Differential determinants for coupling of distinct G proteins with the class B secretin receptor

Gene L. Garcia, Maoqing Dong, Laurence J. Miller


The secretin receptor is a prototypic class B G protein-coupled receptor that is activated by binding of its natural peptide ligand. The signaling effects of this receptor are mediated by coupling with Gs, which activates cAMP production, and Gq, which activates intracellular calcium mobilization. We have explored the molecular basis for the coupling of each of these G proteins to this receptor using systematic site-directed mutagenesis of key residues within each of the intracellular loop regions, and studying ligand binding and secretin-stimulated cAMP and calcium responses. Mutation of a conserved histidine in the first intracellular loop (H157A and H157R) markedly reduced cell surface expression, resulting in marked reduction in cAMP and elimination of measurable calcium responses. Mutation of an arginine (R153A) in the first intracellular loop reduced calcium, but not cAMP responses. Mutation of a dibasic motif in the second intracellular loop (R231A/K232A) had no significant effects on any measured responses. Mutations in the third intracellular loop involving adjacent lysine and leucine residues (K302A/L303A) or two arginine residues separated by a leucine and an alanine (R318A/R321A) significantly reduced cAMP responses, while the latter also reduced calcium responses. Additive effects were elicited by combining the effective mutations, while combining all the effective mutations resulted in a construct that continued to bind secretin normally, but that elicited no significant cAMP or calcium responses. These data suggest that, while some receptor determinants are clearly shared, there are also distinct determinants for coupling with each of these G proteins.

  • G protein-coupled receptor
  • secretin
  • G protein coupling
  • cAMP
  • calcium

peptide hormones and neurotransmitters exert their physiological effects through binding to cell surface receptors, resulting in conformational changes in the receptors and activation of signaling cascades within the target cell. There has been recent recognition that some receptors can interact with various second messengers and initiate a spectrum of biological events, with the possibility of introducing bias in this repertoire by use of unique receptor ligands. This makes it particularly important to understand the impact of each of the intracellular signaling cascades that can be regulated by a given ligand. An understanding of the molecular basis of initiation of these pathways and the ability to differentially activate them in model systems could facilitate the development of potentially useful selective drugs. The present work is focused on the secretin receptor, a prototypic peptide-binding class B G protein-coupled receptor (GPCR) that is known to activate two guanine nucleotide-binding protein (G protein)-initiated pathways.

GPCRs represent the largest group of cell surface receptors, which are present on essentially every regulatable cell in the body. Members of the GPCR superfamily share predicted seven-transmembrane topology and coupling with heterotrimeric guanine nucleotide-binding proteins (G proteins), while possessing a broad variety of themes for binding and activation by a chemically and structurally diverse set of natural agonist ligands. The three-dimensional structures of several members of class A GPCRs have been solved in recent years (35). This has provided insights into the structural basis for binding of their natural ligands, particularly those representing small molecules that dock within the intramembranous helical bundle region of those receptors (2, 3, 17, 29, 50, 70, 73). It has also provided insights into the molecular basis of binding more flexible and larger peptide ligands that dock in extramembranous sites (73).

However, understanding of the molecular basis for G protein association with GPCRs, occurring on the cytosolic face of the membrane, is less well developed. While there are a series of crystal structures of class A GPCRs (35) and heterotrimeric G proteins (49), the first such structure including both molecules within an active-state ternary complex was recently reported (54). Other clues to the molecular basis of this interaction have come from a crystal structure in which the carboxyl-terminal peptide of transducin was used to help stabilize a partially activated conformation of opsin (56) and a structure including an antibody that appears to mimic the effect of G protein coupling with the β2-adrenergic receptor (53, 55). These structures are generally consistent with the survey of residues that could potentially contribute to G protein coupling that have come from receptor mutagenesis studies in which key residues that could contribute to this molecular interaction were modified. These residues typically reside in regions of the intracellular loop and carboxyl-terminal tail of the GPCRs that are predicted to be close to the membrane interface (34, 71, 72). However, this type of approach is indirect, and loss of function may be difficult to interpret, particularly with molecules as flexible as the “shape-shifting” GPCRs (32). Absence of clear consensus sequences for G protein-coupling determinants across this family has also contributed to the difficulty in interpreting these studies.

It is important to recognize that the strongest existing data for G protein association are relevant to class A GPCRs, and that the sequence and structure of the intramembranous helical bundle domain of class A receptors are predicted to be quite distinct from these features of class B GPCRs (23, 37). The signature sequences that are typical of class A receptors are absent in class B GPCRs (71). Additionally, unlike class A GPCRs that tend to couple with one distinct G protein, class B GPCRs often exhibit promiscuous coupling with two or more distinct G proteins (58). Typical of class B GPCRs is the ability of their natural agonists to exhibit potent stimulation of cAMP (through activation of Gαs) and less potent stimulation of intracellular calcium responses (through activation of Gαq) (20). Consistent with this pattern, the secretin receptor has been shown to stimulate both cAMP and intracellular calcium responses through its coupling with these two G proteins (51).

In the present work, the molecular determinants for G protein coupling with the secretin receptor have been carefully explored, taking advantage of the insights coming from studies with other GPCRs. A receptor mutagenesis approach with the in vitro characterization of signaling in recombinant receptor-bearing cells was utilized. Natural ligand binding and agonist-stimulated biological activity through cAMP production (Gαs signaling) and calcium mobilization (Gαq signaling) were studied for receptor constructs in which key residues in the first, second, and third intracellular loop regions of the human secretin receptor had been modified. Additionally, the effects of simultaneously modifying critical determinants for G protein coupling were also examined. These studies demonstrated that residues within each of the intracellular loops contribute to G protein coupling, with some sites contributing disproportionately to Gαs coupling that results in increase in cellular cAMP, and others contributing disproportionately to Gαq coupling that results in increase in intracellular calcium, while still other sites contribute to the coupling with both of these G proteins. As we gain better insights into the tertiary structure of the class B GPCRs, these insights should help to develop and refine the understanding of how distinct complexes might be formed with each of these G proteins. The receptor constructs that are differentially coupled with one or the other G protein can help elucidate the biological effects of each pathway and contribute to a strategy to develop agonists exhibiting signaling bias at a receptor in this important family.



Fetal Clone II serum was from HyClone Laboratories (Logan UT). Dulbecco's modified Eagle medium (DMEM), soybean trypsin inhibitor (STI), and trypsin were purchased from Gibco (Carlsbad, CA). Diethylaminoethyl-dextran (DEAE-dextran), phenylmethylsulfonyl fluoride (PMSF), bacitracin, Triton X-100, probenecid, and pluronic acid F-127 were from Sigma Aldrich (St. Louis, MO). Bovine serum albumin (BSA) Cohn Fraction V was from Equitech Bio (Kerrville, TX). Polyethylamine was from Polyscience (Warrington, PA). Microscint O was from Perkin Elmer (Waltham, MA). Other reagents were analytical grade.


Rat secretin-27 and (Tyr10)rat secretin-27 were synthesized in our laboratory using solid-phase manual techniques, as we have described (52). The (Tyr10)rat secretin-27 was radioiodinated using oxidizing conditions, and purified to homogeneity on reversed-phase HPLC (specific radioactivity ∼2,000 Ci/mmol), as we have described (65).

Mutagenesis and tagged constructs.

Rat Gαs-Rlu was prepared in our laboratory, as previously described (18). Human Gαq-Rlu and Gαi-Rlu were prepared by insertion of the Rlu at the leucine in positions 97 and 91, respectively, as previously described (10).

Site-directed mutagenesis of the human secretin receptor cDNA was performed according to manufacturer's instructions with the QuickChange kit (Stratagene, La Jolla, CA). The template for mutagenesis was the human secretin receptor cDNA construct in pcDNA3 we previously prepared and characterized (9, 30). The human secretin receptor having a carboxyl-terminal fusion with eYFP was described previously (19). The YFP-tagged mutant receptors were prepared by generating PCR products containing the full-length mutant receptors with their stop codons replaced with XhoI restriction sites. These PCR products and the wild-type human secretin receptor-YFP construct were then cut with HindIII and XhoI, and the relevant fragments from the digestion of the PCR products were used to replace the wild-type sequence. The identities of all constructs were confirmed by bidirectional sequencing.

Cell culture and transfection.

African Green Monkey kidney (COS-1) cells were maintained in DMEM containing 5% Fetal Clone II supplement in humidified incubators at 37°C. Semiconfluent cells in 10-cm dishes were transfected using modification of the DEAE-dextran method (43, 61). In brief, COS-1 cells plated 24 h previously were washed once with serum-free DMEM and then incubated with 3 μg DNA and 0.2 mg/ml DEAE-dextran in 1.5 ml DMEM for 2 h. After incubation with 10% DMSO in DMEM for 2 min and with 0.1 mM chloroquine in DMEM for 2 h, cells were then incubated with DMEM containing 5% Fetal Clone II overnight. Cells were lifted using 0.05% trypsin in phosphate-buffered saline, pH 7.4 (PBS) and were then plated at appropriate densities for cAMP and calcium assays, or for membrane preparations.

Membrane preparations.

Cell membranes were prepared from COS-1 cells 72–96 h following transfection, using the method described previously (15). In brief, cells were lifted in PBS from the plates using a Sarstedt 3.1 cm cell scraper. After washing with PBS, cells were resuspended in 0.3 M sucrose with 0.01% STI and 1 mM PMSF and sonicated 3 times for 20 s on ice. The broken cell preparation was brought to a final concentration of 1.3 M sucrose and gently overlaid with 0.3 M sucrose solution. Tubes were then centrifuged in a Beckman Coulter L-100 XP ultracentrifuge in a Ti70 rotor for 1 h at 56,000 rpm at 4°C. The membranes at the interface between the sucrose layers were removed and resuspended in ice-cold water. These were pelletted by centrifugation for 30 min at 56,000 rpm, and the membrane pellet was resuspended in 1.5 ml Krebs Ringer's HEPES (KRH) medium (25 mM HEPES, pH 7.4, 104 mM NaCl, 5 mM KCl, 1 mM KH2PO4, 1.2 mM MgSO4) with 2 mM CaCl2, 0.01% STI, and 1 mM PMSF, and homogenized six times using a Dounce homogenizer. Total protein concentration of this fraction was assayed using the Pierce BCA Protein assay kit (Thermo Scientific, Rockford, IL), and the membranes were stored at −80°C.

Receptor binding assays.

Receptor binding assays were performed using both intact cells and membrane preparations. Intact cell binding was performed as we previously described (9). In brief, 24 h after transfection, cells were plated in 24-well plates at a density of ∼100,000 cells/well. After culture for 48 h, cells were washed twice and incubated with ∼20,000 cpm of the secretin radioligand [125I-(Tyr10)rat secretin-27] and competing concentrations of unlabeled secretin in KRH medium containing 2 mM CaCl2, 0.01% STI, and 0.2% BSA with gentle shaking for 1 h at room temperature. Cells were then washed twice with ice-cold KRH medium before being lysed using 0.5 M NaOH with vigorous shaking for 15 min. Lysates were transferred to 12 × 75 mm tubes, and radioactivity was quantified using an Isodata 20/20 series gamma counter.

Membrane binding assays were set up in Falcon 96-well plates with a final volume of 100 μl KRH medium with 2 mM CaCl2, 0.01% STI, 0.2% BSA, and 1 mM PMSF per well. To each well, 5–10 μg of membrane preparation, ∼10,000 cpm of the secretin radioligand, and increasing concentrations of unlabeled competing rat secretin were added. Assay plates were incubated for 1 h at room temperature. To separate bound radioligand from free radioligand, membranes were transferred to Perkin Elmer Unifilter Plates, which had been soaked in 0.3% (wt/vol) polyethylamine for 1 h, and washed with ice-cold 0.2% BSA using the Perkin Elmer Filtermate Harvester. Plates were dried overnight and 35 μl scintillation fluid was added to each well before being read using the Perkin Elmer Top Count NXT microplate scintillation reader. Nonspecific binding to the wild-type secretin receptor determined in the presence of 320 nM secretin represented <10% of total binding.

cAMP assays.

For the cAMP accumulation assay, transfected cells were plated in 96-well plates at ∼7,000 cells/well 24 h before assay. Cells were washed twice in KRH medium with 2 mM CaCl2 and were stimulated for 30 min using 50 μl secretin dilutions in KRH medium with 2 mM CaCl2, 0.2% BSA, 0.1% bacitracin, 0.01% STI, and 1 mM 3-isobutyl-1-methylxanthine. Stimulation buffer was replaced by 50 μl ice-cold 6% perchloric acid, and plates were shaken on an orbital plate shaker for 15 min. Following this, 25 μl ice-cold water was added to each well, and the pH was adjusted to 6 with 30% KHCO3. Plates were kept at 4°C, and the cAMP content in each well was assayed using the LANCE cAMP assay kit (Perkin Elmer, Waltham, MA) according to the manufacturer's instructions.

Intracellular calcium assays.

For the intracellular calcium assay, transfected cells were plated in black-walled clear-bottom plates (Corning Costar) at a concentration of ∼15,000 cells/well 24 h before assay. Cells were washed once in calcium assay buffer (KRH medium with 1.5 mM CaCl2, 1.2 mM MgCl2, 2.5 mM probenecid, 0.2% BSA, and 0.01% STI), and were then loaded for 1 h with 2 μM fura-2 AM (Molecular Probes, Eugene, OR) in calcium assay buffer containing 0.016% (wt/vol) pluronic acid at 37°C. Cells were then washed twice with calcium assay buffer and incubated for 30 min to allow hydrolysis of the AM ester. The assay was performed using a Flexstation 3 (Molecular Devices, Sunnyvale, CA), allowing the direct addition of secretin with prompt analysis of the cellular responses. Fluorescence readings were acquired for the fraction of calcium-bound fura-2 using excitation at 340 nm and emission at 510 nm, and for the calcium-free fraction using excitation of 380 nm and emission at 510 nm.

Receptor-G protein association studies.

Bioluminescence resonance energy transfer (BRET) assays were performed to examine the association between various receptor constructs and G proteins. Assays were performed as previously described (18) with some modification. COS-1 cells were transfected with 1.5 μg of the Gα-Rlu construct and 1.5 μg of either empty pcDNA3 vector or one of the following constructs with eYFP fused at the carboxyl terminus: wild-type human secretin receptor, each mutant human secretin receptor, or the human type 2 cholecystokinin receptor as a control. Transfected cells were lifted 48 h after transfection using nonenzymatic cell dissociation solution (Sigma); 100,000–200,000 cells per well in 100 μl KRH medium were plated in 96-well white optiplates in triplicate. Coelenterazine h (Biotium, Hayward, CA) was added to each well to a final concentration of 5 μM. Agonist peptides (100 nM cholecystokinin or 50 nM secretin) were then added to stimulate maximal signaling responses. Plates were read immediately after addition of the peptides. A 2103 Envision plate reader (Perkin Elmer, Wellesley, MA) equipped with emission filters for luciferase (460 ± 25 nm) and YFP (535 ± 25 nm) was used to collect the BRET signal. Background signals were subtracted from the data using the formula, {L1exp − [L2exp × (L1Rlu/L2Rlu)]}/L2exp, where L1 represents luminescence at 535 nm, L2 represents luminescence at 460 nm, “exp” conditions represent the averages of three replicates of each experimental condition, and “Rlu” conditions represent the averages of three replicates of the Gα-Rlu co-transfected with empty vector.

Microscopy and Western blotting.

Transfected cells were grown on glass coverslips in six-well dishes for 24 h and fixed with 2% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) in PBS for 15 min. Coverslips were then washed twice with PBS, blocked with 10% normal goat serum in PBS, and incubated for 1 h at room temperature with a polyclonal antibody against the peptide antigen representing amino acids 30–44 of the human secretin receptor [anti-hSecR(30–44)] recently raised in our laboratory (1:1,000 in PBS with 1% normal goat serum). Coverslips were then washed three times with PBS and incubated with 1:1,000 Alexa Fluor 488-conjugated anti-rabbit IgG secondary antibody (Molecular Probes, Eugene, OR) for 30 min. Coverslips were washed three times with PBS and then mounted on microscope slides with Vectashield mounting medium (Vector Laboratories, Burlingame, CA). Cells were visualized with a 40× objective on a Zeiss inverted microscope controlled by QED InVivo software (Media Cybernetics, Bethesda, MD).

Concurrent with the plating of cells onto coverslips for immunostaining as described above, 200,000 cells/well were plated in six-well dishes for Western blotting. Lysates from these cells were prepared during the fixation period for the immunostained cells. The cells were lysed on ice for 10 min with lysis buffer containing 150 mM NaCl, 10 mM Tris (pH 7.4), 1% NP-40, and 0.5% SDS. Protein concentrations were determined using the Pierce BCA protein assay kit. Samples were diluted to equal concentration in lysis buffer and 5× sample buffer (0.02% bromophenol blue, 2% SDS, 10% glycerol, 100 mM DTT, 50 mM Tris pH 6.8). Ten micrograms of samples were run on 10% SDS-PAGE gels and transferred onto Immobilon P 0.45 μm PVDF membranes (Millipore, Billerica, MA). Membranes were blocked for 1 h with 5% nonfat milk in TBST medium containing 150 mM NaCl, 10 mM Tris, pH 7.4, and 0.05% Tween 20. They were then incubated for 1 h with rabbit anti-hSecR(30–44) (1:2,000) diluted in TBST. After being washed three times with TBST, membranes were incubated for 30 min with a 1:10,000 dilution of horseradish peroxidase-conjugated goat-anti-rabbit IgG (Biosource, San Diego, CA) in TBST. Membranes were washed three times with TBST and then developed using Super-signal West Pico Chemiluminescent substrate (Thermo Scientific, Rockford, IL) for 5 min with constant agitation, and exposed to CL-Xposure Film (Thermo Scientific, Rockford, IL).

Data analysis.

Receptor binding kinetics were determined by analysis with the LIGAND program (46), and binding curves were plotted using the nonlinear regression analysis program in Prism 5.0 (GraphPad Software, La Jolla, CA). Data were fit to both one-site and two-site models. Data from functional studies, including cAMP and calcium assays, from adequately efficient transient transfections of COS-1 cells were graphed using nonlinear regression analysis in Prism 5.0. Statistical analyses were performed using one-way ANOVA or paired two-tailed t-tests.


Secretin receptor constructs.

Secretin receptor mutants were designed and prepared based on a careful review of studies that have examined sites critical to G protein coupling with other GPCRs (16, 33, 34, 53, 55, 56, 66, 67, 71, 74, 75), including the analysis of previous studies performed with other class B GPCRs (1, 48, 12, 21, 22, 2528, 38, 44, 45, 57, 6264). Shown in Fig. 1 is the alignment of sequences representing the predicted intracellular loop regions of structurally related class B GPCRs performed using Clustal W2 (39). Residues that have been found to be of importance in previous studies are highlighted in bold, along with the secretin receptor residues that were mutated in the present report. Each of these residues has been mutated to an alanine residue, and His157 was also changed to an arginine. Notation for the identification of these constructs is noted as well in Fig 1. Constructs with mutations of His157 were identified as H157A and H157R. The other alanine-replacement mutants were identified according to their loop locations: first loop R153A (ICL1), second loop R231A/K232A (ICL2), third loop K302A/L303A (ICL3A), and third loop R318A/R321A (ICL3B).

Fig. 1.

Multiple sequence alignment of class B receptors with published intracellular loop mutagenesis studies. Residues outlined by gray boxes indicate the predicted regions of the transmembrane segments. Bold residues indicate residues that have been mutated in one or more previous studies. Residues mutated in the present study are shown in white outlined by black boxes on the bottom line, with the names we use in the text and the mutated residue numbers indicated below. Constructs with multiple mutations are named using the combination of the regions mutated. Sequences were aligned using Clustal W2 (European Bioinformatics Institute website, http://www.ebi.ac.uk/Tools/msa/clustalw2/).

Effects of mutations on secretin binding.

Each of the secretin receptor constructs bound secretin saturably and/or generated signals that were significantly above those representing unstimulated background levels in intact cell assays, providing evidence for normal biosynthesis and trafficking of the mutant constructs. The pharmacological characteristics of receptor binding were determined using membrane preparations enriched in plasma membrane to avoid the potential impact of receptor internalization in intact cells. Data were analyzed using the LIGAND program (46) to give Ki and Bmax values (Table 1). The affinities of secretin binding to those constructs were not different from that of wild-type secretin receptor, except for the position 157 mutants that did not provide adequate saturable binding to analyze (Fig. 2 and Table 1). Of note, the receptor densities were also similar for all of these constructs, except for the position 157 mutants (Table 1).

View this table:
Table 1.

Quantitative analysis of binding and biological activity data

Fig. 2.

Effects of mutations on secretin binding. Shown are curves reflecting the abilities of secretin to complete for binding of the radioligand, 125I-(Tyr10)rat secretin-27, to COS-1 cell membranes expressing the intracellular loop 1 (A), intracellular loop 2 (B), and intracellular loop 3 (C), as well as double (D) and triple/quadruple (E) mutant constructs. Data points represent percentages of maximal saturable binding observed in the absence of the competing peptides for each construct, expressed as means ± SE of duplicate values from a minimum of 3 independent experiments. Of note, the first intracellular loop mutants, H157A and H157R constructs, did not exhibit detectable saturable binding (data not shown). Ki and Bmax values are shown in Table 1.

Effects of mutations on secretin-stimulated cAMP responses.

Each of the secretin receptor constructs, including the position 157 mutants that did not yield a saturable secretin radioligand binding signal, exhibited concentration-dependent, secretin-stimulated cAMP responses (Fig. 3, Table 1). However, the magnitude of the cAMP responses varied considerably among the different mutant receptor constructs. Basal cAMP levels of 5 ± 1 pmol/106 cells were present in the cells expressing wild-type secretin receptor, with this not different for any of the other receptor constructs. Of note, this also included the H157R mutation with basal level of cAMP of 7 ± 2 pmol/106 cells, despite evidence in the literature that this mutation in the rat secretin, human VPAC1, rat glucagon, and rat parathyroid hormone receptors may result in constitutive activity with elevated basal levels of cAMP (11, 12, 22, 57). Stimulation of the wild-type receptor led to an ∼36-fold increase of cAMP levels, achieving an Emax value of 175 ± 2 pmol/106 cells. Each of the mutant secretin receptor constructs, except for that combining all four mutations (ICL1/2/3A/3B), exhibited concentration-dependent increases in cAMP in response to secretin. The ICL1/2/3A/3B construct did not stimulate a significant cAMP response to any concentration of secretin tested (P > 0.05). The potencies of secretin to stimulate cAMP responses were also quite varied among the mutant receptor constructs. The EC50 for the wild-type receptor was 22 ± 3 pM. The majority of the single-mutant region constructs did not exhibit statistically significant changes in EC50 from that of the wild-type receptor; however, the EC50 value for the ICL3B construct was significantly increased relative to the control value (P < 0.05) (Table 1). Each of the double and multiple mutants exhibited significant differences in EC50 values (P < 0.01), except for the ICL2/3A construct, which did not exhibit significant change in EC50 from the wild-type receptor.

Fig. 3.

Effects of mutations on secretin-stimulated cAMP responses. Shown are curves of intracellular cAMP responses to increasing concentrations of secretin in COS-1 cells expressing the intracellular loop 1 (A), intracellular loop 2 (B), and intracellular loop 3 (C), as well as double (D) and triple/quadruple (E) mutant constructs. Data points represent cAMP responses expressed as means ± SE of 4 independent experiments performed in duplicate. EC50 and Emax values are shown in Table 1.

Effects of mutations on secretin-stimulated intracellular calcium responses.

The secretin-stimulated intracellular calcium responses are shown in Fig. 4 and Table 1. The efficacies of the mutant receptors to stimulate calcium responses were quite varied among the constructs. In several cases, there was no significant increase over basal values in response to secretin stimulation. The H157A and H157R receptor constructs did not exhibit significant calcium responses to secretin stimulation that were above basal levels. The calcium response for the ICL1 construct was 23% of that of wild-type secretin receptor, while all of the multiple-site mutant constructs containing this mutation did not significantly increase their calcium responses above their basal levels. Mutation of the ICL2 and ICL3A regions did not significantly affect calcium signaling, while the ICL3B mutation exhibited ∼28% of the calcium response observed in wild-type secretin receptor. Interestingly, the receptor construct with the combination of the ICL2 and ICL3A mutations was still able to activate calcium mobilization to a similar level as was observed for the wild-type secretin receptor. The potencies for secretin to stimulate intracellular calcium responses for each of the constructs that exhibited significant responses were not statistically different from that observed for the wild-type receptor.

Fig. 4.

Effects of mutations on secretin-stimulated intracellular calcium responses. Shown are curves of intracellular calcium responses to increasing concentrations of secretin in COS-1 cells expressing the intracellular loop 1 (A), intracellular loop 2 (B), and intracellular loop 3 (C), as well as double (D) and triple/quadruple (E) mutant constructs. Data points represent the calcium responses expressed as means ± SE of changes in the 380/340 nm ratio of Fura 2-AM from a minimum of 3 independent experiments performed in duplicate. EC50 and Emax values are shown in Table 1.

Direct analysis of G protein coupling.

In an attempt to confirm interactions of the mutated receptors with the relevant heterotrimeric G proteins, we used a BRET assay between the YFP-tagged receptor and Renilla luciferase (Rlu)-tagged Gα subunits co-transfected into COS-1 cells. While we were able to elicit significant BRET signals with the Gαs subunit, the analogous assays with the Gαq and Gαi subunits did not yield significant signals above background levels. This likely reflects the much lower affinity interaction of Gq with the receptor, with the natural agonist-stimulated intracellular calcium responses two orders of magnitude to the right of the cAMP responses, as is typical of class B GPCRs. Pertussis toxin treatment of cells expressing the secretin receptor has been shown to have no effect on signaling (11), providing confirmatory evidence that coupling of this receptor with Gαi does not occur.

As a negative control for the Gαs association assay, we used the type 2 cholecystokinin receptor that does not couple with this G protein (74). We examined association of the Gαs-Rlu with the YFP-tagged receptor constructs after secretin stimulation. Consistent with the observed cAMP data, the H157A, H157R, and ICL1/2/3A/3B mutant secretin receptor constructs did not exhibit significant increases in BRET signals above background (0.010 ± 0.006, 0.003 ± 0.006, and 0.005 ± 0.005, respectively, P > 0.05). The remaining secretin receptor constructs, wild-type (0.038 ± 0.006), ICL1 (0.035 ± 0.005), ICL2 (0.032 ± 0.007), ICL3A (0.033 ± 0.006), and ICL3B (0.037 ± 0.005), all exhibited significant BRET signals above background levels (P < 0.001).

Effects of mutations on receptor localization and expression.

To examine whether the signaling defects observed with some of the mutant receptor constructs could be due to defective biosynthesis and trafficking to the cell surface, we studied receptors expressed in transfected COS-1 cells. Twenty-four hours after transfection, cells were moved to coverslips for microscopy and were plated for preparation of cellular lysates for Western blotting. Cells on coverslips were fixed, but not permeabilized, in preparation for immunostaining. Cell lysates were run on SDS-PAGE gels and transferred to PVDF membranes. The fixed cells and the immunoblot membranes were probed with anti-hSecR(3044) polyclonal antibody. This antibody, which specifically labels the human secretin receptor on fixed cells and in cell lysates, showed no signal in the absence of transfected secretin receptor variants (data not shown). We examined the localization and expression of the single mutants as well as the receptor containing all of the mutations (Fig. 5, A and B). Interestingly, the cell surface expression of the mutant receptors was similar to that of wild-type secretin receptor, except for the H157A and H157R receptor constructs (Fig. 5A). Those receptor constructs exhibited very little cell surface expression, despite their total cellular expression levels being comparable to that of the other receptors (Fig. 5B). This indicates that the H157A and H157R receptor constructs were predominantly trapped within intracellular biosynthetic organelles and were not available for ligand binding or agonist-stimulated signaling. This helps to explain our observations that these constructs did not exhibit significant ligand binding and have greatly diminished cAMP signaling.

Fig. 5.

Effects of receptor mutations on receptor localization and expression. Shown are representative microscopic and immunoblot images. A: localization of receptor surface expression on nonpermeabilized cells using anti-hSecR(3044) antibody. B: expression of human secretin receptor constructs from whole cell lysates using anti-hSecR(3044) antibody. Immunostaining and immunoblot images represented are from the same experiment. Data are representative of at least 3 independent experiments.

Relative effects on signaling pathways.

Relative effects of the individual site mutations on cAMP and calcium signaling were determined by plotting the responses relative to each other, as percentages of responses in wild-type receptor (Fig. 6). The diagonal line represents proportionate responses in these two pathways. Points above or below the line represent disproportionate responses, with points above the line reflecting larger effects on cAMP and those below the line reflecting greater effects on calcium. Interestingly, the ICL2 mutation was nearly on the line indicating that the effects of this construct likely affect both pathways similarly. The ICL1 and ICL3B mutations affected calcium signaling more than cAMP signaling, while the opposite was true for the ICL3A construct.

Fig. 6.

Relationship between cAMP and calcium responses. Shown is the comparison of the cAMP and calcium responses of the single mutants expressed as the percentages of wild-type responses. Data reflect means ± SE of normalized data from a minimum of 3 experiments performed in duplicate.


The present work has explored the molecular basis for G protein coupling with the prototypic class B G protein-coupled secretin receptor. This receptor has a wide variety of biological actions. These include the physiological stimulation of secretion of biliary and pancreatic bicarbonate-rich fluid and inhibition of gastric secretion and gastric emptying (14, 31). Secretin has also been shown to stimulate gastric secretion of mucus and pepsinogen, duodenal secretion from Brunner's glands, and islet cell secretion of insulin, glucagon, pancreatic polypeptide, and somatostatin, as well as to inhibit the secretion of gastric acid (59). This hormone also has positive trophic effects on pancreas and biliary ducts (13, 60) and inhibition of trophic effects of gastrin on the oxyntic, duodenal, and colonic mucosa (69). Roles for secretin as a neuropeptide have been described that include modulation of ANG II-stimulated vasopressin secretion to regulate osmolarity (41) and effects on basket cell communication with Purkinje cells (76). While it is likely that the positive secretory effects of this hormone are mediated by its potent stimulation of cAMP through activation of Gs, mediation of its motility and trophic effects is less clear. Some of these effects might reflect stimulated increase in intracellular calcium through activation of Gq, but they could also be affected by modulatory effects of cAMP. Elucidation of the molecular basis for the coupling of each of these G proteins with the secretin receptor and the development of receptor constructs that differentially couple with each G protein could improve our understanding and facilitate the development of agonists exhibiting signaling bias.

Most of the current insights in G protein coupling to members of the GPCR superfamily have come from studies with members of the class A family (71, 72), believed to be structurally distinct from the class B family (37). Another important difference in these two families is the characteristic tendency of class B GPCRs to couple with more than one G protein, while single predominant coupling is more common in class A GPCRs (7, 8, 20, 48). Nevertheless, it seems clear that ligand binding to receptors in both of these families results in a conformational change in their helical bundles that results in the exposure of critical epitopes for the interaction with their G proteins (71), and the same set of G proteins couple with members of both of these families.

For the class A GPCRs, extensive mutagenesis has been performed for residues throughout regions exposed at their cytosolic face (71, 72). Additionally, recent structural studies have provided direct insights into the conformational changes that occur on activation of some members of this family (5356). These studies provide insights into the potential pockets for G protein interaction. Structural data from the β2-adrenergic receptor show that Gαs and an antibody that stabilizes the active conformation occupy the same pocket of this receptor (53, 55). The same pocket was also identified with a carboxyl-terminal peptide of transducin that was cocrystallized with opsin (56).

Two sequence patterns known to be important for activation of many class A GPCRs are the E/DRY motif in the third transmembrane segment (TM3) and the NPxxY(x5–6)F motif in TM7 (67). There is evidence that these motifs form ionic interactions or contribute to hydrogen bonding networks that help maintain the inactive state of receptors (16, 66, 75), and thus may only affect G protein coupling indirectly. In support of this assertion the structural data do not show any direct interactions of the G proteins with these motifs (5356). Neither of these sequence motifs is present in class B GPCRs.

Mutation of the basic residues in a motif present at the carboxyl-terminal end of the third intracellular loop of many class A GPCRs, having the sequence BBxxB or BxxBB, has been shown to decrease or eliminate agonist-induced G protein activity (24, 36, 40, 42, 47, 68). Some studies have demonstrated that exchanging the third intracellular loop regions of distinct GPCRs can result in a shift in the specificity of the interacting G protein (71, 72). Other studies have also supported contributions of the amino-terminal end of the third intracellular loop and the second intracellular loop for specificity of G protein isotypes that interact with receptors (71, 72). It has been demonstrated that the first intracellular loop also affects G protein coupling in an isotype-specific manner (34, 71, 72). Structural studies indicate that the heterotrimeric G protein complex binds to a fold that is primarily formed by the second and third intracellular loops (54, 56).

More limited mutagenesis of class B GPCRs has also been performed to identify residues involved in G protein coupling to those receptors, recognizing that these are structurally distinct from the class A GPCRs (23, 37, 71). These studies have identified residues in all three intracellular loop regions that affect G protein function, while having minimal effects on their ligand binding. The involvement of the carboxyl-terminal tail of these receptors in G protein activation has been less marked and less consistent (68, 26, 28, 44).

Within the first intracellular loop, an arginine (ICL1) within a conserved motif having the amino acid sequence xCxR and a conserved histidine at the interface with TM2 have been implicated in affecting G protein function, although there has been substantial variation among receptors. Alanine-replacement mutagenesis of the arginine, analogous to Arg153 in the secretin receptor, reduces cAMP responses without interfering with ligand binding in the human VPAC1, rat glucagon, and rat calcitonin receptor-like receptors (4, 8, 45). However, this mutation in the human secretin receptor exhibited both normal binding and cAMP responses, while resulting in complete loss of calcium responses. Alanine-replacement mutagenesis of the histidine at the interface with TM2, analogous to His157 in the secretin receptor, decreases the surface expression, ligand binding, and cAMP responses of rat parathyroid hormone, rat secretin, human VPAC1, and rat glucagon receptors, while replacing this histidine with an arginine in those receptors has been reported to increase basal cAMP levels (11, 12, 22, 57). The latter was not observed in other studies of the rat glucagon-like peptide-1, rat glucose-dependent insulinotropic peptide, and rat glucagon receptors (8, 21, 64). In the present report, both the human secretin H157A and H157R mutants were deficient in ligand binding and showed significant loss in maximal cAMP and calcium signaling, likely reflecting poor surface expression of these receptors.

Within the second intracellular loop, the most consistent effects involved a dibasic motif analogous to Arg231-Lys232 in the secretin receptor (ICL2). Mutating this in the rat parathyroid hormone receptor, rat GLP1 receptor, human VPAC1 receptor, and human calcitonin receptor-like receptor results in normal ligand binding, and cAMP signaling, but loss of IP signaling (5, 7, 27, 44, 45). Unexpectedly, in the present work, alanine-replacement mutagenesis of Arg231 and Lys232 of the secretin receptor resulted in normal ligand binding and almost full cAMP and calcium responses.

Within the third intracellular loop, mutation of a central lysine-leucine motif of the opossum parathyroid hormone receptor, rat GLP1 receptor, and human VPAC1 and VPAC2 receptors, analogous to the human secretin receptor residues Lys302-Leu303 (ICL3A), results in defects in agonist-stimulated cAMP production, calcium mobilization, and ligand binding to varying degrees (1, 7, 25, 38, 45, 62). In the present work, replacing these residues with alanines in the human secretin receptor resulted in normal ligand binding and calcium signaling, but reduced cAMP responses. Another sequence motif consisting of two basic residues interrupted by two small nonpolar residues, analogous to Arg318 and Arg321 in the human secretin receptor (ICL3B), has been shown to have a pronounced effect in the rat GLP1 receptor, rat glucagon receptor, and human VPAC1 receptors (1, 7, 8, 44, 45). Similar effects were observed in the present work for the human secretin receptor.

We also studied these mutations of the secretin receptor, except for the His157 mutations, in combination to elucidate how they might complement each other. With the exception of the ICL2/3A mutation, all of the multiple mutations exhibited significant loss of efficacy for both of the signaling pathways initiated at Gαs and Gαq, as well as a loss of potency for activating the Gαs pathway. Interestingly, the ICL2/3A mutant did not have a significant increase in EC50 for cAMP activation despite having only 50% of the wild-type maximal activity, and this mutant did not exhibit significant changes in calcium signaling. The receptor construct containing all of the noted mutations (ICL1/2/3A/3B) was not able to respond to secretin stimulation with a significant change in cAMP. The calcium responses for all constructs containing either the ICL1 or ICL3B mutations with any of the other mutations studied did not exhibit sufficient stimulation above basal to determine pharmacological parameters.

In the direct association studies of these regions in this study we found that most of the single mutant region constructs (ICL1, ICL2, ICL3A, ICL3B) continue to associate with Gαs. The His157 mutants and the ICL1/2/3A/3B mutant, however, exhibited no signal in the BRET assay with this G protein. This implies that the loss of signaling is a direct result of the loss of G protein binding.

In summary, these data support the existence of shared as well as differential determinants for the coupling of G proteins with the class B secretin receptor. Mutation ICL1 disproportionately reduced calcium stimulation, presumably by disrupting specific interactions with Gαq. Mutation ICL3A disproportionately reduced cAMP responses while maintaining the calcium responses, likely reflecting disruption of interactions with Gαs. Mutations ICL2 and ICL3B had more general effects on both G proteins. This series of mutants provides valuable tools to further examine the signaling events that lead to these defects. In the absence of precise structural data for the class B GPCRs, these receptor constructs provide insights into the relative importance of these regions in coupling to and activating distinct G proteins.


We acknowledge the support of National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-46577, including a supplement to support G. L. Garcia, and the Mayo Clinic.


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: G.L.G., M.D., and L.J.M. conception and design of research; G.L.G. performed experiments; G.L.G. and L.J.M. analyzed data; G.L.G., M.D., and L.J.M. interpreted results of experiments; G.L.G. prepared figures; G.L.G. drafted manuscript; G.L.G., M.D., and L.J.M. edited and revised manuscript; G.L.G., M.D., and L.J.M. approved final version of manuscript.


We acknowledge the excellent technical support of M. L. Augustine and A. M. Ball.


  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.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  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.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
View Abstract