The Amino Terminus of Gα_z is Required for Receptor Recognition, Whereas its α4/β6 Loop Is Essential for Inhibition of Adenylyl Cyclase

Experimental Procedures

Materials.

AC2 and DOR cDNA were gifts from Randall Reed (Johns Hopkins School of Medicine, Baltimore, MD) and Christopher Evans (University of California, Los Angeles, CA), respectively. Pertussis toxin (PTX) was purchased from List Biological Laboratories, Inc. (Campbell, CA). HEK 293 cells were obtained from the American Type Culture Collection (Rockville, MD). [d-Pen^2,5]enkephalin (DPDPE) was from Research Biochemicals Inc. (Natick, MA). Antisera against Gα_z and Gα_t1 were from Gramsch Laboratories (Schwabhausen, Germany) and Transduction Laboratories (Lexington, KY), respectively. [³H]Adenine was purchased from Amersham (Buckinghamshire, UK). Plasmid purification columns were obtained from Qiagen (Hilden, Germany). Taq DNA polymerase, customized primers, restriction endonucleases, and cell culture reagents were obtained from Life Technologies Inc. (Grand Island, NY). All other chemicals were purchased from Sigma (St. Louis, MO).

Construction of Chimeric α Subunits.

A complete list of chimeric and mutational Gα subunits is shown in Table1. The chimeras zt40 and zt43 were constructed by polymerase chain reaction (PCR) using a pair of chimeric primers (Table 2). The full-length PCR product was sequenced using Sequenase Version 2.0 DNA sequencing kit from Amersham. Other chimeras were constructed by making use of convenient unique restriction sites and exchanging the corresponding restriction fragments of Gα_t1, zt40, or zt43 with Gα_z (Table 1). For example, the junctional restriction sites of tz36, tz60, and tz143 are BglII,AflIII, and BamHI, respectively. The insertion mutant +4t and PTX-resistant Gα_t1CG mutant were made by PCR using Gα_t1 as the template, and +4t/CG is generated by combining +4t and Gα_t1CG together through restriction cutting. −4z Is derived from Gα_z with the residues 11 to 14 deleted. For si143 and sz143, the portion of Gα_s was obtained by partial digestion with BamHI and then ligation with the C-terminal portion of either Gα_i2 or Gα_z. GTPase-deficient mutants of various chimeras were made by replacing Gα_t1 and Gα_z with Gα_t1RC (Ho et al., 1999) and Gα_zQL (Wong et al. 1992), respectively in the construction procedures. All the cDNA constructs were subcloned in the mammalian expression vector pcDNAI atEcoRI site.

Transfection of HEK 293 Cells and cAMP Accumulation Assay.

HEK 293 cells were cultured with Eagle's minimal essential medium (MEM) supplemented with 10% fetal calf serum (v/v), 50 U/ml penicillin, and 50 μg/ml streptomycin at 37°C in humidified air with 5% CO₂. They were cotransfected with various recombinant DNA constructs using DEAE-dextran/chloroquine method as described previously (Wong, 1994). Transfected cells were labeled with 1 μCi/ml of [³H]adenine in MEM with 1% fetal calf serum and treated with 100 ng/ml PTX as appropriate. Labeled cells were treated with proper receptor agonists in 20 mM HEPES-buffered MEM with 1 mM 1-methyl-3-isobutylxanthine for 30 min and the reactions were terminated by adding ice-cold 5% trichloroacetic acid with 1 mM ATP. Separation of labeled cAMP from other nucleotides was achieved by sequential ion exchange chromatography as described previously (Salomon, 1991). The cAMP levels were interpreted as the ratios of the counts per minute of [³H]cAMP fractions to those of [³H]ATP fractions and expressed as [cAMP/(cAMP + Total) × 1000]. Absolute values for cAMP accumulation varied between experiments, but variability within a given set of transfection was in general <10%. Data shown in the figures were the mean ± S.E.M. of three to five individual experiments performed in triplicate. ANOVA and paired t test with 95% confidence was used to analyze the significance between different treatment groups.

Western Blotting Analysis.

Crude membrane proteins from HEK 293 cells transfected with various chimeric Gα subunits were extracted as described previously (Ho and Wong, 1997). Each protein sample (50 μg) was resolved in 10% SDS-polyacrylamide gel and transferred to polyvinylidene fluoride membrane. Antiserum 3A-170 (Gramsch Laboratories, Schwabhausen, Germany) against the carboxyl terminus of Gα_z was used for the detection of Gα_z, tz, and ztz chimeras, whereas Gα_t1, zt40, and zt43 were identified with anti-Gα_t1 antiserum (Transduction Laboratories).

Results

The Role of the C-Terminus of Gα_z in Coupling with DOR.

The abilities of chimeric Gα subunits to couple to DOR were monitored by the βγ-mediated stimulation of AC2. This reporter system is advantageous because it could be generally applied to assess the coupling between different categories of receptor-G protein pairs without considering the functions of the Gα subunits. Moreover, AC2 is relatively insensitive to Gα_i-mediated inhibition (Taussig et al., 1994). It eliminates the possibility that Gα_z and βγ complex act on AC2 in an antagonistic fashion. The AC2 system is especially useful for checking the receptor coupling efficiencies of particular chimeric Gα subunits that have lost the integrity of effector interacting domains. A schematic representation of the first series of chimeric Gα subunits was shown in Fig. 1A. We attempted to examine the importance of the C-terminal tail of Gα_z on receptor coupling using this series of chimeras. In the vector control, HEK 293 cells expressing Gα_sQL, DOR, and AC2 were treated with or without PTX. Addition of 100 nM DPDPE stimulated the cAMP production to about two times the basal level in the absence of PTX treatment (Fig.1B). The DPDPE-induced enhancement was abolished in cells treated with PTX, indicating that the coupling between DOR and the endogenous G_i proteins was obstructed. Coexpression of recombinant Gα_z provided a larger pool of G protein trimers capable of coupling to DOR; hence, in the absence of PTX treatment, a 150% enhancement of the cAMP production was observed compared with the corresponding response in vector control. PTX treatment partially reduced the cAMP level of Gα_z-expressing cells, which suggested that G_z could functionally replace the role of endogenous G_i and mediated the DPDPE-induced cAMP production in a PTX-resistant manner. In the case of Gα_t1, the response in the absence of PTX treatment was significantly lower than that of vector control and was close to the basal value. Suppression of βγ-mediated AC2 activity was caused by the strong βγ-scavenging property of Gα_t1. Moreover, there was no increase of cAMP level in cells treated with PTX, which inactivated endogenous G_i and the recombinant G_t1. These results are consistent with those reported earlier (Tsu et al., 1997).

One of our recent studies showed that the Gα_z/Gα_t1 and Gα_z/Gα_i2 chimeras containing the last 36 amino acids of Gα_t1 and Gα_i2, respectively, can couple to DOR efficiently (Tsu et al., 1997). To determine whether the N-terminus is required for receptor coupling, a reversed chimera tz36, an Gα_t1 subunit with the C-terminal 36 residues of Gα_z (Fig. 1A, the third construct), was constructed and examined for its ability to couple to DOR. In this assay system, tz36 did not release βγ subunits for stimulating AC2 upon DOR activation (Fig. 1B). Subsequent substitution of larger C-terminal portions of Gα_t1 with Gα_z sequence up to 143 amino acids (tz60 and tz143) also did not rescue the coupling to DOR. Two other chimeras sz143 and si143 (Fig. 1A, the sixth and seventh constructs) were constructed by replacing the C-terminal tail of Gα_s with that of Gα_z or Gα_i2 portions. These two chimeras resembled tz143 and did not release βγ complex for AC2 stimulation upon DOR activation. These results further suggested neither the C-terminal 143 amino acids of Gα_i2 nor Gα_z could rescue the receptor coupling and the loss of function of tz143, sz143, and si143 may not be related to the choice of parental Gα subunits for chimera construction. Obviously, other essential receptor interacting regions were located on the N-terminal half of Gα_z sequence.

The last three chimeras of this series were constructed as mirror images of tz36, tz60, and tz143. The chimeras were denoted as zt319, zt295, and zt212, respectively (Fig. 1A, the eighth to tenth constructs; numbers refer to the last amino acid of Gα_z). Replacement of the C-terminal tail of Gα_z with that of Gα_t1may alter its coupling to DOR. However, all three zt chimeras seemed to couple to DOR and showed substantial increases of cAMP accumulation in the absence of PTX treatment (enhanced by 60, 50, and 65% for zt319, zt295, and zt212, respectively; see Fig. 1B). These results further reinforced the idea that the C-terminal 143 amino acids of Gα_z are sufficient for receptor recognition and some critical structural elements are located on the N-terminal half of Gα_z. Because these zt chimeras acquired the C-terminal tail of Gα_t1, they should be sensitive to PTX-mediated inactivation. Accordingly, there was no increment of cAMP levels in all three cases when the cells were pretreated with PTX (Fig. 1B).

The N-Terminus of α_z Is Essential for Receptor Coupling.

Recent studies highlighted the importance of the N-terminus of Gα subunit for receptor specificity (Kostenis et al., 1997). A second series of chimeras were thus constructed to verify the role of the N-terminus of Gα_z on DOR coupling. The chimera zt40 was constructed so that the N-terminal helix became Gα_z-like (Fig.2A, the third construct). In the absence of PTX, DPDPE-induced enhancement of cAMP accumulation in zt40-transfected cells resembled that of Gα_z-transfected cells. However, zt40 is PTX-sensitive because it acquired the C-terminus of Gα_t1 and the DPDPE-induced cAMP production was abolished in the presence of PTX (Fig. 2B). This result suggested that the N-terminal region of Gα_z was an essential determinant for receptor coupling. The chimera ztz40/36 was then constructed based on zt40 and tz36 (Fig. 2A, the fourth construct) to check whether the inclusion of the Gα_z-specific C-terminus would confer PTX resistance to the ztz40/36 chimera. The phenotype of ztz40/36 was more or less the same as Gα_z, except that the percentage response over basal value was slightly lower than that of Gα_z(Fig. 2B). It suggested that the C-terminus of Gα_z only conferred PTX-resistance but did not further enhance the receptor coupling efficiency. Two other ztz chimeras, ztz40/60 and ztz40/143 (Fig. 2A, sixth and seventh constructs), were also able to couple to DOR and stimulate AC2 to slightly greater degree than ztz40/36 (Fig. 2B).

Gα_z has distinct biochemical properties in that the loading and hydrolysis of GTP by Gα_z are much slower than other Gα subunits (Casey et al., 1990). The main reason is correlated to the variance of sequence identity in the G-1 GTP-binding region (amino acids 41–43). It is unknown whether the triplet variations affect the receptor coupling efficacies of the chimeras. zt43 and the corresponding series of ztz chimeras (ztz43/36, ztz43/60, and ztz43/143) were made so that the G-1 region resembled the Gα_z sequence (Fig. 2A, the seven to tenth constructs). However, the profile of responses was similar to the zt40 series. All three ztz chimeras with N-terminal 43 residues of Gα_z but not zt43 coupled to DOR and stimulated AC2 in the presence of PTX (Fig. 2B). The results eliminated the assumption that the variation of the G-1 GTP-binding region of Gα_z affected the receptor interaction and subsequent release of βγ subunits.

The Amino Acids 11 to 14 of α_z Are Essential for Receptor Coupling.

The N-terminus of Gα_t1is considerably divergent from those of Gα_i2and Gα_z in two aspects. First, the N-terminus of Gα_z is more homologous to Gα_i2 (47.1%) than to Gα_t1 (33.3%). Second, the length of the theoretical N-terminal helix of Gα_t1 is shorter than Gα_i2 or Gα_z by four amino acids (Fig. 3A). Progressive deletion of the six N-terminal residues of Gα_qbroadened its receptor coupling specificity (Kostenis et al., 1997). It is conceivable that the length of the N-terminal region of the chimeras may affect receptor activation. In the case of zt40, its ability to couple to DOR may be attributable to the elongation of its N-terminal region. To test this possibility, an insertion mutant of Gα_t1 was constructed by inserting the four residues of Gα_z between residues 10 and 11 of Gα_t1. The cysteine residue at −4 position of the C-terminus of Gα_t1 was mutated into glycine and the Gα_t1 mutant (termed +4t/CG) became resistant to PTX-mediated inactivation. This mutation allowed us to examine the coupling of +4t/CG with DOR without the interference caused by endogenous G_i proteins. In the presence of PTX, application of DPDPE can induce a modest increase of cAMP level in +4t/CG-transfected cells, which is significantly higher than that of Gα_t1CG- or vector-transfected cells (Fig. 3B). It indicated that +4t/CG can indeed couple to DOR in a PTX-insensitive manner although the response is relatively weak (∼40% increase as compared with the basal). To further confirm the role of these four amino acids, a deletion mutant of Gα_z was constructed with the amino acids 11 to 14 removed (termed −4z) and the mutant was tested for its ability to recognize DOR. In AC2 assays, −4z-transfected cells showed only slight reduction of DPDPE-induced cAMP accumulation compared with Gα_z (Fig. 3B). The reduction of response was also observed when the cells were pretreated with PTX. It suggested that residues 11 to 14 of Gα_z form one of the minor determinants for DOR coupling. The reduction of DOR-induced inhibitory response in −4z-transfected cells is probably independent to its expression level (Fig. 4) and effector interaction. A constitutively active mutant of −4z, −4z/QL, inhibited isoproterenol-induced cAMP production in the cells cotransfected with β₂-adrenoceptor to a similar extent as Gα_zQL (Fig. 3C). Overall, these results strongly suggested that the N-terminal residues 11 to 14 of Gα_z are involved in receptor recognition.

Protein Expression of Chimeric Gα Subunits.

Membrane proteins of the cells transfected with various chimeras studied here were separated by denaturing gel electrophoresis and the chimeras were detected by specific antisera. The expression levels of zt40 and zt43 were similar to tz36, but lower than Gα_t1 in HEK 293 cells as detected by an anti-Gα_t1antiserum (Fig. 4A). Both Gα_t1CG and +4t/CG were expressed strongly in HEK 293 cells. Detection with an antiserum against the last 15 residues of Gα_z facilitated direct comparison of the expression levels of various chimeras (Fig.4B). All chimeric α subunits were expressed to similar levels in HEK 293 cells as Gα_z. The results indicated that most of the chimeras studied here are expressed to similar levels.

Receptor-Mediated Inhibition of AC by Chimeric α Subunits.

Some of the chimeras contained increasing lengths of C-terminal tails of Gα_z, and could serve as useful tools for localizing the AC-inhibiting regions on Gα_z. Thus, we tested these chimeras for their ability to inhibit AC. HEK 293 cells transfected with β₂-adrenoceptor, DOR, and a parental or chimeric Gα subunit were pretreated with PTX and then stimulated with 10 μM isoproterenol to elevate the intracellular cAMP level. Coadministration of DPDPE to the cells expressing Gα_z showed about 30 to 40% reduction of isoproterenol-induced cAMP level, which was not observed in the cases of vector control or Gα_t1-transfected cells (Fig. 2C). Both zt40 and zt43-transfected cells lacked DPDPE-induced inhibition of AC activity. The N-terminal Gα_z-specific sequence of these two chimeras did not contain sufficient structural elements for AC inhibition. Next, we examined the six ztz chimeras, because it has been shown that the N-terminal region of Gα_z was important for the coupling to DOR. ztz40/36 and ztz43/36 showed no significant inhibitory effect on the cAMP levels (Fig. 2C). Chimeras containing 60 or more C-terminal amino acids of Gα_z inhibited AC significantly. Such results indicated that one of the essential AC inhibitory domains must be located within amino acids 296 to 319 of Gα_z.

Inhibition of AC by Constitutively Active Chimeric Gα Subunits.

Introduction of a point mutation at the GTP-binding regions of Gα subunit created receptor-independent constitutively active mutants (Freissmuth and Gilman, 1989; Graziano and Gilman, 1989). This approach has been successfully applied for studying the effector interacting domains of various Gα subunits even when the receptor interacting domains were disrupted in the chimeras (Medina et al., 1996; Grishina and Berlot, 1997). For the constructs used in this study, Arg-174 of the Gα_t1 domain and Gln-205 of the Gα_z portion were mutated to cysteine and leucine, respectively, and these chimeras were used to examine the receptor-independent constitutive inhibition of AC activity. Results are summarized in Fig. 5. As a positive control, the Gα_z mutant Gα_zQL (Wong et al., 1992) inhibited isoproterenol-stimulated cAMP level by 40 to 50% compared with the vector control, whereas Gα_t1 exhibited no observable inhibition on the AC activity. Among the three tz chimeras, tz60(RC) and tz143(RC) inhibited AC significantly, albeit to a lesser extent than that of Gα_zQL. Among the three zt chimeras, only zt319(QL) inhibited AC efficiently. Of the six ztz chimeras, only four—ztz40/60(RC), ztz40/143(RC), ztz43/60(RC), and ztz43/143(RC)—were able to inhibit AC significantly. The degrees of inhibition for all these chimeras were comparable with each other but slightly weaker than Gα_zQL. The protein expression levels of all these constitutively active chimeras were similar to each other and also comparable with those of their wild-type counterparts (Fig. 4). Collectively, the results confirmed the essential role of amino acids 295 to 319 of Gα_zfor AC inhibition because all chimeras that possess the ability to inhibit AC contained this stretch of residues from Gα_z.

Discussion

G_z is not simply a PTX-resistant substitute of the three G_i subtypes but it may be involved in a number of novel signaling events. The specific functional interactions of Gα_z with newly discovered proteins such as RGSZ1 (Glick et al., 1998), Rap1GAP (Meng et al., 1999), and GRIN1 (Chen et al., 1999) suggested that Gα_z has its distinct roles in signal transduction in addition to the regulation of ACs. Localization of the functional domains of Gα_z could provide valuable information for a better understanding of the roles of Gα_z in different molecular events. The present study investigated the receptor and AC-interacting domains of Gα_z. Two major findings could be concluded. First, the N-terminal sequence of Gα_z seemed to be a critical determinant for the coupling to DOR. Second, one of the AC inhibiting domains of Gα_z is located at amino acids 296–319.

A series of chimera studies indicated that the last five amino acids of Gα subunits are essential for determining the receptor specificity (Conklin et al., 1993, 1996). The last five amino acids of Gα_t1 are identical with those of Gα_i2 (Fig. 6), yet Gα_t1 primarily couples to rhodopsin. Both an evolutionary trace analysis (Lichtarge et al., 1996) and an extensive mutagenesis study (Onrust et al., 1997) suggested that residues spanning the whole α5 helix of the GTPase domain of Gα subunit interacts with receptor. Moreover, a recent study indicated that the intramolecular interaction between the GTPase and helical domains (β4/α3 and αG/α4 loops) of Gα subunit is important for receptor-mediated activation (Marsh et al., 1998). The series of tz chimeras retaining up to 143 C-terminal Gα_zresidues essentially covered all the regions identified in the previous studies for receptor-mediated activation. The lack of DOR coupling with tz chimeras clearly indicated that the N-terminal half of Gα subunit also contain essential elements for receptor coupling.

Alignment of the amino acid sequences of Gα_i2, Gα_z, and Gα_t1 yielded useful information (Fig. 6). The most interesting point is that the N-terminal helix of Gα_t1 is 4 amino acids shorter than Gα_i2 and Gα_z, but the charge distribution of the aligned residues are reasonably similar. A fair prediction would be the length of the N-terminal helix determined the receptor coupling efficiency. The phenotypes of the mutants +4t/CG and −4z (Fig. 3) supported the importance of N-terminus of Gα subunit in receptor coupling. Recent studies on the receptor coupling of Gα_q also support this notion. Deletion of the N-terminal six amino acids of Gα_q allowed the mutant to be activated by G_i-linked receptors (Kostenis et al., 1997). Alignment of the N-terminal sequences of various Gα subunits showed that the first six residues were unique for Gα_q. Indeed, the sequence identities and lengths of the N-terminal helices of different Gα subunits are very divergent and may be related to the peculiarities of the functions of each Gα subunit. For example, the N-terminal residues of Gα_q are involved in membrane attachment (in addition to the palmitate attached) and phospholipase C activation (Hepler et al., 1996). In one of our previous studies (Ho and Wong, 1997), mutation of the two protein kinase C-phosphorylation sites at the N-terminus of Gα_z abolished its constitutive inhibitory effect on AC. Further studies on the roles of N-terminal residues of other Gα subunits should provide more insights on their specific functions.

The crystal structures of trimeric G proteins (Wall et al., 1995;Lambright et al., 1996) provided clear evidence that the N-terminus of Gα subunit is one of the major βγ-interacting sites. An early study of the interaction between Gα_o and βγ subunits suggested that the reduction of the length of the N-terminus of Gα_o by four residues (positions 7–10) diminishes its binding to βγ subunits (Denker et al., 1992). Preservation of the integrity of the N-terminal helix seems to be necessary for binding βγ subunits properly. In the present study, changes in the ability of the chimeras to associate with the βγ subunits might affect their coupling efficiencies to DOR. Moreover, it has been suggested that the N-terminal helix of Gα subunit is sandwiched by Gβ subunit and receptor (Wall et al., 1995; Lichtarge et al., 1996), and it may play a role in transmitting the conformational changes of ligand-bound receptor to the subsequent intra-and intermolecular conformational changes of G protein trimer. Additional studies are required to address these possibilities.

One of the well-characterized representatives in the Gα_i-subfamily in effector interaction is Gα_i2, which is probably the closest member to Gα_z. Berlot and coworkers have identified a region of 79 amino acids of Gα_i2, which was sufficient to convey the AC inhibition (Medina et al., 1996). The residues of Gα_i2 for effector regulation were resolved eventually using alanine mutagenesis by the same laboratory (Grishina and Berlot, 1997) and they are localized on two structural elements of Gα_i2, the Switch II region and the α4/β6 loop. Our results coincided with the findings in Gα_i2, the amino acids 291 to 314 of Gα_z actually corresponded to α4/β6 loop on the GTPase domain of Gα_z. The same loop structure (but not the same residues at the corresponding positions) in Gα_s, Gα_i2, and Gα_t1 has also been shown to be an effector interacting region (Berlot and Bourne, 1992; Spickofsky et al., 1994;Grishina and Berlot, 1997; Natochin et al., 1999). It is interesting that different categories of G protein-regulated effector enzymes have evolved to interact with the similar structural elements of various Gα subunits. The α4/β6 loop may be more than just a structural element bearing a single function. Using Gα_t1/Gα_i1 chimeras, Bae et al. (1997) showed that the area bounded by the α4 helix and the α4/β6 loop of Gα_i1 is important for the coupling of 5-HT_1B receptor. Subsequent studies by the same group (Bae et al., 1999) demonstrated that two α4 helical residues in Gα_i1 (Gln-304 and Glu-308) are critical determinants of receptor-G protein coupling. In the present study, the incorporation of the α4/β6 loop of Gα_z in the Gα_t1backbone (such as tz60 and tz143) did not allow coupling to DOR. Although the α4 helix of Gα_z might be involved in receptor recognition, it alone was insufficient to support coupling to DOR.

We did not investigate further the importance of the Switch II region of Gα_z on AC inhibition because Gα_t1 and Gα_z have sequences identical with Gα_i2 at that region (Fig. 6) and so the corresponding region of Gα_t1 could provide essentially the same effector interacting residues. Inhibition of AC by Gα_z was not related to the N-terminal sequence of Gα_z because the constitutively active chimera tz60(RC) already exerted the inhibitory action (Fig. 4). The contribution of the variation of the G-1 GTP-binding region of Gα_z to the regulation of AC was only minimal (Fig. 2C). It implied that the slow GTP hydrolysis rate of Gα_z might not be related to the abilities of the chimeras to inhibit AC.

In conclusion, this report provides clear evidence that the N-terminal helix of Gα_z (and possibly that of Gα_i) is crucial for receptor-mediated activation. The length of the N-terminal helix is also critical for determining the efficiency of receptor coupling. The AC-inhibiting domains of Gα_z seem to be very similar to those of Gα_i2, including the α4/β6 loop and Switch II region. Our results strongly suggested that Gα_z inhibited AC in a fashion similar to other inhibitory Gα subunits.