Voltage-Dependent Dopamine Potency at D1-Like Dopamine Receptors

In recent years, transmembrane voltage has been found to modify agonist potencies at several G protein-coupled receptors (GPCRs). Whereas the voltage sensitivities of the Gαi/o-coupled dopamine D2-like receptors (D2R, D3R, D4R) have previously been investigated, the putative impact of transmembrane voltage on agonist potency at the mainly Gαs/olf-coupled dopamine D1-like receptors (D1R, D5R) has hitherto not been reported. Here, we assayed the potency of dopamine in activating G protein-coupled inward rectifier potassium (GIRK) channels co-expressed with D1R and D5R in Xenopus oocytes, at -80 mV and at 0 mV. Furthermore, GIRK response deactivation rates upon dopamine washout were measured to estimate dopamine dissociation rate (koff) constants. Depolarization from -80 to 0 mV was found to reduce dopamine potency by about 7-fold at both D1R and D5R. This potency reduction was accompanied by an increase in estimated dopamine koffs at both receptors. While the GIRK response elicited via D1R was insensitive to pertussis toxin (PTX), the response evoked via D5R was reduced by 64% (-80 mV) and 71% (0 mV) in the presence of PTX. Injection of oocytes with Gαs antisense oligonucleotide inhibited the D1R-mediated response by 62% (-80 mV) and 76% (0 mV) and abolished the D5R response when combined with PTX. Our results suggest that depolarization decreases dopamine affinity at D1R and D5R. The voltage-dependent affinities of dopamine at D1R and D5R may be relevant to the functions of these receptors in learning and memory.


INTRODUCTION
Dopamine (DA) receptors are G protein-coupled receptors (GPCRs) subgrouped into the D 1 -like (D 1 R and D 5 R) and the D 2 -like (D 2 R, D 3 R, and D 4 R) receptor families. The D 1 -like receptors are mainly coupled to stimulatory Ga s/olf proteins and expressed in striatum, cortex, amygdala, and hippocampus (Huang et al., 2001;Sarinana et al., 2014) where they are involved in cognitive functions such as memory and attention (Sawaguchi and Goldman-Rakic, 1994;Carr et al., 2017). D 1 R activation has been closely related to prefrontal cortex functioning by a U-shaped relationship; moderate D 1 R activation is required for optimal performance in learning and memory tasks (Goldman-Rakic et al., 2004;Arnsten et al., 2010). Recent data also implicate D 5 R in working memory and prefrontal cortex function (Carr et al., 2017). Accordingly, D 1 R agonists (many of which are also D 5 R agonists) and positive allosteric modulators are being investigated as putative therapeutics for the treatment of cognitive deficits, e.g., in Alzheimer's disease, schizophrenia, and Parkinson's disease (Lewis et al., 2015;Bruns et al., 2018;Luderman et al., 2018;Yang et al., 2018).
Agonist affinities and functional potencies at a number of GPCRs, including D 2 R, D 4 R, and muscarinic M 2 receptors (M 2 R; (Ben-Chaim et al., 2003;Sahlholm et al., 2008a), have been shown to be regulated by the membrane potential. Interestingly, DA potency in D 3 R-mediated GIRK activation was not significantly different between -80 and 0 mV, suggesting that this receptor might be insensitive to voltage (Sahlholm et al., 2008a). However, the influence of the membrane potential on agonist potency at the D 1 -like family of receptors remains unexplored. Several previous experimental investigations of GPCR voltage-dependence have made use of two-electrode voltage-clamp in Xenopus oocytes heterologously expressing GPCRs and G protein-coupled inward-rectifying potassium (GIRK) channels. This has allowed for investigation of GPCR activity by measuring GIRK activation evoked by Ga i/o -coupled GPCRs (Ben-Chaim et al., 2003;Sahlholm et al., 2008a;Sahlholm et al., 2011;Sahlholm et al., 2012) and by Ga q -coupled GPCRs in the presence of a chimeric Ga q-i protein (Ohana et al., 2006).
Although GIRK channel opening is elicited by Gbg, which can be associated with a range of Ga subunits, activation of native GIRK currents is mediated almost exclusively via Ga i/o -and not Ga s/olf -proteins. This specificity may be achieved through the higher rate of turnover of the G protein cycle of Ga i/o -proteins, liberating higher local concentrations of Gbg upon receptor activation (Touhara and MacKinnon, 2018). However, based on previously published work with similarly G s -coupled badrenergic receptors (Lim et al., 1995;Wellner-Kienitz et al., 2001;Hatcher-Solis et al., 2014), we speculated that high expression levels of D 1 R and D 5 R in Xenopus oocytes might allow their downstream Ga s/olf proteins to release sufficient amounts of Gbg to activate GIRK channels, thus allowing us to investigate the putative voltage sensitivities of these receptors. Importantly, previous work has shown that neither G protein dissociation into Ga and Gbg, nor GIRK activation by Gbg, are intrinsically voltage dependent processes, thus making GIRK currents a suitable readout for studies of GPCR voltage sensitivity (Ben-Chaim et al., 2003). Here, we report the differential potencies of DA in D 1 R-and D 5 R-mediated GIRK activation at -80 and 0 mV.

Oocyte Preparation
Oocytes from the African clawed toad, Xenopus laevis, were isolated surgically as described previously . The surgical procedures have been approved by the Swedish National Board for Laboratory Animals and the Stockholm Ethical Committee. Following 24 h of incubation at 12°C, oocytes were injected with 4.5 ng D 1 R receptor cRNA or 25.5 ng D 5 R receptor cRNA and 50 pg of each GIRK1 and GIRK4 cRNA, using the Nanoject II (Drummond Scientific, Broomall, PA) and a volume of 50 nl per oocyte. When used, 3 ng PTX-S1 cRNA was injected, based on previous observations of complete inhibition of D 2 Rinduced GIRK activation with this amount of cRNA (Agren et al., 2018a). In a subset of experiments, 10 pmol/oocyte (in a volume of 50 nl) of an antisense DNA oligonucleotide (sequence: GCTCATATTGGCGCAGGTGCAT) directed against X. laevis Ga s mRNA was injected 48 h before electrophysiology recordings. This treatment has previously been described to abolish G s -dependent signaling in oocytes (de la Pena et al., 1995).

Electrophysiology Methods
Following cRNA injection into oocytes and 7 days of incubation at 12°C, electrophysiological experiments were conducted using the parallel eight-channel, two-electrode voltage-clamp, OpusXpress 6000A (Molecular Devices, San Jose, CA). Continuous perfusion was maintained at 1.5 ml/min. Data were acquired at membrane potentials of -80 mV or 0 mV and sampled at 134 Hz using the OpusXpress 1.10.42 (Molecular Devices) software. To increase the inward rectifier potassium channel current at negative potentials, a high potassium extracellular buffer was used (in mM: 64 NaCl, 25 KCl, 0.8 MgCl 2 , 0.4 CaCl 2 , 15 HEPES, 1 ascorbic acid, adjusted to pH 7.4), yielding a K + reversal potential of about -40 mV. In experiments with 1 mM KCl, the NaCl concentration was 88 mM. DA was purchased from Sigma-Aldrich (St. Louis, MO). Recordings were performed at room temperature (22°C).

Data Analysis
Electrophysiological data were analyzed in Clampfit 10.6 (Molecular Devices). Concentration-response curves were fitted using least squares nonlinear regression in GraphPad Prism 8 (GraphPad Software, San Diego, CA). For each oocyte and each holding voltage, the current response to each concentration of DA tested was normalized to the response to the maximally effective concentration of DA at the same voltage and in the same oocyte. The following equation was fitted to the normalized agonist data: where Y is the normalized response, X the logarithm of DA concentration, and n the Hill slope.
The washout decay time constant, t off , was obtained from single exponential fits (using Levenberg-Marquardt least-squares fitting in Clampfit 10.6) to the washout-induced current deactivation time course. The responses to 1 µM (D 1 R) and 100 nM (D 5 R) DA were used for analysis of response deactivation kinetics. The first 10 s following agonist washout were discarded, and the exponential function was fit to the data over~70% of the response amplitude. k off was calculated from t off using the following relation: Data are represented as mean ± SEM. Concentrationresponse data were analyzed by comparing the fractional responses to DA at individual concentrations at -80 mV and at 0 mV using Student's paired t-test or, if data were not normally distributed, Wilcoxon signed rank test. Normality was assessed using the Shapiro-Wilk test. Current amplitudes were compared using one-way ANOVA with Tukey's multiple comparisons test or, if data were not normally distributed, Kruskal-Wallis test with Dunn's multiple comparisons test. k off rates were compared using the Wilcoxon signed rank test. The significance threshold was p<0.05. Statistical analyses were performed in GraphPad Prism 8.

RESULTS
The effects of DA application on membrane currents were investigated in oocytes co-expressing D 1 R or D 5 R with GIRK1/4 channels. DA was found to elicit inward currents at -80 mV, whereas at 0 mV, outward currents were observed. No appreciable current response to DA could be recorded in oocytes expressing D 1 R or D 5 R in the absence of GIRK channels at either voltage (Supplementary Figure 1), nor did DA elicit any response in oocytes expressing GIRK channels without D 1 R or D 5 R (not shown).
DA potency at D 1 -like receptors was investigated by DA applications (40-s applications, each followed by 100-s washout) of increasing concentration to oocytes injected with D 1 R or D 5 R and GIRK1/4 cRNA. Submaximally effective concentrations were applied, followed by a maximally effective concentration (30 µM for D 1 R and 3 µM for D 5 R), at both -80 and 0 mV ( Figures 1A, B). At D 1 R, the DA EC 50 was 125 nM at -80 mV, increasing significantly to 906 nM at 0 mV ( Figure 1C), a potency shift of about 7-fold. The DA EC 50 at D 5 R was 6.1 nM at -80 mV, again increasing significantly to 44 nM at 0 mV ( Figure 1D), resulting in ã 7-fold decrease in DA potency. The current-voltage relationships of the membrane currents, as assessed by 4-s ramps and normalized to the amplitude at -80 mV, were virtually superimposable in the absence and in the presence of DA (Supplementary Figure 2).
To assess whether GIRK-activation was mediated via Ga i/o proteins, the catalytic subunit of pertussis toxin (PTX-S1), which inactivates Ga i/o proteins by ADP-ribosylation, was expressed in the oocytes (Vivaudou et al., 1997). In addition, oocytes were injected with an antisense oligonucleotide designed to knock down Ga s expression (de la Pena et al., 1995). Co-expression of PTX-S1 with D 1 R and GIRK1/4 did not significantly affect GIRK response amplitudes to 10 µM DA, neither at -80 mV nor at 0 mV. However, injection of the Ga s antisense oligonucleotide strongly and significantly suppressed the DA-evoked current response by 62% (-80 mV) and 76% (0 mV; Figure 1E). In contrast, in oocytes expressing D 5 R and GIRK1/4, the amplitudes of GIRK responses to 10 µM DA were significantly reduced by 64% and 71% compared to control at -80 and 0 mV, respectively, when PTX-S1 was introduced. With the further addition of the Ga s antisense oligonucleotide, the DA response was virtually abolished ( Figure 1F).
The rate of GIRK response deactivation upon removal of agonist has been used to approximate the time course of agonist dissociation from its receptor (Bunemann et al., 2001;Benians et al., 2003) and changes in the rates of GIRK deactivation between hyperpolarized and depolarized potentials have previously been shown to reflect reciprocal increases or decreases in agonist dissociation rates and consequently, affinities (Ben-Chaim et al., 2003;Ohana et al., 2006;Ben Chaim et al., 2013). Exponential functions were fitted to the time courses of GIRK current deactivation upon DA washout (see Methods). The rate of response decay was observed to increase, both at D 1 R and at D 5 R, when the membrane was depolarized from -80 mV to 0 mV ( Figure 2). As judged by the rate of current increase (at -80 mV) upon wash-in of buffer containing 25 mM KCl from a baseline reading in buffer containing 1 mM KCl, the rate of solution exchange was faster than the fastest rate of decrease of the GIRK response upon DA washout (Supplementary Figure 3).
Finally, to visualize the change in DA potency when stepping from one voltage to another, we performed experiments in oocytes co-expressing D 1 R with GIRK1/4 channels where a 40-s prepulse to between -80 to 0 mV was followed by a -80 mV post-pulse, in the absence or presence of an intermediately-(300 nM) or a maximally (30 µM) effective concentration of DA. Subtracting the basal, agonist-independent current from the current evoked in the presence of agonist, a slow current increase at -80 mV was evident following depolarized prepulses in the presence of 300 nM, but not 30 µM DA (Supplementary Figure 4). This behavior of GIRK currents, which has been referred to as "relaxation", has earlier been demonstrated to be a consequence of receptor voltage sensitivity and to reflect the increase in agonist binding to the receptor, at submaximally effective concentrations, upon hyperpolarization of the membrane (Moreno- Galindo et al., 2011;Sahlholm, 2011).
Overexpression of G s -coupled receptors has been reported to support activation of GIRK channels, although much less efficiently than G i/o -coupled receptors, in several expression systems including Xenopus oocytes (Lim et al., 1995;Wellner-Kienitz et al., 2001;Hatcher-Solis et al., 2014). Thus, we believe that the PTX-insensitive D 1 R-mediated GIRK activation observed here is likely to be elicited via G s signaling, although this phenomenon may be a consequence of receptor overexpression and unlikely to take place in native tissue. Indeed, this conclusion is strengthened by the observation that injection of an antisense oligonucleotide directed towards X. laevis mRNA encoding Ga s strongly reduced the current response to DA in oocytes co-expressing D 1 R and GIRK channels.
Interestingly, in contrast to the D 1 R-evoked responses, a major component of the GIRK currents elicited upon D 5 R stimulation was PTX-sensitive, suggesting that D 5 R is able to activate G i/o proteins in addition to G s/olf . While D 5 R-mediated G q signaling has been observed in some heterologous systems (So et al., 2009), G i/o -coupling of D 5 R has, to the best of our knowledge, not been reported previously. In Xenopus oocytes, G q -mediated calcium signaling would typically elicit a characteristic, rapidly desensitizing response mediated through endogenous calcium-activated chloride channels (Hansen and Bräuner-Osborne, 2009); however, we did not observe any such responses upon D 5 R activation.
Contrary to the present observations, D 1 R has previously been reported to couple to PTX-sensitive G o proteins in addition to G s , while D 5 R was found to couple to the PTX-insensitive inhibitory G protein, G z (Sidhu et al., 1998), which efficiently activates GIRK (Vorobiov et al., 2000). However, available evidence suggests that there is no detectable endogenous G z expression in Xenopus oocytes (Vorobiov et al., 2000;Kalinowski et al., 2003), making it unlikely that G z activation underlies the PTX-insensitive component of D 5 R-mediated GIRK activation observed here. Instead, it appears more likely that this component is G s -mediated. Again, this assumption is strengthened by the abolition of the DA-induced current response by injection of Ga s antisense oligonucleotide into oocytes co-expressing D 5 R and GIRK with PTX-S1.
The depolarization-induced decrease in acetylcholine potency at the M 2 R has been related to a corresponding increase in acetylcholine k off at the M 2 R, and a consequent increase in GIRK response deactivation rate at depolarized potentials (Ben-Chaim et al., 2003;Ben Chaim et al., 2013;Agren et al., 2018b;Loṕez-Serrano et al., 2020). Similar findings have been reported for the mGluR3 and histamine H 3 receptors (Ohana et al., 2006;Sahlholm et al., 2012). Likewise, in the present study, we found the rates of GIRK current deactivation upon DA washout to be increased at 0 mV compared to -80 mV. Presumably, these changes in deactivation rates reflect faster DA k off s at D 1 R and D 5 R at 0 mV.

Ågren and Sahlholm
Voltage Dependence of D1-Like Dopamine Receptors Frontiers in Pharmacology | www.frontiersin.org October 2020 | Volume 11 | Article 581151 other tissues. However, this heterologous system provides a well-defined background with no detectable expression of endogenous DA receptors which lends itself well to stable voltage-clamp recordings and detection of even relatively small agonist-induced responses with a high signal-to-noise ratio. The DA EC 50 s for D 1 R-and D 5 R-induced GIRK activation at 0 mV in the present study agree fairly well with the reported values for high-affinity DA binding in isolated membranes in the literature, recapitulating the higher affinity of D 5 R (24 nM; Weinshank et al., 1991) compared to D 1 R (324 nM; Marcellino et al., 2012), suggesting that the DA binding characteristics of these receptors in oocytes are similar to those of D 1 R and D 5 R expressed in mammalian cells.

CONCLUSION
The present results reveal that DA potency at D 1 R and D 5 R is decreased upon depolarization from -80 to 0 mV. Our interpretation is that DA k off rate constants are affected by membrane voltage and contribute to a decrease in DA affinity upon depolarization. This dependence of D 1 R and D 5 Rmediated responses upon transmembrane voltage could allow these receptors, which have been implicated in learning and memory, to function as a sort of "coincidence detectors", responding robustly to low concentrations of DA only at hyperpolarized potentials.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Swedish National Board for Laboratory Animals, Stockholm Ethical Committee (Stockholms djurförsöksetiska nämnd).

AUTHOR CONTRIBUTIONS
RÅ and KS designed the experiments, RÅ conducted the experiments, RÅ and KS analyzed data, RÅ drafted the first version of the manuscript. KS supervised the project. All authors contributed to the article and approved the submitted version.

FUNDING
This study was supported by grants from Stiftelsen Lars Hiertas Minne, Åhleństiftelsen, and Magnus Bergvalls stiftelse (to KS). KS is currently a fellow at the Wallenberg Center for Molecular Medicine at Umeå University.