Differences in Partial Agonist Action at Cholecystokinin Receptors of Mouse and Rat Are Dependent on Parameters Extrinsic to Receptor Structure: Molecular Cloning, Expression and Functional Characterization of the Mouse Type A Cholecystokinin Receptor1
Abstract
The mouse cholecystokinin (CCK) receptor is functionally distinct from the extensively studied rat receptor on the basis of differences in binding and biological activity of phenethyl ester analogs of CCK. These are partial agonists at the rat receptor and full agonists at the mouse pancreatic receptor. To explore this, we cloned the cDNA for the mouse type A CCK receptor, established a receptor-bearing Chinese hamster ovary (CHO) cell line and characterized its binding and biological characteristics. Despite 25 differences in amino acid sequence from the rat receptor, including a seven-amino acid insertion in the third intracellular loop, mouse and rat receptors were functionally indistinguishable when expressed in CHO cells. Of note, in the mouse pancreatic cell environment, a stable analog of guanosine triphosphate significantly inhibited binding of CCK-OPE, whereas it had no effect on binding to the same receptor on the CHO-CCKM cell line or to the rat receptor in either environment of the acinar cell. This likely reflects a difference in coupling of the mouse receptor to its G protein in the natural environment of the acinar cell. This may relate to differences extrinsic to the receptor, in the stoichiometry or character of G proteins or in the composition or organization of the lipid environment of the mouse acinar cell membrane. Although this may require complementation of the unique sequence of the mouse receptor, that structure alone is insufficient to explain this phenomenon. Receptor microenvironment makes an important, yet often ignored, contribution to receptor function.
Similarities and differences between structurally related but functionally distinct molecules have been extremely useful for elucidating the functional roles of structural domains of those proteins. Related receptor molecules that are functionally distinct provide a good opportunity for this type of analysis and indeed have been the basis for a large number of chimeric receptor studies (Holtmann et al., 1995). Theoretically, the more structurally similar two molecules are to each other while retaining functional differences, the more focused the insights might be into the responsible structural element. When the same receptor expressed in different species has an identifiable functional difference, it represents an ideal setting in which to apply this approach because the structural differences between such molecules are usually minimal.
The mouse type A CCK receptor provides this type of opportunity because it is reported to possess novel and interesting characteristics (Matozaki et al., 1989; Bianchi et al., 1994). At this target, native CCK has been reported to elicit a normal biological response, identical to that in the extensively studied rat pancreatic acinar cell. In contrast, in the mouse pancreatic acinar cell, phenethyl ester analogs of CCK [CCK-JMV-180 (Galas et al., 1988) and CCK-OPE (Gaisano et al., 1989)] appear to be full agonists, eliciting a full concentration-response curve that includes the supramaximal inhibitory region (Matozaki et al., 1989), whereas they are only partial agonists at the rat pancreatic receptor, at which they are equally active to natural CCK but elicit no supramaximal inhibition (Gaisano et al., 1989). This observation has been confirmed and correlated with differences in degrees of stimulation of phoshatidyl inositol hydrolysis by this CCK analog in these two species (Bianchi et al., 1994). A binding correlate to this has also been reported in which this analog binds to rat cells with a single affinity and binds to mouse cells with two distinct affinities (Matozaki et al., 1989). The molecular basis for these differences in receptor behavior in these two species has not been determined.
The CCK receptor is a G protein-coupled receptor in the betaadrenergic receptor family that has physiological functions to stimulate pancreatic exocrine secretion, contraction of smooth muscle in the gallbladder and regions of the digestive tract and neuronal activity in select sites within the central and peripheral nervous systems (Mutt, 1980). It is a phospholipase C agonist, eliciting a prominent intracellular calcium response. To date, the sequence of the mouse CCK receptor cDNA has not been reported, despite cloning of cDNAs for this receptor in rat (Wank et al., 1992), human (Ulrichet al., 1993), guinea pig (De Weerth et al., 1993) and rabbit (Reuben et al., 1994). These receptor sequences have been highly homologous (95% similar, with paired identities in excess of 87%).
Our goals in the present study were to clone the mouse CCK receptor cDNA, compare the sequence with that of the other known CCK receptor sequences from other species, establish a recombinant mouse CCK receptor-bearing cell line and functionally characterize this receptor expressed in that model cellular environment. We also explored the GTP sensitivity of CCK-OPE binding to the mouse and rat type A CCK receptors in their natural cellular environment and in the model cell system to further elucidate the molecular basis for this species-specific effect.
Methods
Materials.
Synthetic sulfated CCK-8 was purchased from Peninsula Laboratories (Belmont, CA). Other analogs of CCK were synthesized and purified in our laboratory, as we have reported (Powerset al., 1988; Gaisano et al., 1989). The stable analogue of guanosine triphosphate, GppNHp, was purchased from Sigma Chemical (St. Louis, MO).
CHO-K1 cells that do not express endogenous type A or B CCK receptors (Hadac et al., 1996) were purchased from American Type Culture Collection (Rockville, MD). This also was the cell line previously transfected to express the rat type A CCK receptor (Hadacet al., 1996). The cell line stably expressing that receptor (CHO-CCKR) has been extensively characterized (Hadac et al., 1996). Cells were cultured onto tissue culture plastic in Ham’s F-12 medium in a humidified incubator at 37°C in an atmosphere of 5% CO2. Cells were typically passaged twice weekly. For studies, cells were lifted mechanically, triturated, and washed with the relevant medium in preparation for use.
Cloning the mouse CCK receptor cDNA.
Pancreata were freshly harvested from female C57BL mice according to procedures and approval of the Mayo Clinic Institutional Animal Care and Use Committee. For cDNA cloning, tissue was flash-frozen in liquid nitrogen, and total RNA was extracted according to the guanidinium thiocyanate-guanidine hydrochloride method of Han et al. (1987). First-strand cDNA was synthesized with 12.5 μg of RNA as template and through the use of oligo(dT) primers with AMV-reverse transcriptase (Boehringer-Mannheim, Indianapolis, IN) according to the manufacturer’s recommended conditions. The reaction was allowed to proceed at 42°C for 1 hr and at 65°C for 10 min before cooling to 4°C.
The CCK receptor cDNA was cloned by polymerase chain reaction using degenerate oligonucleotide primers for both 5′ and 3′ sequences on the basis of existing known CCK receptor sequences and modified on the basis of mouse codon preferences. Primer sequences were 5′-ATG GAY GTS GTS GAY WSH YT-3′ and 5′-TCA DGG DGG DGG NRC DSW DGY DSW CAT-3′. Amplification was performed using first-strand cDNA as template withTaq DNA polymerase (Boehringer-Mannheim) according to the manufacturer’s recommended conditions, with denaturation at 95°C for 10 min and proceeding with 35 cycles of 94°C for 1 min, 55°C for 2 min and 72°C for 3 min. The expected ∼1.3-kb product was visualized with ethidium bromide on an agarose gel, excised and extracted with Qiaex (Qiagen, Chatsworth, CA). This product was cloned into pT7blue (Novagen, Madison, WI) and subcloned into pBK-CMV (Stratagene, La Jolla, CA) at the BamH1 and SalI sites. It was then sequenced in both directions using the dideoxynucleotide chain-termination method of Sanger et al. (1977). Three independent clones were isolated from separate amplifications to ensure the integrity of the product.
The cDNA sequence was translated using GCG software (Genetics Computer Group, Madison, WI). The seven putative transmembrane domains were determined by hydropathy analysis of the predicted amino acid sequence and by analogy with other CCK receptor sequences.
Establishment of the mouse CCK receptor-bearing cell line.
The mouse cDNA in the eukaryotic expression vector that incorporates the sequence for neomycin resistance was used to transfect the CHO-K1 cells described above. This was performed through lipofection with Lipofectin (Life Technologies, Gaithersburg, MD). After 48 hr, neomycin-resistant cells were selected with 1 mg/ml G418. Resistant colonies were selected and screened using fluorescence-activated cell sorting with a fluorescein-conjugated analog of CCK as we previously described (Hadac et al., 1996). This line was cultured under conditions described above for the other CHO cell lines.
CCK receptor binding characteristics.
Binding studies were performed with intact cells and enriched plasma membranes of cells as we previously described (Miller et al., 1981; Gaisanoet al., 1989). The radioligand125I-d-Tyr-Gly-[(Nle28,31)CCK26-33] has been previously demonstrated to behave identically to natural CCK-8 (Powers et al., 1988). In select studies, a CCK-OPE-like radioligand was also used. This represented125I-d-Tyr-Gly-[(Nle28,31)CCK26–32]phenethyl ester (Gaisano et al., 1989). Approximately 0.5 million cells or membranes containing 1 to 10 μg of protein were incubated with a constant amount of radioligand (3–5 pM) in the presence of increasing amounts of competing ligands for 1 hr at 22°C in Krebs-Ringer-HEPES medium (KRH) containing 25 mM HEPES, pH 7.4, 1 mM KH2PO4, 104 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1.2 mM MgSO4, 0.2% bovine serum albumin, 1 mM phenylmethylsulfonyl fluoride and 0.01% soybean trypsin inhibitor. CHO cell line membranes and pancreatic membranes were prepared by the methods previously described (Hadac et al., 1996; Lutz et al., 1993). Bound and free radioligands were separated using a Skatron cell harvester with receptor-binding filtermats. Bound radioactivity was quantified with a gamma spectrometer. Nonspecific binding was determined in the presence of 1 μM CCK.
The sensitivity of the binding of the CCK-like and CCK-OPE-like radioligands to stable analogues of guanosine triphosphate was determined using membrane preparations and Gpp(NH)p. Binding was performed as described above.
CCK-stimulated cell signaling.
Signaling was determined by observing intracellular calcium responses. This was achieved with cells grown onto glass coverslips to 50% to 75% confluence. They were then loaded with 5 μM Fura2-AM (Molecular Probes, Eugene, OR) for 10 min at 22°C, followed by a 10-min incubation at 37°C. Cells were then washed and studied on the heated stage (37°C) of a Zeiss Axiovert inverted fluorescence microscope. Excitation was performed at 340 and 380 nm, with emission at 520 nm, and data were analyzed using the Attofluor digital imaging system with Ratio Vision software (Atto Instruments, Rockville, MD).
Statistical analysis.
Observations were performed in duplicate a minimum of three times in independent experiments. Statistical differences were assessed using the Mann-Whitney test for unpaired values, with P < .05 considered to be significant.
Results
The mouse type A CCK receptor cDNA was cloned from the mouse pancreas. The nucleotide sequence and translation are shown in figure1, with putative transmembrane domains highlighted. At the nucleotide level, this sequence was 94% identical to that of the rat, with the protein sequence 95% identical and 98% similar in the two species. Of particular interest, there was a seven-amino acid insertion (GGGGGGS) in the predicted third intracellular loop of the mouse receptor that has not been seen in CCK receptors from any other species cloned to date. There were 25 differences in amino acid sequence between the mouse and rat, with eight of these present in at least one other cloned CCK receptor (fig.2). The major site of changes in sequence was in predicted intracellular domains, with the third intracellular loop predominant, and with the carboxyl-terminal tail also harboring several changes in sequence. Of note, there were only two unique mouse residues within predicted transmembrane domains and no unique residues in any site predicted to be in the ectodomain of the receptor.
Nucleotide sequence and translation of mouse type A CCK receptor cDNA, with predicted transmembrane domains identified by underlined and bold residues.
Alignment of amino acid sequence of the mouse type A CCK receptor with those of the same receptor from other species that have been reported to date. Identities representing a consensus are noted in black, and similarities are noted in gray.
CCK receptor binding characteristics.
When expressed in membranes from the CHO-CCKM cell line, the mouse receptor bound the CCK radioligand saturably and with high affinity (Ki= 1.7 ± 1.5 nM) (fig. 3). This was not different from binding to membranes from the CHO-CCKR cell line that expresses the rat CCK receptor (Ki = 1.4 ± 0.9 nM) or to membranes prepared from mouse pancreas (Ki = 1.1 ± 0.3 nM). Figure4 illustrates the structural specificity of the mouse CCK receptor, showing the ability of a series of agonists and antagonists to compete for binding of the CCK radioligand to intact CHO-CCKM cells. Relative affinities of each are not different from their binding properties to the rat pancreatic receptor (Gaisanoet al., 1989; Hadac et al., 1996). The CHO-CCKM cell line expressed ∼105,000 sites/cell, as determined by LIGAND analysis of Munson and Rodbard (1980), which was not significantly different from the receptor density on the CHO-CCKR cell line (Hadacet al., 1996).
Competition of CCK for binding of a CCK radioligand to membranes from the CHO-CCKM cell line and from mouse pancreas. Binding data are expressed as percentages of maximal binding that was observed in the absence of competitor. Values represent mean ± S.E.M. of at least three independent experiments performed in duplicate.
Competition of receptor ligands for binding of a CCK radioligand to intact CHO-CCKM cells. Binding data are expressed as percentages of maximal binding which was observed in the absence of competitor. Values represent mean ± S.E.M. of at least three independent experiments performed in duplicate.
The binding of the CCK-OPE radioligand to the mouse CCK receptor expressed on the CHO-CCKM cells was carefully characterized. Shown in figure 5 are competition-binding curves for displacement of this radioligand by CCK-OPE and for CCK-8. This also was analogous to that we reported for the rat CCK receptor (Gaisano et al., 1989).
Competition of receptor ligands for binding of a CCK-OPE radioligand to membranes from the CHO-CCKM cell line and from mouse pancreas. Binding data are expressed as percentages of maximal binding that was observed in the absence of competitor. Values represent mean ± S.E.M. of at least three independent experiments performed in duplicate.
The GTP sensitivity of ligand binding to CCK receptors was also determined. We previously demonstrated that CCK binding, but not CCK-OPE binding, to rat pancreatic membranes is sensitive to the stable guanosine triphosphate analog GppNHp (Gaisano et al., 1989). The rat receptor expressed on CHO-CCKR cells expressed similar characteristics (Hadac et al., 1996). Membranes prepared from mouse pancreas had different characteristics. The binding of both CCK and CCK-OPE to these membranes was significantly inhibited by GppNHp (p < .02) (fig.6). Of note, when the same mouse receptor was expressed on the CHO-CCKM cells, the binding of the CCK-OPE radioligand was no longer significantly affected by the Gpp(NH)p (fig.6).
Sensitivity of radioligand binding to a stable analog of GTP. Shown are curves representing the ability of Gpp(NH)p to inhibit the binding of CCK and CCK-OPE radioligands to membranes from the CHO-CCKM cell line and from mouse pancreas. Binding data are expressed as percentages of maximal binding that was observed in the absence of competitor. Values represent mean ± S.E.M. of at least three independent experiments performed in duplicate. Asterisks mark inhibition of binding of the CCK-OPE radioligand that was significant at p < .02.
CCK-stimulated cell signaling.
The mouse CCK receptor-bearing CHO-CCKM cells had clear and appropriate intracellular calcium responses to both CCK and CCK-OPE, as have been described for rat pancreatic acini and rat CCK receptor-bearing CHO cell lines (Matozakiet al., 1990; Hadac et al., 1996) (fig.7). In these cells, the pattern of responses was dependent on both the agonist used and its concentration, as has been previously observed (Matozaki et al., 1990).
Typical intracellular calcium responses of the CHO-CCKM cells stimulated by noted concentrations of CCK and CCK-OPE. Values were determined by the fluorescent ratio imaging of cells loaded with Fura-2, as described in Methods. Responses were typical of a minimum of three independent experiments.
Discussion
In this study, we cloned the mouse type A CCK receptor cDNA, established a CHO-CCKM cell line stably expressing this receptor and characterized its binding and biological activity. Although highly homologous with the CCK receptors previously cloned from other species (Wank et al., 1992; Ulrich et al., 1993; De Weerth et al., 1993; Reuben et al., 1994), the mouse receptor had a number of interesting and potentially unique features. The most noteworthy was a seven-amino acid insertion in the third intracellular loop. All except four of the 25 amino acid differences between the mouse and rat sequences were present within predicted intracellular domains, with the third loop the site of most of these (17 differences). The four amino acid differences in nonintracellular domains were all predicted to be within transmembrane domains, with two of these present in CCK receptors from all species except the rat. The unique mouse residues were present within the predicted first and fifth transmembrane domains. Both of these represented residues that were quite homologous to those present in the rat (I65F and I238V) and in each of the CCK receptors from other species previously cloned.
The specific receptor domains and residues that are critical for CCK binding to its receptor have not yet been defined. By analogy with other structurally related G protein-coupled receptors that bind peptide agonists, critical sites likely reside within the extracellular loop regions near the membrane and possibly within the outer third of key transmembrane domains (Schwartz and Rosenkilde, 1996). The latter sites have been particularly important for binding of nonpeptidyl ligands (Schwartz and Rosenkilde, 1996), and this appears to also be true for CCK family receptors (Beinborn et al., 1993). Given the paucity of changes in regions likely to be involved in binding, it is not surprising that the recombinant mouse CCK receptor had ligand binding characteristics that were indistinguishable from those of the rat receptor.
The predominant changes in structure between the mouse and rat CCK receptors reside within domains that have been implicated in regulating coupling between receptor and G protein proximal effectors (Wadeet al., 1994, 1996). Such an impact could clearly explain the functional differences previously observed between mouse and rat CCK receptors (Matozaki et al., 1989; Bianchi et al., 1994). The partial agonist activity of phenethyl ester analogs of CCK acting at the rat CCK receptor could reflect different, and possibly less efficient, coupling in that species than in the mouse, in which it is a full agonist (Matozaki et al., 1989). More efficient coupling in the mouse pancreas could similarly explain the two-state binding (Matozaki et al., 1989) and the greater effect on phosphatidyl inositol hydrolysis (Bianchiet al., 1994) previously observed.
It was perhaps most surprising that the functional differences unique to the mouse CCK receptor were not reflected in its activity in the CHO-CCKM cell line. That cellular environment is fully capable of exhibiting full agonist activity at the CCK receptor because it displays all known responses to stimulation with the natural hormonal form, CCK-8. The inability of GppNHp to inhibit the binding of CCK-OPE to the mouse receptor expressed in the CHO-CCKM cell line while inhibiting the binding of this ligand to the same receptor when expressed in its natural pancreatic cellular environment suggests that there is a key difference extrinsic to the receptor. Possible explanations include differences between the CHO cell and the mouse pancreatic acinar cell in the stoichiometry or complement of G proteins or in the composition or organization of the lipid bilayer itself. It is possible that this represents the only difference between the function of mouse and rat pancreatic CCK receptors, but it is likely that the mouse receptor sequence critical for G protein coupling is also different, and possibly more efficient, than that of the rat receptor. This question will have to wait until the CCK receptors unique to each species can be expressed in the pancreas from another species. Little data now exist that provide insights into these possibilities. It is clear, however, that although the mouse CCK receptor sequence may be necessary to exhibit the unusual manifestations, it is not itself sufficient.
As expected, the mouse type A CCK receptor is structurally quite homologous to this receptor present in other species. This could have provided an ideal setting to focus in on the structural domain of this receptor responsible for the functional differences observed in the mouse CCK receptor, relative to CCK receptors from other species studied. The differences, however, must be subtle because they are only apparent using a ligand that is a partial agonist in one species and a full agonist in another. Binding and biological activity of the natural hormonal agonist ligand is indistinguishable in the mouse and all other species. Focusing on this subtle difference, in this study we identified a difference that is at least partially extrinsic to the receptor itself, requiring the natural milieu of the mouse pancreatic plasmalemma to express this difference. As we learn more about the determinants for the coupling of the CCK receptor to its G protein proximal effector(s), the explanations for this may become clearer.
Acknowledgments
The authors acknowledge the excellent technical assistance of D. Pinon and I. Ferber and the excellent secretarial help of S. Erickson.
Footnotes
-
Send reprint requests to: Laurence J. Miller, M.D., Center for Basic Research in Digestive Diseases, Guggenheim 17, Mayo Clinic, Rochester, MN 55905.
-
↵1 This work was supported by grants from the National Institutes of Health (DK32878) and the Fiterman Foundation.
- Abbreviations:
- CCK
- cholecystokinin
- CHO
- Chinese hamster ovary
- CCK-OPE
- Tyr-Gly-[(Nle28,31)CCK26–32]phenethyl ester
- CHO-CCKR
- rat cholecystokinin receptor-bearing Chinese hamster ovary cell line
- GTP
- guanosine triphosphate
- Gpp(NH)p
- 5′-guanylylimido diphosphate
-
- Received January 31, 1997.
- Accepted May 21, 1997.
- The American Society for Pharmacology and Experimental Therapeutics
