PAQR10 (progestin and adipoQ receptor 10) is a Golgi-localized protein that is able to enhance the retention and activation of Ras proteins in the Golgi apparatus, subsequently leading to a sustained ERK (extracellular-signal-regulated kinase) signalling. However, little is known about the topology and functional domains of PAQR10. In the present study, we extensively dissected and analysed the structure of PAQR10. The topology analysis reveals that PAQR10 is an integral membrane protein with its N-terminus facing the cytosol. Multiple domains, including the membrane-proximal region at the N-terminus, the membrane-proximal region at the C-terminus and the three loops facing the cytosol, were found to be required for PAQR10 to reside in the Golgi apparatus, to stimulate ERK phosphorylation and to tether Ras to the Golgi apparatus. Furthermore, when PAQR10 was artificially forced to be expressed in the endoplasmic reticulum, it could neither mobilize Ras to the Golgi apparatus nor increase ERK phosphorylation. Finally, the PAQR10 mutants that lost Golgi localization failed to promote differentiation of PC12 cells. Collectively, the results of the present study indicate that Golgi localization is indispensable for PAQR10 to implement its regulatory functions in the Ras signalling cascade.

INTRODUCTION

The PAQR (progestin and adipoQ receptor) family was first identified in 2005, defined by an ancient seven-transmembrane pass motif [1]. This family consists of 11 members (PAQR1– PAQR11) in the human genome, and all of them share significant sequence identity with ancient bacterial haemolysin-like proteins, suggesting that they play important roles during evolution [1]. All members of the PAQR family are predicted to contain seven TMs (transmembrane domains) with an N-terminus facing the cytosol, with a topology distinct from the traditional GPCR (G-protein-coupled receptor) proteins [2]. The subcellular localization and physiological functions of some members of the PAQR family have been characterized. PAQR1 and PAQR2 are adiponectin receptors predominantly localized on the plasma membrane [3]. Adiponectin is an adipocyte-derived cytokine that plays an important role in glucose and lipid metabolism [4]. PAQR5, PAQR7 and PAQR8 were reported to serve as receptors for progesterone and might play important roles in ovarian physiology [5,6]. Our group revealed previously that PAQR3 was exclusively localized in the Golgi apparatus [7,8]. PAQR3 was able to inhibit Ras signalling by sequestrating Raf kinase to the Golgi apparatus and was renamed RKTG (Raf kinase trapping to Golgi). Owing to the negative role of PAQR3/RKTG on Ras to ERK (extracellular-signal-regulated kinase) signalling, PAQR3/RKTG has been found to function as a tumour suppressor in vivo [911]. PAQR3/RKTG also plays an important role in angiogenesis and is significantly down-regulated in human clear cell renal cell carcinomas [12].

Recently, our group demonstrated that PAQR10 and PAQR11, two highly homologous proteins within the PAQR family, are exclusively localized in the Golgi apparatus [13]. Furthermore, we found that PAQR10 and PAQR11 play a pivotal role in the spatial regulation of Ras signalling [13]. Overexpression of PAQR10/PAQR11 can stimulate basal and EGF (epidermal growth factor)-induced ERK phosphorylation and markedly elevate Golgi localization of H-Ras, N-Ras and K-Ras4A, but not K-Ras4B. Consistently, PAQR10 and PAQR11 are able to interact with H-Ras, N-Ras and K-Ras4A, but not K-Ras4B. The increased Ras protein in the Golgi apparatus by overexpression of PAQR10/PAQR11 is in an active state. In addition, PAQR10 and PAQR11 are able to interact with RasGRP1 (Ras guanine-nucleotide-releasing protein 1), a guanine-nucleotide-exchange protein of Ras, and increase Golgi localization of RasGRP1. The simulation of ERK phosphorylation by overexpressed PAQR10/PAQR11 is abrogated by down-regulation of RasGRP1. Furthermore, differentiation of PC12 cells is significantly enhanced by overexpression of PAQR10/PAQR11. Collectively, this study revealed a new paradigm of spatial regulation of Ras signalling in the Golgi apparatus by PAQR10 and PAQR11 [13].

The Golgi apparatus is a central organelle that provides a major site for protein modification and vesicular transport in eukaryotic cells [14]. Unlike the ER (endoplasmic reticulum), there is, at present, no evidence for soluble proteins residing in the Golgi lumen, but instead all of the proteins in the Golgi apparatus are either integral membrane proteins or membrane-peripheral proteins [15]. The Golgi-targeting signals of these proteins have been found in all three domains, i.e. extracytoplasmic, transmembrane and cytoplasmic domains [15,16]. Besides its classical roles in protein modification and sorting, the Golgi apparatus has been reported to be involved in signal transduction pathways traditionally thought to emanate from the plasma membrane. For example, numerous studies have demonstrated that Golgi-localized Ras proteins could actively trigger the MAPK (mitogen-activated protein kinase) signalling pathway and eventually induce cell differentiation [17,18]. Our previous study also revealed that PAQR10 and PAQR11 are actively involved in Golgi-localized Ras signalling [13]. In the present study, we investigated the topology and structural domains of PAQR10 and identified the critical motifs required for Golgi localization and modulation in Ras signalling.

EXPERIMENTAL

Cell culture

HEK (human embryonic kidney)-293T and HeLa human cervical carcinoma cells were cultured in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% (v/v) FBS (fetal bovine serum) (Biochrom), penicillin (100 units/ml; Gibco-BRL) and streptomycin (100 μg/ml; Gibco-BRL). Cells were maintained at 37°C in 5% CO2 in humidified air and subcultured every 2 or 3 days.

Cell transfection

HeLa cells were seeded in 12-well plates at a density of 2×105 cells/ml. After culturing for 16–24 h, the cells were switched to fresh DMEM without serum and transfected with plasmids using the PEI (polyethyleneimine) method [19] with PEI at a molecular mass of 25000 Da (Polysciences). HEK-293T cells were transfected using the calcium phosphate precipitation method.

Plasmids and antibodies

The HA (haemagglutinin)-tagged TGN38 (trans-Golgi network protein 38) was kindly provided by Dr Juan S. Bonifacino (National Institute of Child Health and Human Development, Bethesda, MD, U.S.A.) [20]. The full-length cDNA of human PAQR10 was isolated as reported previously [13]. The wild-type human PAQR10 and its deletion mutants used in the present study were cloned into pRc/CMV-FLAG vectors to fuse with one FLAG epitope tag at the N-terminus. All PAQR10 mutants were generated by a PCR-based method and were confirmed by DNA sequencing. H-Ras was cloned into pEGFP-C1 (Clontech) to fuse with an EGFP (enhanced green fluorescent protein) at the N-terminus [13]. PAQR10-KDEL plasmid was constructed by adding nucleotide AAA (encoding lysine), GAT (encoding aspartate), GAG (encoding glutamate) and TTG (encoding leucine) sequences to the C-terminal end of PAQR10 immediately before the stop codon by a PCR-based method. The chimaeras between PAQR10 and PAQR1/AdipoR1 (adiponectin receptor 1) were generated by swapping the three putative cytoplasmic loops between the two proteins. The swapped regions for PAQR10 were those between TM2 and TM3, between TM4 and TM5, and between TM6 and TM7 as shown in Figure 2(A). The swapped regions for PAQR1/AdipoR1 corresponded to the amino acid sequences TVYCHSEKVSRTFSKLDY (for loop1), RFATPKHRQTR (for loop2) and AARIPERFFPGKFD (for loop3) respectively. The antibodies were as follows: monoclonal anti-FLAG antibody, anti-calnexin antibody and anti-tubulin antibody were from Sigma–Aldrich; antibodies against HA and total ERK1/2 were from Santa Cruz Biotechnology; anti-phospho-ERK1/2 antibody was from Cell Signaling Technology; anti-catalase antibody was from Abcam; anti-golgin-97 monoclonal antibody, Alexa Fluor® 488-conjugated donkey anti-(mouse IgG), Alexa Fluor® 546-conjugated goat anti-(rabbit IgG), Alexa Fluor® 546-conjugated goat anti-(mouse IgG) and Hoechst 33342 were from Molecular Probes; and Cy5 (indodicarbocyanine)-labelled goat anti-mouse IgG was from Jackson ImmunoResearch Laboratories. The polyclonal anti-PAQR10 antibody was generated as reported previously [13].

Solubility assay of PAQR10

The solubility assay was carried out and analysed as described previously [8] with the following modifications. Briefly, after transfection for 24 h, HEK-293T cells were lysed in a buffer (0.25 M sucrose, 10 mM Tris/HCl, pH 8.0, and 0.1 mM EDTA), and homogenized by ten strokes in a Dounce homogenizer. The samples were kept on ice throughout the duration of the procedure. Then the cells were centrifuged at 750 g for 15 min at 4°C to remove nuclei and cell debris. The supernatants were then centrifuged at 43000 rev./min for 30 min at 4°C using an S140-AT rotor in a Hitachi CS-150 GX ultracentrifuge to separate the cellular membrane and cytoplasmic fractions. The pellet fractions were subjected to treatment with different solvents (including 0.1 M Na2CO3, pH 11.5, 1 M NaCl and 1% SDS or 1% Triton X-100) and then centrifuged at 43000 rev./min for 30 min at 4°C using an S140-AT rotor in a Hitachi CS-150 GX ultracentrifuge. The supernatant and pellet were analysed by Western blotting.

Western blotting

After separation by SDS/PAGE (10% gels), the proteins were transferred on to a PVDF membrane. Blots were blocked overnight at 4°C in blocking solution [5% (w/v) dried skimmed milk powder in TBST (Tris-buffered saline with Tween 20: 20 mM Tris/HCl, pH 7.4, 150 mM NaCl and 0.1% Tween 20)] and the proteins were detected using different primary antibodies [anti-FLAG, 1:10000 dilution; anti-tubulin, 1:5000 dilution; anti-calnexin, 1:1000 dilution; anti-catalase, 1:1000 dilution; anti-golgin-97, 1:1000 dilution; anti-phospho-ERK1/2, 1:1000 dilution; and anti-(total ERK1/2), 1:3000 dilution], followed by a horseradish-peroxidase-conjugated secondary antibody (GE Healthcare) at a dilution of 1:5000 and enhanced chemiluminescence reagents from Pierce.

Selective permeabilization of plasma membranes and immunofluorescence microscopy

At 48 h after transfection, HeLa cells were washed twice with PBS. For complete permeabilization, the cells were fixed with 4% (w/v) paraformaldehyde in PBS for 10 min, permeabilized with 0.1% Triton X-100 in PBS for 10 min, blocked with 3% (w/v) BSA in PBS for 1 h and then incubated with primary antibody for 1 h (antibodies were diluted as follows: anti-HA, 1:1000; anti-golgin-97, 1:200; and anti-FLAG, 1:5000). After washing three times, the cells were incubated with secondary antibodies (the dilution of each secondary antibody was 1:500) for 1 h and then washed three times. Selective permeabilization of the plasma membrane was performed as described previously [8]. Briefly, cells were incubated at 4°C for 45 min in sucrose buffer (1% BSA, 0.3 M sucrose, 0.1 M KCl, 2.5 mM MgCl2, 1 mM EDTA and 10 mM Hepes at pH 7.4) containing various concentrations of digitonin, and then incubated with primary antibody in blocking solution for 1 h, followed by washing with PBS. The cells were fixed with methanol at 4°C for 6 min, incubated in blocking solution for 30 min and then incubated with secondary antibodies for 30 min in blocking solution. The nucleus was stained with 1 μg/ml Hoechst 33342 for 10 min. Confocal images were captured using an LSM 510 confocal microscope with a ×64 1.4 NA (numerical aperture) apochromat objective (Carl Zeiss). An argon laser was used at 488 nm for fluorescence excitation of EGFP and Alexa Fluor® 488-conjugated antibodies. A helium/neon laser (543 nm) was used for excitation of Alexa Fluor® 546-conjugated antibodies, a helium/neon laser (350 nm) was used for excitation of Cy5-labelled antibodies, and a two-photon laser (720 nm) was used for excitation of Hoechst 33342. After data acquisition, RGB (red/green/blue) images were processed using LSM 510 software.

PC12 cell differentiation

The experiment of PC12 cell differentiation was performed as described previously [17]. In brief, PC12 cells were cultured in RPMI 1640 medium containing 10% (v/v) horse serum and 5% (v/v) FBS, and co-transfected with pEGFP vector and a 5-fold excess of PAQR10 mutants using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. Neurite outgrowth was measured 48 h later and quantified as the number of transfected cells (marked by EGFP expression) with at least one neurite with a length equal to or longer than the cell body diameter.

Statistic analysis

Student's t test was used for all statistical analysis.

RESULTS

Topological analysis of PAQR10

We reported previously that PAQR10 was predicted to contain seven TMs by hydrophobicity analysis, similar to the topological structure of other identified members in PAQR family [1,3,7,13]. To analyse whether PAQR10 is indeed an integral membrane protein, we prepared membrane fractions and treated with different reagents to discriminate between transmembrane association and peripheral membrane association. According to previous papers [8,21,22], 1 M NaCl (high concentrate salt solvent) and 0.1 M Na2CO3 at pH 11.5 (high pH solvent) were able to dissolve peripheral proteins, whereas 1% SDS or 1% Triton X-100 could release both peripheral and integral membrane proteins. As shown in Figure 1(A), PAQR10 remained in the particulate fraction after treatment with 1 M NaCl or 0.1 M Na2CO3, but was released into the soluble fraction upon treatment with 1% SDS or 1% Triton X-100, a pattern similar to that of calnexin, which is a type I integral membrane protein [23,24]. As a control, tubulin, a cytosolic protein, was found to exclusively remain in the supernatant fraction of homogenate. Taken together, these results indicate that PAQR10 is an integral membrane protein.

Topological analysis of PAQR10

Figure 1
Topological analysis of PAQR10

(A) PAQR10 is a integral membrane protein. HEK-293T cells were transiently transfected with FLAG-tagged PAQR10 for 24 h and then homogenized. The membrane fractions were extracted separately with the indicated buffers, centrifuged to obtain the supernatant and pellet fractions and then analysed by Western blotting with the antibodies indicated. (B) The N-terminus of PAQR10 faces the cytosol. HeLa cells were transiently transfected with the plasmids indicated. After transfection for 48 h, cells were used for selective permeabilization treatment as indicated, followed by immunofluorescence analysis. The cells were stained with respective antibodies (red) and the cell nucleus was stained with Hoechst 33342 (blue). Scale bar, 20 μm.

Figure 1
Topological analysis of PAQR10

(A) PAQR10 is a integral membrane protein. HEK-293T cells were transiently transfected with FLAG-tagged PAQR10 for 24 h and then homogenized. The membrane fractions were extracted separately with the indicated buffers, centrifuged to obtain the supernatant and pellet fractions and then analysed by Western blotting with the antibodies indicated. (B) The N-terminus of PAQR10 faces the cytosol. HeLa cells were transiently transfected with the plasmids indicated. After transfection for 48 h, cells were used for selective permeabilization treatment as indicated, followed by immunofluorescence analysis. The cells were stained with respective antibodies (red) and the cell nucleus was stained with Hoechst 33342 (blue). Scale bar, 20 μm.

We next investigated whether the N-terminus of PAQR10 faces the cytosol, similar to the topology of PAQR3/RKTG [8]. We expressed the epitope-tagged PAQR10 protein in HeLa cells and investigated epitope accessibility in selectively permeabilized cells. At 48 h after transfection, the cells were subjected to immunofluorescence analysis under one of three conditions: no permeabilization (in the absence of detergent), selective permeabilization by digitonin treatment or complete permeabilization by Triton X-100 treatment. As a control, we used a Golgi integral membrane protein TGN38 that has its N-terminus facing the Golgi lumen [25]. As expected, TGN38 with an HA tag at the N-terminus clearly displayed plasma membrane localization without detergent treatment (Figure 1B, left-hand panel). Our pilot study indicated that digitonin at a concentration below 3 μg/ml specifically permeabilized the plasma membrane without disrupting membranes of intracellular organelles. Under conditions of 1 μg/ml digitonin with the plasma membrane, but not Golgi membrane, being permeabilized, TGN38 had a punctate staining pattern in the plasma membrane and cytosol, but not in the Golgi apparatus (Figure 1B, left-hand panel). However, when both the plasma membrane and Golgi membrane were permeabilized with 0.1% Triton X-100, TGN38 showed a clear Golgi localization pattern (Figure 1B, left-hand panel). Without detergent treatment, we did not detect any fluorescence signal of PAQR10 protein using an antibody against its N-terminus-tagged FLAG epitope. Under conditions of 1 μg/ml digitonin, PAQR10 clearly displayed a Golgi-localized pattern (Figure 1B, middle panel). The Golgi localization of PAQR10 was also found when 0.1% Triton X-100 was applied to permeabilize both plasma and Golgi membranes (Figure 1B, middle panel). As another control, golgin-97, a peripheral membrane protein located at the cytoplasmic side of the Golgi membrane, could be detected in the Golgi apparatus when either digitonin or Triton X-100 was used (Figure 1B, right-hand panel). Collectively, these results reveal that PAQR10 is an integral membrane protein with its N-terminus facing the cytosol.

Structural determinants of PAQR10 at the N-terminus required for Golgi localization

We analysed the protein sequence of PAQR10 and predicted the seven-transmembrane spans by using the program TMpred (http://www.ch.embnet.org/software/TMPRED_form.html). The predicted topological model of PAQR10 is illustrated in Figure 2(A). We first investigated whether the sequence(s) in the N-terminus of PAQR10 is required for its Golgi localization. To address this issue, we generated a series of N-terminus-truncated PAQR10 constructs (Figure 2B) and used them in localization studies in HeLa cells. Deletion of the N-terminal 35 amino acid residues (Δ1–35) located upstream of the first TM led to a complete loss of Golgi localization (Figure 2C). However, deletion of the first 11 amino acid residues (Δ1–11) had no effect on the Golgi localization of PAQR10 (Figure 2C). In contrast, deletion of amino acids 12–35 (Δ12–35) resulted in a loss of Golgi localization of the protein, suggesting that the 24 amino acids residing in the membrane-proximal region contain a Golgi-targeting signal for PAQR10.

Identification of structural determinants in the N-terminus of PAQR10 for Golgi localization and ERK stimulation

Figure 2
Identification of structural determinants in the N-terminus of PAQR10 for Golgi localization and ERK stimulation

(A) Amino acid sequence of human PAQR10 and hydrophobicity analysis. The locations of the seven predicted TMs are in red and underlined. (B) Schematic diagram of PAQR10 with a FLAG tag fused at the N-terminus. The site of the amino acids used in the deletion studies is marked. (C) Subcellular localization of various PAQR10 mutants at the N-terminus. HeLa cells were transfected with the plasmids as indicated, fixed and used in confocal analysis. The PAQR10 proteins are labelled with an anti-FLAG antibody (red) and the Golgi was marked with an anti-golgin-97 antibody (green). The arrow indicates apparent Golgi localization of PAQR10 proteins. Scale bar, 10 μm. (D) Effect of various deletions at the N-terminus of PAQR10 on ERK1/2 activation. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting using the antibodies indicated.

Figure 2
Identification of structural determinants in the N-terminus of PAQR10 for Golgi localization and ERK stimulation

(A) Amino acid sequence of human PAQR10 and hydrophobicity analysis. The locations of the seven predicted TMs are in red and underlined. (B) Schematic diagram of PAQR10 with a FLAG tag fused at the N-terminus. The site of the amino acids used in the deletion studies is marked. (C) Subcellular localization of various PAQR10 mutants at the N-terminus. HeLa cells were transfected with the plasmids as indicated, fixed and used in confocal analysis. The PAQR10 proteins are labelled with an anti-FLAG antibody (red) and the Golgi was marked with an anti-golgin-97 antibody (green). The arrow indicates apparent Golgi localization of PAQR10 proteins. Scale bar, 10 μm. (D) Effect of various deletions at the N-terminus of PAQR10 on ERK1/2 activation. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting using the antibodies indicated.

We also investigated the activities of these mutant proteins to stimulate ERK phosphorylation. As expected, overexpression of the full-length PAQR10 was able to activate ERK1/2 phosphorylation (Figure 2D), similar to our previous study [13]. Whereas PAQR10 with deletion of the N-terminal 11 amino acids retained its activity to stimulate ERK1/2 phosphorylation, the mutants with deletion of amino acids 1–35 or 12–35 markedly lost their effect to elevate ERK1/2 phosphorylation (Figure 2D). Taken together, these results indicate that the N-terminal 12–35 amino acids of PAQR10 are not only required for the Golgi localization, but also indispensable for the activity to stimulate the ERK signalling pathway.

Structural domains of the C-terminus of PAQR10 responsible for Golgi localization

We next investigated the structural determinants in the C-terminus of PAQR10 required for its Golgi localization by using a series of deletion mutants in the C-terminus (Figure 3A). Deletions of the C-terminus (Δ232–246) and the membrane-proximal portion (Δ232–236) led to a complete loss of Golgi localization, whereas the deletion of the very end of PAQR10 (Δ237–246) did not affect Golgi localization (Figure 3B). Furthermore, deletion of the C-terminus (Δ233–246) also caused a complete loss of Golgi localization (Supplementary Figure S1A at http://www.BiochemJ.org/bj/443/bj4430643add.htm). In addition, the two deletion mutants (Δ232–246 and Δ232–236) lost their activity to stimulate ERK1/2 phosphorylation (Figure 3C). Together, these results indicate that the membrane-proximal four amino acids (233–236) are indispensable for both Golgi localization and ERK stimulation of PAQR10.

Determination of structural domains of the C-terminus responsible for Golgi localization and ERK stimulation of PAQR10

Figure 3
Determination of structural domains of the C-terminus responsible for Golgi localization and ERK stimulation of PAQR10

(A) Schematic diagram of PAQR10 and its C-terminal deletion mutants. The sites of amino acids used in the deletion or truncation studies are indicated. (B) Effect of the C-terminal mutants of PAQR10 on Golgi localization. HeLa cells transfected with FLAG-tagged PAQR10 mutants were used in confocal analysis. Scale bar, 10 μm. (C) Effect of various deletions at the C-terminus of PAQR10 on ERK1/2 phosphorylation. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting using the antibodies indicated.

Figure 3
Determination of structural domains of the C-terminus responsible for Golgi localization and ERK stimulation of PAQR10

(A) Schematic diagram of PAQR10 and its C-terminal deletion mutants. The sites of amino acids used in the deletion or truncation studies are indicated. (B) Effect of the C-terminal mutants of PAQR10 on Golgi localization. HeLa cells transfected with FLAG-tagged PAQR10 mutants were used in confocal analysis. Scale bar, 10 μm. (C) Effect of various deletions at the C-terminus of PAQR10 on ERK1/2 phosphorylation. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting using the antibodies indicated.

Analysis of the three loops of PAQR10 facing the cytosol

On the basis of the topology analysis, PAQR10 contains three loops facing the cytosol (Figure 4A). We replaced these three loops of PAQR10 with the corresponding sequence of PAGR1/AdipoR1 so as to maintain its basic topological structure (Figure 4A), as PAQR1/AdipoR1 was reported to contain seven TMs [3]. As shown in Figure 4(B), swapping these loops with those of PAQR1/AdipoR1 disrupted Golgi localization of PAQR10. Furthermore, all three loop mutants lost their ability to stimulate ERK phosphorylation (Figure 4C). Collectively, these results indicate that all three loops facing the cytosol are required for Golgi localization and ERK stimulation of PAQR10 protein.

Analysis of three extra-cytoplasmic loops involved in Golgi localization and ERK stimulation

Figure 4
Analysis of three extra-cytoplasmic loops involved in Golgi localization and ERK stimulation

(A) Schematic diagram of PAQR10 with three cytosolic loops replaced by the corresponding sequences of PAQR1/AdipoR1 respectively. (B) Subcellular localization studies. HeLa cells transfected with FLAG-tagged PAQR10 chimaeric mutants were used in confocal analysis. The cells were stained with an anti-FLAG antibody for PAQR10 (red) and an anti-golgin-97 antibody (green). Scale bar, 10 μm. (C) ERK1/2 phosphorylation assay. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting.

Figure 4
Analysis of three extra-cytoplasmic loops involved in Golgi localization and ERK stimulation

(A) Schematic diagram of PAQR10 with three cytosolic loops replaced by the corresponding sequences of PAQR1/AdipoR1 respectively. (B) Subcellular localization studies. HeLa cells transfected with FLAG-tagged PAQR10 chimaeric mutants were used in confocal analysis. The cells were stained with an anti-FLAG antibody for PAQR10 (red) and an anti-golgin-97 antibody (green). Scale bar, 10 μm. (C) ERK1/2 phosphorylation assay. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting.

Structural determinants of PAQR10 required for tethering H-Ras to the Golgi apparatus

As we reported previously [13], the Golgi-localized PAQR10 could significantly increase retention of H-Ras protein in the Golgi apparatus, thereby triggering ERK stimulation in situ. We next examined the co-localization of H-Ras with a series of PAQR10 mutants by performing an immunofluorescence assay. As shown in Figure 5, overexpression of PAQR10 significantly increased the localization of H-Ras in the Golgi apparatus marked by a Golgi marker golgin-97, consistent with the previous study [13]. Meanwhile, the Golgi localization of H-Ras was also markedly increased by overexpression of PAQR10(Δ1–11) and PAQR10(Δ237–246), with a clear co-localization with golgin-97 (Figure 5). However, all other PAQR10 mutants that lost Golgi localization (Figures 2–4) were unable to mobilize H-Ras to the Golgi apparatus (Figure 5). Deletion of the C-terminus (Δ233–246) of PAQR10 also led to a loss of co-localization of PAQR10 with H-Ras (Supplementary Figure S1B). Together, these findings suggest that the essential motifs of PAQR10 responsible for its Golgi localization are implicated in the mobilization of Ras protein to the Golgi apparatus and the subsequent activation of downstream ERK signalling.

Co-localization studies of various PAQR10 mutants with H-Ras

Figure 5
Co-localization studies of various PAQR10 mutants with H-Ras

HeLa cells were co-transfected with FLAG-fused PAQR10 and EGFP-tagged H-Ras and used in confocal analysis. The PAQR10 proteins were labelled with an anti-FLAG antibody (red) and the Golgi apparatus was marked with an anti-golgin-97 antibody (purple). The nuclei were stained with Hoechst 33342 (blue). The arrow indicates apparent co-localization of H-Ras with PAQR10 proteins in the Golgi apparatus. Scale bar, 10 μm.

Figure 5
Co-localization studies of various PAQR10 mutants with H-Ras

HeLa cells were co-transfected with FLAG-fused PAQR10 and EGFP-tagged H-Ras and used in confocal analysis. The PAQR10 proteins were labelled with an anti-FLAG antibody (red) and the Golgi apparatus was marked with an anti-golgin-97 antibody (purple). The nuclei were stained with Hoechst 33342 (blue). The arrow indicates apparent co-localization of H-Ras with PAQR10 proteins in the Golgi apparatus. Scale bar, 10 μm.

To corroborate further the importance of Golgi localization for the functions of PAQR10, we analysed another PAQR10 mutant with misplaced localization. We fused an ER-retention signal to the C-terminus of PAQR10 (Figure 6A). Previous studies have shown that the addition of the KDEL motif to the C-terminal end of various proteins leads to the retention of these proteins in the ER [26]. As expected, overexpressed PAQR10-KDEL was exclusively localized in the ER with a clear co-localization with an ER marker calnexin, but without co-localization with Golgi marker golgin-97 (Figure 6B). Moreover, this ER-localized PAQR10 mutant could not increase the mobilization of H-Ras in either the ER or the Golgi apparatus (Figure 6C). Meanwhile, unlike the wild-type PAQR10, overexpression of PAQR10-KDEL failed to stimulate ERK1/2 phosphorylation (Figure 6D). These results therefore provide additional evidence that anchoring of PAQR10 to the Golgi apparatus is indispensable for its activity to affect the Ras/ERK signalling cascade.

Anchoring to the Golgi apparatus is important for the functions of PAQR10

Figure 6
Anchoring to the Golgi apparatus is important for the functions of PAQR10

(A) Schematic diagram of the chimaeric mutant of PAQR10 targeted to the ER by adding a KDEL motif to the C-terminus of the protein. (B) Subcellular localization studies. HeLa cells were transfected with FLAG-tagged PAQR10-KDEL and used in confocal analysis. The PAQR10 proteins were labelled with an anti-FLAG antibody (red). Golgin-97 and calnexin were used as markers for Golgi apparatus and ER respectively (green). The arrow indicates apparent co-localization of PAQR10-KDEL with calnexin. Scale bar, 10 μm. (C) Co-localization studies of PAQR10 proteins with H-Ras. HeLa cells were transfected with FLAG-tagged PAQR10-KDEL (red) and EGFP–H-Ras (green), followed by immunostaining and confocal analysis. The Golgi was labelled using an anti-golgin-97 antibody (purple). The nuclei were stained with Hoechst 33342 (blue). The arrow indicates apparent co-localization of H-Ras and PAQR10 in the Golgi apparatus. Scale bar, 10 μm. (D) PAQR10-KDEL could not stimulate ERK1/2 phosphorylation. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting.

Figure 6
Anchoring to the Golgi apparatus is important for the functions of PAQR10

(A) Schematic diagram of the chimaeric mutant of PAQR10 targeted to the ER by adding a KDEL motif to the C-terminus of the protein. (B) Subcellular localization studies. HeLa cells were transfected with FLAG-tagged PAQR10-KDEL and used in confocal analysis. The PAQR10 proteins were labelled with an anti-FLAG antibody (red). Golgin-97 and calnexin were used as markers for Golgi apparatus and ER respectively (green). The arrow indicates apparent co-localization of PAQR10-KDEL with calnexin. Scale bar, 10 μm. (C) Co-localization studies of PAQR10 proteins with H-Ras. HeLa cells were transfected with FLAG-tagged PAQR10-KDEL (red) and EGFP–H-Ras (green), followed by immunostaining and confocal analysis. The Golgi was labelled using an anti-golgin-97 antibody (purple). The nuclei were stained with Hoechst 33342 (blue). The arrow indicates apparent co-localization of H-Ras and PAQR10 in the Golgi apparatus. Scale bar, 10 μm. (D) PAQR10-KDEL could not stimulate ERK1/2 phosphorylation. HEK-293T cells were transiently transfected with the plasmids as indicated. At 24 h after transfection, the cells were serum-starved for 6 h, and the cell lysate was used in Western blotting.

The effect of PAQR10 mutants on the differentiation of PC12 cells

We next used PC12 cells as a cellular model to investigate the structural determinants of PAQR10 required for sustained ERK signalling in PC12 cells, as it has been reported that activation of Golgi-localized Ras signalling is able to promote differentiation of PC12 cells [18]. Compared with the vector control, PAQR10 was able to significantly stimulate neurite outgrowth (Figure 7), a marker of PC12 cell differentiation. Interestingly, two other Golgi-localizing mutants, PAQR10(Δ1–11) and PAQR10(Δ237–246), were also able to stimulate neurite outgrowth (Figure 7). In contrast, none of the PAQR10 mutants that were not localized in the Golgi apparatus could affect PC12 cell differentiation (Figure 7). Collectively, these results reveal that the structural determinants required for PAQR10 to correctly localize to the Golgi apparatus are also required for PAQR10 to stimulate a sustained Ras signalling in PC12 cells.

Effect of PAQR10 mutations on PC12 cell differentiation

Figure 7
Effect of PAQR10 mutations on PC12 cell differentiation

PC12 cells were co-transfected with pEGFP and excess amounts of PAQR10 or PAQR10 mutants as indicated and neurite outgrowth (arrowed) was analysed 48 h later. Representative images are shown and the quantification is shown in the lower right-hand panel. **P<0.01 compared with the group without PAQR10 expression. The cell nucleus was stained with Hoechst 33342 (blue). Scale bar, 10 μm.

Figure 7
Effect of PAQR10 mutations on PC12 cell differentiation

PC12 cells were co-transfected with pEGFP and excess amounts of PAQR10 or PAQR10 mutants as indicated and neurite outgrowth (arrowed) was analysed 48 h later. Representative images are shown and the quantification is shown in the lower right-hand panel. **P<0.01 compared with the group without PAQR10 expression. The cell nucleus was stained with Hoechst 33342 (blue). Scale bar, 10 μm.

DISCUSSION

In the present study, we extensively characterized the topology and functional domains of PAQR10. We showed that PAQR10 is a Golgi integral membrane protein with its N-terminus facing the cytosol, in part consistent with the predicted topological model of PAQR family members [1]. In addition, we identified multiple structural domains that are critical for Golgi localization of PAQR10. We found that at least five motifs are required for the Golgi localization of PAQR10, including the N-terminal membrane-proximal region (amino acids 12–35), the C-terminal membrane-proximal region (amino acids 232–236) and the three cytosolic loops. Intriguingly, we observed that these critical Golgi-targeting motifs are also essential for the biological function of PAQR10 to stimulate ERK signalling, as the proteins depleted of the five critical motifs can neither stimulate ERK1/2 phosphorylation (Figures 2–4) nor potentiate differentiation of PC12 cells (Figure 7). Such an observation was supported further by the finding that artificial compartmentalization of PAQR10 in the ER failed to stimulate ERK1/2 phosphorylation. Furthermore, the PAQR10 mutants that could not be tethered to the Golgi apparatus were co-localized with an ER marker, calnexin (Supplementary Figure S2 at http://www.BiochemJ.org/bj/443/bj4430643add.htm), suggesting that these mutant proteins are retained in the ER after protein synthesis, probably due to misfolding. Collectively, our results indicate that the correct subcellular localization of PAQR10 is prerequisite for its actions to mobilize Ras to the Golgi apparatus and activate the Ras/ERK signalling cascade.

The overall topology of PAQR family members is similar to GPCR family members as both of them contain seven-transmembrane motifs. However, all GPCR proteins have a characteristic topology with an extracellular N-terminus and an intracellular C-terminal domain (type I topology) [27]. Although PAQR family members share similar seven-transmembrane characteristics, their N-termini are predicted to be localized intracellularly [1]. However, conflicting evidence exists regarding the membrane topology of PAQR family members. Paradoxically, the predicted progesterone receptors PAQR5, PAQR7 and PAQR8 were predicted to locate on the plasma membrane with a similar topology to that of GPCRs [5]. Another study also revealed that PAQR7 is predominantly localized in the ER with both N- and C-termini localized intracellularly [28,29]. Moreover, adiponectin receptors PAQR1/AdipoR1 and PAQR2/AdipoR2 were reported to reside on the plasma membrane with the N-termini facing the cytosol [3]. Our previous study demonstrated that PAQR3/RKTG is a specific Golgi-localized protein with its N-terminus facing the cytosolic side of the Golgi apparatus [8]. In the present study, we found that the topology of PAQR10 is very similar to that of PAQR3/RKTG. Interestingly, it is noteworthy that both PAQR3/RKTG and PAQR10 are implicated in the spatial regulation of the Ras signalling cascade, although in different manners [7,13].

The Golgi-targeting motifs of Golgi-resident proteins vary greatly without a consensus sequence(s). Unlike the majority of ER-resident proteins that have a common retention sequence of KDEL on the C-termini [26], the localization determinants of Golgi proteins are complicated [15]. It has been shown that the TM is essential and sufficient for specifying the Golgi localization of glycosyltransferases [30]. In addition, the sequences flanking the TM or the luminal portion contribute to the Golgi localization. A localization determinant has also been identified in the cytoplasmic tail of some Golgi membrane proteins [16]. On the other hand, multiple motifs are required for both PAQR3/RKTG and PAQR10 to localize in the Golgi apparatus (the present study and [8]). For example, both the membrane-proximal regions in the N-termini and C-termini are essential for targeting PAQR3/RKTG and PAQR10 to the Golgi apparatus (the present study and [8]. However, these membrane-proximal regions of PAQR3/RKTG and PAQR10 share no sequence similarity with each other. Therefore it is worth exploring in the future the precise molecular mechanism of how these Golgi-targeting motifs direct PAQR3/RKTG and PAQR10 to correctly compartmentalize these proteins to the Golgi apparatus.

Abbreviations

     
  • AdipoR

    adiponectin receptor

  •  
  • Cy5

    indodicarbocyanine

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • EGFP

    enhanced green fluorescent protein

  •  
  • ER

    endoplasmic reticulum

  •  
  • ERK

    extracellular-signal-regulated kinase

  •  
  • FBS

    fetal bovine serum

  •  
  • GPCR

    G-protein-coupled receptor

  •  
  • HA

    haemagglutinin

  •  
  • HEK

    human embryonic kidney

  •  
  • PAQR

    progestin and adipoQ receptor

  •  
  • PEI

    polyethyleneimine

  •  
  • RasGRP1

    Ras guanine-nucleotide-releasing protein 1

  •  
  • RKTG

    Raf kinase trapping to Golgi

  •  
  • TGN38

    trans-Golgi network protein 38

  •  
  • TM

    transmembrane domain

AUTHOR CONTRIBUTION

Ting Jin, Qiurong Ding and Yan Chen designed the experiments. Ting Jin, Daqian Xu and Qiurong Ding performed the experiments. Yixuan Zhang, Chenqian Mao, Yi Pan and Zhenzhen Wang provided technical assistance. Ting Jin and Yan Chen wrote the paper.

FUNDING

This work was supported by the Ministry of Science and Technology of China [grant numbers 2012CB524900 (to Y.C.) and 2010CB529506 (to Y.P. and Z.W.)], National Natural Science Foundation of China [grant numbers 30830037, 81021002 and 81130077 (to Y.C.) and 30971660 (to Y.P.)] and Chinese Academy of Sciences [grant number KSCX2-EW-R-08 (to Y.C.)].

References

References
1
Tang
Y. T.
Hu
T. H.
Arterburn
M.
Boyle
B.
Bright
J. M.
Emtage
P. C.
Funk
W. D.
PAQR proteins: a novel membrane receptor family defined by an ancient 7-transmembrane pass motif
J. Mol. Evol.
2005
, vol. 
61
 (pg. 
372
-
380
)
2
Vassilatis
D. K.
Hohmann
J. G.
Zeng
H.
Li
F. S.
Ranchalis
J. E.
Mortrud
M. T.
Brown
A.
Rodriguez
S. S.
Weller
J. R.
Wright
A. C.
, et al. 
The G protein-coupled receptor repertoires of human and mouse
Proc. Natl. Acad. Sci. U.S.A.
2003
, vol. 
100
 (pg. 
4903
-
4908
)
3
Yamauchi
T.
Kamon
J.
Ito
Y.
Tsuchida
A.
Yokomizo
T.
Kita
S.
Sugiyama
T.
Miyagishi
M.
Hara
K.
Tsunoda
M.
, et al. 
Cloning of adiponectin receptors that mediate antidiabetic metabolic effects
Nature
2003
, vol. 
423
 (pg. 
762
-
769
)
4
Kadowaki
T.
Yamauchi
T.
Kubota
N.
Hara
K.
Ueki
K.
Tobe
K.
Adiponectin and adiponectin receptors in insulin resistance, diabetes, and the metabolic syndrome
J. Clin. Invest.
2006
, vol. 
116
 (pg. 
1784
-
1792
)
5
Zhu
Y.
Bond
J.
Thomas
P.
Identification, classification, and partial characterization of genes in humans and other vertebrates homologous to a fish membrane progestin receptor
Proc. Natl. Acad. Sci. U.S.A.
2003
, vol. 
100
 (pg. 
2237
-
2242
)
6
Zhu
Y.
Rice
C. D.
Pang
Y. F.
Pace
M.
Thomas
P.
Cloning, expression, and characterization of a membrane progestin receptor and evidence it is an intermediary in meiotic maturation of fish oocytes
Proc. Natl. Acad. Sci. U.S.A.
2003
, vol. 
100
 (pg. 
2231
-
2236
)
7
Feng
L.
Xie
X. D.
Ding
Q. R.
Luo
X. L.
He
J.
Fan
F. J.
Liu
W. Z.
Wang
Z. Z.
Chen
Y.
Spatial regulation of Raf kinase signaling by RKTG
Proc. Natl. Acad. Sci. U.S.A.
2007
, vol. 
104
 (pg. 
14348
-
14353
)
8
Luo
X.
Feng
L.
Jiang
X.
Xiao
F.
Wang
Z.
Feng
G. S.
Chen
Y.
Characterization of the topology and functional domains of RKTG
Biochem. J.
2008
, vol. 
414
 (pg. 
399
-
406
)
9
Fan
F. J.
Feng
L.
He
J.
Wang
X.
Jiang
X. M.
Zhang
Y. X.
Wang
Z. Z.
Chen
Y.
RKTG sequesters B-Raf to the Golgi apparatus and inhibits the proliferation and tumorigenicity of human malignant melanoma cells
Carcinogenesis
2008
, vol. 
29
 (pg. 
1157
-
1163
)
10
Xie
X. D.
Zhang
Y. X.
Jiang
Y. H.
Liu
W. Z.
Ma
H.
Wang
Z. Z.
Chen
Y.
Suppressive function of RKTG on chemical carcinogen-induced skin carcinogenesis in mouse
Carcinogenesis
2008
, vol. 
29
 (pg. 
1632
-
1638
)
11
Jiang
Y. H.
Xie
X. D.
Li
Z. G.
Wang
Z.
Zhang
Y. X.
Ling
Z. Q.
Pan
Y.
Wang
Z. Z.
Chen
Y.
Functional cooperation of RKTG with p53 in tumorigenesis and epithelial–mesenchymal transition
Cancer Res.
2011
, vol. 
71
 (pg. 
2959
-
2968
)
12
Zhang
Y.
Jiang
X.
Qin
X.
Ye
D.
Yi
Z.
Liu
M.
Bai
O.
Liu
W.
Xie
X.
Wang
Z.
, et al. 
RKTG inhibits angiogenesis by suppressing MAPK-mediated autocrine VEGF signaling and is downregulated in clear-cell renal cell carcinoma
Oncogene
2010
, vol. 
29
 (pg. 
5404
-
5415
)
13
Jin
T.
Ding
Q.
Huang
H.
Xu
D.
Jiang
Y.
Zhou
B.
Li
Z.
Jiang
X.
He
J.
Liu
W.
, et al. 
PAQR10 and PAQR11 mediate Ras signaling in the Golgi apparatus
Cell Res.
2011
 
doi:10.1038/cr.2011.161
14
De Matteis
M. A.
Luini
A.
Exiting the Golgi complex
Nat. Rev. Mol. Cell Biol.
2008
, vol. 
9
 (pg. 
273
-
284
)
15
Munro
S.
Localization of proteins to the Golgi apparatus
Trends Cell Biol.
1998
, vol. 
8
 (pg. 
11
-
15
)
16
Bos
K.
Wraight
C.
Stanley
K. K.
Tgn38 is maintained in the trans-Golgi network by a tyrosine-containing motif in the cytoplasmic domain
EMBO J.
1993
, vol. 
12
 (pg. 
2219
-
2228
)
17
Chiu
V. K.
Bivona
T.
Hach
A.
Sajous
J. B.
Silletti
J.
Wiener
H.
Johnson
R. L.
Cox
A. D.
Philips
M. R.
Ras signalling on the endoplasmic reticulum and the Golgi
Nat. Cell Biol.
2002
, vol. 
4
 (pg. 
343
-
350
)
18
Bivona
T. G.
Perez de Castro
I.
Ahearn
I. M.
Grana
T. M.
Chiu
V. K.
Lockyer
P. J.
Cullen
P. J.
Pellicer
A.
Cox
A. D.
Philips
M. R.
Phospholipase Cγ activates Ras on the Golgi apparatus by means of RasGRP1
Nature
2003
, vol. 
424
 (pg. 
694
-
698
)
19
Sun
X.
Han
R.
Wang
Z.
Chen
Y.
Regulation of adiponectin receptors in hepatocytes by the peroxisome proliferator-activated receptor-γ agonist rosiglitazone
Diabetologia
2006
, vol. 
49
 (pg. 
1303
-
1310
)
20
Ohno
H.
Stewart
J.
Fournier
M. C.
Bosshart
H.
Rhee
I.
Miyatake
S.
Saito
T.
Gallusser
A.
Kirchhausen
T.
Bonifacino
J. S.
Interaction of tyrosine-based sorting signals with clathrin-associated proteins
Science
1995
, vol. 
269
 (pg. 
1872
-
1875
)
21
Yu
M. R.
Liu
X. L.
Sun
L.
Chen
C. W.
Ma
G. P.
Kitamura
Y.
Gao
G. F.
Liu
W. J.
Subcellular localization and topology of porcine reproductive and respiratory syndrome virus E protein
Virus Res.
2010
, vol. 
152
 (pg. 
104
-
114
)
22
Feldheim
D.
Rothblatt
J.
Schekman
R.
Topology and functional domains of Sec63p, an endoplasmic-reticulum membrane-protein required for secretory protein translocation
Mol. Cell. Biol.
1992
, vol. 
12
 (pg. 
3288
-
3296
)
23
Bergeron
J. J. M.
Brenner
M. B.
Thomas
D. Y.
Williams
D. B.
Calnexin: a membrane-bound chaperone of the endoplasmic reticulum
Trends Biochem. Sci.
1994
, vol. 
19
 (pg. 
124
-
128
)
24
Ou
W. J.
Bergeron
J. J. M.
Li
Y.
Kang
C. Y.
Thomas
D. Y.
Conformational changes induced in the endoplasmic-reticulum luminal domain of calnexin by Mg-ATP and Ca2+
J. Biol. Chem.
1995
, vol. 
270
 (pg. 
18051
-
18059
)
25
Wilde
A.
Reaves
B.
Banting
G.
Epitope mapping of two isoforms of a trans-Golgi network specific integral membrane protein Tgn38/41
FEBS Lett.
1992
, vol. 
313
 (pg. 
235
-
238
)
26
Pagny
S.
Lerouge
P.
Faye
L.
Gomord
V.
Signals and mechanisms for protein retention in the endoplasmic reticulum
J. Exp. Bot.
1999
, vol. 
50
 (pg. 
157
-
164
)
27
Ritter
S. L.
Hall
R. A.
Fine-tuning of GPCR activity by receptor-interacting proteins
Nat. Rev. Mol. Cell Biol.
2009
, vol. 
10
 (pg. 
819
-
830
)
28
Ashley
R. L.
Clay
C. M.
Farmerie
T. A.
Niswender
G. D.
Nett
T. M.
Cloning and characterization of an ovine intracellular seven transmembrane receptor for progesterone that mediates calcium mobilization
Endocrinology
2006
, vol. 
147
 (pg. 
4151
-
4159
)
29
Krietsch
T.
Fernandes
M. S.
Kero
J.
Losel
R.
Heyens
M.
Lam
E. W. F.
Huhtaniemi
I.
Brosens
J. J.
Gellersen
B.
Human homologs of the putative G protein-coupled membrane progestin receptors (mPRα, β, and γ) localize to the endoplasmic reticulum and are not activated by progesterone
Mol. Endocrinol.
2006
, vol. 
20
 (pg. 
3146
-
3164
)
30
Colley
K. J.
Golgi localization of glycosyltransferases: more questions than answers
Glycobiology
1997
, vol. 
7
 (pg. 
1
-
13
)

Author notes

1

These authors contributed equally to this work.

Supplementary data