1. Biochemistry and Chemical Biology
Download icon

Gq activity- and β-arrestin-1 scaffolding-mediated ADGRG2/CFTR coupling are required for male fertility

  1. Dao-Lai Zhang
  2. Yu-Jing Sun
  3. Ming-Liang Ma
  4. Yi-jing Wang
  5. Hui Lin
  6. Rui-Rui Li
  7. Zong-Lai Liang
  8. Yuan Gao
  9. Zhao Yang
  10. Dong-Fang He
  11. Amy Lin
  12. Hui Mo
  13. Yu-Jing Lu
  14. Meng-Jing Li
  15. Wei Kong
  16. Ka Young Chung
  17. Fan Yi
  18. Jian-Yuan Li
  19. Ying-Ying Qin
  20. Jingxin Li
  21. Alex R B Thomsen
  22. Alem W Kahsai
  23. Zi-Jiang Chen
  24. Zhi-Gang Xu
  25. Mingyao Liu
  26. Dali Li  Is a corresponding author
  27. Xiao Yu  Is a corresponding author
  28. Jin-Peng Sun  Is a corresponding author
  1. Shandong University School of Medicine, China
  2. Binzhou Medical University, China
  3. School of Medicine, Duke University, United States
  4. School of Basic Medical Sciences, Peking University, China
  5. Sungkyunkwan University, Korea
  6. National Health and Family Planning Commission, China
  7. Shandong University, China
  8. Shandong University School of Life Sciences, China
  9. Institute of Biomedical Sciences, East China Normal University, China
  10. Texas A&M University Health Science Center, United States
Research Article
  • Cited 0
  • Views 813
  • Annotations
Cite as: eLife 2018;7:e33432 doi: 10.7554/eLife.33432

Abstract

Luminal fluid reabsorption plays a fundamental role in male fertility. We demonstrated that the ubiquitous GPCR signaling proteins Gq and β-arrestin-1 are essential for fluid reabsorption because they mediate coupling between an orphan receptor ADGRG2 (GPR64) and the ion channel CFTR. A reduction in protein level or deficiency of ADGRG2, Gq or β-arrestin-1 in a mouse model led to an imbalance in pH homeostasis in the efferent ductules due to decreased constitutive CFTR currents. Efferent ductule dysfunction was rescued by the specific activation of another GPCR, AGTR2. Further mechanistic analysis revealed that β-arrestin-1 acts as a scaffold for ADGRG2/CFTR complex formation in apical membranes, whereas specific residues of ADGRG2 confer coupling specificity for different G protein subtypes, this specificity is critical for male fertility. Therefore, manipulation of the signaling components of the ADGRG2-Gq/β-arrestin-1/CFTR complex by small molecules may be an effective therapeutic strategy for male infertility.

https://doi.org/10.7554/eLife.33432.001

Introduction

Male infertility is transforming from a personal issue to a public health problem because approximately 15% of reproductive-age couples are infertile, and male infertility accounts for approximately 50% of this sterility (Hamada et al., 2012; Jodar et al., 2015). The unique structure of the male reproductive system increases the difficulty of determining the working mechanisms. Among male reproductive system, the efferent ductules of the male testis play important roles during sperm transportation and maturation by reabsorbing the fluid of the rete testis and maintaining the homeostasis of water and ion metabolism (Hess et al., 1997). Whereas a dysfunction of the efferent ductule reabsorption capacity caused by a developmental defect that produces improper signaling results in epididymal obstructions and abnormal spermiostasis, which ultimately lead to infertility in both humans and other mammals (Hendry et al., 1990; Nistal et al., 1999), manipulating the reabsorption function in the efferent ductules could be developed into a useful contraceptive method for males (Gottwald et al., 2006).

Receptors play key roles in the regulation of fluid reabsorption in tissues such as the proximal tubules and alveoli (Haithcock et al., 1999; Thomson et al., 2006). In contrast, only a few receptor functions in the efferent ductules have been characterized. Nuclear estrogen receptor α (ERα) must be activated for male reproductive tract development and reabsorption function maintenance to occur (Hess et al., 1997). However, the mechanism by which fluid reabsorption is regulated by cell surface receptors in the efferent ductules is only beginning to be appreciated (Shum et al., 2008). Knockout of an orphan G-protein-coupled receptor (GPCR), ADGRG2 (adhesion G-protein-coupled receptor G2), results in male infertility due to dysregulated fluid reabsorption in the efferent ductules, suggesting an active role for this cell surface receptor in regulating these processes (Davies et al., 2004). However, how ADGRG2 regulates water-ion homeostasis and fluid reabsorption remains elusive.

ADGRG2 belongs to the seven transmembrane receptor superfamily (Hamann et al., 2015), which regulates approximately 80% of signal transduction across the plasma membrane and accounts for 30% of current clinical prescription drug targets. Five different types of G proteins and arrestins act as signaling hubs downstream of these GPCRs, mediating most of their functions (Alvarez-Curto et al., 2016; Cahill et al., 2017; Dong et al., 2017; Li et al., 2018; Liu et al., 2017; Nuber et al., 2016; Thomsen et al., 2016; Yang et al., 2015). In the efferent ductules, it remains unclear how G proteins and their parallel signaling molecules, the arrestins, regulate reabsorption as well as fertility.

Here, we developed a new labeling method utilizing specific red fluorescent protein (RFP) expression driven by the ADGRG2 promoter, which enabled a detailed mechanistic study of efferent ductule functions. By exploiting Adgrg2-/Y, Gnaq+/-, Arrb1-/- and Arrb2-/- knockout mouse models, together with the combination of pharmacological interventions and electrophysiological approaches, we have identified the importance of the ubiquitous Gq protein and β-arrestin-1, which confer the ADGRG2 constitutive activity to a basic cystic fibrosis transmembrane conductance regulator (CFTR) current, in fluid reabsorption in the efferent ductules. Both specific Gq activity- and β-arrestin-1 scaffolding-mediated ADGRG2/CFTR coupling are required for male fertility and Cl-/acid-base homeostasis in the efferent ductules. Our results not only reveal how fluid reabsorption in the male efferent ductules is precisely controlled by a specific subcellular signaling compartment encompassing ADGRG2, CFTR, β-arrestin-1 and Gq in non-ciliated cells but also provide a foundation for the development of new therapeutic approaches to control male fertility.

Results

Gq activity is required for fluid reabsorption and male fertility

Previous studies have found that knockout of the orphan receptor ADGRG2 causes infertility and fluid reabsorption dysfunction in the efferent ductules, indicating important roles for GPCR signaling in male reproductive functions. Downstream of GPCRs, there are 16 Gα proteins that mediate diverse GPCR functions (DeVree et al., 2016). However, the expression of these G protein subtypes and their functions in the efferent ductules have not been investigated. Here, we show that Gs is more enriched, while G11 and Gi3 have expression levels in the efferent ductules similar to those in brain tissue, whereas all other 11 tested G protein subtypes have detectable expression levels in the efferent ductules (Figure 1A). ADGRG2 localizes in cells devoid of acetylated-tubulin staining, suggesting that it is specifically expressed in non-ciliated cells (Figure 1B and Figure 1—figure supplement 1A–C). We next used the promoter region of ADGRG2 to direct the expression of the fluorescent protein RFP, which enabled the specific labeling of ADGRG2-expressing non-ciliated cells in the efferent ductules (Figure 1C and Figure 1—figure supplement 2A–C). After fluorescence-activated cell sorting (FACS), quantitative RT-PCR (qRT-PCR) results indicated that ADGRG2-expressing non-ciliated cells have expression levels of Golf, Gi2, Gq, G11, and G13 that are higher than those in brain tissue and expression levels of Gs, G12 and Gz that are similar to those in brain tissue (Figure 1D).

Figure 1 with 2 supplements see all
The expression of G protein subtypes in the efferent ductules and ADGRG2 promoter-labeled non-ciliated cells.

(A) qRT-PCR analysis of mRNA transcription profiles of G proteins in brain tissues and the efferent ductules of WT (n = 3) male mice. Expression levels were normalized to GAPDH levels. *p<0.05, **p<0.01, ***p<0.001, efferent ductules compared with brain tissue. (B) Co-localization analysis of ADGRG2 (red fluorescence) and acetylated-tubulin (green fluorescence) in the efferent ductules of WT mice. Scale bars, 50 μm. (C) Co-localization of ADGRG2 (green fluorescence) and RFP (red fluorescence) in the same cells of male murine efferent ductules infected with the ADGRG2 promoter RFP adenovirus in WT mice. Scale bars, 50 μm. (D) qRT-PCR analysis of mRNA transcription profiles of G protein subtypes in brain tissues and isolated ADGRG2 promoter-labeled non-ciliated cells derived from the efferent ductules of WT (n = 3) male mice. Expression levels were normalized to GAPDH levels. *p<0.05, **p<0.01, ***p<0.001, ADGRG2 promoter-labeled efferent ductule cells compared with brain tissues. n.s., no significant difference. At least three independent biological replicates were performed for Figure 1A and D.

https://doi.org/10.7554/eLife.33432.002

We next investigated the contribution of different G protein subtype signaling pathways to fluid reabsorption in the efferent ductules using specific pharmacological interventions and knockout models. An ADGRG2 knockout mouse was produced by introducing an 11-nucleotide sequence into the first exon of the ADGRG2 gene (Figure 2—figure supplement 1), thereby creating a positive control for fluid reabsorption dysfunction in the efferent ductules (Davies et al., 2004). The wild-type (WT) mice did not show size alterations due to the normal reabsorption of luminal fluid, but the ligated efferent ductules derived from the ADGRG2 knockout mice displayed a 40% increase in luminal area after 72 hr of in vitro culture (Figure 2A). Application of the Gi inhibitor pertussis toxin (PTX) or the MEK-ERK signaling inhibitor U0126 did not have a significant effect on the efferent ductules (Figure 2B and C). In contrast, a 50% reduction in Gq protein levels in Gnaq+/- mice or the application of the protein kinase C(PKC) inhibitor Ro 31–8220 significantly impaired fluid reabsorption in the efferent ductules, which mimicked the phenotype of the ductules derived from Adgrg2-/Y mice (Figure 2A and D–E and Figure 2—figure supplement 1F–G). The contribution of Gs-PKA (protein kinase A) signaling to fluid reabsorption of the efferent ductules is confounded. While the application of the Gs inhibitor NF449 or the PKA inhibitors PKI14-22 or H89 to the efferent ductules derived from WT mice slightly increased the volume of the efferent ductules (Figure 2F–H), cAMP regulators, such as the adenyl cyclase activator forskolin (FSK) and the phosphodiesterase (PDE) inhibitor 3-isobutyl-1-methylxanthine(IBMX), increased the volume of the efferent ductules in an acute manner in both Adgrg2-/Y mice and WT littermates (Figure 2—figure supplement 2A–B). These results suggested that Gs-PKA signaling is finely tuned in the efferent ductules to maintain its fluid reabsorption function because both increasing and decreasing its activity caused detrimental effects. In conclusion, Gi and MEK-ERK signaling exerted no significant effects, whereas Gq-PKC signaling was required for efficient fluid reabsorption in the efferent ductules.

Figure 2 with 2 supplements see all
Gq activity is required for fluid reabsorption.

(A) Images of cultured ligated efferent ductules derived from WT male mice, Adgrg2-/Y mice and Gnaq+/- male mice. Ductule segments were selected by examination of the ciliary beat, which is a marker of cell integrity. Ductule pieces from Adgrg2-/Y, Gnaq+/- or WT mice were ligated, microdissected and cultured for up to 72 hr. Scale bars, 200 μm. (B–C, E–H) Effects of pharmacological intervention on the diameters of ligated efferent ductules derived from WT or Adgrg2-/Y mice. (B) PTX (100 ng/ml), a Gi inhibitory protein. WT (n = 9) or Adgrg2-/Y (n = 8); (C) U0126 (10 μM), a MEK inhibitor (ERK pathway blockade), WT (n = 12) or Adgrg2-/Y (n = 12). (E) Ro 31–8220 (500 nM), a protein kinase C (PKC) inhibitor, WT (n = 12) or Adgrg2-/Y (n = 10); (F) NF449 (1 μM), a Gs inhibitor, WT (n = 9) or Adgrg2-/Y (n = 9); (G) PKI14-22 (300 nM), a PKA inhibitor, WT (n = 9) or Adgrg2-/Y (n = 9); (H) H89 (500 nM), a non-selective PKA inhibitor, WT (n = 9) or Adgrg2-/Y (n = 9). (D) Diameters of the luminal ductules derived from WT (n = 27) mice remained unchanged over 72 hr, whereas the lumens of the ductules derived from Adgrg2-/Y (n = 21) mice and Gnaq+/- (n = 16) mice were significantly increased, indicating fluid reabsorption dysfunction. (2B-2H) *p<0.05, **p<0.01, ***p<0.001, Adgrg2-/Y mice or Gnaq+/- mice were compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001, treatment with selective inhibitors or stimulators was compared with control vehicles. n.s., no significant difference. At least three independent biological replicates were performed for Figure 2B–H.

https://doi.org/10.7554/eLife.33432.005

The efferent ductules of the Gnaq+/- animals consistently showed the accumulation of obstructed spermatozoa compared with those of WT mice, whereas the lumen of the initial segment and caput region in Gnaq+/- mice contained significantly reduced sperm levels (Figure 3A–D). Sperm numbers prepared from the caudal epididymis and the birth rate of the Gnaq+/- mice were also significantly decreased compared with their WT littermates (Figure 3E–G). Taken together, these data demonstrated that among different G protein subtypes, Gq activity is required for fluid reabsorption and male fertility.

Gq expression is required for sperm transportation and male fertility.

(A) Representative hematoxylin and eosin staining of WT, Adgrg2-/Y or Gnaq+/- mice. Scale bars, 200 μm. (B–D) Corresponding bar graphs demonstrating the accumulation of spermatozoa according to the hematoxylin and eosin staining of WT (n = 8), Adgrg2-/Y (n = 9) or Gnaq+/- (n = 9) mice. ED: efferent ductules; IS: epididymal initial segment; CA: caput epididymis. (E) Representative photographs of caudal sperm preparation from the caudal epididymis of WT, Adgrg2-/Y or Gnaq+/- mice. Scale bars, 50 μm. (F) Bar graph depicting the quantitative analysis of the number of sperm shown in (Figure 1E) of WT (n = 8), Adgrg2-/Y (n = 10) or Gnaq+/- (n = 10) mice. (G) Line graph depicting the fertility of Gnaq+/- (n = 6) and WT (n = 6) male mice at various ages, as measured by the median number of embryos. (3B-D and 3 F-G): *p<0.05, **p<0.01, ***p<0.001, Adgrg2-/Y mice or Gnaq+/- mice were compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001. Gnaq+/- mice were compared with Adgrg2-/Y mice. n.s., no significant difference. At least three independent biological replicates were performed for Figure 3B–D and and F–G.

https://doi.org/10.7554/eLife.33432.008

ADGRG2 and CFTR coupling in the efferent ductules and its function in fluid reabsorption

Membrane proteins, including bicarbonate and chloride transporters, sodium/potassium pumps and specific ion channels, are potential osmotic drivers for fluid secretion and reabsorption in the efferent ductules (Estévez et al., 2001; Harvey, 1992; Liu et al., 2015; Park et al., 2001; Russell, 2000; Xiao et al., 2012; Xiao et al., 2011; Zhou et al., 2001). Therefore, we examined the expression levels of these membrane proteins in the efferent ductules and ADGRG2 promoter-labeled ductule cells (Figure 4A and Figure 4—figure supplement 1A). Specifically, Na+-K+-Cl- cotransporter (NKCC), down-regulated in adenoma (DRA), CFTR, solute carrier family 26 member 9(SLC26a9), Na+/H+ exchanger 3(NHE3) and the L-type voltage dependent calcium channel Cav1.3 levels were readily measured in ADGRG2 promoter-labeled non-ciliated ductule cells; Na+/H+ exchanger 1(NHE1), carbonic anhydrase II(CAII), Short transient receptor potential channel 3(TRPC3), chloride channel accessory 1(CLCA1) and Cav1.2 had lower but detectable expression levels, whereas anoctamin-1 (ANO1), V-ATPase and Cav2.2 demonstrated very little expression (Figure 4A and Figure 4—figure supplement 1A). Notably, we used the ADGRG2 promoter to label the non-ciliated cells, as the ADGRG2 receptor is specifically expressed on the apical membrane of these cells in efferent ductules (Figure 1B–C and Figure 1—figure supplements 12). A higher expression level of a particular membrane protein, such as CFTR, in the ADGRG2 promoter-labeled cells indicated that these membrane proteins are enriched in non-ciliated cells in efferent ductules but does not indicate that the expression of these proteins is dependent on ADGRG2. For example, the CFTR expression level in ADGRG2 promoter-labeled efferent ductule cells derived from Adgrg2-/Y mice did not differ significantly from that in the cells derived from their WT littermates (Figure 4—figure supplement 1D).

Figure 4 with 1 supplement see all
Inhibition of CFTR activity in the efferent ductules pheno-copied the activity in Adgrg2-/Y mice.

(A) qRT-PCR analysis of the mRNA transcription profiles of potential osmotic drivers including selective ion channels and transporters in ADGRG2 promoter-labeled cells, non-ADGRG2 promoter-labeled cells and brain tissues of WT (n = 3) male mice. Expression levels were normalized to GAPDH levels. *p<0.05, **p<0.01, ***p<0.001, ADGRG2 promoter-labeled cells were compared with brain tissues. #p<0.05, ##p<0.01, ###p<0.001, non-ADGRG2 promoter-labeled cells were compared with brain tissues. (B–M) Effects of different channel blockers on the diameters of luminal ductules derived from WT or Adgrg2-/Y mice. (B) Bumetanide (10 μM), an NKCC blocker, WT (n = 9) or Adgrg2-/Y (n = 10); (C) Ani9 (150 nM), an ANO1 inhibitor, WT (n = 9) or Adgrg2-/Y (n = 9); (D) NFA (20 μM), a CaCC inhibitor, WT (n = 9) or Adgrg2-/Y (n = 10); (E) ruthenium red (10 μM), a non-specific TRP channel blocker, WT (n = 12) or Adgrg2-/Y (n = 12); (F) SKF96365 (10 μM), a TRPC channel inhibitor, WT (n = 12) or Adgrg2-/Y (n = 9); (G) nicardipine (20 μM), an L-type calcium channel blocker, WT (n = 12) or Adgrg2-/Y (n = 12); (H) EGTA (5 mM), an extracellular calcium chelator, WT (n = 9) or Adgrg2-/Y (n = 9); (I) DIDS (20 μM), a chloride-bicarbonate exchanger blocker, WT (n = 9) or Adgrg2-/Y (n = 10); (J) GlyH-101 (25 μM), a non-specific CFTR inhibitor, WT (n = 17) or Adgrg2-/Y (n = 15); (K) CFTRinh-172(10 μM), a specific CFTR inhibitor, WT (n = 12) or Adgrg2-/Y (n = 10). (L) Effects of angiotensin II (100 nM, an angiotensin receptor agonist) and PD123319 (1 μM, an AT2 receptor antagonist) on the diameters of luminal ductules derived from WT or Adgrg2-/Y mice (n ≥ 12). (M) Effects of angiotensin II (100 nM) and candesartan (1 μM, an AT1 receptor antagonist) on the diameters of luminal ductules derived from WT or Adgrg2-/Y mice (n ≥ 12). Application of GlyH-101 and CFTRinh-172 to ligated ductules derived from WT mice recapitulated the phenotype of the ductules derived from Adgrg2-/Y mice. (4A-M)*p<0.05, **p<0.01, ***p<0.001; Adgrg2-/Y mice compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001. Treatment with selective inhibitors or stimulators was compared with control vehicles. n.s., no significant difference. At least three independent biological replicates were performed for Figure 4A–M.

https://doi.org/10.7554/eLife.33432.009

We next used a panel of pharmacological blockers to examine whether the inappropriate regulation of these membrane protein functions was involved in the ADGRG2- or Gq-mediated regulation of fluid reabsorption in the efferent ductules. Importantly, application of the NKCC blocker bumetanide, the ANO1 inhibitor Ani9, the calcium-dependent chloride channel (CaCC) inhibitor niflumic acid (NFA), TRP channel inhibitors including ruthenium red, SKF96365 and LaCl3, the L-type calcium channel blocker nicardipine or chelating extracellular calcium with EGTA showed no significant effects on fluid reabsorption in the efferent ductules in ligation experiments (Figure 4B–H and Figure 4—figure supplement 1B). Application of 4,4'-diisothiocyano-2,2'-stilbenedisulfonic acid (DIDS) to block the chloride-bicarbonate exchanger exerted a small effect only after 60 hr (Figure 4I), and the application of amiloride to inhibit sodium/hydrogen antiporter NHE1 activity exerted an acute effect on fluid reabsorption (Figure 4—figure supplement 1C), an outcome different from that observed in Adgrg2-/Y or Gnaq+/- mice (Figure 2D). In contrast, blocking CFTR activity either with GlyH-101 or CFTRinh-172 had significant effects on fluid reabsorption in the efferent ductules and pheno-copied the Adgrg2-/Y mice (Figure 4J–K). Collectively, the phenotype caused by inactivating ADGRG2 and administering a CFTR channel blocker in WT mice suggested that CFTR and ADGRG2 may be functionally connected to the regulation of fluid reabsorption.

CFTR is the key regulator of pH homeostasis and chloride in the reproductive and renal systems and has important functions in fluid reabsorption (Chen et al., 2012). Therefore, we measured the pH value of the efferent ductules. The pH homeostasis was impaired in Adgrg2-/Y mice, with a pH value of 7.6 for the inner solution in the efferent ductules, compared to a pH of 7.2 in WT littermates (Figure 5A and Figure 5—figure supplement 2). This dysfunction was not caused by decreased CFTR expression because the mRNA levels of CFTR in the Adgrg2-/Y mice were not reduced compared with those of their WT littermates (Figure 5B and Figure 5—figure supplement 1). Moreover, application of the CFTR inhibitor CFTRinh-172 increased the pH value of the efferent ductules in WT mice by approximately 0.3 but did not have a significant effect in Adgrg2-/Y mice, suggesting that CFTR dysfunction in Adgrg2-/Y mice influences pH homeostasis (Figure 5A–B). Importantly, the pH imbalance in Adgrg2-/Y mice was rescued by bicarb-free media or application of the carbonic anhydrase inhibitor acetazolamide (Figure 5—figure supplement 2B–C).

Figure 5 with 4 supplements see all
Functional coupling and co-localization of CFTR and ADGRG2 on the apical membrane in the efferent ductules.

(A) Intracellular pH (pHi) of the ligated efferent ductules from WT (n = 9) mice and Adgrg2-/Y (n = 9) mice were measured by carboxy-SNARF (5 μM), with or without incubation with the CFTR inhibitor CFTRinh-172. (B) qRT-PCR analysis of CFTR levels in the efferent ductules of WT (n = 3) or Adgrg2-/Y (n = 3) mice. (C) Co-localization of ADGRG2 (red fluorescence) and CFTR (sc-8909, Santa Cruz, green fluorescence) in the male efferent ductules of WT mice. Scale bars, 50 μm. (D) Analysis of ADGRG2 and CFTR fluorescence intensities by Pearson’s correlation analysis. The Pearson's correlation coefficient was 0.76. (E) Immunofluorescence staining of ADGRG2 (red fluorescence) and CFTR (sc-8909, Santa Cruz, green fluorescence) in the efferent ductules of Adgrg2-/Y mice. Scale bars, 50 μm. (F) Co-localization of ADGRG2 (red fluorescence) and ezrin (green fluorescence) in the male efferent ductules of WT mice. Scale bars, 50 μm. (G) Analysis of ADGRG2 and ezrin fluorescence intensities by Pearson’s correlation analysis. The Pearson's correlation coefficient was 0.69. (H) ADGRG2 was immunoprecipitated with an anti-ADGRG2 antibody from the male efferent ductules of WT mice or Adgrg2-/Y mice, and co-precipitated CFTR, Gs, Gq, β-arrestin-1, β-arrestin-2 and Gi-1/2/3 levels were examined by using specific corresponding antibodies (CFTR antibody:20738–1-AP, Proteintech). (5A-5B) *p<0.05, **p<0.01, ***p<0.001, Adgrg2-/Y mice compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001. Treatment with selective inhibitors or stimulators was compared with control vehicles. n.s., no significant difference. At least three independent biological replicates were performed for Figure 5A–B.

https://doi.org/10.7554/eLife.33432.011

In particular, unambiguous co-localization of ADGRG2 and CFTR on the apical membrane was detected (Figure 5C–G and Figure 5—figure supplement 3) and ADGRG2 was associated with CFTR in co-immunoprecipitation assays (Figure 5H and Figure 5—figure supplement 4). Taken together, these results suggest a complex formation and functional coupling of ADGRG2 and CFTR in the non-ciliated cells of the efferent ductules.

The outwardly rectifying whole-cell Cl- current (IADGRG2-ED) of ADGRG2 promoter-labeled efferent ductule cells

We then performed whole-cell Cl- recording of primary ADGRG2 promoter-labeled efferent ductule cells derived from WT and Adgrg2-/Y mice with normal Cl- concentrations or by substituting Cl- with gluconate (Gluc-) in the bath solution (Figure 6A–E and Table 1). Patch-clamp recording on ADGRG2 promoter-labeled non-ciliated cells derived from WT mice revealed a reversible whole-cell Cl- current (IADGRG2-ED), which was significantly diminished in response to substitution of the bath Cl- solution with Gluc- (148.5 mM Cl- was replaced by 48.5 mM Cl- and 100 mM Gluc-) (Figure 6A–B). This whole-cell Cl- current (IADGRG2-ED) was recovered once Gluc- was substituted with Cl- solution (Figure 6A–B). Further I-V analysis identified an outwardly rectifying whole-cell Cl- current (IADGRG2-ED) of wild type mice, which was significantly reduced in response to Gluc- substitution (Figure 6C–E and Table 1). The change in the reversal potential (Erev) with Gluc- replacement followed the Nernst equation (Figure 6C and Table 1). In contrast, the IADGRG2-ED of Adgrg2-/Y mice was substantially lower than the IADGRG2-ED of their WT littermates, which showed no significant changes in response to substitution of the bath Cl- solution with Gluc- (Figure 6A–6E and Table 1). These results suggested that ADGRG2 deficiency in the efferent ductules significantly reduced the whole-cell Cl- current of ADGRG2 promoter-labeled non-ciliated cells.

The whole-cell Cl- current recording of ADGRG2 promoter-labeled efferent ductule cells.

(A) Time course of whole-cell Cl- current (IADGRG2-ED) at +100 and −100 mV in ADGRG2 promoter-labeled efferent ductule cells derived from Adgrg2-/Y mice or their littermates. An ‘a’ or ‘d’ indicates the substitution of the Cl- bath solution with Gluc- (148.5 mM Cl- was replaced by 48.5 mM Cl- and 100 mM Gluc-); and ‘b’ or ‘e’ indicates the substitution of the Gluc- bath solution with Cl- (148.5 mM Cl-). ‘a’,”b’ and ‘c’ belong to WT mice. ‘d’,”e’ and ‘f’ belong to Adgrg2-/Y mice. (B) The current-voltage relationship of IADGRG2-ED at specific time points (from 6A) is shown. (C) The whole cell Cl- current of IADGRG2-ED elicited by voltage steps between −100 mV and +100 mV in a representative ADGRG2-promoter-RFP labeled efferent ductule cells derived from the Adgrg2-/Y mice and their wild type littermates. The outwardly rectifying IADGRG2-ED was significantly diminished when bath Cl- was substituted for gluconate (Gluc-). (D) Representative whole-cell Cl- current of ADGRG2 promoter-labeled efferent ductule cells; IADGRG2-ED versus voltage (I–V) relationships in response to voltage ramps recorded with a CsCl pipette solution in Adgrg2-/Y (n = 8) or WT mice (n = 8). The outwardly rectifying IADGRG2-ED was significantly diminished, and its reversal potential (Erev) shifted to the positive direction when Cl- was substituted for Gluc-. (E) Average current densities (pA/pF) measured at 100 mV of (C). Inset: average Erev (±s.e.m., n = 8 for each condition). **p<0.01, IADGRG2-ED in Gluc- solution was compared with IADGRG2-ED in Cl- solution. ns, no significant difference. At least three independent biological replicates were performed.

https://doi.org/10.7554/eLife.33432.016
Table 1
Average reversal potential calculated at different Cl- concentrations for Figure 6C.

Average reversal potential(Erev) (±s.e.m., n = 8 for each condition) in Figure 6C and calculated Nernst potential at different Cl- concentrations. The Nernst equation was: Erev=-RT/Z [Ln (Cl-)in/(Cl-)out].

https://doi.org/10.7554/eLife.33432.017
GroupErev[Cl-]o148.5 mM(mV)Erev[Cl-]o48.5 mM(mV)
Nernst−4.625.3
WT−4.0 ± 0.5120.1 ± 2.52
Adgrg2-/Y−4.1 ± 0.3619.4 ± 2.47

CFTR mediates the whole-cell Cl- current of ADGRG2 promoter-labeled efferent ductule cells

We next examined the effects of different Cl- channel and transporter inhibitors on the IADGRG2-ED of efferent ductule cells derived from Adgrg2-/Y mice and their WT littermates. Although application of the ANO1 inhibitor Ani9 or the chloride-bicarbonate exchanger inhibitor DIDS exerted no significant effects on the IADGRG2-ED of WT mice, the specific CFTR inhibitor CFTRinh-172 significantly reduced the IADGRG2-ED current (Figure 7A–B and Figure 7—figure supplement 1). Moreover, the difference in the IADGRG2-ED between Adgrg2-/Y mice and their WT littermates was eliminated by the application of CFTRinh-172(Figure 7A–B). After the application of CFTRinh-172, the IADGRG2-ED showed no significant response to Gluc- substitution in the bath solution (Figure 7—figure supplement 2). Consistently, when we knocked down CFTR expression in efferent ductules (Figure 7C), the whole-cell Cl- current (IADGRG2-ED) of WT mice was significantly reduced (Figure 7D–E and Figure 7—figure supplement 3). These results suggested that CFTR is essentially activated in ADGRG2 promoter-labeled efferent ductule cells, which mediate the observed outwardly rectifying whole-cell Cl- current, and ADGRG2 is required for the basic activation of CFTR in these cells.

Figure 7 with 3 supplements see all
Cl- currents in the non-ciliated cells of the efferent ductules through CFTR.

(A, D and F) Corresponding I-V curves of the whole-cell Cl- IADGRG2-ED currents recorded in Figure 6 and (A, D and F) Corresponding I-V curves of the whole-cell Cl- IADGRG2-ED currents recorded in Figure 7—figure supplement 1(A,F) and Figure 7—figure supplement 3(D). WT (n = 6), Adgrg2-/Y (n = 6); WT +CFTRinh-172 (n = 6), Adgrg2-/Y+CFTRinh-172 (n = 6), WT +ANI9 (n = 6), Adgrg2-/Y+ANI9 (n = 6), WT +DIDS (n = 6), Adgrg2-/Y+DIDS (n = 6); WT +Control RNAi (n = 6), WT +CFTR RNAi (n = 6), Adgrg2-/Y+Control RNAi (n = 6), Adgrg2-/Y+CFTR RNAi (n = 6); WT +FSK + IBMX (n = 6), Adgrg2-/Y+FSK+IBMX (n = 6). (B,E and G) Corresponding bar graph depicting the average current densities (pA/pF) measured at 100 mV in (A), (D) and (F). (C) qRT-PCR analysis of CFTR levels in the efferent ductules treated with CFTR siRNA (n = 3) or control RNAi (n = 3). (B, E and G) *p<0.05, **p<0.01, ***p<0.001, Adgrg2-/Y mice compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001. Treatment with selective inhibitors, stimulators or CFTR RNAi was compared with control vehicles or control RNAi. n.s., no significant difference. At least three independent biological replicates were performed for Figure 7B,E and G.

https://doi.org/10.7554/eLife.33432.018

CFTR is activated by FSK and IBMX (Lu et al., 2010). In response to FSK and IBMX stimulation, the IADGRG2-ED of both Adgrg2-/Y and WT mice significantly increased to similar levels (Figure 7F–G), consistent with the western blot results, indicating that CFTR expression levels did not change in Adgrg2-/Y mice. The results also indicated that basic CFTR activation in ADGRG2 promoter-labeled efferent ductule cells does not represent the full activation state (Figure 7F–G).

Gq activity is required for ADGRG2/CFTR coupling in the efferent ductules

Similar to Adgrg2-/Y mice, the efferent ductules derived from Gnaq+/- mice exhibited imbalances in pH homeostasis (Figure 8A). We utilized Gnaq+/- mice because Gnaq-/- mice were not available due to the infertility of the Gnaq+/- mice. Consistently, we observed a significantly decreased whole-cell Cl- IADGRG2-ED current of the ADGRG2 promoter-RFP-labeled primary non-ciliated cells in Gnaq+/- mice compared with that observed in their WT littermates (Figure 8B–D and Figure 8—figure supplement 1A–B). The application of Ro 31–8220, an inhibitor of the Gq downstream effector PKC, further inhibited the observed IADGRG2-ED and showed much stronger effects than the PKA inhibitor PKI 14–22 (Figure 8—figure supplement 1D–G). These results indicated that the Gq-PKC pathway plays critical roles in basic CFTR activation in the efferent ductules, which controls Cl- and pH homeostasis for efficient fluid reabsorption.

Figure 8 with 3 supplements see all
Gq activity regulated Cl- current and pH homeostasis in the efferent ductules.

(A) Intracellular pH (pHi) of the ligated efferent ductules from WT (n = 9) mice or Gnaq+/- (n = 9) mice was measured by carboxy-SNARF. (B). The whole-cell Cl- current of the IADGRG2-ED elicited by voltage steps between −100 mV and +100 mV in representative ADGRG2 promoter-RFP-labeled efferent ductule cells derived from Gnaq+/- mice, their WT littermates, or WT murine cells incubated with the PKC inhibitor Ro 31–8220 (500 nM). The whole-cell Cl- IADGRG2-ED current was recorded with a CsCl pipette solution (101 mM CsCl, 10 mM EGTA, 10 mM Hepes, 20 mM TEACl, 2 mM MgATP, 2 mM MgCl2, 5.8 mM glucose, pH7.2, with D-mannitol compensated for osm 290) and a bath solution containing 138 mM NaCl, 4.5 mM KCI, 2 mM CaCl2, 1 mM MgCl2, 5 mM glucose, and 10 mM HEPES, pH 7.4 with D-mannitol compensated for osm 310. (C) Corresponding I-V curves of the whole-cell Cl- currents recorded in (B). WT (n = 6), Gnaq+/- (n = 6), WT +Ro 31–8220 (n = 6). (D) Corresponding bar graph of the average current densities (pA/pF) measured at 100 mV according to (C). (E) Co-localization of ADGRG2 (red) and Gq (green) in the male efferent ductules. Scale bars, 50 μm. (F) Co-localization of Gq (red) and acetylated-tubulin (yellow) in the male efferent ductules. Scale bars, 50 μm. (G) IP1 levels in the brain tissues, ligated efferent ductules, and livers of WT (n = 9) or Adgrg2-/Y (n = 9) mice in response to ATP (5 mM) or control vehicles, measured by ELISA. (H) cAMP concentrations in the brains, ligated efferent ductules, and livers of WT (n = 9) or Adgrg2-/Y (n = 9) mice were measured using ELISA. (8A,8D,8G-H) *p<0.05, **p<0.01, ***p<0.001, Adgrg2-/Y mice or Gnaq+/-mice compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001, ATP- or Ro 31–8220-treated cells were compared with control vehicles. n.s., no significant difference. At least three independent biological replicates were performed for Figure 8A,D,G and H.

https://doi.org/10.7554/eLife.33432.022

We next investigated whether Gq activation by ADGRG2 is required for CFTR function, as both Gq and ADGRG2 are required for normal CFTR currents in the efferent ductules. In the efferent ductules, the Gq is localized in ADGRG2-expressing cells but not acetylated tubulin-labeled cells (Figure 8E–F and Figure 8—figure supplement 2). Consistently, Gq was readily detected in ADGRG2 antibody immuno-precipitated complexes, whereas Gi was not detectable, suggesting a physical interaction of ADGRG2 with Gq in the efferent ductules (Figure 5H and Figure 5—figure supplement 4). Moreover, the endogenous resting IP1 and cAMP levels of the ligated efferent ductules derived from the Adgrg2-/Y mice were significantly lower than those of their WT littermates (Figure 8G and H). These decreases were not caused by changes in the expression of the Gs-Adenyl-cyclase or Gq-PLC (Phospholipase C) system because Gs and Gq protein levels were similar (Figure 8—figure supplement 3), and the application of ATP induced similar levels of IP3 accumulation in the Adgrg2-/Y mice and their WT littermates (Figure 8G). Taken together, these data indicate that Gq regulates fluid reabsorption by mediating ADGRG2/CFTR coupling, and both the Gq-IP3-PKC pathway and the Gs-cAMP pathway were activated in ADGRG2 promoter-labeled efferent ductule cells.

Previous studies have shown that the activation of Angiotensin II receptor, type 2(AGTR2) increases proton secretion (Shum et al., 2008). We therefore stimulated the efferent ductules with different concentrations of angiotensin II and evaluated whether they rescued the fluid reabsorption dysfunction in Adgrg2-/Y mice by restoring pH homeostasis in the efferent ductules. Although applying 1 μM angiotensin II had no significant effect, administering 100 nM angiotensin II restored fluid reabsorption in the efferent ductules derived from Adgrg2-/Y mice (Figure 4L–M). This rescue was blocked by only the AGTR2 antagonist PD123319 (Figure 4L) but not by the Angiotensin II receptor, type 1(AGTR1) antagonist candesartan (Figure 4M). In summary, Gq and ADGRG2 regulated fluid reabsorption by maintaining pH and chloride homeostasis. The pharmacological activation of AGTR2 rescued the ADGRG2 or Gq dysfunction involved in fluid reabsorption in the efferent ductules.

ADGRG2/CFTR complex formation mediated by β-arrestin-1 but not β-arrestin-2 is essential for fluid reabsorption in the efferent ductules

In parallel with G protein signaling, arrestins mediate important functions downstream of many GPCRs, including the connection of GPCR activation to channel functions (Alvarez-Curto et al., 2016; Dong et al., 2017; Liu et al., 2017; Thomsen et al., 2016). We therefore examined the fluid reabsorption in Arrb1-/- and Arrb2-/- knockout mice. Whereas the efferent ductules derived from Arrb2-/- knockout mice showed normal fluid reabsorption as well as pH homeostasis compared to their WT littermates, these functions of the efferent ductules derived from Arrb1-/- knockout mice were significantly impaired (Figure 9A–C and Figure 9—figure supplement 1). Moreover, whereas ADGRG2 and CFTR co-localized in the apical membrane regions of the non-ciliated cells of the efferent ductules derived from Arrb2-/- or WT mice, they were separated in Arrb1-/- mice (Figure 9D–K). In β-arrestin-1-deficient efferent ductules, CFTR localized away from ezrin (Figure 9F–K), an apical membrane marker, suggesting that β-arrestin-1 is required for the correct localization of CFTR. Consistently, whereas CFTR was co-immunoprecipitated with ADGRG2 in WT and Arrb2-/- mice, it was not found in ADGRG2-immunoprecipitated complexes from the efferent ductules derived from Arrb1-/- mice, further suggesting that β-arrestin-1 is an essential component in a signaling complex encompassing ADGRG2 and CFTR in the efferent ductules (Figures 5H and 9L and Figure 9—figure supplement 2).

Figure 9 with 3 supplements see all
β-arrestin-1 is required for fluid reabsorption in the efferent ductules via scaffolding ADGRG2/CFTR complex formation.

(A) Diameters of the luminal ductules derived from WT (n = 12), Adgrg2-/Y (n = 12) or Arrb1-/- (n = 15) mice. (B) Diameters of the luminal ductules derived from WT (n = 12), Adgrg2-/Y (n = 12) or Arrb2-/- (n = 15) mice. (C) Intracellular pH (pHi) of the ligated efferent ductules derived from WT (n = 9), Arrb1-/- (n = 9) or Arrb2-/- (n = 9) mice were measured by carboxy-SNARF. (D) Co-localization of ADGRG2 (red fluorescence) and CFTR (sc-8909, Santa Cruz, green fluorescence) in the male efferent ductules of Arrb2-/- mice. (E) Analysis of ADGRG2 and CFTR fluorescence intensities in Arrb2-/- mice by Pearson’s correlation analysis. The Pearson's correlation coefficient was 0.62. (F) Localization of ADGRG2 (red fluorescence) and CFTR (sc-8909, Santa Cruz, green fluorescence) in the male efferent ductules of Arrb1-/- mice. (G) Analysis of ADGRG2 and CFTR fluorescence intensities in Arrb1-/- mice by Pearson’s correlation analysis. The Pearson's correlation coefficient was −0.15. (H) Co-localization of ezrin (red fluorescence) and CFTR (sc-8909, Santa Cruz, green fluorescence) in the male efferent ductules of Arrb2-/- mice. (I) Analysis of ezrin and CFTR fluorescence intensities in Arrb2-/- mice by Pearson’s correlation analysis. The Pearson's correlation coefficient was 0.66. (J) Co-localization of ezrin (red fluorescence) and CFTR (sc-8909, Santa Cruz, green fluorescence) in the male efferent ductules of Arrb1-/- mice. (K) Analysis of ezrin and CFTR fluorescence intensities in Arrb1-/- mice by Pearson’s correlation analysis. The Pearson's correlation coefficient was −0.15. (L) ADGRG2 was immunoprecipitated by an anti-ADGRG2 antibody in the male efferent ductules of Arrb1-/- mice or Arrb2-/- mice, and co-precipitates with CFTR, β-arrestin-1, and β-arrestin-2 were examined by using specific corresponding antibodies (CFTR antibody:20738–1-AP, Proteintech). (9A-C) *p<0.05, **p<0.01, ***p<0.001, Adgrg2-/Y mice compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001, Arrb1-/- mice or Arrb2-/- mice compared with WT mice. ns, no significant difference. At least three independent biological replicates were performed for Figure 9A–C and L.

https://doi.org/10.7554/eLife.33432.026

We therefore used HEK293 cells to investigate the in vitro role of β-arrestins in ADGRG2/CFTR complex formation. Overexpression of β-arrestin-1 but not β-arrestin-2 promoted the interaction between ADGRG2 and CFTR (Figure 9—figure supplement 3), confirming the essential role of β-arrestin-1 in assembly of ADGRG2/CFTR coupling.

Molecular determinants of ADGRG2 coupling with G protein subtypes and their contribution to the regulation of CFTR activity in vitro

ADGRG2 belongs to the adhesion GPCR group of the GPCR superfamily (Purcell and Hall, 2018; Monk et al., 2015). Whereas the endogenous ligand of ADGRG2 in the testis is unknown, several members of the same adhesion GPCR subfamily, such as VLGR1 and GPR56, showed constitutive activity via overexpression in a heterologous system (Purcell and Hall, 2018; Hu et al., 2014; Paavola et al., 2011). To dissect the molecular mechanism underlying ADGRG2 signaling in the modulation of CFTR functions, we overexpressed ADGRG2 and CFTR in HEK293 cells (Figure 10—figure supplement 1). In vitro, the overexpression of ADGRG2 causes constitutive Gs and Gq coupling activity; a stronger effect is observed with ADGRG2β (Figure 10—figure supplements 25). Whole-cell recordings were performed to examine the effects of ADGRG2 and CFTR co-expression on membrane currents by using an I-V analysis (Figure 10A). The co-expression of ADGRG2 and CFTR significantly increased the amplitude and slope of the current responses, which were significantly reduced by the CFTR inhibitor CFTRinh-172, compared with cells transfected with CFTR alone, indicating that CFTR channels are activated by ADGRG2 in a recombinant system (Figure 10B–D). Similar to primary efferent ductule cells (Figure 7F–G), the application of FSK and IBMX further increased the whole-cell Cl- current in the presence of both ADGRG2 and CFTR, confirming that ADGRG2 increased the basal activity of CFTR but did not stimulate CFTR to a full activation state (Figure 10B–C and Figure 10—figure supplement 5A).

Figure 10 with 5 supplements see all
ADGRG2 upregulates CFTR Cl- currents through G protein signaling.

(A) Whole-cell Cl- currents recorded with a CsCl pipette solution in HEK293 cells transfected with plasmids encoding ADGRG2 or/and CFTR, with or without CFTR inhibitor CFTRinh-172(10 μM) or its activator (FSK (10 μM)+IBMX (100 μM)). (B) Corresponding I-V curves of the whole-cell Cl- currents recorded in (C). ADGRG2 (n = 6), CFTR (n = 6), CFTR + CFTRinh-172(n = 6), CFTR + FSK + IBMX (n = 6), CFTR + ADGRG2 (n = 6), CFTR + ADGRG2+CFTRinh-172(n = 6), CFTR + ADGRG2+FSK + IBMX (n = 6). (C and D) Bar graph representation of average current densities (pA/pF) measured at 100 mV according to (B) and Figure 9; (C and D) Bar graph representation of average current densities (pA/pF) measured at100mVaccording to (B) and Figure 10–figure supplement 5C. (10C-10D) *p<0.05, **p<0.01, ***p<0.001, HEK293 cells transfected with CFTR compared with cells transfected with pCDNA3.1. #p<0.05, ##p<0.01, ###p<0.001, HEK293 cells transfected with ADGRG2 compared with non-ADGRG2 transfected cells. $p<0.05, $$, p<0.01, $$$, p<0.001, CFTRinh-172, FSK, NF449, U73122 or Ro 31–8220 compared with control vehicle. n.s., no significant difference. At least three independent biological replicates were performed for Figure 10C–D.

https://doi.org/10.7554/eLife.33432.030

Importantly, increased CFTR activity induced by ADGRG2 was significantly diminished by the PKC inhibitor Ro 31-8220 (Figure 10D and Figure 10—figure supplement 5B–D). Taken together, these data demonstrate that ADGRG2 increases CFTR Cl- currents through the activation of Gq-PLC-PKC signaling.

Previous crystallographic studies have shown that the intracellular loop 2 of the β2-adrenergic receptor is important for Gs coupling, and mutations in the intracellular loop three affect G protein coupling activity by receptors (Hu et al., 2014; Rasmussen et al., 2011). We therefore selected mutations in intracellular loops 2 and 3 and examined their effects on the constitutive activity of ADGRG2 in Gs or Gq signaling, as detected by cAMP or NFAT-dual-luciferase reporter (DLR) luciferase measurements (Figure 11A–C and Figure 11—figure supplement 1) in HEK293 cells. Under the equal expression of these mutants in the cell membrane, a double mutation in the ‘DRY’ motif H696A/M697A of ADGRG2 eliminated coupling activity with both Gs and Gq (Figure 11B–C and Figure 11—figure supplement 2). Three mutations in intracellular loop 2, specifically Y698A and F705A, significantly impaired the Gs coupling activity of ADGRG2 but did not exert significant effects on NFAT-DLR activity (Figure 11B–C). However, Y708A in intracellular loop 2 and R803E/K804E in intracellular loop 3 nearly abolished the Gq coupling activity of ADGRG2 but did not have significant effects on intracellular cAMP levels compared with the WT ADGRG2. Thus, the ‘DRY/HMY’ motif mutant is a G-protein dysfunctional mutant for both Gs and Gq signaling, Y698A and F705A are specific Gs-defective mutants, and Y708A and R803E/K804E are specific Gq-defective mutants of ADGRG2 (Figure 11B–C).

Figure 11 with 3 supplements see all
Key mutations of ADGRG2 downregulates CFTR Cl- currents through G protein signaling.

(A) Schematic representation of the location of the selected ADGRG2 mutants in intracellular loop 2 and loop 3 of ADGRG2. (B) Effects of the overexpression of ADGRG2 (n = 6) and its mutations (n = 6) on cAMP levels. (C) Effects of the overexpression of ADGRG2 (n = 6) and its mutations (n = 6) on NFAT-DLR activation. (D) Whole-cell Cl- currents recorded with a CsCl pipette solution in HEK293 cells overexpressing CFTR, CFTR and ADGRG2-WT, CFTR and ADGRG2-HM696AA, CFTR and ADGRG2-Y698A, CFTR and ADGRG2-F705A, CFTR and ADGRG2-Y708A or CFTR and ADGRG2-RK803EE. (E) Corresponding I-V curves for the whole-cell Cl- currents recorded in (D). (F) Bar graph representation of average current densities (pA/pF) measured at 100 mV according to (E). (11B-11C and 11F) *p<0.05, **p<0.01, ***p<0.001, cells transfected with ADGRG2-WT or mutants compared with the control plasmid (pCDNA3.1). #p<0.05, ##p<0.01, ###p<0.001, cells overexpressing ADGRG2 mutants compared with ADGRG2-WT. n.s., no significant difference. At least three independent biological replicates were performed for Figure 11B–C,F.

https://doi.org/10.7554/eLife.33432.036

The coupling of these ADGRG2 mutants to CFTR activity was then examined using the whole-cell recording technique. Voltage clamps were used to generate the I-V relationships of the CFTR currents in cells co-transfected with CFTR and ADGRG2 (Figure 11D–F and Figure 11—figure supplement 3). Interestingly, although the mutant with a specific Gs signaling defect showed decreased coupling of ADGRG2 to CFTR, the Gq-dysfunctional mutant and the H696A/M697A double Gs/Gq signaling-defective mutant did not demonstrate coupling between ADGRG2 and CFTR (Figure 11D–F). Taken together, these results demonstrate that specific residues in intracellular loops 2 and 3 are determinants of the G protein subtype coupling of ADGRG2. Furthermore, downstream of ADGRG2, Gq signaling is essential for CFTR activation in recombinant in vitro systems.

Effects of the conditional expression of WT-ADGRG2 or its selective G-subtype signaling mutants on the rescue of reproductive defects in Adgrg2-/Y mice

We next examined how the molecular determinants of ADGRG2/G protein subtype interactions contribute to the function of ADGRG2 infertility in vivo. Both ADGRG2 WT and G protein subtype mutants were conditionally expressed in the efferent ductules via virus infection under the 1 kb ADGRG2 promoter (Figure 12A). Similar to ADGRG2 WT mice, exogenously introduced ADGRG2 WT and mutants specifically localized to the inner surface of the non-ciliated cells of the efferent ductules (Figure 12—figure supplement 1A).

Figure 12 with 1 supplement see all
Conditional expression of ADGRG2 wild-type or its selective G-subtype signaling mutants in the efferent ductules in Adgrg2-/Y mice and their effects on the morphology, sperm maturation of efferent ductules.

(A) Schematic representation of the mouse ADGRG2 promoters used in the rescue experiment. (B) Representative hematoxylin-eosin staining of the WT mice, Adgrg2-/Y mice or Adgrg2-/Y mice infected with lentivirus encoding ADGRG2-WT or different G-subtype mutants at the efferent ductules, initial segment or caput of the epididymis. Scale bars, 200 μm. (C) Bar graph representing the quantitative analysis of the number of sperm shown in Figure 12B from at least four independent experiments. (D–E) The corresponding bar graph of the accumulation of spermatozoa according to the hematoxyline-eosin staining of the WT, Adgrg2-/Y mice or Adgrg2-/Y mice infected with lentivirus encoding GRP64-WT or different G subtype mutants at the efferent ductules (D) or initial segment (E) of epididymis. (C–E) *p<0.05, **p<0.01, ***p<0.001; Adgrg2-/Y mice compared with WT mice. #p<0.05, ##p<0.01, ###p<0.001; Adgrg2-/Ymice infected with the lentivirus encoding different ADGRG2 constructs compared with Adgrg2-/Y mice infected with the control lentivirus. $, p<0.05, $$$, p<0.001; Adgrg2-/Ymice infected with the lentivirus encoding different ADGRG2 constructs compared with Adgrg2-/Y mice infected with the ADGRG2-WT lentivirus. n.s., no significant difference. At least three independent biological replicates were performed for Figure 12C–E.

https://doi.org/10.7554/eLife.33432.040

The efferent ductules of Adgrg2-/Y animals frequently exhibited the accumulation of obstructed spermatozoa compared with observations in WT mice (Figure 12B). The conditional expression of ADGRG2 in non-ciliated cells in Adgrg2-/Y mice significantly reduced this obstruction, whereas the expression of G protein signaling-deficient mutants of ADGRG2, including Y698A, F705A, Y708A, RK803EE and HM696AA, significantly reduced this rescue effect (Figure 12B–E and Figure 12—figure supplement 1B–C). Specifically, conditional infection of the Gs/Gq double signaling-deficient mutant ADGRG2-HM696AA or the Gq signaling-deficient mutants Y708A and RK803EE did not result in differing levels of accumulation in the efferent ductules compared with those in Adgrg2-/Y mice infected with a control virus. The Gs signaling-deficient mutants Y698A and F705A exhibited improved rescue activity compared with the Gq mutants (Figure 12B–E and Figure 12—figure supplement 1B–C).

Consistent with observations in the efferent ductules, the lumen of the initial segment and caput region in Adgrg2-/Y mice showed reduced sperm numbers compared with those in WT mice (Figure 12B,D–E and Figure 12—figure supplement 1C). The exogenous introduction of WT ADGRG2 to non-ciliated cells nearly restored the appearance of sperm in the initial segment and significantly increased sperm numbers in the caput (Figure 12B and E and Figure 12—figure supplement 1C). However, introducing any of the Gs or Gq signaling-deficient mutants into Adgrg2-/Y mice did not induce a significant effect on sperm number restoration in these regions (Figure 12B and E and Figure 12—figure supplement 1C).

Sperm prepared from the caudal epididymis were then examined. Adgrg2-/Y mice exhibited significantly reduced sperm numbers and presented morphologically abnormal sperm compared with those of WT mice (Figure 12C and Figure 12—figure supplement 1B). Conditional expression of WT ADGRG2 in the efferent ductules restored sperm numbers in the caudal epididymis by more than half, whereas exogenous introduction of the two Gs-deficient mutants Y698A and F705A into Adgrg2-/Y mice increased sperm numbers by 5–10% compared with those in Adgrg2-/Y mice. Expression of the other 4 Gs-, Gq- or double-deficient mutants did not rescue the phenotype (Figure 12C).

To investigate whether the sperm production phenotype was related to fluid reabsorption, we isolated the efferent ductules after virus infection with the WT ADGRG2 or one of the mutants and measured the luminal area after ligation. Interestingly, conditional expression of the Gs-deficient mutations Y698A and F705A marginally reduced the inflation of the efferent ductules of Adgrg2-/Y mice, whereas the Gq signaling mutants did not exert significant effects on the luminal volume (Figure 13A–F). This result is consistent with the effects of these mutants on sperm numbers in the caudal epididymis, thereby suggesting a direct correlation between efferent ductule reabsorption ability and mature sperm numbers (Figures 12C and 13A–F). Taken together, our results demonstrate that Gq activity is required downstream of ADGRG2, and Gs function contributes to fluid reabsorption in the efferent ductules and sperm transportation.

Effects of conditional expression of ADGRG2 wild-type or its selective G-subtype signaling mutants in Adgrg2-/Y mice on the fluid reabsorption of efferent ductules.

(A) Effects of the expression of the ADGRG2-WT adenovirus on the diameter of the ligated efferent ductules derived from the WT or Adgrg2-/Y mice. (B–F) Effects of the expression of adenovirus encoding different ADGRG2 mutants on the diameter of the ligated efferent ductules derived from the WT (n = 12) or Adgrg2-/Y (n = 12) mice. (A–F) *p<0.05, **p<0.01, ***p<0.001; Adgrg2-/Y mice infected with the empty adenovirus compared with WT mice infected with the empty adenovirus. #p<0.05, ##p<0.01, ###p<0.001; Adgrg2-/Y mice infected with the adenovirus encoding different ADGRG2 constructs compared with Adgrg2-/Y mice infected with the control adenovirus. n.s., no significant difference.

https://doi.org/10.7554/eLife.33432.042

Discussion

Fluid reabsorption is the main function of the efferent ductules and is essential for sperm maturation; it therefore serves as a promising target for the development of new contraceptive methods for men (Hess, 2002). The cell surface orphan receptor ADGRG2 is an X-linked gene specifically expressed in the reproductive system, and recent studies have found that its deficiency results in the dysfunction of fluid reabsorption and male fertility. However, the mechanism by which fluid reabsorption is regulated by ADGRG2 in the efferent ductules remains unclear (Davies et al., 2004). ADGRG2 belongs to the adhesion GPCR (aGPCRs) subfamily, whose members are either structurally essential in specific tissues (VLGR1 participates in forming the ankle link) or critical signaling molecules in the nervous and immune systems (GPR56, CD97 and EMRs) (Purcell and Hall, 2018; Sun et al., 2013; Sun et al., 2016). Although the efferent ductules of Adgrg2-/Y mice exhibit normal morphology, our results here have identified essential signaling roles for ADGRG2 in non-ciliated cells of the efferent ductules to maintain pH homeostasis as well as the basic CFTR outward-rectifying current, which is required for fluid reabsorption and sperm maturation. Currently, there have been no reported endogenous ADGRG2 ligands. While an unknown ADGRG2 agonist may be responsible for ADGRG2 function in the efferent ductules, it is also likely that the constitutive activity of ADGRG2 in non-ciliated cells is sufficient to maintain the basic CFTR current and pH homeostasis, which is supported by our data using both primary ADGRG2 promoter-labeled efferent ductule cells and a recombinant heterologous HEK293 system (Figures 57 and Figure 10). Therefore, our results provide an example of the functional relevance of the constitutive activity of aGPCRs. Moreover, there are several examples indicating the constitutive activity of aGPCRs is tunable by mechanical stimulation (Purcell and Hall, 2018; Petersen et al., 2015; Scholz et al., 2015). As ADGRG2 was expressed in efferent ductules that were controlled by extensive tension, it will be interesting to investigate the effects of tension on ADGRG2 functions in future studies.

Downstream of GPCRs, 16 different G protein subtypes and arrestins play important roles in almost every aspect of human physiological processes (Liu et al., 2017; Ning et al., 2015; Yang et al., 2015, 2017). However, the expression and function of five different G protein subtypes as well as arrestins in the efferent ductules have never been systematically investigated. Here, we have determined that the majority of G protein subtypes are expressed in the efferent ductules (Figure 1A and D). Gq activity is essential for male fertility by maintaining basic CFTR activity and pH homeostasis in the efferent ductules (Figure 14). In particular, specific residues in intracellular loops 2 and 3 are structural determinants of the ADGRG2/Gq interaction (Figures 1114), which mediates the constitutive activity of Gq-PLC-IP3 signaling in non-ciliated cells of the efferent ductules.

Schematic diagram depicting the GPCR signaling pathway in the regulation fluid reabsorption in the efferent ductules.

The ADGRG2 and CFTR localized at cell plasma membrane, whereas Gs and Gq localize at the inner surface of non-ciliated cells. Deficiency of ADGRG2 in Adgrg2-/Y mice, reducing the Gq protein level by half in Gnaq+/- mice or PKC inhibitor Ro 31–8220 significantly destroyed the coupling of ADGRG2 to CFTR, thus impaired Cl- and H+ homeostasis and fluid reabsorption of efferent ductules. Structurally, residues in intracellular loops 2 and 3 of ADGRG2 are required for the specific interactions between ADGRG2 and Gq, which are required for CFTR and ADGRG2 coupling and fluid reabsorption. In addition to G protein signaling, β-arrestin-1 is also required for fluid reabsorption in efferent ductules by scaffolding the ADGRG2 and CFTR coupling and complex formation. Therefore, a signaling complex including ADGRG2, Gq, β-arrestin-1 and CFTR that specifically localizes in non-ciliated cells is responsible for the regulation of Cl- and H+ homeostasis and fluid reabsorption in the efferent ductules; thus, these functions are important for male fertility. Moreover, activation of the AGTR2 could rescue the H+ metabolic disorder caused by ADGRG2 deficiency, which restored the ability of fluid reabsorption in efferent ductules, providing a potential therapeutic strategy in treatment of male infertility caused by dysfunction of GPCR-CFTR signaling in non-ciliated cells.

https://doi.org/10.7554/eLife.33432.043

Notably, we found that ADGRG2 and Gq regulate fluid reabsorption in the efferent ductules via the activation of CFTR, an important ion channel whose mutation leads to cystic fibrosis (CF). One of the hallmarks of CF is infertility (Cutting, 2015; Massie et al., 2014), which has a 97–98% incidence rate in male CF patients (Chen et al., 2012). CFTR knockout and the application of specific CFTR inhibitors in animal models indicate that CFTR plays important roles in spermatogenesis and sperm capacitation (Chen et al., 2012). Here, we demonstrated the specific co-localization of ADGRG2 and CFTR in the apical membrane in the non-ciliated cells of the efferent ductules (Figure 5 and Figure 9). CFTR was basically active in ADGRG2 promoter-labeled efferent ductule cells, and this activity was significantly decreased by ADGRG2 or Gq deficiency. The application of a specific CFTR inhibitor, CFTRinh-172, consistently pheno-copied the ligated efferent ductules of Adgrg2-/Y mice (Figure 4K). Further pharmacological intervention in the efferent ductules and recombinant experiments in vitro confirmed the coupling of ADGRG2 and CFTR activity through Gq. Moreover, previous studies have shown that PKC phosphorylation is required for subsequent PKA phosphorylation to fully activate CFTR (Chappe et al., 2004; Jia et al., 1997). Our study not only agreed with the observation that PKC and PKA lie downstream of Gq and Gs, respectively, but also suggested that ADGRG2-activated Gq primes the full activation of CFTR in the efferent ductules. Therefore, our results demonstrate that the physiological and functional coupling of ADGRG2 and CFTR mediated by Gq in the non-ciliated cells of the efferent ductules primes the basic activity of CFTR, which is essential for fluid reabsorption. The ADGRG2-Gq-CFTR signaling axis is important to maintain male reproductive functions (Figure 14).

Parallel to G protein signaling, β-arrestins are known to play important roles in almost all GPCR functions (Cahill et al., 2017; Dong et al., 2017; Liu et al., 2017; Yang et al., 2017). Knockout of β-arrestin-1 but not β-arrestin-2 abolished the co-localization of ADGRG2 and CFTR, demonstrating the essential role of β-arrestin-1 in assembling ADGRG2/CFTR/Gq signaling compartmentalization to regulate Cl- and pH homeostasis during fluid reabsorption in the efferent ductules. For decades, GPCR/β-arrestin complexes were thought to play fundamental roles in the internalization and desensitization of G protein signaling. Recently, a mega complex encompassing the GPCR, G trimer proteins and β-arrestins was identified by using an in vitro reconstruction system in HEK293 cells to provide a new paradigm of GPCR signaling (Thomsen et al., 2016). Consistently, we identified the ability of β-arrestin-1 to facilitate ADGRG2/Gq/CFTR signaling compartmentalization, which indicated that such a receptor/G protein/β-arrestin mega complex plays important roles in the regulation of important physiological processes, such as fluid reabsorption in the efferent ductules.

Finally, our results suggest that the inhibition of either CFTR or ADGRG2 impairs the resorptive function of the efferent ductules, which may confer a contraceptive function. Indeed, anti-spermatogenic agents, such as indazole compounds, block CFTR activity (Chen et al., 2005; Gong et al., 2002). Compared with CFTR, which is broadly expressed and has important functions in many tissues, ADGRG2 is specifically expressed in the efferent ductules and epididymis. Contraceptive compounds targeting ADGRG2 may have fewer side effects. Moreover, the dysfunction of ADGRG2 or CFTR is rescued by the activation of AGTR2 in the efferent ductules (Shum et al., 2008). Therefore, a specific agonist of AGTR2 should be considered for the development of therapeutic methods to treat male infertility caused by impaired ADGRG2-Gq-CFTR signaling, such as that observed in CF patients.

Materials and methods

Key resources table
Reagent type (species)
or resource
DesignationSource or referenceIdentifiersAdditional information
Chemical compound, drugPTXEnzoCat#:BML-G100100 ng/ml
Chemical compound, drugU0126SigmaCat#:U12010 μM
Chemical compound, drugRo 31–8220AdooqCat#:A13514500 nM
Chemical compound, drugNF449TocrisCat#:13911 μM
Chemical compound, drugPKI14-22AdooqCat#:A16031300 nM
Chemical compound, drugH89BeyotimeCat#:S1643500 nM
Chemical compound, drugbumetanideAladdinCat#:B12994210 μM
Chemical compound, drugAni9SigmaCat#:SML1813150 nM
Chemical compound, drugNiflumic acid (NFA)AladdinCat#:N12959720 μM
Chemical compound, drugDIDSSigmaCat#:D351420 μM
Chemical compound, drugGlyH-101AdooqCat#:A1372310 μM
Chemical compound, drugCFTRinh-172AdooqCat#:A1289710 μM
Chemical compound, drugEGTAAladdinCat#:E1044345 mM
Chemical compound, drugSKF96365SigmaCat#:S780910 μM
Chemical compound, drugRuthenium redSigmaCat#:R275110 μM
Chemical compound, drugNicardipineSigmaCat#:N751020 μM
Chemical compound, drugLaCl3SigmaCat#:449830100 μM
Chemical compound, drugIBMXSigmaCat#:I7018100 μM
Chemical compound, drugU73122SigmaCat#:U675610 μM
Chemical compound, drugForskolinBeyotimeCat#:S161210 μM
Chemical compound, drugPD123319AdooqCat#:A132011 μM
Chemical compound, drugCandesartanAdooqCat#:A101751 μM
Chemical compound, drugAmilorideAladdinCat#:A1295451 mM
Chemical compound, drugAcetazolamideMedchem expressCat#:HY-B0782500 μM
Peptide, recombinant proteinANGIIChina Peptides100 nM
Commercial assay or kitCarboxy SNARF−1, acetoxymethyl esterInvitrogenCat#:C-12725 μM
Commercial assay or kitLipofectamine TM2000InvitrogenCat#:11668–019
Commercial assay or kitCollagenase IsigmaCat#:C0130
Commercial assay or kitcAMP ELISA kitR and D systemsCat#:KGE012B
Commercial assay or kitIP1 ELISA assayShanghai Lanpai Biotechnology Co., LtdCat#:lp034186
Commercial assay or kitThe dual-luciferase reporter assay systemPromegaCat#:E1960
AntibodyADGRG2 antibody(rabbit polyclonal)SigmaRRID:AB_1078923
AntibodyADGRG2 antibody(rabbit polyclonal)SigmaRRID:AB_2722557
AntibodyADGRG2 antibody(sheep polyclonal)R and D systemsRRID:AB_2722556
AntibodyCFTR antibody(goat polyclonal)Santa CruzRRID:AB_638427
AntibodyCFTR antibody(rabbit polyclonal)ProteintechRRID:AB_2722558
AntibodyGq antibody(goat polyclonal)Santa CruzRRID:AB_2279038
AntibodyGq antibody(rabbit polyclonal)ProteintechRRID:AB_2111647
AntibodyFlag antibody(mouse monoclonal)SigmaRRID:AB_259529
AntibodyHA antibody(mouse monoclonal)Santa CruzRRID:AB_627809
AntibodyGAPDH(rabbit monoclonal)Cell SignalingRRID:AB_10622025
AntibodyGs antibody(rabbit polyclonal)ProteintechRRID:AB_2111668
AntibodyGi antibody(mouse monoclonal)Santa CruzRRID:AB_2722559
Antibodyβ-arrestin-1 antibody(rabbit polyclonal)Dr R.J. LefkowitzA1CT
Antibodyβ-arrestin-2 antibody(rabbit polyclonal)Dr R.J. LefkowitzA2CT
AntibodyANO1 antibody(rabbit polyclonal)ProteintechRRID:AB_2722560
AntibodyEzrin antibody(rabbit polyclonal)ProteintechRRID:AB_2722561
AntibodyAcetylated Tubulin(Lys40) Antibody(mouse monoclonal)ProteintechRRID:AB_2722562
AntibodyDonkey anti-sheep IgG(H + L) (secondary antibody)AbcamRRID:AB_2716768
AntibodyDonkey anti-rabbit IgG(H + L) (secondary antibody)InvitrogenRRID:AB_2534017
AntibodyDonkey anti-mouse IgG(H + L) (secondary antibody)InvitrogenRRID:AB_141607
AntibodyDonkey anti-goat IgG(H + L) (secondary antibody)InvitrogenRRID:AB_142672, RRID:AB_141788
AntibodyHRP-conjugated Affinipure Rabbit Anti-Sheep IgG(H + L)ProteintechRRID:AB_2722563
AntibodyHRP-conjugated Affinipure Goat Anti-Rabbit IgG(H + L)ProteintechRRID:AB_2722564
AntibodyHRP-conjugated Affinipure Goat Anti-Rabbit IgG(H + L)ProteintechRRID:AB_2722565
  1. All other chemicals or reagents were from Sigma unless otherwise specified.

Mice

Mice were individually housed in the Shandong University on a 12:12 light: dark cycle with access to food and water ad libitum. The use of mice were approved by the animal ethics committee of Shandong university medical school (protocol LL-201502036). All animal care and experiments were reviewed and approved by the Animal Use Committee of Shandong University, School of Medicine. Adgrg2+/- mice were obtained from Dr DLL and MYL at East China Normal University, Shanghai, China. Adgrg2-/Y mice and WT mice were generated by crossing WT (C57BL/6J) males mice and Adgrg2+/- females mice. Arrb1-/- and Arrb2-/- mice were obtained from Dr RJ Lefkowitz (Duke University, Durham, NC); Arrb1-/- and WT mice were generated by crossing Arrb1+/- male mice and Arrb1+/- female mice. Arrb2-/- and WT mice mice were generated by crossing Arrb2+/- male mice and Arrb2+/- female mice. Gnaq+/- mice were obtained from Dr JL Liu at Shanghai Jiao Tong University. Gnaq+/- mice and WT mice were generated by crossing Gnaq+/- male mice and Gnaq+/- female mice. All C57BL/6J male mice were purchased from Beijing Vital River Laboratory Animal Technology.

Genotyping the Adgrg2-/Y KO mice

Genotyping of the intercrossed mice were examined using following primers: Fcon (Forward-control): TTTCATAGCCAGTGCTCACCTG, Fwt (Forward-wild-type): CCTGTTGGCAGACCTGAAG, Fmut (Forward-mutant): CTGTTGGCAGACCTTTTGTATATC, R (Reverse-general): CTTCCTAACATGTGCCATGGC. For the wild-type Adgrg2+/Y mice, Fcon, Fwt and R primers were used to generate two PCR products (189 bp, 397 bp); and Fcon, Fmut and R primers were used to generate one PCR product (397 bp). For the mutant Adgrg2-/Y, Fcon, Fwt and R primers were used to generate one PCR product (405 bp); and Fcon, Fmut and R primers were used to generate two PCR products (196 bp, 405 bp). The female mice were genotyped by the same method. The knockout of ADGRG2 in these mice was confirmed by western blotting.

Preparation of the membrane fraction of the epididymis and efferent ductules

The membrane fraction of the epididymis or efferent ductules was prepared from pooled mouse tissues (n = 4–6). These tissues (epididymis or efferent ductules) were dounced in a glass tube within ten volumes of homogenization buffer (75 mM Tris-Cl, pH 7.4; 2 mM EDTA, and 1 mM DTT supplemented with protease inhibitor cocktail). The dounced suspension was centrifuged at 1000 rpm for 15 min to discard the unbroken tissues. The collected suspensions were then centrifuged at 17,000 rpm for 1 hr to prepare the plasma membrane fraction. For the western blot or immunoprecipitation assays, the membranes were re-suspended in lysis buffer (50 mM Tris pH 8.0; 150 mM NaCl; 10% glycerol; 0.5% NP-40; 0.5 mM EDTA; and 0.01% DDM supplemented with protease inhibitor cocktail (Roche, Basel Switzerland) for 30 min.

Isolation and ligation of efferent ductules

The efferent ductules were microdissected into 1–1.5 mm lengths and incubated for 24 hr in M199 culture medium containing nonessential amino acids (0.1 mM), sodium pyruvate (1 mM), glutamine (4 mM), 5α-dihydrotestosterone (1 nM), 10% fetal bovine serum, penicillin (100 IU/ml), and streptomycin (100 μg/ml) at 34°C in 95% humidified air and 5% CO2. The segments were then ligated on two ends to exclude the entry and exit of fluids. Digital images of the ductules were analyzed at 0, 3, 12, 24, 36, 48, 60 and 72 hr after ligation. Damaged ductal segments were discarded. A rapid ciliary beat and clear lumens were used as evaluation standards for ductile segments that had undergone ligation. Between 9 and 36 total ductal segments from at least three mice were analyzed for each group. The differences between the means were calculated by one-way or two-way ANOVA.

Recombinant adenovirus construction (Wang et al., 2009)

The recombinant adenovirus carrying the RFP or ADGRG2 gene with the ADGRG2 promoter (pm-ADGRG2) from the epididymal genome was produced in our laboratory using the AdEasy system for the rapid generation of recombinant adenoviruses according to the established protocol (Luo et al., 2007). An adenovirus carrying green fluorescent protein (GFP) was used as a control. For the in vivo studies, a single exposure to 5 × 108 plaque-forming units (pfu) of pm-RFP or pm-ADGRG2 adenovirus was delivered to isolated efferent ductules and incubated for 24 hr to allow for sufficient infection. Epididymal efferent ductules or epididymal efferent ductule epithelium were prepared for further experiments.

Measurement of intracellular pH (pHi) with carboxy-SNARF−1

Digital images of the ductules were analyzed at 36 hr after ligation. Intracellular pH is examined with SNARF-1, a pH-sensitive fluorophore with a pKa of about 7.5. To load SNARF-1, cultured ductules were incubated with 5 μM SNARF-1-AM (diluted from a 1 mM stock solution in DMSO) for 45 min in culture medium at 37°C, 5% CO2. The cells are washed twice with buffer containing 110 mM NaCl, 5 mM KCl, 1.25 mM CaCl2, 1.0 mM Mg2SO4, 0.5 mM Na2HPO4, 0.5 mM KH2PO4, and 20 mM HEPES, pH 7.4, then placed on the microscope stage in buffer containing 5 mM KCl, 110 mM NaCl, 1.2 mM NaH2PO4, 25 mM NaHCO3, 30 mM glucose, 10 U/ml penicillin, 10 μg/ml streptomycin, and 25 mM HEPES, pH 7.30. The fluorescence was examined using an LSM 780 laser confocal fluorescence microscope (Carl Zeiss) with the excitation wavelength at 488 nm. The emissions of SNARF-1 at 590 and 635 nm were captured in the first two consecutive scans.

Intracellular pH calibration (Seksek et al., 1991)

In vivo pH calibration was performed according to the method developed by Seksek et al. Briefly, after incubation with the fluorescent probe, cells were washed in a buffer containing 10 mM Hepes, 130 mM KCl, 20 mM NaCl, 1 mM CaCl2, 1 mM KH2PO4, 0.5 mM MgSO4, at various pH values obtained by addition of small amounts of 0.1 M solutions of KOH or HCl. The pH changes of the external buffer of the cell suspension were followed with a Tacussel Isis 20000 pH-meter. Addition of nigericin (1 pg/ml) and valinomycin (5 pM) allowed an exchange of K+ for H+ which resulted in a rapid equilibration of external and internal pH. The fluorescence of the probe was excited at 488 nm, then the emission of SNARF-1 at 590 and 635 nm were captured in the first two consecutive scans. The fluorescent ratio values obtained for each pH point were used for the calibration curve obtained with Prism software, from which pHi values of the samples (6.0–8.5) were determined. Determinations were performed in quintuplicate. The sensor does not have significant effects on cell viability.

The effect of bicarbonate on intracellular pH was determined by incubating ductules in culture medium containing 25 mM bicarbonate for 40 min at 37°C, and then transferring these ductules into bicarbonate-free salt solution and then the fluorescence of the SNARF-1 probe was examined (Teti et al., 1989). Bicarbonate-free solutions were prepared by substituting NaHCO3 with Na- gluconate and equilibrating with air.

1 mM amiloride or 500 μM acetazolamide were added 100 s after the beginning of the measurement to examine the effects of acetazolamide and amiloride.

Quantitative real-time PCR

Total RNA from the mouse efferent ductules was extracted using a standard TRIzol RNA isolation method (Invitrogen, Carlsbad, CA) as previously described (Wang et al., 2014). The reverse transcription and PCR experiments were performed with the Revertra Ace qPCR RT Kit (TOYOBO FSQ-101) using 0.5 μg of each sample, according to the manufacturer’s protocols. The quantitative real-time PCR was conducted in the LightCycler apparatus (Bio-Rad) using the FastStart Universal SYBR Green Master (Roche). The qPCR protocol was as follows: 95°C for 10 min; 40 cycles of 95°C for 15 s and 60°C for 1 min; and then increasing temperatures from 65°C to 95°C at 0.1 °C/s. The mRNA level was normalized to GAPDH in the same sample and then compared with the control. All primers are listed in Supplementary file 1 and Supplementary file 2.

Immunofluorescence staining

The mice were decapitated, and the epididymis and efferent ductules were removed immediately. After dissection, the epididymis and efferent ductules were fixed in 4% paraformaldehyde by immersion overnight at 4°C. The fixed tissues were then rinsed for 4 hr at 4°C in PBS containing 10% sucrose, for 8 hr in 20% sucrose, and then overnight in 30% sucrose. The tissues were embedded in Tissue-Tek OCT compound (Sakura Fintek USA, Inc., Torrance, CA) and then mounted and frozen at −25°C. Subsequently, 8-μm-thick coronal serial sections were cut at the level of the efferent ductules and mounted on poly-D-lysine-coated slides. The slides were incubated in citrate buffer solution for antigen retrieval. Non-specific binding sites were blocked with 2.5% (wt/vol) BSA, 1% (vol/vol) donkey serum and 0.1% (vol/vol) Triton X-100 in PBS for 1 hr. After blocking, the slides were incubated in primary antibody against ADGRG2 (1:300), CFTR (1:50), Gs (1:20), Gq (1:20), ANO1(1:50), Anti-ezrin(1:50) or Anti-Acetylated Tubulin(Lys40)(1:50) at 4°C overnight. Subsequently, the slides were incubated for 1.5 hr with the secondary antibody (1:500, Invitrogen) at room temperature. For nuclear staining, the slides were incubated with DAPI (1:2000, Beyotime) for 15 min at room temperature. The immunofluorescence results were examined using a LSM 780 laser confocal fluorescence microscope (Carl Zeiss). The normal saline group was treated as the control.

Culture of mouse epididymal efferent duct epithelium (Leung et al., 2001)

After opening the lower abdomen, the efferent ductules were isolated under sterile conditions to remove fat or connective tissue. The ductules were severed into small segments and then transferred to Hanks balanced salt solution (HBSS) containing 0.2% (w/v) collagenase I and 0.1% (w/v) trypsin. Subsequently, the ductules were incubated at 34°C for 1 hr with vigorous shaking (150 strokes/min) and then separated by centrifugation at 800 g for 5 min. The pellets were re-suspended in HBSS containing collagenase I 0.2% (w/v) for 30 min at 34°C with vigorous shaking. The solutions were then centrifuged again at 800 g for 5 min, and the cell pellets were re-suspended in HBSS buffer containing 0.2% (w/v) collagenase I and then subjected to repeated pipetting for 15 min. Finally, the cells were centrifuged at 800 × g again for 5 min and resuspended in M199 medium. The cell suspension was incubated at 34°C for 5–6 hr in 5% CO2. The resulting fibroblasts and smooth muscle cells were attached to the bottom of the culture flask, whereas the epithelial cells were in suspension. The suspensions were collected, and the epithelial cells were seeded on culture flasks.

Constructs

The wild-type ADGRG2 full-length (ADGRG2FL) plasmid was obtained from Professor Xu Z. G. at Shandong University School of Life Sciences, Jinan, Shandong, China. ADGRG2 was cloned from mouse total cDNA libraries using the following primers: forward, ATTCTCGAGGATGCTTTTCTCTGGTGGG; and reverse, ATTGAATTCCATTTGCTCGATAAAGTG. The sequences were inserted into the mammalian pEGFP-N2 expression vector, and then ADGRG2FL and ADGRG2 C-terminal truncations (ADGRG2β) were subcloned into the pcDNA3.1 expression vector, with the flag sequence added at the N-terminus. The ADGRG2FL mutants (HM696AA, H696A, M697A, Y698A, K703A, V704A, F705A, Y708A, QL798AA, RK803EE) were generated using a QuikChange Mutagenesis Kit (Stratagene). All of the mutations were verified by DNA sequencing. All primers are listed in Supplementary file 3.

Cell culture, transfection, and western blotting

HEK293 cells were obtained from Cell Resource Center of Shanghai Institute for Biological Sciences(Chinese Academy of Sciences, Shanghai, China). The cell line was validated by STR profiling (Shanghai Biowing Applied Biotechnology (SBWAB) Co. Ltd.) and was negative for mycoplasma as measured by MycoAlert Mycoplasma Detection Kit (Lonza). HEK293 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (Hyclone Thermo Scientific, Scoresby, Victoria, Australia), penicillin (100 IU/ml), and streptomycin (100 μg/ml) as previously described (Hu et al., 2014; Wang et al., 2014). For receptor or other protein expression, plasmids carrying the desired genes were transfected into cells using Lipofectamine TM 2000 (Invitrogen). To monitor the protein expression levels, cells were collected 48–72 hr post-transfection with lysis buffer (50 mM Tris, pH 8.0; 150 mM NaCl; 1 mM NaF; 1% NP-40; 2 mM EDTA; Tris-HCl, pH 8.0; 10% glycerol; 0.25% sodium deoxycholate; 1 mM Na3VO4; 0.3 μM aprotinin; 130 μM bestatin; 1 μM leupeptin; 1 μM repstatin; and 0.5% IAA). The cell lysates were subjected to end-to-end rotation for 20 min and spun at 12,000 rpm for 20 min at 4°C. Then, an equal volume of 2 × loading buffer was added. Proteins were denatured in the loading buffer and subjected to western blot analysis. The protein bands from the western blot were quantified using ImageJ software (National Institutes of Health, Bethesda MD). Each experiment was repeated at least in triplicate. A data analysis was conducted using GraphPad software.

Co-immunoprecipitation

The efferent ductules of WT or Adgrg2-/Y mice were dissected into small pieces. The interaction between proteins is stabilized by addition of 1 ml of cross-linker buffer (D-PBS containing 10 mM HEPES and 2.5 mM DSP in 1:1 (v/v) dimethyl sulfoxide (DMSO)) as previously described(Ning et al., 2015; Yang et al., 2015). After continuous slow agitation for 30 min at room temperature, crosslinking was stopped by adding 25 mM Tris-HCl (pH 7.5) and incubated for another 15 min. The tissue were washed with cold PBS and then lysed in cold lysis buffer with protease inhibitors. After centrifugation, the supernatants were incubated with anti-ADGRG2 antibody (AF7977, R and D systems) for at least 2 hr at 4°C. Next, Protein A/G PLUS-Agarose (sc-2003, Santa Cruz) was added, and the complexes were incubated overnight at 4°C. The beads were washed with PBS buffer several times, and proteins were denatured in the SDS-PAGE loading buffer and subjected to western blot analysis with the indicated antibodies.

Whole-cell patch-clamp recording (Guo et al., 2014)

The efferent ductules infected by adenovirus with the ADGRG2 promoter were isolated, and epithelial cells were purified and cultured on coverslips before the patch-clamp recording. ADGRG2-promoter labeling was achieved by observation of the RFP fluorescence with the microscope. HEK293 cells transfected with plasmids encoding CFTR together with or without the ADGRG2 wild type or its mutants were cultured on coverslips before the patch-clamp recording. Borosilicate glass-made patch pipettes (Vitrex, Modulohm A/S, Herlev, Denmark) were pulled with a micropipette puller (P-97, Sutter Instrument Co.) to a resistance of 5–7 MΩ after they were filled with pipette solution. The ionic current was recorded with a data acquisition system (DigiData 1322A, Axon Instruments) and an amplifier (Axopatch-200B, Axon Instruments, Foster City, CA). The command voltages were controlled by a computer equipped with pClamp Version nine software. For the whole cell Cl- current measurement, cells were bathed in a solution of NaCl at 130 mM, KCl at 5 mM, MgCl2 at 1 mM, CaCl2 at 2.5 mM, and HEPES 20 mM, and D-mannitol was added to an osmolarity of 310 (pH 7.4). Pipettes were filled with a solution of 101 mM CsCl, 10 mM EGTA, 10 mM Hepes, 20 mM TEACl, 2 mM MgATP, 2 mM MgCl2, 5.8 mM glucose, pH7.2, with D-mannitol compensated for osm 290. When the whole-cell giga-seal was formed, the capacitance of the cell was measured. The whole-cell current was obtained by a voltage clamp with the commanding voltage elevated from −100 mV to +100 mV in 20 mV increments (Yu et al., 2011). Further validation of these observed currents were Cl- selective was provided by experiments in which 100 mM of the extracellular Cl- was replaced by gluconate.

cAMP ELISA

The efferent ductules were carefully microdissected under sterile conditions to remove fat or connective tissue and then were ligated on two ends to exclude the entry and exit of fluids. After 24 hr, these tissues were rinsed with PBS and homogenized with a tissue homogenizer in cold 0.1 N HCl containing 500 μM IBMX at a 1:5 ratio (w/v). The supernatants were collected after the centrifugation of the tissue lysates at 10,000 × g and then neutralized with 1 N NaOH. The supernatant was collected for the cAMP determination by ELISA according to the manufacturer’s instructions.

IP1 ELISA

The efferent ductules were ligated on two ends for 24 hr, and then were added 5 mM ATP or control vehicles to the tissues for 30 min. After half an hour, the tissues were homogenized with a tissue homogenizer in an assay buffer (10 mM HEPES, 1 mM CaCl2, 0.5 mM MgCl2, 4.2 mM KCl, 146 mM NaCl, 5.5 mM glucose, 50 mM LiCl, pH 7.4). The 50 mM LiCl was added to block the IP1 degradation. The lysates were centrifuged at 10,000 × g to remove insoluble components, and the supernatant was then collected for IP1 determination by ELISA (lp034186) according to the manufacturer’s instructions.

GloSensor cAMP assay

The GloSensor cAMP assay was performed as previously described (Binkowski et al., 2009; Fan et al., 2008; Hu et al., 2014; Kimple et al., 2009). HEK293 cells were transfected with the GloSensor plasmid and the desired expression plasmids (0.8 μg of total DNA) with Lipofectamine 2000 in 24-well dishes. Twenty-four hours after transfection, the cells were plated on 96-well plates at a cell density of 20,000 cells/well. The cells were maintained in DMEM for another 24 hr, washed with PBS and then incubated with 100 μl of solution containing 10% FBS, 2% (v/v) GloSensor cAMP reagent and 88% CO2-independent medium in each well for 2 hr. The cAMP signal was examined using a luminescence counter (Mithras LB 940).

NFAT dual-luciferase reporter(DLR) assay (Hu et al., 2014)

HEK293 cells in 24-well dishes were co-transfected with plasmids encoding ADGRG2 or its mutants, pGL4.16-NFAT luciferase or pGL4.16-basic luciferase, and pRL-TK Renilla using Lipofectamine 2000. These cells were cultured for approximately 48 hr and then harvested by the addition of 1 × passive lysis buffer. After incubation for 15 min at room temperature with shaking, the cell lysates were centrifuged for 10 min at 12,000 rpm at 4°C. NFAT-DLR activity was quantified by a standard luciferase reporter gene assay and then normalized to Renilla luciferase activity (Promega) as previously described (Wang et al., 2014). At least three independent experiments were executed for each dual-luciferase reporter (DLR) assay.

Recombinant lentivirus construction and lentivirus injection

Recombinant lentiviruses containing the ADGRG2 gene and its mutants (HM696AA, H696A, M697A, Y698A, K703A, V704A, F705A, Y708A, QL798AA, RK803EE) under the ADGRG2 promoter were produced according to standard procedures (Tiscornia et al., 2006; Ye et al., 2008). The lentivirus titer was 1 × 109 TU/ml. Mice were anesthetized with 10% chloral hydrate and then the conditional expression of ADGRG2-WT or its selective G-subtype signaling mutants’ lentivirus were microinjected into the interstitial space of the efferent ductules and the initial segment of epididymis at a multiplicity of infection of 100. After 14–21 days, the epididymis transfected with lentivirus were collected for use in further experiments.

Histology (Mendive et al., 2006)

The epididymis and efferent ductules were removed and fixed overnight at 4°C in 4% paraformaldehyde and stored in 70% ethanol until further use. The tissues were dehydrated, embedded in paraffin, and then sectioned into 10 μm slices. In most cases, the whole epididymis was sectioned, and representative samples throughout the organ were mounted on slides for hematoxylin and eosin staining. Hematoxylin and eosin staining was performed according to standard procedures.

Analysis of spermatozoa (Davies et al., 2004)

Spermatozoa from the caudal epididymis of the wild-type (n = 13) or Adgrg2-/Y knockout (n = 12) mice (ages between 15 and 20 weeks) were collected. The caudal region from the epididymis was open and incubated for 10 min in PBS at 34°C to allow the spermatozoa to appear. The spermatozoa were counted and analyzed by spreading the diluted homogenous suspension over a microscope slide.

Treatment of mice efferent ductules with CFTR siRNA dicer

CFTR siRNA was designed as described before (Ruan et al., 2012; Wang et al., 2006) and chemically modified by the manufacturer (GenePharma). Sequences corresponding to the siRNA of scrambled were: sense, 5’-CUUCCUCUCU UUCUCUCCCU UGUGA-3’; and antisense, 5’- TCACA AGGGAGAGAA AGAGAGGAAG-3’ or CFTR-specific siRNA-CFTR, dicer-1: sense, 5’-GUGCAAAUUCAGAGCUUUGUGGAACAG-3’; and antisense, 5’- CUGUUCCACAAA GCUCTGAAUUUGCAC-3’; CFTR-specific siRNA-CFTR, dicer-2: sense,5’-GACAACUUGUUAGUCUUCUUUCCAA-3’; and antisense, 5’- UUGGAAAGAAGACUAACAAGUUGUC-3’; CFTR-specific siRNA-CFTR, dicer-3: sense, 5’-GAGAUUGAU GGUGUCUCAUGGAAUU-3’; and antisense, 5’-AAUUCCAUGAGACACCAUCAAUCUC-3’; For in vivo studies, 15 μg of the siRNA dissolved in 30% pluronic gel (Pluronic F-127, Sigma) solution was delivered to the mice efferent ductules immediately as previously described (Wang et al., 2009). After 7 days, the epididymis transfected with siRNA were collected for further experiments.

Statistics

All the western blots were performed independently for at least three times, and the representative experimental results were shown in the main or supplementary figure. All the data are presented as the mean ±SD from at least three independent experiments. Statistical comparisons were performed using an ANOVA with GraphPad Prism5. Significant differences were accepted at p<0.05. The sequence alignments were performed using T-coffee.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
    Renal proximal tubular AT2 receptor: signaling and transport
    1. D Haithcock
    2. H Jiao
    3. XL Cui
    4. U Hopfer
    5. JG Douglas
    (1999)
    Journal of the American Society of Nephrology : JASN 10:S69–74.
  17. 17
  18. 18
  19. 19
    Physiology of V-ATPases
    1. WR Harvey
    (1992)
    The Journal of Experimental Biology 172:1–17.
  20. 20
    Testicular obstruction: clinicopathological studies
    1. WF Hendry
    2. DA Levison
    3. MC Parkinson
    4. JM Parslow
    5. MG Royle
    (1990)
    Annals of the Royal College of Surgeons of England 72:396–407.
  21. 21
  22. 22
    The Efferent Ductules: Structure and Functions
    1. RA Hess
    (2002)
    The Efferent Ductules: Structure and Functions.
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64

Decision letter

  1. Michel Bagnat
    Reviewing Editor; Duke University, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "Gq activity and β-arrestin-1 scaffolding are required for male fertility through mediating GPR64/CFTR coupling" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Peter Haggie (Reviewer #3).

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

All three reviewers thought the connection between GPR64 and CFTR was highly interesting and that mouse phenotype was also relevant in the context of efferent ductule physiology and cystic fibrosis. However, all reviewers also pointed out significant technical concerns that affect key results. In particular, the use of a relatively unspecific CFTR inhibitor that also targets Slc26a9, and concerns with the electrophysiology experiments raised questions about the primary data and the interpretation of key results. Moreover, the probes and methods used for measuring pH and Cl- levels also present technical problems. These significant technical issues and the need for extensive experimental work to provide proper controls throughout make a revision of this manuscript unfeasible within a reasonable time frame.

Reviewer #1:

In this paper, Zhang et al. link GPR64-dependent activation of CFTR with male reproductive tract physiology and link in Gq/b-arrestin. Finding that GPR64 can regulate CFTR is novel and interesting. The major problem with this paper, however, is that there is too much data, yet not enough controls. My recommendation would be to focus the paper on GPR64 and CFTR story and try to keep in the data that is really needed to justify these conclusions. Then, with the space that you've saved, perform the controls that are needed and save the additional data for other papers. The mouse experiments are nicely done and believable, but you fall down on the immunohistochemistry, western blots and patch clamping due to lack of controls. This paper is highly focused on CFTR, GPR64, Gq – it'd be nice to see for example, that GPR64 does not interact/colocalize with other ion channels (e.g. Ano1 ro whatever) etc. My comments are below.

General

What is the natural ligand for GPR64 in the testes? This is not discussed. For GPR64, this might not be known, but that's ok and it should be pointed out not avoided. GPR64 belongs to a superfamily of "adhesion GPCRs". From Wikipedia, "The defining feature of adhesion GPCRs that distinguishes them from other GPCRs is their hybrid molecular structure. The extracellular region of adhesion GPCRs can be exceptionally long and contain a variety of structural domains that are known for the ability to facilitate cell and matrix interactions." This has been reviewed and might give you a clue or at least give you something to discuss.

Patch clamp. As discussed below, more needs to be done by patch clamp to ensure that this is CFTR that you are recording (more inhibitors, demonstrate that the reversal potential changes appropriately). Also, is this CFTR that's basally active? More needs to be done to define this. What happens with forskolin or inhibitors of PKA/PKC to CFTR in WT and KO cells? Why is CFTR basally active in this cell type? How is GPR64/Gq regulating CFTR? Is it affecting N or Po? Single channel records would be nice to see, as would surface biotinylation/western blots to see if this process affects CFTR trafficking.

WBs are too cropped and often lack controls to demonstrate specificity (except those with knockout mice). Also, integrated densitometry should be included for all WBs. There are also specific issues with the CFTR westerns (i.e. I think that they're inverted). Here, the reviewers should indicate which is the mature band.

Immnuo – Not enough info is presented to convince me that what you are seeing takes place in non-ciliated cells. Please provide light images and more extensive co-localization markers to verify this.

Figure 1 - Going from A to R is excessive and many images are too small. Please consider splitting up this figure into 3.

Figure 1 A–C - Why did you only do brain and testes? Why not do other organs that expresses CFTR? This would fit better into the theme of the paper (that is expanded upon in the discussion) that CFTR is highly expressed but GPR64 gives specificity by being more highly expressed in testes.

Figure 1B - The images are too small and there is not enough information. Also, there are no controls. Also, the authors claim to that GPR64 is localized to non-ciliated cells, but I cannot tell this from the information provided.

Figure 1D - The images are too small. Scale bars are hard to read.

Figure 1P- I can barely see the cells on these images.

Figure 1I - H89 is only PKA-specific at nM levels. At 10 μm it will hit many other kinases – this needs to be addressed.

IBMX is a phosphodiesterase inhibitor – This should be stated in both the Results section and discussion section rather than calling it a "cAMP motivator".

Figure 2.

In Figure 2–E - The authors measure duct width as an indicator of fluid secretion. The first add a series of Ca2+ channel inhibitors and then a CFTR-specific inhibitor. This is a strange juxtaposition. Ca2+ is not present in biological fluids at a sufficiently high concentration to be an osmotic driver of fluid secretion and secretion is mediated by Cl, HCO3-, Na+ and K+. As such, a more conventional approach where inhibits of Cl, HCO3-, Na+ and K+ transport should be employed. For example, as well as GLYH101, DIDS, bumetanide, niflumic acid could be added as inhibitors of Cl- transport. Similarly, amiloride (Na+) etc. should also be added. Inhibiting Ca2+ channels is a different question, and if they are going to go this route, they should also inhibit Orai1/STIM1, chelate Ca2+ etc.

Figure 2F - I suspect that something is wrong with your measurements. Inhibition of CFTR does not usually lead to changes in intracellular Cl- measurements since there are many other anion channels/exchangers that can modulate intracellular Cl- homeostasis. Indeed, cells from CF patients have normal intracellular Cl- levels. MQAE is a non-rationmetric dye that has all the problems associated with this type of dye – for example, changes in cell size can concentrate or dilute the dye which will change fluorescence. Or, is dye loading normal in these cells? What if GPR64 affects xenobiotic pumps like MRP1 that can extrude fluorescent dyes?

Figure 2H - Are you really seeing a chronic pH change of >1? This is huge and likely incompatible with normal cellular function. See above comments about dye extrusion. Also, give the potential importance of this, where are the controls? What happens in bicarb free media? With acetazolamide or inhibition of Na/H or H/K exchange?

Figure 2I–J - You really should be using Pearson's correlation to measure and quantify colocalization. Controls are needed – it'd be nice to show an ion channel that does not co- colocalize. Also, showing some antibody specificity by showing secondary only.

Figure 2K–L - The current traces look like CFTR but a sizeable amount of current remains after addition of GLYH101. Is there leak or another background current. To prove that this is CFTR other anion channel inhibitors should be used (e.g. DIDs etc.). Also, these experiments are performed in near-symmetrical solutions. The authors should reduce the extracellular solution by 100 mM or more and show that they can get a Nernstian shift in reversal potential.

Figure 3

Figure 3A–B - I have the same issues with the ic Cl/pH measurements. Here, the pH difference is smaller, but WT is closer to 7.

Figure 3C–E - The above statement that not enough is done to.

Figure 3F - No controls (i.e. other proteins) – also they should include ciliated cell-specific markers such as anti-acetylated a tubulin which is only found in cilia. I'm not convinced by their claim that this effect doesn't occur in cilia. Also, there is no quantification.

Figure 3G – Why did they do spleen as a control here (brain was used in Figure 1). I think that a more rational approach should be used – i.e. some other CFTR-expressing and some non-CFTR expressing tissues.

Figure 3G - Please include a positive control that can alter IP levels such as ATP which will activate Gq via purinergic receptors. Since IP metabolism can vary depending on the cells needs, a common approach to meaure IP levels is to add lithium to prevent IP degradation and look for total IP levels of all species.

Figure 3 - See above but WBs are too cropped and don't have controls like cells that do not express GPR64.

Figure 4

Figure 4C–D - Same comments as above.

Figure 4E–J - Same comments as above. (needs controls and better quantification).

Figure 4K - CFTR is usually seen as Band C (mature, glycosylated) and Band B (immature, minimally glycosylated). Normally Band C is the predominant form in non-CF cells. Thus, unless your cells are CF or very different, I strongly suspect that your CFTR blots are upside down.

Figure 5

Figure 5A - Same comments as above.

5B-C - Same comments as above. (needs controls and better quantification).

Figure 5D–F - For the CFTR electrophys in this paper, the reversal potential is negative (unlike Figure 2 where it is 0). This is surprising given the predicted reversal potential for Cl- is ~0. Why is this? Did you change your conditions relative to Figure 2?

Figure 5G–L - Given the length of the paper, this in my opinion is too much. This data should be removed and saved for another paper.

Figure 6

Figure 6B - Please spell out what ED, IS stand for on the figures. There's room. Also, please add arrows pointing to sperm accumulation etc.

Supplement – many of the same concerns arise. The figures are very small, immunofluorescence doesn't have adequate controls and Western blots are too cropped.

Discussion section – given all of the data presented. The discussion is very short. I would like to see more consideration of GPR64s physiological role as well as more discussion of CFTR regulation.

Reviewer #2:

Gq activity and B-arrestin1-scaffolding are required for male fertility through mediating CPR64/CFTR coupling.

Overview: This is a comprehensive study evaluating the role for a CFTR/GPR64 complex in fluid reabsorption across efferent ducts of the testis – a function that is vital for male fertility. Further the authors interrogated the regulation of this complex by Gq and B-arrestin. For this most part- these studies are robust and convincing with respect to the role of GPR64 in this function. However- the role of CFTR in this function could use additional supportive data.

Specific comments:

Figure 1 focuses on the tubule fluid transport properties of efferent ductules- studied ex-vivo. These studies clearly show that disruption of GPR64 and Gq impairs fluid reabsorption, morphology and sperm count.

On the other hand- the role for CFTR is not clear as increases and decreases in activation (via kinase activators and inhibitors respectively) leads to the same effect of luminal swelling.

Figure 2 shows that disruption of GPR64 alters the sensitivity of the ductular fluid transport to ion channel blockers (calcium transport proteins in addition of a CFTR channel blocker). They also show that chloride channel activity contributes to fluid transport. Their data falls somewhat short of proving that this activity is CFTR mediated. The regulatory properties of the conductance (i.e. regulation by PKA) was not shown and this is important. Also, while GlyH-101 is a well-known CFTR channel blocker- it is not specific and inhibits other chloride channels, including the SLC26A9 channel. The authors should include another inhibitor (i.e. CFTRinh-172) to test specificity.

Figure 4 aims to show that Β-arrestin1 contributes to fluid reabsorption in these tubes and its expression promotes co-localization of GPR64 and CFTR. However, labeling of the diagrams in Figure 4 seems somewhat confusing- does 4J show a line scan of GPR64 and CFTR localization or BArr1 and GPR64? The pattern of staining for GPR64 and CFTR looks similar but not overlapping- what compartments are the two proteins localized after BArr1 KO?

Figure 5 shows the consequence of co-expressing CFTR with GPR64 Wt or mutants bearing substitution in intracellular loops 2 and 3. This is a comprehensive set of studies supporting the role for these loops in mediating functional interaction between GRP64 and CFTR. However- an important control would include single transfections with GPR64 (no CFTR) to ensure that it is not modulating a distinct chloride channel. Western blotting to ensure expression of the each of the mutant GPR64 proteins would also be helpful to support the conclusion that there are site specific effects in the interaction- rather than reporting differences in protein abundance.

Reviewer #3:

The manuscript under review considers interaction between GPR64, CFTR and arrestins in non-cilliated cells of the male reproductive system. In general, the content of this study is interesting and relevant. However, a number of significant technical concerns with the presented study raise doubts about key results.

Major points:

Several studies have shown that GlyH-101 is not entirely specific for CFTR, for instance at 50 microM GlyH-101 inhibits SLC26A9. A panel of CFTR inhibitors, including CFTRinh172 should have been considered for studies presented in Figure 2. In addition, qPCR or similar should be employed to unambiguously determine that SLC26A9 is not a relevant player in cellular system under consideration.

In terms of the intracellular measurements of [Cl-], MQAE is essentially completely insensitive to [Cl-] above 100 mM, so it is hard to see how intracellular [Cl-] could be determined to be ~140 mM. In this regard, it is notable that the calibration curve for MQAE versus [Cl-] presented in the supplementary data is only extended to ~80 mM [Cl-] (i.e., well below the reported value). As such, it is difficult to have any confidence in the presented values of [Cl-].

In terms of the GlyH-101 studies presented in Figure 2F, do driving forces predict that CFTR inhibition would mediate accumulation of cytoplasmic [Cl-]? What are the consequences of such inappropriately high intracellular [Cl-]? What happens to the concentration of cations, and the membrane potential of cells with such non-physiological levels of [Cl-]?

In terms of the pH measurements presented in Figure 2, sensitivity of BCECF to determine pH above ~pH 7.8 is limited. In addition, no calibration curve is presented – it is critical to demonstrate that the employed method would accurately determine the reported pKa for the fluorescent probe employed, i.e., BCECF, to have confidence that a pH of 8.4 could be accurately determined. The presented methods are unclear, for instance a 25 mM bicarbonate solution would require gassing with 5% CO2, but this is not mentioned. The description of what was measured in the Results was vague, inner solution of efferent ductules does not imply that cytoplasmic pH was determined.

In terms of the patch clamp analysis presented in Figure 2, prior studies by Muanprasat and colleagues have demonstrated that GlyH-101 inhibition of CFTR is strongly dependent on membrane potential – as would be expected for a charged molecule with a pore occluding mechanism of action. As such, GlyH-101 inhibition alters CFTR current-voltage curves from being linear to showing inward rectification. This is apparently not observed in the data presented in Figure 2L (where I-V curve remains linear). There is a concentration dependence of this phenomenon, however, I was unable to find information about how much GlyH-101 was used in the Legend or Material and Methods section for the presented data. In general, most studies used 25 microM GlyH-101. If this concentration was used in Figure 2, then inward rectification of CFTR I-V relationship would definitely be anticipated. Consideration of submaximal concentrations of GlyH-101, per Muanprasat and colleagues, should be considered to provide confidence that CFTR currents are really being observed in the reported data. In addition, delivery of PKA (in the pipette) is typically used for excised patch data, as such, experimental data should be presented for whole cell recordings with consideration of an alternative CFTR stimulant such as forskolin. In addition, for patch clamp data presented in Figure 3D, pharmacological validation that currents are CFTR -dependent should be presented.

For the data presented in Figure 5, does stimulation of CFTR with an alternative agonist, such as forskolin, mediate similar cytoplasmic chloride concentration reduction? Molecular details of how GPR64 is activated have been elucidated and are considered by the authors in the supplemental data, for instance in regard to GPR64beta elevating of cAMP. Does the carboxy-terminal fragment of GPR64 also reduce cytoplasmic [Cl-]?

In Figure 2I–J, the mere co-localization of two proteins to a membrane determined imaged by confocal microscopy does not indicate or suggest that a complex with functional coupling exists. It indicates that two proteins are targeted to the same membrane. This same concern is relevant for data presented in Figure 3, Figure 4, and Figure 5.

For co-IP experiments shown in Figure 4K, there is insufficient explanation (in Materials and methods section, Results section, Legend etc.) to comprehend what is being done. By elimination, I assume the anti-HA blot was against arrestins, but, this is not detailed (for instance, I cannot find details of HA-tagged arrestin constructs).

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for resubmitting your work entitled "Gq activity- and β-arrestin-1 scaffolding-mediated GPR64/CFTR coupling are required for male fertility" for further consideration at eLife. Your revised article has been evaluated by Didier Stainier (Senior editor), a Reviewing editor, and three reviewers.

The manuscript has been improved but there are some remaining problematic issues that need to be addressed, as outlined below:

1) The intracellular Cl- measurements remain problematic due to the lack of a radiometric method. These measurements should be either replaced by suitable radiometric or electrophysiological measurements or removed from the manuscript.

2) Western blots need controls and markers.

3) The CFTR currents are extremely small. The i/v shift with gluconate is too small and not typical for CFTR. Original whole cell overlay currents or continuous recordings should be shown. What is the proof in addition to CFTRinh172 that the authors truly measured CFTR currents?

4) Figure 4A suggests a GPR64-dependent expression of CFTR-mRNA. How does this fit to the expression data shown in Figure 5D/I and to the data shown in Figure 6?

5) Figure 10: In what cells were these data obtained? No control for expression of the various GPR64 mutants is provided.

https://doi.org/10.7554/eLife.33432.050

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

Reviewer #1:

In this paper, Zhang et al. link GPR64-dependent activation of CFTR with male reproductive tract physiology and link in Gq/b-arrestin. Finding that GPR64 can regulate CFTR is novel and interesting.

Thank you for your positive comments.

The major problem with this paper, however, is that there is too much data, yet not enough controls. My recommendation would be to focus the paper on GPR64 and CFTR story and try to keep in the data that is really needed to justify these conclusions. Then, with the space that you've saved, perform the controls that are needed and save the additional data for other papers. The mouse experiments are nicely done and believable, but you fall down on the immunohistochemistry, western blots and patch clamping due to lack of controls. This paper is highly focused on CFTR, GPR64, Gq – it'd be nice to see for example, that GPR64 does not interact/colocalize with other ion channels (e.g. Ano1 ro whatever) etc. My comments are below.

Thank you for your helpful suggestions, which have improved the flow of our story and made the manuscript more compelling by including appropriate controls. We have added additional controls, including immunohistochemistry, Western blots and patch clamp experiments. For example, we have used another ion channel, Ano1, as a negative control for the co-localization of GPR64 with CFTR (Figure 1—figure supplement 1 and Figure 5—figure supplement 4). We have also reorganized the data to clarify the results. Please see the following point-by-point responses.

General

What is the natural ligand for GPR64 in the testes? This is not discussed. For GPR64, this might not be known, but that's ok and it should be pointed out not avoided. GPR64 belongs to a superfamily of "adhesion GPCRs". From Wikipedia, "The defining feature of adhesion GPCRs that distinguishes them from other GPCRs is their hybrid molecular structure. The extracellular region of adhesion GPCRs can be exceptionally long and contain a variety of structural domains that are known for the ability to facilitate cell and matrix interactions." This has been reviewed and might give you a clue or at least give you something to discuss.

Thank you for your helpful suggestion. We have now included an introduction of GPR64 as an adhesion GPCR member in subsection “Molecular determinants of GPR64 coupling with G protein subtypes and their contribution to the regulation of CFTR activity in vitro”, and a discussion of the potential GPR64 activation mechanism in the Discussion section in the revised manuscript.

Patch clamp. As discussed below, more needs to be done by patch clamp to ensure that this is CFTR that you are recording (more inhibitors, demonstrate that the reversal potential changes appropriately). Also, is this CFTR that's basally active? More needs to be done to define this. What happens with forskolin or inhibitors of PKA/PKC to CFTR in WT and KO cells? Why is CFTR basally active in this cell type? How is GPR64/Gq regulating CFTR? Is it affecting N or Po? Single channel records would be nice to see, as would surface biotinylation/western blots to see if this process affects CFTR trafficking.

Thank you for your helpful comments regarding the working mechanism of CFTR in the efferent ductules. We have performed substantial new experiments, and a detailed response follows:

1) We have performed new electrophysiology whole-cell Cl- recordings of primary GPR64 promoter-labeled efferent ductule cells derived from wild-type and GPR64-/Y mice at normal Cl- concentrations or with the substitution of Cl- by gluconate (Gluc-) in the bath solution, with or without different agonists or pharmacological inhibitors, including forskolin, the PKC inhibitor Ro 31-8220, the PKA inhibitor PKI 14-22, the ANO1 inhibitor Ani9 (Namkung, 2016), the chloride-bicarbonate exchanger inhibitor DIDS and the specific CFTR inhibitor CFTR-Inh-172 (Figure 6A–6F, Figure 7C–7E, Figure 6—figure supplement 1 to Figure 6—figure supplement 3, and Figure 7—figure supplement 1).

In GPR64 promoter-labeled wild-type efferent ductule cells, we observed an outwardly-rectifying whole-cell Cl- current (IGPR64-ED), which was significantly diminished in response to substitution of the bath Cl- solution with Gluc- (148.5 mM Cl- was replaced with 48.5 mM Cl- and 100 mM Gluc-) (Figure 6A–B and Figure 6—figure supplement 1). The change in the reversal potential (Erev) followed the Nernst equation (Figure 6B and Figure 6—figure supplement 1). However, the whole-cell Cl- current of the GPR64 promoter-labeled efferent ductule cells derived from GPR64-/Y mice was significantly lower than that in cells derived from their wild-type littermates, which showed only a modest difference in response to substitution of the bath Cl- solution with Gluc- (Figure 6A–6B). These results suggested that GPR64 deficiency in the efferent ductules significantly reduced the whole-cell Cl- current of GPR64 promoter-labeled non-ciliated cells.

2) Our Western blot data showed that the ablation of GPR64 had no effect on CFTR protein expression levels (Figure 5B and I and Figure 5—figure supplement 5). To examine whether CFTR is basically activated in GPR64 promoter-labeled cells, we utilized the CFTR-specific inhibitor CFTR-Inh-172, the CFTR agonist FSK+IBMX (Lu et al., 2010) and inhibitors for other Cl- transporters, including ANI9 and DIDS. The whole-cell Cl- current of GPR64 promoter-labeled wild-type efferent ductule cells was significantly reduced by the application of the CFTR inhibitor CFTR-Inh-172 but not by the ANO1 inhibitor ANI9 or the chloride-bicarbonate exchanger inhibitor DIDS, suggesting that CFTR is likely the main mediator of the Cl- current of GPR64 promoter-labeled efferent ductule cells (Figure 6C–D and Figure 6—figure supplement 3). Moreover, the CFTR agonist FSK+IBMX induced similar Cl- currents in cells derived from both wild-type and GPR64-/Y mice (Figure 6E–F and Figure 6—figure supplement 2). Moreover, the application of CFTR-Inh-172 to GPR64 promoter-labeled cells derived from wild-type mice decreased the IGPR64-ED to a level closer to the IGPR64-ED of cells derived from GPR64-/Y mice (Figure 6D). These results suggested that CFTR in GPR64 promoter-labeled wild-type efferent ductule cells was generally active, and this activity was dependent on the expression of GPR64 in these cells.

3) To define the mechanism underlying CFTR activity in the efferent ductules, we performed a series of cell biology experiments. The GPR64 promoter-labeled efferent ductule cells derived from GPR64-/Y cells had much lower IP1 and cAMP levels compared with cells derived from wild-type cells (Figure 7H–I in the revised manuscript). These increases potentially cause the basal activity of CFTR, which is activated by PKA and PKC downstream of IP1 and cAMP (Chappe et al., 2003; Guggino and Stanton, 2006), and our data showed the constitutive Gq and Gs activity of GPR64 following its expression in HEK293 cells, which promoted CFTR activity in a heterologous system (Figure 9, Figure 10 and Figure 9—figure supplement 1 to Figure 9—figure supplement 6). The inhibitory effects of the PKC inhibitor Ro 31-8220 and the PKA inhibitor PKI 14–22 in the efferent ductules further supported this hypothesis (Figure 7C–E and Figure 7—figure supplement 1D–G). The increased IP1 and cAMP levels were attributed to the constitutive activity of GPR64, a property shared by many adhesion GPCR members (Gupte et al., 2012; Paavola and Hall, 2012). Currently, single channel recording of GPR64-labeled efferent ductules is technologically challenging as we are unable to decrease the noise, which is essential for a good signal-to-noise ratio. We are still optimizing conditions for these experiments and may include these results in a future manuscript.

4) To monitor CFTR trafficking in the efferent ductule cells, we performed co-immunostaining experiments. In both GPR64-/Y and wild-type mice, CFTR colocalized with ezrin, an apical membrane marker for the efferent ductules (Figure 8—figure supplement 2), suggesting that GPR64 deficiency does not affect CFTR trafficking. However, in β-arrestin-1-deficient mice, CFTR was localized away from both GPR64 and ezrin (Figure 8G and Figure 8K), suggesting that its proper localization in the apical membrane is dependent on β-arrestin-1 but not on β-arrestin-2 or GPR64.

WBs are too cropped and often lack controls to demonstrate specificity (except those with knockout mice). Also, integrated densitometry should be included for all WBs. There are also specific issues with the CFTR westerns (ie I think that they're inverted). Here, the reviewers should indicate which is the mature band.

Thank you for these helpful suggestions. We have included larger gel images and statistical analysis of the integrated densitometry (Figure 8—figure supplement 3, Figure 9—figure supplement 2 to Figure 9—figure supplement 4, the Figure 5I, the Figure 8M, the Figure 7—figure supplement 3 and the Figure 10—figure supplement 2). The mature band of CFTR is indicated in Figures 5I, 8M, Figure 7—figure supplement 3 and Figure 8—figure supplement 3 of the revised manuscript.

Immnuo – not enough info is presented to convince me that what you are seeing takes place in non-ciliated cells. Please provide light images and more extensive co-localization markers to verify this.

Thank you for these helpful suggestions. We have performed new immunofluorescence experiments using α-acetylated tubulin as a marker for the ciliated cells in the efferent ductules. As shown in Figure 1B and Figure 1—figure supplement 1 in the revised manuscript, the results indicated that GPR64 localized in the non-ciliated cells.

Figure 1 - Going from A to R is excessive and many images are too small. Please consider splitting up this figure into 3.

Thank you for your suggestion. We have divided the contents of the previous Figure 1 into three figures in the revised manuscript (Figures 13).

Figure 1 A–C. Why did you only do brain and testes? Why not do other organs that expresses CFTR? This would fit better into the theme of the paper (that is expanded upon in the discussion) that CFTR is highly expressed but GPR64 gives specificity by being more highly expressed in testes.

Thank you for your helpful consideration. GPR64 is a G protein coupled receptor with selective cell subtype expression in the efferent ductules (Figure 1B). Therefore, we initially labeled GPR64-expressing cells in the efferent ductules and examined the immediate downstream effectors of G protein subtype expression in Figure 1A and 1D (Figure 1A and C in our original manuscript). Later, when we characterized the mechanism underlying GPR64-mediated suppression of fluid reabsorption dysfunction in the efferent ductules, according to the reviewer’s suggestion, we performed new experiments with liver and brain tissues. As shown in the revised Figure 4A and Figure 4—Supplemental Figure 1, we examined the expression of important channels and transporters related to fluid reabsorption in both brain and liver tissues. The liver demonstrated high expression of CFTR but little GPR64 expression.

Figure 1B - The images are too small and there is not enough information. Also, there are no controls. Also, the authors claim to that GPR64 is localized to non-ciliated cells, but I cannot tell this from the information provided.

Thank you for your helpful suggestions. We have enlarged image 1B so that it is more visible. We have performed new immunofluorescence experiments using α-acetylated tubulin as a marker for ciliated cells in the efferent ductules. As shown in Figure 1B and Figure 1—figure supplement 1 in the revised manuscript, GPR64 is localized in the non-ciliated cells of the efferent ducts.

Figure 1D – the images are too small. Scale bars are hard to read.

Thank you for your helpful suggestion. We have enlarged the figure (new Figure 2A in the revised manuscript) accordingly.

Figure 1P - I can barely see the cells on these images.

Thank you for your helpful suggestion. We have enlarged the figure (new Figure 3E in the revised manuscript) accordingly.

Figure 1I - H89 is only PKA-specific at nM levels. At 10 μm it will hit many other kinases – this needs to be addressed.

Thank you for these helpful suggestions. We have changed the H89 concentration to 500 nM, and we also used other drugs, such as PKI14-22 (a specific PKA inhibitor) and NF449 (a Gs inhibitor), to characterize the Gs-PKA pathway in fluid reabsorption in the efferent ductules (for example, new Figure 2G-H in the revised manuscript).

IBMX is a phosphodiesterase inhibitor – this should be stated in both the results (p5) and discussion rather than calling it a "cAMP motivator".

Thank you for your suggestions. We have clarified that IBMX is a phosphodiesterase inhibitor in the figure legend and results (line 146 in the revised manuscript).

Figure 2.

In Figure 2A-E, the authors measure duct width as an indicator of fluid secretion. The first add a series of Ca2+ channel inhibitors and then a CFTR-specific inhibitor. This is a strange juxtaposition. Ca2+ is not present in biological fluids at a sufficiently high concentration to be an osmotic driver of fluid secretion and secretion is mediated by Cl, HCO3-, Na+ and K+. As such, a more conventional approach where inhibits of Cl, HCO3-, Na+ and K+ transport should be employed. For example, as well as GLYH101, DIDS, bumetanide, niflumic acid could be added as inhibitors of Cl- transport. Similarly, amiloride (Na+) etc should also be added. Inhibiting Ca2+ channels is a different question, and if they are going to go this route, they should also inhibit Orai1/STIM1, chelate Ca2+ etc.

Thank you for your very helpful suggestions to improve our paper. We have incorporated all of the drugs you have suggested in new experiments to confirm our results.

We performed the following new experiments: first, we examined the expression of different chloride channels as well as HCO3-, Na+ and K+ transporters in GPR64-expressing efferent ductule cells and compared them with non-GPR64 promoter-labeled cells (Figure 4A in the revised manuscript). We found that CFTR, NHE1, NHE3, NKCC, DRA, SLC26a9, CAII, CLCA1, TRPC3, Cav1.2 and Cav1.3 are expressed in GPR64 promoter-labeled cells, whereas ANO1, V-ATPase and Cav2.2 demonstrated relatively lower expression levels (Figure 4A).

We next used bumetanide to block the Na-K-Cl cotransporter NKCC, nifumic acid (NFA) to block the calcium-dependent chloride channel CaCC, and ANI9 to block anoctamin-1 (ANO1) (Figure 4B-D). None of these inhibitors had significant effects on fluid reabsorption in the efferent ductules in ligation experiments. Similarly, chelating calcium with EGTA or the administration of TRP channel blockers had no significant effects (Figure 4E-H). The application of 4,4'-diisothiocyano-2,2'-stilbenedisulfonic acid (DIDS) to block chloride-bicarbonate exchange induced small effects only after 60 hours (Figure 4I). The application of amiloride to block sodium/hydrogen antiporter NHE1 activity resulted in different phenotypes compared to those of GPR64-/Y mice (Figure 4—figure supplement 1C). In contrast, blocking CFTR activity with either GlyH-101 or CFTR-Inh-172 exerted significant effects on fluid reabsorption in the efferent ductules and pheno-copied GPR64 -/Y mice (Figure 4J–4K). Collectively, these results indicated that CFTR plays an important role in fluid reabsorption and is a potential downstream effector of GPR64 in the efferent ductules.

Figure 2F - I suspect that something is wrong with your measurements. Inhibition of CFTR does not usually lead to changes in intracellular Cl- measurements since there are many other anion channels/exchangers that can modulate intracellular Cl- homeostasis. Indeed, cells from CF patients have normal intracellular Cl- levels. MQAE is a non-rationmetric dye that has all the problems associated with this type of dye – for example, changes in cell size can concentrate or dilute the dye which will change fluorescence. Or, is dye loading normal in these cells? What if GPR64 affects xenobiotic pumps like MRP1 that can extrude fluorescent dyes?

Thank you for your helpful suggestions. We have changed our method for intracellular Cl- measurement by implementing the use of specific GFP variants in the YFP family (mCIY-8M) whose fluorescence is affected by halide concentrations in the cell (Figure 5—figure supplement 1; Figures 5A and Figure 7A in the revised manuscript).

Figure 2H - Are you really seeing a chronic pH change of >1? This is huge and likely incompatible with normal cellular function. See above comments about dye extrusion. Also, give the potential importance of this, where are the controls? What happens in bicarb free media? With acetazolamide or inhibition of Na/H or H/K exchange?

Thank you for your very helpful comments. We have performed new experiments with another pH indicator, 5'(and 6')-carboxy-10-dimethylamino-3-hydroxy-spiro[7H-benzo[c]xanthene-7,1'(3H)-isobe nzofuran]-3'-one (carboxy SNARF-1) (Figure 5—figure supplement 3). Newly acquired data is shown in Figure 5C, Figure 7B and Figure 8D in the revised manuscript. We did see a chronic pH change, which was below 1, which agreed well with your prediction. The pH imbalance in GPR64-/Y mice was rescued by bicarb-free media or application of the carbonic anhydrase inhibitor acetazolamide (Figure 5—figure supplement 3B–C). Application of the Na/H exchanger inhibitor amiloride is confounding, as it decreased the pH but caused acute fluid reabsorption dysfunction (Figure 4—figure supplement 1C). The role of the Na/H exchanger in fluid reabsorption may be the focus of a future study.

Figure 2I–J - You really should be using Pearson's correlation to measure and quantify colocalization. Controls are needed – it'd be nice to show an ion channel that does not co- colocalize. Also, showing some antibody specificity by showing secondary only.

Thank you for your helpful suggestions. We changed our co-localization measurement method to Pearson’s correlation analysis to better explain the results that we acquired (Figure 5E, Figure 5H, Figure 8F, Figure 8H, Figure 8J and Figure 8L; Figure 8—figure supplement 2A–2B and Figure 9—figure supplement 1). We included another ion channel, Ano1, as a negative control (Figure 5—figure supplement 4). We also demonstrated the specificity of our secondary antibody, which is shown in Figure 1—figure supplement 1A–1B.

Figure 2K–L - The current traces look like CFTR but a sizeable amount of current remains after addition of GLYH101. Is there leak or another background current. To prove that this is CFTR other anion channel inhibitors should be used (e.g. DIDs etc.). Also, these experiments are performed in near-symmetrical solutions. The authors should reduce the extracellular solution by 100 mM or more and show that they can get a Nernstian shift in reversal potential.

Thank you for your helpful suggestions. We have examined the effects of the

Ano1 inhibitor Ani9, the non-selective chloride-bicarbonate exchanger DIDS and the CFTR-specific inhibitor CFTR-Inh-172 on the electrophysiological recording of GPR64 promoter-labeled efferent ductule cells (Figure 6C-D and Figure 6—figure supplement 2 to Figure 6—figure supplement 3). The results indicated that only the blockade of CFTR activity significantly inhibited the IGPR64-ED. Considering the sizable current in wild-type mice after application of the CFTR-specific inhibitor CFTR-Inh-172, a plausible explanation is that the Cl- current is mediated by neither CFTR nor GPR64 in these efferent ductule cells. When we performed these experiments by substituting Cl- in the bath buffer with Gluc- (148.5 mM Cl- was replaced by 48.5 mM Cl- and 100 mM Gluc-) (Figure 6A–6B), accompanied by the inhibitory effects of CFTR inh-172, we still observed a decrease in the current and a change in the reversal potential (Erev) that followed the Nernst equation (Figure 6—figure supplement 3C). However, these currents did not contribute to fluid reabsorption dysfunction in response to GPR64 deficiency, as there were no significant differences between GPR64-/Y mice and their wild-type littermates in terms of the IGPR64-ED after the application of CFTR-Inh-172 (Figure 6D).

Figure 3

Figure 3A–B- I have the same issues with the ic Cl/pH measurements. Here, the pH difference is smaller, but WT is closer to 7. Figure 3C-E. The above statement that not enough is done to.

Thank you for your helpful suggestions. We have used alternative intracellular Cl/pH level sensors (mCIY-8M and the carboxy SNARF-1) for superior measurement (Figure 5A and C and Figure 5—figure supplement 1 and Figure 5—figure supplement 3).

Figure 3F No controls (i.e. other proteins) – also they should include ciliated cell-specific markers such as anti-acetylated a tubulin which is only found in cilia. I'm not convinced by their claim that this effect doesn't occur in cilia. Also, there is no quantification.

Thank you for your helpful comments. We used an acetylated α-tubulin antibody to label the ciliated cells (Figure 1B and Figure 1—figure supplement 1). Co-localization experiments suggested that Gq is mainly expressed in non-ciliated cells (Figure 7F–G in the revised manuscript), and Gq and GPR64 are colocalized in the same cells (Figure 7F in the revised manuscript). The quantification is shown in Figure 7—figure supplement 2 in the revised manuscript

Figure 3G – why did they do spleen as a control here (brain was used in Figure 1). I think that a more rational approach should be used – i.e. some other CFTR-expressing and some non-CFTR expressing tissues.

Thank you for your suggestions. We have included the liver (which has higher CFTR expression) and brain (relative lower CFTR expression) as controls for the efferent ductules in our revised manuscript (Figure 7H-I and Figure 7—figure supplement 3).

Figure 3G. Please include a positive control that can alter IP levels such as ATP which will activate Gq via purinergic receptors. Since IP metabolism can vary depending on the cells needs, a common approach to measure IP levels is to add lithium to prevent IP degradation and look for total IP levels of all species.

Thank you for your suggestions. We have included ATP stimulation in Figure 7H in the revised manuscript, and all IP levels were assayed with lithium incubation in our assays accordingly (see Materials and methods section).

Figure 3I. See above but WBs are too cropped and don't have controls like cells that do not express GPR64.

We have included new WB results in Figure 8—figure supplement 3, Figure 9—figure supplement 2C and Figure 9—figure supplement 3F to demonstrate GPR64 expression.

Figure 4

Figure 4C–D - Same comments as above.

Figure 4E–J - Same comments as above. (needs controls and better quantification).

Figure 4K. CFTR is usually seen as Band C (mature, glycosylated) and Band B (immature, minimally glycosylated). Normally Band C is the predominant form in non-CF cells. Thus, unless your cells are CF or very different, I strongly suspect that your CFTR blots are upside down.

Thank you for your helpful suggestions. We have used new sensors to measure the Cl- concentration and pH (Figure 6A-B and Figure 5—figure supplement 1 and Figure 5—figure supplement 3). The interactions of GPR64 with β-arrestin1/β-arrestin2 or GPR64 with CFTR were confirmed by co-immunoprecipitation experiments (Figure 5I and Figure 5—figure supplement 5). The co-localization of GPR64 with CFTR was confirmed by immunostaining and analyzed by Pearson’s correlation analysis (Figure 5D and 5E). The localization of CFTR/GPR64 with the apical membrane or microvilli was assayed by co-immunostaining with the marker ezrin (Figure 5G, Figure 8—figure supplement 2 and Figure 5—figure supplement 4).

As indicated correctly by the reviewer, we incorrectly included the CFTR blots upside down in our previous manuscript. We have corrected this accordingly and have indicated the mature CFTR band in Figure 5I and Figure 8M in the revised manuscript.

Figure 5

Figure 5A - Same comments as above.

5B–C - Same comments as above. (needs controls and better quantification).

Figure 5D-F. For the CFTR electrophys in this paper, the reversal potential is negative (unlike Figure 2 where it is 0). This is surprising given the predicted reversal potential for Cl- is ~0. Why is this? Did you change your conditions relative to Figure 2?

Thank you for your helpful suggestions. For co-localization, we have included a negative control (secondary antibody only in Figure 1—figure supplement 1 in the revised manuscript). We used an alternative chloride sensor in Figure 9A and 9B (Figure 5B in the original manuscript) for more accurate measurements. We performed new sets of electrophysiological recording experiments, and the reverse potentials of these measurements, which are summarized in Figure 9—figure supplement 6, almost completely followed Nernst function predictions

Figure 5G–L - Given the length of the paper, this in my opinion is too much. This data should be removed and saved for another paper.

Thank you for your helpful suggestions. These data are helpful for providing direct evidence of the coupling of GPR64 to Gs and Gq as well as how they function in fluid reabsorption in the efferent ductules.

Figure 6.

Figure 6B - Please spell out what ED, IS stand for on the figures. There's room. Also, please add arrows pointing to sperm accumulation etc.

Thank you for your helpful suggestions. We have included the full names of the structures and tissues used in the revised manuscript (Figure 11B). We have also added arrows pointing to sperm accumulation.

Supplement – many of the same concerns arise. The figures are very small, immunofluorescence doesn't have adequate controls and Western blots are too cropped.

Thank you for your helpful suggestions. We have included larger immunostained images and reduced cropping of the WB results in the supplemental data accordingly.

Discussion section – given all of the data presented. The discussion is very short. I would like to see more consideration of GPR64s physiological role as well as more discussion of CFTR regulation.

Thank you for your helpful suggestions. We have included a discussion of the physiological roles of GPR64 and CFTR regulation in the Discussion section.

Reviewer #2:

Gq activity and B-arrestin1-scaffolding are required for male fertility through mediating CPR64/CFTR coupling.

Overview: This is a comprehensive study evaluating the role for a CFTR/GPR64 complex in fluid reabsorption across efferent ducts of the testis – a function that is vital for male fertility. Further the authors interrogated the regulation of this complex by Gq and B-arrestin. For this most part- these studies are robust and convincing with respect to the role of GPR64 in this function. However- the role of CFTR in this function could use additional supportive data.

Thank you for your positive comments.

Specific comments:

Figure 1 focuses on the tubule fluid transport properties of efferent ductules- studied ex-vivo. These studies clearly show that disruption of GPR64 and Gq impairs fluid reabsorption, morphology and sperm count.

On the other hand- the role for CFTR is not clear as increases and decreases in activation (via kinase activators and inhibitors respectively) leads to the same effect of luminal swelling.

Thank you for your helpful comments. Our data have clearly identified an essential role for the Gq-PKC pathway in fluid reabsorption and the whole-cell Cl- current of the efferent ductules; however, Gs-PKA signaling is confounded. We therefore only drew conclusions regarding the role of Gq-PKC signaling in fluid reabsorption of the efferent ductules in the current manuscript.

According to our current data, we have demonstrated that the Gq protein is mainly expressed in GPR64-expressing cells, whereas the Gs protein is expressed in both GPR64-expressing and non-GPR64-expressing efferent ductule cells, as shown in Author response image 1.

Author response image 1
Expression and localization of Gs in different types of efferent ductule cells.

(A) The Gs was expressed in both GPR64-expressed cells and non-GPR64-expressed cells revealed by co-immunostaining. (B) The Gs expression in GPR64-promoter labeled cells and non GPR64-promoter labeled cells were examined by qRT-PCR after FACS. The results indicate that the similar mRNA level of Gs were detected in GPR64-promoter labeled cells and non GPR64-promoter labeled cells..

https://doi.org/10.7554/eLife.33432.049

The Gs-PKA signaling in GPR64-labeled cells facilitated the analysis of CFTR function in maintaining Cl- homeostasis, as supported by the inhibitory effects of the PKA inhibitor PKI14-22 and the Gs inhibitor NF-449 (Figure 2F–H). However, activation of Gs-PKA signaling in non-GPR64-pm-labeled cells may have also impaired fluid reabsorption but in a different manner via an unknown mechanism (Figure 2—figure supplement 2). Therefore, distinct functions of Gs-PKA signaling in different efferent ductule cells may account for the confounding observations. Both acute activation of the Gs-cAMP pathway by forskolin/IBMX and the inhibition of Gs by NF449 impaired fluid reabsorption, suggesting a delicate function for Gs in the efferent ductules. The function of Gs in the efferent ductules may be investigated in future studies, but this is not the focus of our current manuscript.

Figure 2: Shows that disruption of GPR64 alters the sensitivity of the ductular fluid transport to ion channel blockers (calcium transport proteins in addition of a CFTR channel blocker). They also show that chloride channel activity contributes to fluid transport. Their data falls somewhat short of proving that this activity is CFTR mediated. The regulatory properties of the conductance (i.e. regulation by PKA) was not shown and this is important. Also, while GlyH-101 is a well-known CFTR channel blocker- it is not specific and inhibits other chloride channels, including the SLC26A9 channel. The authors should include another inhibitor (i.e. CFTRinh-172) to test specificity.

Thank you for your helpful suggestions. We have performed new experiments to clarify that these effects, shown in Figure 4 (previous Figure 2), are CFTR-mediated. All GlyH-101 data were confirmed by the CFTR-specific inhibitor CFTR-Inh-172 and were compared with the data obtained with other Cl- transporter inhibitors (Figure 4B–4D, Figure 4I and K; Figure 6C–D and Figure 6—figure supplement 2). We used the specific CFTR inhibitor CFTR-Inh-172, the PKA inhibitor PKI14-22 and the cAMP activator Forskolin for electrophysiological measurements of our primary efferent ductule cells (Figure 6C–F and Figure 6—figure supplement 1 to Figure 6—figure supplement 3). We have also included the CFTR inhibitor CFTR-Inh-172 in the chloride measurements in Figure 5A in the revised manuscript (previous Figure 2F–H).

Figure 4 aims to show that Β-arrestin1 contributes to fluid reabsorption in these tubes and its expression promotes co-localization of GPR64 and CFTR. However, labeling of the diagrams in figure 4 seems somewhat confusing- does 4J show a line scan of GPR64 and CFTR localization or BArr1 and GPR64? The pattern of staining for GPR64 and CFTR looks similar but not overlapping- what compartments are the two proteins localized after BArr1 KO?

Thank you for your very helpful comments. Compared to other receptors, β-arrestins are abundant proteins and their co-localization with receptors in resting states are not easy to be assayed, due to the quality of the antibody. We therefore used co-immunoprecipitation experiments of efferent ductules to examine the interaction of GPR64 with arrestins (new Figure 5I in the revised manuscript). The β-arrestin-1 was co-immunoprecipitated with GPR64 in the efferent ductules derived from wild-type mice, but not GPR64-/Y mice (new Figure 5I and span class="jrnlFigRef" data-citation-string=" F5-S5 ">Figure 5—figure supplement 5 in the revised manuscript). Moreover, in the wild-type efferent ductules, both the GPR64 and CFTR are co-localized with Ezrin on the apical membrane (Figure 5G-5H and Figure 8—figure supplement 2). In the β-arrestin-1-/- mice, while the GPR64 still localized at the apical membrane, the CFTR was mislocalized relative to GPR64 and Ezrin in the efferent ductules (Figure 8E–8L). The disruption of the interaction of GPR64 with CFTR in the β-arrestin-1-/- mice was also confirmed by co-immunoprecipitation results (Figure 8M in the revised manuscript).

Figure 5 shows the consequence of co-expressing CFTR with GPR64 Wt or mutants bearing substitution in intracellular loops 2 and 3. This is a comprehensive set of studies supporting the role for these loops in mediating functional interaction between GRP64 and CFTR. However- an important control would include single transfections with GPR64 (no CFTR) to ensure that it is not modulating a distinct chloride channel. Western blotting to ensure expression of the each of the mutant GPR64 proteins would also be helpful to support the conclusion that there are site specific effects in the interaction- rather than reporting differences in protein abundance.

Thank you for your helpful suggestions. We have included new data for GPR64 transfected alone in in vitro electrophysiological experiments and chloride concentration measurements in the revised manuscript (Figure 9A and C–E). We have also included the expression levels of different GPR64 mutations in Figure 10—figure supplement 2.

Reviewer #3:

The manuscript under review considers interaction between GPR64, CFTR and arrestins in non-cilliated cells of the male reproductive system. In general, the content of this study is interesting and relevant. However, a number of significant technical concerns with the presented study raise doubts about key results.

Thank you for your positive comments.

Major points:

Several studies have shown that GlyH-101 is not entirely specific for CFTR, for instance at 50 microM GlyH-101 inhibits SLC26A9. A panel of CFTR inhibitors, including CFTRinh172 should have been considered for studies presented in Figure 2. In addition, qPCR or similar should be employed to unambiguously determine that SLC26A9 is not a relevant player in cellular system under consideration.

Thank you for your helpful suggestions. We have performed new experiments using CFTR-Inh-172 to specifically block CFTR in all studies related to CFTR functions (Figure 4K, Figure 5A, Figure 6C–D, Figure 9A and Figure 9C–E in the revised manuscript). We have performed qRT-PCR experiments, and the results indicated that SLC26A9 is highly expressed in GPR64-promoter-labeled cells (Figure 4A in the revised manuscript). However, DIDS, an inhibitor known to inhibit SLC26A9, had little effect on the inhibition of fluid reabsorption in the efferent ductules (Figure 4I), whereas the CFTR inhibitor CFTR-Inh-172 exerted prominent effects (Figure 4K). Further electrophysiological experiments support this conclusion (Figure 6C–D). Therefore, our results suggested that CFTR is the major player downstream of GPR64 for the regulation of fluid reabsorption in the efferent ductules. SLC26A9 may also contribute but potentially plays a more minor role. We have included these considerations in our Discussion section in the revised manuscript.

In terms of the intracellular measurements of [Cl-], MQAE is essentially completely insensitive to [Cl-] above 100 mM, so it is hard to see how intracellular [Cl-] could be determined to be ~140 mM. In this regard, it is notable that the calibration curve for MQAE versus [Cl-] presented in the supplementary data is only extended to ~80 mM [Cl-] (i.e., well below the reported value). As such, it is difficult to have any confidence in the presented values of [Cl-].

Thank you for your helpful suggestions. We have measured Cl- concentrations using a new sensor, mCIY-8M (Figure 5—figure supplement 1). We have compared mCIY-H148Q and newly reported 8-mutations of mCIY, which show slightly better signals and calibration curves. We therefore used mCIY-8M to measure Cl- concentration changes (Jayaraman et al., 2000; Zhong et al., 2014). The new experimental results are presented in revised Figure 5A, Figure 7A and Figure 9A–B).

In terms of the GlyH-101 studies presented in Figure 2F, do driving forces predict that CFTR inhibition would mediate accumulation of cytoplasmic [Cl-]? What are the consequences of such inappropriately high intracellular [Cl-]? What happens to the concentration of cations, and the membrane potential of cells with such non-physiological levels of [Cl-]?

Thank you for your helpful comments. With the new Cl- indicator mCIY-8M, the Cl- concentration was more accurately measured and was much lower than previously measured. The changes in Cl- and pH homeostasis may account for the fluid reabsorption dysfunction, and we did observe changes in the efferent ductule volume in ligation experiments (Figure 2 and Figure 4B–K). Moreover, ATGR2 activation, which reportedly restores pH homeostasis, rescued the fluid reabsorption dysfunction (Figure 4L–M). These results support the critical roles of Cl- and pH homeostasis in fluid reabsorption in the efferent ductules.

In terms of the pH measurements presented in Figure 2, sensitivity of BCECF to determine pH above ~pH 7.8 is limited. In addition, no calibration curve is presented – it is critical to demonstrate that the employed method would accurately determine the reported pKa for the fluorescent probe employed, i.e., BCECF, to have confidence that a pH of 8.4 could be accurately determined. The presented methods are unclear, for instance a 25 mM bicarbonate solution would require gassing with 5% CO2, but this is not mentioned. The description of what was measured in the Results was vague, inner solution of efferent ductules does not imply that cytoplasmic pH was determined.

Thank you for your helpful comments. We have used a new pH sensor, carboxy SNARF-1, to determine the pH under different conditions presented in the manuscript (Figure 5—figure supplement 3) (Buckler and Vaughanjones, 1990; Thornell et al., 2017). A calibration curve is presented in Figure 5—figure supplement 3. For the 25 mM bicarbonate solution, we did perform gassing with 5% CO2.

We have revised the Materials and methods section accordingly.

In terms of the patch clamp analysis presented in Figure 2, prior studies by Muanprasat and colleagues have demonstrated that GlyH-101 inhibition of CFTR is strongly dependent on membrane potential – as would be expected for a charged molecule with a pore occluding mechanism of action. As such, GlyH-101 inhibition alters CFTR current-voltage curves from being linear to showing inward rectification. This is apparently not observed in the data presented in Figure 2L (where I-V curve remains linear). There is a concentration dependence of this phenomenon, however, I was unable to find information about how much GlyH-101 was used in the Legend or Material and Methods section for the presented data. In general, most studies used 25 microM GlyH-101. If this concentration was used in Figure 2, then inward rectification of CFTR I-V relationship would definitely be anticipated. Consideration of submaximal concentrations of GlyH-101, per Muanprasat and colleagues, should be considered to provide confidence that CFTR currents are really being observed in the reported data. In addition, delivery of PKA (in the pipette) is typically used for excised patch data, as such, experimental data should be presented for whole cell recordings with consideration of an alternative CFTR stimulant such as forskolin. In addition, for patch clamp data presented in Figure 3D, pharmacological validation that currents are CFTR -dependent should be presented.

Thank you for your helpful comments. In the GPR64 promoter-labeled wild-type efferent ductule cells, we observed an outwardly-rectifying whole-cell Cl- current (IGPR64-ED), which was significantly diminished in response to substitution of the bath Cl- solution with Gluc- (148.5 mM Cl- was replaced by 48.5 mM Cl- and 100 mM Gluc-) (Figure 6A–B). The change in the reversal potential (Erev) followed the Nernst equation (Figure 6B and Figure 6—figure supplement 1 to Figure 6—figure supplement 3). However, the whole-cell Cl- current of the GPR64 promoter-labeled efferent ductule cells derived from GPR64-/Y mice was significantly lower than that of cells derived from their wild-type littermates, which showed a slight difference in response to substitution of the bath Cl- solution with Gluc- (Figure 6A–B). These results suggested that GPR64 deficiency in the efferent ductules significantly reduced the whole-cell Cl- current of GPR64 promoter-labeled non-ciliated cells. Moreover, we characterized the electrophysiological measurements by applying the specific CFTR inhibitor CFTR-Inh-172 and forskolin in primary GPR64 promoter-labeled efferent ductule cells derived from wild-type and GPR64-/Y mice (new Figure 6A–F). In response to CFTR-Inh-172 inhibition, we did observe an inwardly-rectifying IGPR64-ED current (green color in Figure 6C) instead of the original outwardly-rectifying current (Figure 6A). The forskolin+IBMX data are shown in Figure 6E–6F. We added new data to previous Figure 3D to include the PKC inhibitor Ro 31-8220 (new Figure 7C–E in the revised manuscript). Application of these selective pharmacological blockers supported the notion that the observed whole-cell Cl- current of GPR64 promoter-labeled efferent ductule cells was mediated by CFTR.

For the data presented in Figure 5, does stimulation of CFTR with an alternative agonist, such as forskolin, mediate similar cytoplasmic chloride concentration reduction? Molecular details of how GPR64 is activated have been elucidated and are considered by the authors in the supplemental data, for instance in regard to GPR64beta elevating of cAMP. Does the carboxy-terminal fragment of GPR64 also reduce cytoplasmic [Cl-]?

Thank you for your helpful comments. We have measured the Cl- by overexpressing CFTR and stimulating with forskolin (FSK) (Figure 9A). The results indicated that the application of FSK significantly lowered the intracellular Cl- in HEK293 cells transfected with CFTR. The results obtained with GPR64-C-terminal (GPR64β) co-expression caused more significant Cl- reduction in HEK293 cells compared with that in CFTR/GPR64 full-length co-transfected cells (Figure 9—figure supplement 5).

In Figure 2I–J, the mere co-localization of two proteins to a membrane determined imaged by confocal microscopy does not indicate or suggest that a complex with functional coupling exists. It indicates that two proteins are targeted to the same membrane. This same concern is relevant for data presented in Figure 3, Figure 4, and Figure 5.

Thank you for your helpful comments. In addition to co-localization, we also performed co-immunoprecipitation experiments. The results are shown in Figure 5I, Figure 5—figure supplement 5, Figure 8M and Figure 8—figure supplement 2B in the revised manuscript.

For co-IP experiments shown in Figure 4K, there is insufficient explanation (in Methods, Results, Legend etc.) to comprehend what is being done. By elimination, I assume the anti-HA blot was against arrestins, but, this is not detailed (for instance, I cannot find details of HA-tagged arrestin constructs).

Thank you for these helpful suggestions. We have revised the labeling of each lane in the new Figure 8—figure supplement 3 (previous Figure 4K) and the corresponding figure legends accordingly.

[Editors' note: the author responses to the re-review follow.]

The manuscript has been improved but there are some remaining problematic issues that need to be addressed, as outlined below:

1) The intracellular Cl- measurements remain problematic due to the lack of a radiometric method. These measurements should be either replaced by suitable radiometric or electrophysiological measurements or removed from the manuscript.

We thank the reviewer for these helpful suggestions. The Cl- measurements and the corresponding statements were removed from the current manuscript (Figure 5A, Figure 7A, Figure 8C, Figure 9A–9B, Figure 5—figure supplement 1 and Figure 9—figure supplement 5 in the previous version of the manuscript).

2) Western blots need controls and markers.

Thank you for these helpful suggestions. The controls and markers of Figure 5H, Figure 9L, Figure 2—figure supplement 1, Figure 8—figure supplement 3, Figure 9—figure supplement 1, Figure 9—figure supplement 3, Figure 10— figure supplement 2–4 and Figure 11—figure supplement 2 in the revised manuscript were added accordingly

3) The CFTR currents are extremely small. The i/v shift with gluconate is too small and not typical for CFTR. Original whole cell overlay currents or continuous recordings should be shown. What is the proof in addition to CFTRinh172 that the authors truly measured CFTR currents?

We thank the reviewer for these helpful suggestions. A continuous recording of the whole-cell Cl- current in response to Cl- concentration changes is shown in Figure 6A–6B. In addition to the pharmacological intervention by the application of CFTRinh172, we knocked down the CFTR expression in efferent ductules using si-RNA; this knockdown was verified by qRT-PCR (Figure 7C). The genetically reduced CFTR expression significantly reduced the Cl- current measured in GPR64 promoter-labeled cells compared to that in control si-RNA-treated efferent ductule cells (Figure 7D-7E). Both the pharmacological intervention and genetic approach supported that CFTR mediated the observed Cl- currents. Additionally, by checking more detail in the published literature, we found that the magnitude of CFTR current is sperm is similar to our observation in efferent ductules (Figueiras-Fierro et al., 2013).

4) Figure 4A suggests a GPR64-dependent expression of CFTR-mRNA. How does this fit to the expression data shown in Figure 5D/I and to the data shown in Figure 6?

We thank the reviewer for this comment. The qRT-PCR result in Figure 4A indicates that both CFTR and GPR64 mRNA are highly enriched in GPR64 promoter-labeled efferent ductules. Two potential mechanisms may explain this observation. (1) The first mechanism is that CFTR expression is dependent on GPR64 expression. However, in GPR64 knockout mice, the CFTR expression level did not change significantly (Figure 5H in the revised manuscript). Furthermore, when we compared the GPR64 promoter-labeled efferent ductule cells derived from the GPR64-/Y mice with their wild-type littermates, the CFTR mRNA level did not change significantly despite the significant difference in the GPR64 mRNA expression level (Figure 4—figure supplement 1D in the revised manuscript, new data). Therefore, CFTR expression is not dependent on GPR64 expression. (2) The second mechanism is that both GPR64 and CFTR were specifically expressed in non-ciliated cells in efferent ductules. GPR64 and CFTR may share a similar transcriptional or epigenetic regulatory mechanism. We added these considerations to the revised manuscript (subsection “ADGRG2 and CFTR coupling in the efferent ductules and its function in fluid reabsorption”).

In addition, we used GPR64-promoter-RFP to label the non-ciliated cells in efferent ductules, which, compared to ciliated cells (non-GRP64-RFP-labeled cells), bear distinct morphological properties (Figure 1B); we also changed the label in Figure 4A to clarify the description.

5) Figure 10: In what cells were these data obtained? No control for expression of the various GPR64 mutants is provided.

We thank the reviewer for his helpful suggestion. The experiments of Figure 10 (Figure 11 in the revised manuscript) were performed using HEK293 cells. We have added corresponding statement in subsection “Molecular determinants of ADGRG2coupling with G protein subtypes and their contribution to the regulation of CFTR activity in vitro” in the revised manuscript. The expression control was added in Figure 11—figure supplement 2 accordingly.

https://doi.org/10.7554/eLife.33432.051

Article and author information

Author details

  1. Dao-Lai Zhang

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    3. School of Pharmacy, Binzhou Medical University, Yantai, China
    Contribution
    Data curation, Software, Formal analysis, Funding acquisition, Investigation, Visualization, Methodology, Writing—original draft
    Contributed equally with
    Yu-Jing Sun, Ming-Liang Ma and Yi-jing Wang
    Competing interests
    No competing interests declared
    ORCID icon 0000-0003-4428-8731
  2. Yu-Jing Sun

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Data curation, Software, Formal analysis, Investigation, Visualization, Methodology, Writing—original draft
    Contributed equally with
    Dao-Lai Zhang, Ming-Liang Ma and Yi-jing Wang
    Competing interests
    No competing interests declared
  3. Ming-Liang Ma

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Software, Investigation, Visualization, Methodology
    Contributed equally with
    Dao-Lai Zhang, Yu-Jing Sun and Yi-jing Wang
    Competing interests
    No competing interests declared
  4. Yi-jing Wang

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Software, Investigation, Visualization, Methodology
    Contributed equally with
    Dao-Lai Zhang, Yu-Jing Sun and Ming-Liang Ma
    Competing interests
    No competing interests declared
  5. Hui Lin

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Software, Investigation, Methodology
    Competing interests
    No competing interests declared
  6. Rui-Rui Li

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
  7. Zong-Lai Liang

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
  8. Yuan Gao

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  9. Zhao Yang

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Formal analysis
    Competing interests
    No competing interests declared
  10. Dong-Fang He

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Supervision
    Competing interests
    No competing interests declared
  11. Amy Lin

    Department of Biochemistry, School of Medicine, Duke University, Durham, United States
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon 0000-0001-6723-5443
  12. Hui Mo

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  13. Yu-Jing Lu

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  14. Meng-Jing Li

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  15. Wei Kong

    Key Laboratory of Molecular Cardiovascular Science, Department of Physiology and Pathophysiology, School of Basic Medical Sciences, Peking University, Beijing, China
    Contribution
    Methodology
    Competing interests
    No competing interests declared
  16. Ka Young Chung

    School of Pharmacy, Sungkyunkwan University, Suwon, Korea
    Contribution
    Methodology
    Competing interests
    No competing interests declared
  17. Fan Yi

    Department of Pharmacology, Shandong University School of Medicine, Jinan, China
    Contribution
    Methodology
    Competing interests
    No competing interests declared
  18. Jian-Yuan Li

    Key Laboratory of Male Reproductive Health, National Research Institute for Family Planning, National Health and Family Planning Commission, Beijing, China
    Contribution
    Methodology
    Competing interests
    No competing interests declared
  19. Ying-Ying Qin

    National Research Center for Assisted Reproductive Technology and Reproductive Genetics, Shandong University, Jinan, China
    Contribution
    Methodology
    Competing interests
    No competing interests declared
  20. Jingxin Li

    Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Methodology
    Competing interests
    No competing interests declared
  21. Alex R B Thomsen

    Department of Biochemistry, School of Medicine, Duke University, Durham, United States
    Contribution
    Data curation
    Competing interests
    No competing interests declared
  22. Alem W Kahsai

    Department of Biochemistry, School of Medicine, Duke University, Durham, United States
    Contribution
    Data curation
    Competing interests
    No competing interests declared
  23. Zi-Jiang Chen

    National Research Center for Assisted Reproductive Technology and Reproductive Genetics, Shandong University, Jinan, China
    Contribution
    Resources
    Competing interests
    No competing interests declared
  24. Zhi-Gang Xu

    Shandong Provincial Key Laboratory of Animal Cells and Developmental Biology, Shandong University School of Life Sciences, Jinan, China
    Contribution
    Data curation
    Competing interests
    No competing interests declared
  25. Mingyao Liu

    1. Shanghai Key Laboratory of Regulatory Biology, School of Life Sciences, Institute of Biomedical Sciences, East China Normal University, Shanghai, China
    2. Department of Molecular and Cellular Medicine, Institute of Biosciences and Technology, Texas A&M University Health Science Center, Houston, United States
    Contribution
    Resources
    Competing interests
    No competing interests declared
  26. Dali Li

    Shanghai Key Laboratory of Regulatory Biology, School of Life Sciences, Institute of Biomedical Sciences, East China Normal University, Shanghai, China
    Contribution
    Resources
    For correspondence
    dlli@bio.ecnu.edu.cn
    Competing interests
    No competing interests declared
  27. Xiao Yu

    Department of Physiology, Shandong University School of Medicine, Jinan, China
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Methodology, Project administration
    For correspondence
    yuxiao@sdu.edu.cn
    Competing interests
    No competing interests declared
  28. Jin-Peng Sun

    1. Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, Shandong University School of Medicine, Jinan, China
    2. Department of Biochemistry, School of Medicine, Duke University, Durham, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Methodology, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    sunjinpeng@sdu.edu.cn
    Competing interests
    No competing interests declared
    ORCID icon 0000-0003-3572-1580

Funding

National Natural Science Foundation of China (31470789)

  • Jin-Peng Sun

National Natural Science Foundation of China (31611540337)

  • Ka Young Chung
  • Jin-Peng Sun

National Natural Science Foundation of China (81773704)

  • Jin-Peng Sun

Shandong Natural Science Fund for Distinguished Young Scholars (JQ201517)

  • Jin-Peng Sun

Shandong Provincial Natural Science Foundation (ZR2014CP007)

  • Dao-Lai Zhang

National Natural Science Foundation of China (31671197)

  • Xiao Yu

The Program for Changjiang Scholars and Innovative Research Team in University (IRT13028)

  • Xiao Yu

National Natural Science Foundation of China (31471102)

  • Xiao Yu

National Science Fund for Distinguished Young Scholars (81525005)

  • Fan Yi

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Ethics

Animal experimentation: Mice were individually housed in the Shandong university on a 12:12 light:dark cycle with access to food and water ad libitum.The use of mice was approved by the animal ethics committee of Shandong university medical school (protocol LL-201502036). All animal care and experiments were reviewed and approved by the Animal Use Committee of Shandong University School of Medicine.

Reviewing Editor

  1. Michel Bagnat, Reviewing Editor, Duke University, United States

Publication history

  1. Received: November 8, 2017
  2. Accepted: January 30, 2018
  3. Accepted Manuscript published: February 2, 2018 (version 1)
  4. Version of Record published: March 6, 2018 (version 2)

Copyright

© 2018, Zhang et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 813
    Page views
  • 154
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Biochemistry and Chemical Biology
    2. Chromosomes and Gene Expression
    Minxian Qian et al.
    Research Article Updated
    1. Biochemistry and Chemical Biology
    James Attwater et al.
    Research Article