D2 Δ776, D2 Δ804, D2 Δ900, D2 Δ587, D2 Δ475, D2 Δ500) with reverse transcription for the trout saccular hair cell layer with sequence confirmation. S = standards; H + = hair cell sheet with reverse transcriptase; H - = hair cell sheet without reverse transcriptase; Br = trout brain with reverse transcriptase, a positive control. B = water blank.

Fig. 2. Sequence from cloning of the combined AC isoform PCR products from application of degenerate primers

Amino acid sequences are presented for AC isoforms expressed in the saccular hair cell layer including three apparent splice variants of AC9 (lines 1-3), AC7 (line 4) and AC5/6 (line 5). The relative number of clones for AC9:AC7:AC5/6 was 11:5:2.

Fig. 3. Amino acid sequences for Gαs/olf expressed in trout saccular hair cell preparation compared with sequences for Gαs and Gαolf in human

Line 1, human Gαs (NP_000517). Line 2, trout hair cell Gαs/olf-1 (ABU45775). Line 3, trout hair cell Gαs/olf-2 (ABU45776). Line 4, human Gαolf-2 (NP_002062). Line 5, human Gαolf-1 (NP_892023).

Fig. 4. Western blot and immunohistochemical localization of Gαs/olf proteins in the trout saccule

a. Western blot for Gαs/olf protein in the trout saccular hair cell layer. Hair-cell layer lysate was probed with Gαs/olf antibody. Lane 1, molecular size standard (Magic Mark, Invitrogen); Lane 2, trout hair cell layer. The band (arrow) corresponds to the estimated molecular mass for Gαs/olf of 45 kDa, based on the deduced trout Gαs/olf-1/Gαs/olf-2 amino acid sequence. b. Immunohistochemical localizations were conducted with transverse sections cut perpendicular to the longitudinal axis of the trout saccule. At low magnification, Gαs/olf immunoreactivity was localized to the upper two-thirds of the sensory epithelium, with immunostaining of hair cells and neural components. Gαs/olf immunoreactivity was found in hair cell stereocilia throughout the sensory epithelium (arrows) and not observed associated with supporting cells whose nuclei lie just above the basal lamina in the lower third of the epithelium. Scale bar = 50 μm. c-d. At higher magnification, Gαs/olf immunoreactivity was localized to the lateral membrane (intermediate-length arrows) and stereocilia (short arrows) of saccular hair cells. Immunoreactivity was also associated with large afferent fibers (c, single arrowheads encompass large afferent) with circles of immunoreactivity at sites of afferent nerve contacts on hair cells (c, double arrowhead). Small-caliber efferent fibers were immunoreactive for Gαs/olf (d, arrowheads). Scale bar in c, applicable also for d = 10 μm.

Fig. 5. Gαi amino acid sequences for the trout saccular hair cell preparation were deduced from nucleotide sequencing of cloned 1,016 bp and 771 bp PCR products

The 1,016 bp cloned product (line 1), representing 95% of the coding region for a Gαi protein, bore highest amino acid identity to medaka Gαi1 (line 2, NP_001116402) and rat Gαi1 (line 3, NP_037277). A second cloned product 771 bp in length (line 4) was highest in amino acid identity to Salmo salar Gαi3 sequence (line 5) (NP_001135256) and rat Gαi3 (line 6, NP_037238).

Fig. 6. Western blot and immunohistochemical localization for Gαi protein in the trout saccule

a. Western blot for Gαi1/ Gαi3 protein in the trout saccular hair cell layer. Lane 1, molecular size standards. Lane 2, trout saccular hair cell layer protein. The band at 40 kDa (arrow) indicates an immunolabeled protein corresponding to calculated molecular mass for cloned saccular hair cell Gαi1. b. Low magnification of immunoreactivity for Gαi (Gαi1/Gαi3) in the trout saccule. Scale bar = 100 μm. Relative positions in the saccule corresponding to more highly-magnified micrographs c-h are indicated by arrows. c. Immunoreactivity for Gαi1/Gαi3 was localized to stereocilia (asterisks) and to the lateral membranes (short arrows) and basal membranes of saccular hair cells (intermediate length arrow). Scale bar = 10 μm. d. Immunoreactivity for Gαi1/Gαi3 was localized to a small diameter presumptive efferent fiber positioned close to the hair cell plasma membrane (short arrow) juxtaposed to the ending of an unreactive nerve fiber (inset: arrowheads and dotted line point to bow-shaped unreactive nerve fiber). Scale bar = 10 μm. e. Gαi immunoreactivity was observed in small nerve fibers entering the sensory epithelium just above the basal lamina passing between supporting cell nuclei (arrows). Scale bar = 10 μm. f. Immunoreactivity for Gαi1/Gαi3 filled the cell bodies of flask-shaped hair cells located close to the border of sensory and non-sensory epithelium. Scale bar = 50 μm. g,h. Immunoreactivity for Gαi at one focal plane was observed in hair cell stereocilia (g, asterisks) and small foci along the basal membrane of hair cell (g, arrow) close to large afferent fiber ending (g, arrowhead). At a lower focal plane, immunoreactivity was associated with efferents (h, arrow). The emergence of the large afferent is indicated (h, arrowheads). Scale bar = 10 μm, for g,h. i. Preabsorption of the antibody with peptide antigen essentially eliminated immunostaining. Scale bar = 10 μm.

Fig. 7. Amino acid sequence for dopamine D1 expressed in the trout saccular hair cell layer

OMIGA alignment of sequence for dopamine D1A4 expressed in the trout saccular hair cell layer (line 1, EU371401) with sequences for carp dopamine D1A4 (line 2, Y14627), rat dopamine D1A (line 3, S46131), and human dopamine D1A (line 4, CH471062).

Fig. 8. Amino acid sequence for dopamine D2 receptors expressed in the trout saccular hair cell layer

Line 1, trout hair cell (thc) dopamine D2L-1 sequence. Line 2, variant 1 of dopamine D2-1 sequence. Line 3, variant 2 of dopamine D2-1 sequence expressed in hair cell layer. Line 4, trout dopamine D2L-2 (CAC79663, Oncorhynchus mykiss dopamine D2 receptor 2). Line 5, trout saccular hair cell dopamine D2L-2. Line 6, trout saccular hair cell D2S-2. Line 7, human dopamine D2L (NP_000786).

Fig. 9. Immunoreactivity for dopamine D1A, D2L and tyrosine hydroxylase in the sensory epithelia of rat saccule and mouse utricle detected with 3,3′-diaminobenzidine

A. Dopamine D1A immunoreactivity in the rat saccule and mouse utricle. 1. Dopamine D1A immunoreactivity in rat saccule was observed on the inner face of an afferent calyx at its interface with a type I vestibular hair cell (arrow + dot) and in presumptive efferents (intermediate length arrow) at the base of a type II vestibular hair cell identified on basis of the large afferent ending (short arrow; compare afferent in Fig. 9A panel 1 for a type II hair cell with afferent endings in Fig. 4c, Fig. 6c,g,h designated by arrowheads). 2. Dopamine D1A immunoreactivity was found in hair cell subcuticular sites in rat saccule (short arrow). 3. D1A-positive efferents (long arrow) were associated with large afferents (short arrow) at base of putative type I vestibular hair cell in the mouse utricle. Scale bar = 10 μm for 1-3. B. Dopamine D2L immunolocalization in mouse utricle. 1. Immunoreactivity for dopamine D2L was observed at apical, subcuticular sites on type I vestibular hair cells (long arrow) and at both the outer face (intermediate-length arrow) and inner face (short arrow) of calyceal afferent endings, with the inner face corresponding to the interface of calyx and hair cell membrane. 2,3. Dopamine D2L immunoreactivity at subcuticular sites on vestibular hair cells (arrows). Scale bar = 10 μm for 1-3. C. Tyrosine hydroxylase immunoreactivity in rat saccule. Immunoreactivity for tyrosine hydroxylase was found within the saccular sensory epithelia in small nerve fibers traveling between supporting cell nuclei just above the basal lamina (short black arrows). Small-diameter immunoreactive fibers contacted the outer membrane of a putative type II vestibular hair cell (long black arrow with dot) and were observed (in cross section, white arrow) to be in close proximity to a larger unlabeled fiber. Small-diameter tyrosine hydroxylase-positive nerve fibers made contact with the outer face of a calyx associated with a type I hair cell (long black arrow) and hair cell body (intermediate-length black arrow). Inset. A grouping of three small-caliber tyrosine hydroxylase-positive nerve fibers (arrow) are identified making contact with the outer face of the afferent calyx, the cell membrane of a type I hair cell, and the cell membrane of an adjacent type II vestibular hair cell. Scale bar = 10 μm.

Fig. 10. Dopamine D1 and D2L and tyrosine hydroxylase immunofluorescence in the rat saccule and utricle in comparison with calyceal afferent marker calretinin and AC5/6

a. Dopamine D1A-positive nerve fibers passed from the connective tissue through the basal lamina to the saccular sensory epithelium of the rat saccule (arrows). b. Dopamine D2L immunofluorescence was localized to nerve fibers which traveled from the connective tissue through the basal lamina to the sensory epithelium of rat saccule (arrows). c. Dopamine D1A immunofluorescence (green) was compared with immunofluorescence for calretinin (red). D1A-positive nerve fibers (long arrow) shadow the calretinin-positive fibers (short arrow) within the connective tissue before passing into the sensory epithelium. Within the sensory epithelium, overlap of dopamine D1A and calretinin immunoreactivity (merge, yellow) occurred on calretinin-containing calyces (double arrowheads) and on calyces at the base of type I vestibular hair cells (intermediate-length arrow). D1A immunoreactivity was found on the cell membrane of the type 1 vestibular hair cell (arrowhead) and at two supranuclear sites (asterisks), one close to the reticular lamina (larger asterisk). d1. D1A immunoreactivity and overlapping differential interference contrast (DIC) image of a type I vestibular hair cell illustrated in c, with immunoreactivity on both faces of the calcyceal afferent (short arrows), on basolateral aspects of the hair cell membrane (arrowhead), and at both intermediate and apical (subcuticular) supranuclear sites. d2. Calretinin immunoreactivity for calyx surrounding the type I vestibular hair cell depicted in d1. e. Dopamine D1 and AC5/6 in rat utricle. AC5 immunofluorescence (green) overlapped/co-localized with dopamine D1 immunofluorescence (red) at a calyceal ending on type I vestibular hair cell (long arrow) (merge, yellow). Co-localization of AC5/6 with dopamine D1 was observed on the basal cell membrane of a type II vestibular hair cell (short arrows) close to a putative efferent nerve ending. f. AC5/6 immunofluorescence image of vestibular type II hair cell illustrated in e overlapping DIC images in adjacent section. g. Dopamine D2L and AC5/6 in rat saccule. Dopamine D2L immunofluorescence (green) co-localized with AC5/6 immunofluorescence (red) at calyceal afferent endings on type I vestibular hair cells (arrow), at the base of type I vestibular hair cell (arrowhead) and at supranuclear sites (asterisks) (merge, yellow). h. DIC and D2L immunoreactivity for the same section. Positions of D2L immunofluorescence at subcuticular sites are marked by asterisks corresponding to first three asterisks in g (left to right). i. Immunofluorescence (green) for tyrosine hydroxylase, a dopaminergic/catecholaminergic (efferent) neuronal marker, was compared to immunofluorescence (red) for calretinin, a calyceal afferent marker. The merged image indicates a small-diameter, tyrosine-positive fiber (green, intermediate-length arrow) in the connective tissue passing through the basal lamina into the sensory epithelium to end (short arrow) on a calyceal afferent (red) of a type I vestibular hair cell (overlap/co-localization in yellow). Tyrosine hydroxylase-positive nerve fibers (presumptive efferents) passed through the connective tissue en route to the sensory epithelium (long arrows). j. DIC and tyrosine hydroxylase immunoreactivity in same section with arrowhead pointing to same site of accumulation of tyrosine hydroxylase immunoreactivity on cell body of a type I vestibular hair cell in i and j. Scale bars = 10 μm, except for d (5 μm).

Fig. 11. Proposed dopaminergic innervation of type I and type II vestibular hair cells of the mammalian saccule/utricle

Localizations of dopamine D1 and D2 receptors and AC5/6 and tyrosine hydroxylase-positive efferent nerve fibers are indicated for type I and type II vestibular hair cells of the saccule/utricle in rat/mouse. Labels reflect localizations for which there is direct evidence from the present study. Putative neurotransmitters in nerve fibers are indicated in italics. E, efferent; A, afferent.