Dopamine HCl (cat. no. H8502), fenoldopam (cat. no. SML0198), phenylthiourea (cat. no. P7629) and proteinase K (cat. no. 3115836001) were obtained from Merck KGaA, SCH39166 hydrobromide was purchased from Biomol (cat. no. Cay25331), and tenapanor was purchased from BIOZOL Diagnostica Vertrieb GmbH (cat. no. ADQ-A14011-2). Poly-D-lysine was obtained from Merck KGaA. Veterinarian-grade tricaine mesylate was obtained from PharmaQ.

The following fish lines were used: AB and EK wild-type lines, Grk4^mut (4) and Tg(wt1b:GFP) (9). Adult zebrafish were housed in a Zebtec zebrafish holding rack system with removable tanks, permanent water cleanup and recycling and automatic monitoring and adjustment of water parameters such as pH, temperature and conductivity (Tecniplast). The fish were kept under a 14-h light and 10-h dark cycle and were fed three times a day with pelleted dry food (Sparos) and once a day with living artemia. Zebrafish husbandry and propagation were approved by local authorities at Ulm University and University of Tuebingen, as well as the regional board for scientific animal experiments. Experiments performed herein did not require an additional permit and were carried out in agreement with the European Communities Council Directive of September 22, 2010 (2010/63/UE).

Fertilized eggs were generated by natural mating and embryos were incubated at 28.5°C in embryo medium (13.7 mM NaCl, 0.5 mM KCl, 0.025 mM HNa₂PO₄, 0.044 mM KH₂PO₄, 1.3 mM CaCl₂, 1 mM MgSO₄ × 7 H₂O, 0.42 mM NaHCO₃, pH 7.2) until the desired stage. Manipulation by microinjection was performed at the 1-cell stage using an antisense morpholino oligonucleotide (MO) targeting the translation start site of zebrafish Grk4 as previously characterized and published (4): Grk4 MO: 5'-GTC AGA AGA AAG ATC CCA TGC AGT C. All injected embryos were controlled against embryos injected with a control MO (CTRL MO: 5'-GTG ACA ACA AAC ATC CCA TCC AGT C) harboring 5 mismatches compared to the Grk4 MO (4). In addition, uninjected wild-type embryos of the same clutch were incubated along to assess clutch quality. All MOs were custom-synthetized by Gene Tools.

Zebrafish embryos homozygous for a hypomorphic allele of Grk4 (named Grk4^mut) were generated by heterozygous mating. Embryos were raised until the desired stage, which was either 20 somite stage or 48 hpf, processed for analysis and subsequently genotyped (please see below).

Manipulations with small chemical compounds were performed from the tailbud stage until the final analysis was completed. All compound treatments were controlled against the vehicle DMSO, in which the compounds were dissolved. Fenoldopam, SCH39166 and tenapanor were dissolved in DMSO and used at 5 μM (tenapanor) and 10 μM (fenoldopam, SCH39166) final concentration.

Embryos undergoing subsequent fluorescent analyses were furthermore treated from tailbud stage with phenylthiourea to block pigmentation. Analyses (i.e. image analysis or phenotypic scoring) did not include the blinding of the experimenter.

Embryos were digested in lysis buffer (100 mM Tris HCl pH 8.5, 5 mM EDTA, 0.2% SDS, 200 mM NaCl) containing 1 mg/ml proteinase K at 55°C overnight. An equal amount of isopropanol was added and the DNA was pelleted by centrifugation at room temperature, 11,000 × g for 15 min. The DNA pellet was dissolved in ultrapure water and directly used for polymerase chain reaction (PCR). PCR was performed with the following primers amplifying exon 7 along with parts of the flanking introns selected based on the Ensembl sequence ENSDART00000066638.7: WT forward, 5'-TGGGATTTACGGTGCATTGTTATTAGTGAA and Mut forward, 5'-GACATTACAGGGTATTGGGCGGAT and reverse, 5'-ACCACTGATAACGCCCATTCAATCAAATG, and Taq polymerase from NEB (cat. no. M0267). Following an initial denaturation at 94°C for 1 min, 35 cycles of 12 sec 94°C, 15 sec 68°C and 30 sec elongation at 68°C were run followed by an additional 3 min at 68°C. Bands (WT: 493 bp, mutant: 487 and 385 bp) were resolved on a 2% agarose gel.

Capped RNA encoding dnSTAT3 was generated by the linearization of pcDNA3-STAT3-Y705F (Addgene no. 74434), which has been previously validated as dominant negative (10) using DraIII and subsequent in vitro transcription using the T7 mMessage mMachine Kit of Ambion (cat. no. AM1344; Thermo Fisher Scientific, Inc.) following the manufacturer's instructions.

For the generation of new in situ probes, TOPO TA cloning was performed (cat. no. K460001; Thermo Fisher Scientific, Inc.). As inserts, coding sequence fragments were generated using PCR using the Expand™ Long Template PCR-System (cat. no. 11681834001; Merck KGaA) as polymerase. Following an initial denaturation at 94°C for 1 min, 35 cycles of 12 sec at 94°C, 15 sec at 56°C and 30-60 sec elongation at 68°C were run followed by an additional 3 min at 68°C. Primers and GenBank accession numbers are listed in Table I. A 1400 fragment of trpm7 from clone cb495 provided by ZIRC was subcloned via BamHI and EcoRI into pCS2+. In situ probes were then generated by the in vitro transcription of linearized plasmids using SP6, T3 or T7 polymerase (NEB cat. nos. M0207, M0378 and M0251), respectively and the RNA DIG labeling mix (cat. no. 11277073910; Merck KGaA).

A plasmid encoding human GRK4 ΔNLS, in which amino acids 224-226 were mutated from KKR to LLE (Table II), was generated by site-directed mutagenesis from pCS2+Flag-hGRK4 (4).

Embryos were fixed in 4% buffered paraformaldehyde (cat. no. P6148; Merck KGaA) at 4°C overnight, dehydrated in a graded methanol series and stored at −20°C for at least 1 h before processing by in situ hybridization. In situ hybridization was performed with DIG-labeled in situ probes essentially as described in great detail in the study by Thisse and Thisse (11). Plasmids not listed above were kindly provided by Professor Jeroen Bakkers (tbx2b) (12) and ZIRC (slc20a1a/clone cb1011).

293T cells were cultured in high-glucose DMEM containing 10% heat-inactivated fetal bovine serum and antibiotics (all from Gibco; Thermo Fisher Scientific, Inc.). Transfections with a plasmid encoding C-terminally GFP-tagged GRK4 variants were performed in six-well plates using Lipofectamine 3000 according to the manufacturer's recommendations (cat. no. L3000008; Thermo Fisher Scientific, Inc.). After 1 day, the cells were transferred to poly-D-lysine-coated coverslips. At 2 days following transfection, the cells were processed for immunofluorescence.

The cells were routinely and regularly examined for mycoplasma contamination using the LookOut Mycoplasma PCR Detection kit and according to the manufacturer's recommendations (cat. no. MP0035; Merck KGaA). Cells were authenticated and matched to the Cellosaurus databases.

For the visualization of cilia and apical cell borders of the proximal tubule, embryos at 48 h post-fertilization (hpf) were fixed in 4% buffered paraformaldehyde at 4°C overnight. Subsequently, the staining protocol provided by Jaffe et al (13) was followed. The primary antibodies used were mouse anti-acetylated tubulin (1:500, cat. no. T6793; Merck KGaA) and rabbit anti-PKCz (1:500, cat. no. sc-216; Santa Cruz Biotechnology, Inc.). Alexa-coupled secondary antibodies were used for detection (1:1,000, cat. nos. A10037 and A11008; Invitrogen, Thermo Fisher Scientific, Inc.). Following immunostaining, the tail of the embryos was cut off using forceps and embedded between two coverslips using Vectashield with DAPI (cat. no. VEC-H-1200; Biozol). The same staining protocol was used to detect p-Stat3 (1:100, clone no. PS3/1; MBL International Corporation) and total STAT3 (1:500, cat. no. 10253-2-AP, Proteintech Group, Inc.; or 1:500, cat. no. ab226942, Abcam).

Transiently transfected 293T cells expressing GFP-tagged GRK4 variants were fixed for 10 min using 4% buffered paraformaldehyde and mounted using Vectashield with DAPI.

Living embryos were anesthetized by immersion in 0.02% (m/v) buffered tricaine and imaged with a Leica M125 stereomicroscope and a Leica IC80 HD camera or a Flexacam C1 color camera. The same setup was used for imaging embryos processed by in situ hybridization. A Leica M205FA and a K8 sCMOS camera were used for epifluorescence analyses. Confocal z-stacks were acquired with a Leica SP5 or a Leica Stellaris 5 system, respectively and maximum projections were generated.

FIJI (14) was used to perform all measurements: The line tool was used to measure the length of distinct pronephric tubule parts in in situ hybridizations and for the measurement of the proximal tubule diameter in maximum projections of z-stacks. The freehand selection tool was used to measure the area of the glomerulus in in situ hybridizations. Cilia length was determined using the Simple Neurite Tracer within the neuroanatomy tool set. In detail, the beginning and the end of a given cilium was selected in 3D and the whole length was traced and measured.

To assess p-Stat3 levels, all embryos processed by antibody staining were imaged and images were compared side-by-side for staining being stronger than the average control-injected embryo in each experiment.

Statistical analyses were performed with the aid of Prism 10 (Dotmatics). First, data were analyzed for normal distribution using a Shapiro-Wilk-test before the respective parametric or non-parametric test was applied. An α level of 0.05 was considered to indicate a statistically significant difference. The specific tests applied, as well as P-values and experimental repeats are described in the respective figure legends. In detail, the following tests were applied for data with normal distribution: A two-tailed t-test with Welch's correction, one-way ANOVA with Sidak's multiple comparison test and two-way ANOVA with Sidak's multiple comparison post-test. Data not following a normal distribution were analyzed using the following tests: A two-tailed Mann-Whitney test or the Kruskal-Wallis test with Dunn's multiple comparison post-test. All Fisher's exact tests were two-sided and the α levels were Bonferroni-corrected for the number of comparisons. Values of P<0.025 were considered to indicate statistically significant differences in the case of two comparisons.

Zebrafish serve as a simple, yet elegant model of human kidney development, as they develop only two nephrons, one on each side of the midline. The zebrafish pronephros furthermore has a very similar composition as the mammalian nephron and displays a comparable functionality during embryonic development (15). In the present study, in order to assess Grk4-dependent kidney development in greater detail, two models were used, which were previously validated and analyzed by the authors: A knockdown approach by an antisense MO injection into the yolk of fertilized eggs and a mutant, in which two amino acids at position 194 and 195 are deleted producing a hypomorphic allele (4). Whole mount in situ hybridization for commonly used marker genes was performed to visualize distinct regions of the pronephros (Fig. 1A): Using a probe against wt1a revealed an enlargement of the glomerular area in Grk4-deficient embryos compared to control embryos (Fig. 1B), whereas the length of the proximal convoluted tubule, as detected by slc20a1a expression, was reduced (Fig. 1C). By contrast, the depletion of Grk4 increased the length of the proximal straight tubule, which corresponds to the proximal tubule in the mammalian nephron (Fig. 1D). A stronger signal of rfx2 (Fig. 1E) was also observed at the expense of the distal late tubule visualized by tbx2b expression (Fig. 1F). Taken together, these results suggest that Grk4 is involved in pronephros patterning.

In mammals, GRK4 modulates renal sodium reabsorption through dopamine receptor desensitization in the proximal tubule (16,17). As an extension of the proximal straight tubule was observed in Grk4-knockdown embryos, the authors wished to determine whether Nhe3a, the sodium transport protein, controlled by dopamine receptors in the mammalian proximal tubule (18), was affected by Grk4 LOF. Hence, in situ hybridization was performed at several developmental stages for nhe3a, the close zebrafish homolog of mammalian NHE3. It was observed that nhe3a is maternally provided and ubiquitously expressed, as shown for the 256 cell stage (Fig. 2A), but appears to be downregulated when embryonic gene expression begins; thus, at the shield stage, almost no expression remains. Towards the end of gastrulation (tailbud stage), and with the beginning of somitogenesis (as shown for the 16 somites stage), nhe3a begins being expressed again. From these stages on, when the pronephric mesoderm has formed, nhe3a can be exclusively detected in pronephric cells with the highest abundance in a region corresponding to the proximal straight tubule in wild-type embryos and also in the distal late segment or the cloaca, respectively (Fig. 2A). In embryos lacking functional Grk4, however, the expression domain of nhe3a was elevated at the 20 somites stage (Fig. 2B). At 48 hpf, when the pronephros has become functional, a strong nhe3a expression and a larger expression domain along the pronephric tubule could be detected (Fig. 2C). Hence, Grk4 appears to regulate Nhe3a expression during kidney development.

To examine whether the observed nhe3a upregulation contributes to the pronephros phenotypes evoked by Grk4 LOF, experiments using the specific NHE3 inhibitor, tenapanor, were performed. Zebrafish eggs were injected with either a control or the Grk4 MO and subsequently treated during organogenesis with 5 μM tenapanor or the vehicle. Of note, tenapanor partially, yet significantly, rescued the disrupted kidney patterning, as shown for the expression domains of wt1a, trpm7 and rfx2 (Fig. 3A-C). Furthermore, significantly fewer Grk4-knockdown embryos developing cystic glomeruli were observed compared to vehicle-treated counterparts (Fig. 3D), with less dilated pronephric tubules (Fig. 3E) and the significant shortening of pronephric cilia (Fig. 3F). In addition, hydrocephalus, which develops as a consequence of the pronephric dysfunction (4), occurred to a lesser extent when Grk4 knockdown embryos were treated with tenapanor (Fig. 3G). By contrast, in the control-injected embryos, tenapanor had no or only very minor effects on the assessed phenotypes. Taken together, these data suggest a Grk4-Nhe3a axis contributing to the Grk4 LOF phenotype in zebrafish embryos.

Subsequently, the present study assessed whether the observed phenotypes were coupled to dopaminergic receptor activity as Grk4 LOF may prevent receptor desensitization and cause elevated dopaminergic receptor signaling. Fertilized eggs were injected with control or Grk4 MO and treated with the vehicle or 10 μM of the dopamine receptor antagonist, SCH39166, from the tailbud stage on until analysis at 48 hpf. Notably, no significant change was detected in the expression domain of nhe3a in either the control-injected or Grk4 MO-injected embryos (Fig. 4A). SCH39166 also failed to reduce the percentage of embryos developing glomerular cysts (Fig. 4B). To further exclude unleashed dopaminergic receptor signaling as a driver of aberrant pronephric development, wild-type zebrafish embryos were treated with dopamine. Yet again, the length of the nhe3a⁺ tubule was not altered (Fig. 4C and D). In addition, dopaminergic receptor stimulation using fenoldopam did not induce glomerular cysts or hydrocephalus (Fig. 4E and F). Hence, defects in pronephric development and function in Grk4-depleted embryos are unlikely due to abnormal dopaminergic receptor signaling.

Previously and consistent with the data shown herein for dopaminergic receptor signaling, the authors found that the function of GRK4 during pronephros development does not depend on its kinase activity (4). In addition to the kinase domain, however, GRK4 also contains a nuclear localization sequence (NLS), which entails a low, yet reproducible GRK4 expression in the nucleus of transfected 293T cells (Fig. 5A). Mutation of the NLS (GRK4 ΔNLS) results in the exclusion of GRK4 from the nucleus and sole cytosolic distribution (Fig. 5A). Of note, rescue experiments with GRK4 ΔNLS normalized the expression of nhe3a and reduced both hydrocephalus, as well as cyst formation in Grk4 LOF embryos, suggesting an improvement in kidney function (Fig. 5B-E). Therefore, it can be concluded that the location of Grk4 action resides in the cytosol and does not depend on the nuclear localization of Grk4.

In addition to dopaminergic receptors, NHE3 has been reported to be controlled by other means. The transcriptional activity of STAT3 directs the RNA synthesis of NHE3 in kidney epithelial cells by binding to its promoter (8). Moreover, STAT3 activity based on the phosphorylation of tyrosine 705 (corresponding to Y708 in zebrafish Stat3) is higher in cystic than in the healthy kidneys of patients diagnosed with autosomal polycystic kidney disease (19). Mutation of that same tyrosine 705 to phenylalanine abrogates the transcriptional activity of human STAT3 (20). Hence, the present study assessed STAT3 activity using an antibody specific for phosphorylated Stat3 in zebrafish and observed indeed a brighter fluorescence signal in Grk4-depleted embryos than in control embryos, while total Stat3 levels appeared to be the same (Fig. 6A and B). Co-injection of the Grk4 MO with RNA encoding for dominant negative STAT3 (dnSTAT3) revealed a significant reduction of the expression domain of nhe3a compared to embryos injected only with the Grk4 MO (Fig. 6C and D) and consistent with the transcriptional control of NHE3 by STAT3, many more embryos displaying a very weak nhe3a signal were counted (Fig. 6E). Last, but not least, the expression of dnSTAT3 partially, yet significantly reduced the percentage of embryos displaying kidney dysfunction as detected by hydrocephalus development (Fig. 6F) and glomerular cysts (Fig. 6G). Taken together, these findings suggest that the loss of Grk4 induces Stat3 activation in zebrafish embryos, which in turn results in higher rates of nhe3a transcription and subsequently the ciliopathy-like malformation of the developing pronephros, which was previously reported by the authors (4) and shown again herein.

The previously identified regulatory action of GRK4 on NHE3 subcellular localization through the phosphorylation of dopaminergic receptors has ranked GRK4 as a potential therapeutic target in hypertension. The data presented herein classify GRK4 additionally as a candidate gene in polycystic kidney diseases, such as nephronophtisis or autosomal dominant polycystic kidney disease. It further suggests that the mechanistic impact of GRK4 on NHE3 is considerably beyond G protein-coupled receptors and places GRK4 in the center of a much more complex signaling network, which includes STAT3. STAT3 is involved in a rather large variety of pathologies ranging from immune disorders and inflammation over cardiovascular diseases and neurodegenerative conditions to several types of cancer (37). Hence, potential therapeutic strategies targeting GRK4 may require careful design in order to avoid interference with the full GRK4 signaling network.

The present study has certain limitations, which should be mentioned. The data presented herein solely rely on zebrafish embryos. While the authors demonstrated in a previous study that the abrogation of GRK4 in human fibroblasts, as well as kidney organoids derived from murine inner medullary collecting duct cells results in the same cilia defects as in zebrafish embryos (4), further studies using NHE3-expressiong mammalian cells or animal models are warranted to test the conservation of the newly identified Grk4-Stat3-Nhe3 axis across species.