European Journal of Pharmacology - Molecular Pharmacology Section, 246 (1993) 129-133 129 © 1993 Elsevier Science Publishers B.V. All rights reserved 0922-4106/93/$06.00

EJPMOL 90462

Functional expression of rat a2B-adrenoceptor in Escherichia

Y u n Xia, Vijay Chhaj lan i and Jarl E.S. Wikbe rg

Department of Pharmaceutical Pharmacology, Uppsala University, Biomedicum, Uppsala, Sweden

Received 11 November 1992, revised MS received 1 March 1993, accepted 23 March 1993

coli

Rat a2B-adrenoceptor was expressed in Escherichia coli using 'ATG vector' containing cDNA encoding the 'non-glycosylated rat az-adrenoceptor' (RNG~2). The highest receptor binding activity (using the a2-adrenoceptor ligand [3H]MK 912) was found when transfected bacteria cultures were grown at 30°C for about 4 h after induction with isopropyl /3-D-thiogalactopyranoside (IPTG). Saturation experiments showed that the radioligand bound to a single saturable site with a K d of 1.42 _+ 0.09 nM and capacity of 281 _+ 6 fmol/mg protein. Binding constants of 14 compounds for the rat a2B-adrenoceptor expressed in E. coli were determined and compared to the values previously obtained for the rat a2B-adrenoceptor when expressed in COS cells as well as for native neonatal rat lung azB-adrenoceptors. The results indicate that when the rat azB-adrenoceptor is expressed in E. coli it retains identical ligand binding properties to those found when the receptor is present in the eukaryotic system. Expressing a2a-adrenoceptors in E. coli would, therefore, seem to constitute a valid alternative in, e.g., drug screening and structure analysis of the c~2B-adrenoceptors.

a2B-Adrenoceptors; Expression; Radioligand binding; (E. coli)

1. Introduction

a2-Adrenoceptors participate in a variety of physio- logical processes such as in the regulation of blood pressure, control of nociception, regulation of lipid metabolism, regulation of hormone release and regula- tion of platelet aggregability. These functions may be mediated by different a2-adrenoceptor subtypes, since recent studies have indicated the existence of at least three different subtypes: a2A , Ot2B , a2C (Harrison et al., 1991b; Bylund, 1992). These receptor subtypes show distinctly different pharmacological properties when they are expressed in mammalian cell lines. However, expression of proteins in such eukaryotic systems is expensive. When large amounts of receptor protein are required as in routine drug screening, it would be advantageous to have an inexpensive and simple ex- pression system for a2-adrenoceptors. Previously, /31- and /32-adrenoceptor genes have been successfully ex- pressed in E. coli and shown to retain their pharmaco- logical properties (Marullo et al., 1988, 1989; Chapot et al. 1990). Like /3-adrenoceptors, a2-adrenoceptors be- long to G-protein coupled receptors, which possess a common topological characteristic of seven transmem-

Correspondence to: Dr. Jarl Wikberg, Department of Pharmaceuti- cal Pharmacology, Box 591, Biomedicum, S-751 24 Uppsala, Sweden.

brane domains spanning the plasma membrane, In this study we report the expression of a rat a2B-adrenoce p- tor in E. coli. Our data show that the pharmacology displayed by this receptor is identical when expressed in the prokaryotic system and in a mammalian cell line. Moreover, the pharmacology of the expressed receptor shows identical binding properties to that of the native a2B-adrenoceptor present in the neonatal rat lung; the latter is regarded as the prototypic source for the a2B-adrenoceptor.

Materials and methods

2.1. Isotopes and drugs

[3H]MK 912 ((2S, 12bs)l',3'-dimethylspiro(1,3,4,5', 6,6',7,12b-octahydro-2H-benzo[b]furo[2,3-a]quinazol- ine)-2,4'-pyrimidin-2'-one; 79 Ci /mmol) was a kind gift from NEN; ( - ) -adrena l ine , ( - ) -noradrena l ine , yohim- bine, prazosin, chlorpromazine and isopropyl /3-D- thiogalactopyranoside (IPTG) were from Sigma Chemi- cal Co.; (+)-adrenal ine was from Sterling-Winthrop Research Institute, Rensselaer, NY; ARC 239 from Thomae, Biberach, Germany; WB 4101 from Research Biochemicals, Natick, MA; BDF 8933 from Beiersdorf, Hamburg, Germany; rauwolscine from Carl Roth, Karlsruhe, Germany; BRL 41992 from Beecham, Es-

130

sex, UK; oxymetazoline from Draco, Lund, Sweden; SKF 104078 from Smith Kline & French, Swedeland, PA, USA; guanfacine was from Astra, S6dert~ilje, Sweden.

2.2. Construction of plasmid containing rat a2B-adreno- ceptor gene

The rat a2a-adrenoceptor cDNA (RNG,,2; 'rat non-glycosylated a2-adrenoceptor ') , cloned in the EcoR1 site of P G E M 7 Z f ( + ) vector (Zeng et al., 1990) was a kind gift from Dr. K.R. Lynch, University of Virginia School of Medicine, Charlottesville, VA. The cDNA was cleaved with Ncol and Xbal restrictive enzymes, and the isolated 1.9 Kb fragment containing the entire coding region of the azB-adrenoceptor was subcloned into the 'ATG vector' pKK388-1 (Clontech).

2.3. Expression of rat a2B-adrenoceptor in E. coli

E. coli, JM109 strain, was transformed with the recombinant plasmid. Clones containing the plasmid were identified and grown at 37°C in L-broth contain- ing 100 /zg/ml of ampicillin up to an A600 value of 0.2-0.3. After addition of 1 mM IPTG, the cultures were, unless otherwise stated, grown for 4 h at 30°C (see results). E. coli cells were pelleted at 1500 × g and resuspended in an ice-cold incubation buffer (33 mM Tris-HCl, 140 mM NaC1, 1 mM EDTA, pH 7.5). Bacte- ria were centrifuged again, resuspended in the same buffer and used for radioligand binding.

2.4. Binding studies

Binding assays were performed in a final volume of 150 /xl of incubation buffer using aliquots containing approximately 1 × 10 s cells with appropriate concen- trations of [3H]MK 912 and drugs. Following incuba- tion at 25°C for 1 h, cells were collected by filtering and washing on Whatman G F / C fibre glass filter pa- per with 20 ml ice-cold 50 mM Tris-HC1, pH 7.5, using a cell harvester (Brandel, Gaithersburg, MD, USA). The filters were counted in a Beckman LS 1801 /3- counter using 3.5 ml of Ready Protein scintillation cocktail (Beckman Instruments). The counting effi- ciency was determined by using an external standard and the Beckman Instruments H-number method. Non-specific binding was defined in the presence of 1 /zM BDF 8933. All assays were performed with dupli- cate determinations. The results were analyzed by computer modelling using a non-linear curve fitting programme suitable for analysis of ligand binding data (Uhl6n and Wikberg, 1991a). Experimentally deter- mined values are given as the mean + S.E. mean of 2 -4 independent experiments. Protein was measured according to Lowry et al. (1951).

3. Results

The influence of different culture temperatures and incubation times for the expression of the rat azB- adrenoceptor were evaluated by using radioligand binding on the E. coli cells. Preliminary experiments showed that when the cultures were maintained at 37°C after addition of IPTG, the binding of [3H]MK 912 to the ceils was extremely low. In contrast, when the culture temperature was lowered to 30°C, the spe- cific binding on E. coli increased dramatically and peaked at 4 h. A similar situation was present when the cultures were grown at 23°C, the only difference being that the response was delayed (fig. 1).

Computer modelling of saturation curves obtained using the az-adrenoceptor ligand [3H]MK 912 (Uhl6n et al., 1992; Xia et al., 1993) with intact E. coli cells expressing the rat azB-adrenoceptor showed that the ligand bound to a single saturable site with a K d of 1.42 _+ 0.09 nM and a Bma x of 281 + 6 fmol /mg total bacterial protein (n = 3) (data not shown graphically). The Rosenthal plot of the saturation curves was straight, indicating that [3H]MK 912 labelled a single site (data not shown). In control experiments, using E. coli which had not been transformed with the recombi- nant plasmid, the cells showed only negligible non- specific binding for [3H]MK 912 (data not shown). Competition curves of drugs for the azB-adrenoceptor in E. coli obtained by using ~ 2.5 nM [3H]MK 912 were uniphasic and best modelled into one site fits (fig. 2). The K d values of 14 drugs obtained by computer modelling the competition curves are given in table 1. As shown in the table the binding is stereospecific since the 1-isomer of adrenaline was much more potent than the d-isomer. The binding sites also show the expected a2a-adrenoceptor specificity; i.e., high affinity

30O

200-

.¢,

100- I

o3 O ' z

Time (h) Fig. 1. Effect of culture temperature and time on the binding of [3H]MK 912 to E. coli expressing rat t~EB-adrenoceptor. At time zero, IPTG was added, after which aliquots were taken fromthe cultures of each group and the binding of [3H]MK 912 estimated. Cultures grown at 23°C (13), 30°C (e) and 37°C ( • ) are shown. The

concentration of [3H]MK 912 used in the assay was 2.7 nM.

110-

90-

70- O,I

,~ 50-

30-

10 i i i

-11 "9 -7 -5 -3

Competitor Log(M)

Fig. 2. Competition curves of ARC 239 (11), prazosin (o), and guanfacine (©) in E. coli expressing the rat a2n-adrenoceptor. In these experiments a fixed concentration of ~ 2.5 nM [3H]MK 912 was used. The lines represent the computer drawn best fits assuming

that ligands bound to one site.

for ARC 239 and prazosin, but low affinity for oxymetazoline and guanfacine. Table 1 also gives val- ues for rat azB-adrenoceptors expressed in COS cells as well as for azB-adrenoceptors of neonatal rat lung tissue. As can be seen, the K d values are virtually identical for all three systems.

4. Discussion

The results of the present study indicate that the culture temperature is an important factor for influenc- ing the level of expression of the rat a2u-adrenoceptor in E. coil. Previous studies have shown that foreign

131

proteins produced in E. coil can be subjected to rapid intracellular degradation. This seems to be due to a proteolytic process requiring energy which is mainly regulated by several heat shock genes, such as the lon gene. The Ion gene encodes an ATP-dependent pro- tease termed protease La (Chung and Goldberg, 1981; Goldberg et al., 1982; Miller, 1987). It has been ob- served that protease La may be induced two to three fold after a shift from 30°C to 42°C (Phillips et ai., 1984; Goff et al., 1984; Goff and Goldberg 1985). The low binding activity observed in the present study at higher temperature might therefore reflect a heat-shock response by which the degradation of expressed a2B- adrenoceptor proteins is increased. In addition, heat- shock response can be induced by other factors, such as production of abnormal proteins and upon exposure to 4% ethanol (Goff et al., 1984; Goff and Goldberg, 1985; Lindler et al. 1991). Mutation in htpR, whose product is a sigma factor for heat-shock promoters, has been shown to decrease the degradation of eukaryotic protein in E. coil (Goff et al., 1984; Lindler et al., 1991). Accordingly, a potential further increase in the expression of the a2B-adrenoceptor in E. coli might be obtained by using E. coli mutants that are devoid of the heat-shock response.

The Kd values obtained for the a2B-adrenoceptor expressed in E. coli. was virtually identical to the Kd-Values that we had previously obtained for native a2n-adrenoceptors of the neonatal rat lung, as well as for rat a2B-adrenoceptor expressed in COS7-cells. Our results, therefore, strongly indicate that the a2B-adren- oceptor becomes correctly folded in the E. coil cell membrane. Moreover, the a2B-adrenoceptor does not

TABLE 1

Drug K d values for rat a2a-adrenoceptors expressed in E. coli (E. coli-a2B), rat a2B-adrenoceptors expressed in COS7 cells (COS7-a2B) and for native a2B-adrenoceptors in neonatal rat lung (NRL-a2B).

Values for rat a2n-adrenoceptor expressed in COS cells and for neonatal rat lung a2a-adrenoceptor were taken from our previous study (Xia et al., 1993), in which the drug Kdva were determined using [3H]MK 912 binding with an assay buffer similar to that used in the present study; they are reproduced by courtesy of Pharmacology & Toxicology.

Ligands E. coli-a2B COS7-a2B NRI-a2B K a (nM) K a (nM) K a (nM)

( - ) -Norad rena l i ne 1010 + 100 1220 + 90 ( - ) - A d r e n a l i n e 2820 +- 180 1790 +- 100 (+) -Adrena l ine 10600 _+300 15200 _+ 1900

BDF 8933 0.57+- 0.1 0.97+- 0.15 A R C 239 4.54+- 0.6 6.96+- 0.53 Rauwolscine 9.96+- 1.30 9.66+- 0.10 Yohimbine 10.5 + 1.0 10.5 + 2.8 BRL41992 21.3 +- 0.78 13.1 + 3.2 Prazosin 26.7 + 4.3 55.5 +_ 11.1 WB4101 31.2 +- 2.8 24.9 _+ 5.5 SKF 104078 255 +- 48 - Chlorpromazine 169 +- 53 - Oxymetazoline 1320 +- 39 789 + 95 Guanfacine 1410 + 47 1170 +- 100

3000 + 410 3 710 + 240

34500 +_4500

0.57 +- 0.09 8.76 +- 0.58 9.27+- 0.28

10.0 + 0.4 22.3 +_ 1.5 19.3 _+ 1.0 26.3 +- 1.0

213 + 48 152 +- 4 857 +- 142

1620 + 50

132

have any sites for glycosylation (otherwise a common feature of many G-protein coupled receptors), thus eliminating the possibility of a change in ligand binding or structural properties due to a difference in glyco- sylation, between the prokaryotic and eukaryotic ex- pression systems. In addition to our previous studies in which we characterized expressed a2B-adrenoceptors, Harrison et al. (1991) and Bylund et al. (1992) also reported drug affinities for a2B-adrenoceptors ex- pressed in eukaryotic cells. The drug affinities reported by these authors compare favourably to both our pre- sent and previous data. However, in the interpretation of these data one must remember that whereas we have used an EDTA/NaC1 containing assay buffer for our present and previous studies, the assay buffers of Harrison et al. (1991) and Bylund et al. (1992) either included MgCI 2 and/or excluded EDTA. The latter buffers will allow the formation of the agonist high-af- finity states of the adrenoceptors, whereas our assay buffer will prevent the formation of such high-affinity states (see Uhl6n and Wikberg 1991b). This seems to be the reason that the agonist affinities for a2B-adren- oceptors reported by Harrison et al. (1991) and Bylund et al. (1992) are consistently higher than the agonist affinities reported by us.

Previous studies have demonstrated that the/3~- and /32-adrenoceptor genes can be expressed in E. coli by fusing them with /3-galactosidase (Marullo et al., 1988) or LamB gene (Marullo et al., 1989). Using photo- affinity labelling techniques it has been shown that for both fusion approaches it is required that the endeoge- nous proteolytic system of the bacteria cleaves the /3-adrenoceptors from their hybrid partners, in order for the receptor to acquire binding activity (Marullo et al., 1988; Marullo et al., 1989; Chapot et al., 1990). However, the cleavage of the receptor from the fusion protein may also alter the receptor function if the cleaved product is not identical to the receptor protein in situ (Chapot et al., 1990). Based on these considera- tions it was interesting to test the possibility of express- ing the a2B-adrenoceptor in E. coli without fusing it to a bacterial protein. Moreover, there is another exam- ple of the successful functional expression of a G-pro- tein coupled receptor in a prokaryotic system, namely the human /32-adrenoceptor (Breyer et al., 1990). The present study shows that the rat c~2B-adrenoceptor can successfully be expressed in E. coli without fusion, giving levels of a2B-adrenoceptor-protein which are sufficient for drug screening purposes using radioli- gand binding techniques.

Acknowledgements

We are indebted to Dr. K.R. Lynch, University of Virginia School of Medicine, Charlottesville, VA, for supplying us with the rat

a2B-adrenoceptor gene. Supported by the Swedish MRC (04X-05957), CFN, the Swedish National Board for Technical Development (89- 02211P) and the ,~tke Wiberg, Magnus Bergvall and Groschinsky Foundations.

References

Breyer, R.M., A.D. Strosberg and J.G. Guillet, 1990, Mutational analysis of ligand binding activity of beta-2-adrenergic receptor expressed in Escherichia coli. EMBO J. 9, 2679.

Bylund, D.B., 1992, Subtypes of alpha-l- and alpha-2-adrenergic receptors. FASEB J. 6, 832.

Bylund, D.B., H.S. Blaxall, L.J. Iversen, M.G. Caron, R.J. Lefkowitz and J.W. Lomasney, 1992, Pharmacological characteristics of alpha-2-adrenergic receptors: comparison of pharmacologically defined subtypes with subtypes identified by molecular cloning. Mol. Pharmacol. 42, 1.

Chapot, M.P., Y. Eshdat, S. Marullo, J.G. Guillet, A. Charbit, A.D. Strosberg and C. Delavier-Klutchko, 1990, Localization and char- acterization of three different beta-adrenergic receptors ex- pressed in Escherichia coli. Eur J Biochem. 187, 137.

Chung, C.H. and A.L. Goldberg, 1981, The product of the Ion (capR) gene in Escherichia coli is the ATP-dependent protease, protease La. Proc. Natl. Acad. Sci. USA 78, 4931.

Goff, S.A., L.P. Casson and A.L. Goldberg, 1984, Heat stock regula- tory gene htpR influences rates of protein degradation and ex- pression of the Ion gene in Escherichia coli. Proc. Natl. Acad. Sci. USA 81, 6647.

Goff, S.A. and A.L. Goldberg, 1985, Production of abnormal pro- teins in E. coli stimulates transcription of Ion and other heat shock genes. Cell. 41,857.

Goldberg, A.L., K.H.S. Swamy, C.H. Chung and F.S. Larimore, 1982, Proteases in Escherichia coli. in: Methods in Enzymology. ed. L. Lorand (Academic Press, New York) Vol. 80, p. 680.

Harrison, J.K., D.D. D'Angelo, D. Zeng and K.R. Lynch, 1991, Pharmacological characterization of rat a2-adrenergic receptors. Mol. Pharmacol. 40, 407.

Harrison, J.K., W.R. Pearson and K.R. Lynch, 1991b, Molecular characterization of a 1- and a2-adrenoceptors. TIPS 12, 62.

Lindler, L.E., J.C. Anders and W.E. Herman, 1991, Mutation in the Escherichia coli htpR locus results in stabilization of recombinant expression products that are susceptible to proteolytic degrada- tion. Protein Exp. Purif. 2, 321.

Lowry, O.H., N.J. Rosebrough, A.L. Farr and R.J. Randall, 1951, Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265.

Marullo, S., C. Delavier-Klutchko, Y. Eshdat, A.D. Strosberg and L. Emorine, 1988, Human beta-2-adrenergic receptors expressed in Escherichia coli membranes retain their pharmacological proper- ties. Proc. Natl. Acad. Sci. USA 85, 7551.

Marullo, S., C. Delavier-Klutchko, J.-G-Guillet, A. Charbit, A.D. Strosberg and L.J. Emorine, 1989, Expression of human ill- and /3a-adrenergic receptors in E. Coli as a new tool for ligand screening. Biotechnology 7, 923.

Miller, C.G., 1987, Protein degradation and proteolytic modificauon. In: Escherichia coli and Salmonella Typhimurium, ed. F.C. Neid- hardt et al. (Am. Society Microbiol., Washington) Vol. 44, p. 680.

Phillips, T.A., R.A. VanBogelen and F.C. Neidhardt, 1984, lon gene product of Escherichia coli is a heat-shock protein. J. Bacteriol. 159, 283.

Uhl6n, S. and J.E.S. Wikberg, 1991a, Delineation of rat kidney te2A- and a2B-adrenoceptors with [3H]RX821002 radioligand binding: computer modelling reveals that guanfacine is an a2A-selective compound. Eur. J. Pharmacol. 202, 235.

Uhl6n, S. and J.E.S. Wikberg, 1991b, Rat spinal cord a2-adrenoce p- tots are of the C~2A-Sublype: Comparison with azn- and aZB- adrenoceptors in rat spleen, cerebral cortex and kidney using [3H]RX821002 ligand binding. Pharmacol. Toxicol. 69, 341.

Uhl~n, S., Y. Xia, V. Chhajlani, C.C. Felder and J.E.S. Wikberg, 1992, [3H]MK 912 binding delineates two az-adrenoceptor sub- types in rat CNS one of which is identical with the cloned pA2d az-adrenoceptor. Br. J. Pharmacol. 106, 986.

133

Xia, Y., S. Uhl6n, V. Chhajlani, E.J. Lien and J.E.S. Wikberg, 1993, Further evidence for the existence of two forms of a2B-adrenoc- eptors in the rat. Pharmacol. Toxicol. 72, 40.

Zeng, D.W., J.K. Harrison, D.D. D'Angelo, C.M. Barber, A.L. Tucker, Z. Lu and K.R. Lynch, 1990, Molecular characterization of a rat a2a-adrenergic receptor. Proc. Natl. Acad. Sci. USA 87, 3102.