Transient-expression technologies, their ... - Semantic Scholar

Biochemical Society Transactions ( 1999) Volume 27, part 6

Transient-expression technologies, their application and scale-up : 5-HT, serotonin receptor case study H.D. Blasey*', R. Hoviusj-I, H.Vogelt and A. R. Bernard* *Serono Pharmaceutical Research Institute, CH- I 228 Geneva, Switzerland, and t Ecole Polytechnique Federale Lausanne, CH- I0 I 5 Lausanne, Switzerland

Introduction

posed of 2 4 different subunits. This makes it attractive for structural and functional studies. Expression of the 5-HT, receptor was first attempted in the yeast Saccharomyces cerevisiae. Radioligand binding showed that 0.154.75 pmol of receptor protein per mg of membrane protein was produced. T h e rank order of potencies of several receptor ligands was similar to that of the native receptor. However, the functionality of the channel could not be demonstrated, as no agonistinduced (serotonin up to 100p M ) [14C]guanidium influx could be observed. Therefore mammalian expression systems were pursued.

Historically, animal tissue was used for receptor studies within the pharmaceutical drug-discovery process before molecular biology allowed the cloning of the receptor genes. T h e latter led to the expression of large quantities of receptors, providing ample protein for various studies. In the pharmaceutical industry this permits the screening for chemical compounds with whole cells or membrane preparations as well as structural and functional analyses with the purified protein. Further progress in heterologous protein expression was achieved with the advent and the scaling-up of transient gene expression. About a decade ago the baculovirus system made it possible to produce large quantities of recombinant protein at reactor scale within 1 or 2 months. Recently, mammalian transient-expression systems were successfully scaled-up : (i) with viral systems such as adenovirus, Vaccinia virus and the Semliki Forest virus (SFV) expression system and (ii) with D N A vector-mediated systems, in particular the simian virus 40-derived (transformed green monkey kidney cells, or COS cells, large T antigen) and the Ebstein Barr virus (EBV)-derived expression systems. Here we report on our experience with the HEK293EBNA (HEK293 is human embryonic kidney cell line 293 and EBNA is EBV nuclear antigen) and the SFV systems for the expression of the 5-HT, (5-hydroxytryptamine, or serotonin) receptor. This receptor is a ligand-gated cation channel, implicated in several physiological disorders such as those related to drug abuse, emesis, schizophrenia and diarrhoea, and is thus of great pharmacological interest [11. Moreover, the 5HT, receptor is assumed to be a homopentamer [2], in contrast with most of the other ligand-gated ion channels, which are heteropentamers com-

SFV expression system T h e SFV expression system [3] is a helperdependent system that is very appealing for the expression of heterologous genes in a broad range of mammalian cells. T h e main advantages are high expression levels and the speed at which proteins can be generated when starting from the cDNA. T h e gene of interest is subcloned into the plasmid pSFV. Co-transfection of in vitro-transcribed recombinant and helper RNA into BHK (baby hamster kidney) cells leads to the in vivo packaging of the recombinant RNA and the production of non-infectious, recombinant virus particles (approximately 10' virus particles/ml). After activation with chymotrypsin, this virus infects nearly all mammalian, insect and amphibian cells. T h e SFV expression system has been used to express various proteins and among those a large number of different receptor types (neurokinin, dopamine, serotonin, adenosine, prostaglandin, glutamate, purinergic P,X, adrenergic, opioid, galanin and histamine receptors [4]). Since safety is a concern, it is possible to alternatively transfect cells with the in vitro-generated recombinant RNA. T h e S F V can also be used in the form of a D N A vectorbased system [S].

HEK293EBNA expression system

Abbreviations used: cmc, critical micelle concentration; EBV, EbsteinBarrvirus; EBNA, EBV nuclearantigen; BHK.baby hamster kidney; CHO, Chinese hamster ovary; SN Semliki Forest virus; HEK, human embryonic kidney; SPAP, secreted placental alkaline phosphatase; 5-HT. 5-hydroxytryptamine; NTA, nitrilotriacetic acid: MOI, multiplicty of infection. 'To whom correspondence should be addressed.

Q 1999 Biochemical Society

Based on the EBV, an increasingly popular expression system that allows episomal D N A vector replication was developed. HEK293 cells, which constitutively express the EBV nuclear antigen-1 (EBNA-I), are transfected with a plasmid that

956

Expression and Purification of Membrane Proteins

by making in vitro RNA transcripts from the recombinant pSFVl and the pSFVHelper plasmid. RNA products were co-electroporated into BHK cells [9]. T h e in vivo packaging of virus particles was completed 24 h after electroporation. Activation of virus stocks was achieved with chymotrypsin followed by enzyme inactivation with aprotinin. T h e virus titre estimation was described earlier [lo]. For recombinant protein expression, cells were grown to the end of the logarithmic growth phase, the medium was exchanged and virus was added so as to reach multiplicity of infection (MOI) 30. T h e cultures were harvested at 24 h post infection by centrifugation (250 g, 10 min) and the cell pellet, after washing in PBS, was stored at -880°C. Initial 5-HT, receptor expression was done with Chinese hamster ovary (CHO) cells using stirred glass spinner bottles (Bellco). Cells (4.4 x 10’) were pelleted by centrifugation and infected in a shaker at 250 ml with an estimated MOI of 35. Then, 2 h later, medium was added to yield a 5000-ml working volume and the culture was passaged in two suspension spinners. Harvesting was performed at 20 h post infection.

Receptors expressed by HEK293EBNA cells with plasmids containing the OriP of EBV Soluble recepton

Membrane recepton

Soluble interleukin receptors [ 191 Soluble tumour necrosis factor receptor [20] Mouse 5-HT3 receptor [ 191 Macrophage scavenger receptor

PI1 Human corticotropin-releasing hormone CRH, and CRH

[221 Human prostanoid DP receptor

~ 3 1

carries the replication origin region (OriP) from EBV. An example of such a vector is pCEP4 (Invitrogen), which contains a multiple cloning site, the cytomegalovirus promotor, bacterial ampicillin-resistance marker and the mammalian hygromycin B-resistance gene. HEK293EBNA cells can be transfected with relatively small amounts of plasmid DNA carrying the gene of interest. Addition of antibiotic enables the selection of transfected cells to yield a cell population that can be scaled-up in a production process. Cultures stably maintain expression for up to 6 months [6]. Both genomic and cDNA sequences can be expressed in this system. In combination with different plasmids which carry OriP, HEK293EBNA cells have been used to express receptors and other proteins at levels between 0.2 and 10pg/ml (Table 1) [6].

Expression of the 5-HT3, receptor by HEK293EBNA cells T h e His-tagged 5-HT, gene, cut from the pSFV15-HT, by digesting with BamH1, was inserted into pCEP4 (Invitrogen). HEK293EBNA cells were electroporated with 4 0 p g of the plasmid pCEP4-5-HT3 (0.4-mm cuvettes, 290 V, 960 pF, Gene Pulser, Bio-Rad) and the culture was selected with 300 pg/ml hygromycin B (Boehringer Mannheim) from day 2 post transfection. A cell bank was frozen at passage 3 post transfection. For large-scale production we used the Cellcube system as previously described [ l l ] and a 12-litre stirred-tank reactor from MBR (Wetzikon, Switzerland) equipped with a Vibromix, for expression in suspension culture (inoculation density was 4 x lo4 cells per ml).

Materials and methods Cell-culture conditions T h e medium for BHK and HEK293EBNA cells was Iscove’s and Ham’s F12 (1:1, v/v), 10 yo fetal calf serum and 2 m M Gln. For spinner cultures, 18 m M Hepes was added.

Analysis Secreted placental alkaline phosphatase (SPAP) activity was determined by a classical assay with p nitrophenyl phosphate as substrate [12].

Expression of the 5-HT3, receptor in the SFV expression system T h e SFV expression plasmids pSFV3-LacZ and pSFVHelper-2 have been described elsewhere [3,7]. T h e pSFVl-5-HT3 plasmid was described earlier by Lundstrom et al. [8], who introduced a hexa-histidine tag at the 3’ end of the 5-HT, coding sequence to facilitate receptor purification. Generation of recombinant protein was initiated

Membrane solubilization and receptor purification A suspension of CHO- or BHK-cell membranes was incubated in detergent in 1 0 m M Hepes/

957

0 I999 Biochemical Society


square using the software package Igor 3.0 (WaveMetrics, Lake Oswego, OR, U.S.A.).

1 mM EDTA (pH 7.4) for 1 h. Non-solubilized material was removed by centrifugation at 100000 g.The receptor was purified in one step by chromatography on nitrilotriacetic acid (NTA)agarose (Qiagen, Hilden, Germany)-immobilized metal ions as described in detail elsewhere [13]. For further analysis of the eluted protein, the imidazole element was removed by gel filtration on a G-25 column (NAP-10, Pharmacia, Uppsala, Sweden) equilibrated and eluted with 10 mM Hepes/0.4 mM nonaethyleneglycol monododecyl ether (C,,E,) (pH 7.4).

Results Expression in mammalian cells with the SFV system The SFV system allowed very rapid protein production. Starting from the plasmid DNA it took 2 days to generate virus and 1-2 days to get the protein expressed. We identified a number of parameters which had a significant influence on the expression level [4,14] : the culture medium needed to be changed before infection because cells had a significantly increased metabolism when virus was added; infection in small volumes improved expression levels ; we observed different pH optima for infection (pH 6.9) and for protein expression (pH 7.3); and the expression level of recombinant protein by SFV was clearly dependent on the MOI. Initial 5-HT, receptor expression by CHO cells in suspension spinners yielded 7 x lo5 receptors per cell. Expression in CHO cells can drop significantly when infection and expression is done in suspension instead of in T-flasks. This phenomenon was not observed for BHK cells. Furthermore, recombinant protein yield is significantly dependent on the host cell line, as shown for the expression of the 5-HT, receptor and for SPAP (Figure 1). BHK and CHO cells are the best-expressing cell lines amongst all the ones tested. We used BHK cells for the expression of the 5-HT, receptor in a fermentor [4]. Expression was

Radioligandhindingassay The total number of ligand-binding sites, the inhibition of ligand binding by detergents and the binding affinities of different pharmacologically active compounds were determined as detailed elsewhere [131. Samples were incubated for 60 min at room temperature in 10 mM Hepes (pH 7.4) with the appropriate amounts of 3-( 5-methyl-1Himidazol-4-y1)- 1-( 1-[,H]-methyl- 1H-indol-3-y1)propanone ([,H]GR65630) (NEN-Du Pont, Boston, MA, U.S.A.). The incubation was terminated by rapid filtration through Whatman GF/B filters [presoaked in 0.5 yo (w/v) polyethyleneimine] followed by two washes with 3 ml of icecold 10 mM Hepes (pH 7.4). The radioactivity on the filters was measured by liquid scintillation counting. Dissociation constants, total number of binding sites and concentrations of half inhibition were evaluated from experimental binding isotherms by curve fitting with an iterative Levenberg-Marquardt algorithm minimizing Chi-

Expression of the 5-HT3receptor and SPAP by different cell lines Cells from eight cell lines were infected w t h the recombinant 5-HT3-SFV and SPAP-SN at a MOI of 30. The level of expression for each protein is normalized with respect t o the highest expression observed (100% expression equals 3 . 8 ~lo6 5-HT recepton per cell and 0.22 unrtdml SPAP, respectively). Abbreviation' NSO, mouse myeloma.

1

C

.-0 u) E v)

::

w

n

i 4

0 1999 Biochemical Society

0+c

1

958

1

5-HT3

SPAP

Expression and Purification of Membrane Proteins

Solubilization T h e first step in membrane-protein purification regards the solubilization of membranes with detergents, with careful attention paid to receptor integrity and solubilization efficiency. A series of detergents were tested for inhibition of radioligand binding to the receptor. Membranes from C H O cells transfected with SFV encoding the 5-HT, receptor were incubated for 1 h with increasing concentrations of different detergents, followed by a radioligand-binding assay performed at the same detergent concentration. Detergents with critical micelle concentrations (cmcs) above 1 m M virtually abolished ligand binding at concentrations above the cmc. Low-cmc detergents were less detrimental; CI2E, was found to be optimal, decreasing binding by only 10% [13]. T h e efficiency of this detergent at solubilizing the receptor from the membranes was 50-60% at concentrations of 5-50 times the cmc, and this was independent of the protein concentrations up to 2 mg of protein per ml.

at (3-4) x lo6 receptors per cell. By applying an optimal p H for expression, the recombinant protein level was increased to 8 x lo6 receptors per cell. With a 12-litre culture and lo6 cells/ml, 10-25 mg of receptor protein was expressed per reactor run.

Expression in mammalian cells with the HEK293EBNA system HEK293EBNA cells were grown after transfection in T-flasks under selection pressure. For the production of larger quantities of receptor they were transferred into a Cellcube reactor (adherent growth) or a stirred bioreactor (suspension growth). T h e expression levels were (1-3) x lo6 5-HT, receptors per cell when grown attached to a surface and 9.7 x lo6 receptors in the stirred reactor. T h e reason for the increased productivity in the stirred reactor is under investigation. T h e time elapsed between transfection and harvest of a 12-litre reactor was about 5 weeks. In summary, the two transient-expression systems SFV and HEK293EBNA allow very highlevel 5-HT, receptor expression in a short period of days to weeks. T h e expressed receptor featured ligand-binding [15] and channel properties [16] similar to the native receptor.

Purification Ligand-affinity chromatography was used initially to attempt purification. However, no receptor activity could be recovered upon elution, despite

Purification of the 5-HT, receptor (A) Solubilized BHK-cell membranes were loaded on to a Ni-NTA column. Non-specifically bound proteins were removed by washing with 20 mM irnidazole/500mM NaCl in I0 mM Hepes/O4 rnM C,,E, (pH 7.4) ( I ) The receptor was then eluted as a sharp peak with 80 rnM imidazole in the same buffer (2). The receptor was quantified by radioligand-binding assays (B) The punty of the isolated fraction was shown by electrophoresis on SDS/polyacrylarnide ( 10% gel) and visualization by silver stain

80-1- 1

B

2

n

70

Y

F

8

.k

60

(kD)

\

850 E

8

83

40

.P 30 z F

g)

front 101

51

20

10

36

3

0 0

40 Elution (ml)

20

60

959



the use of different resins (Affi-Gel 10 or 15, BioRad) and several ligands {serotonin, quipazin, 1,2,3,9-tetrahydro-3-[( 5-methyl- 1-H imidazol-4yl)methyl]-9-( 3-aminopropyl)-4H-carbazol-4one (GR119566X)} at different degrees of coupling ( 1-1 00 Yo). A hexa-histidine tag was therefore introduced at the C-terminus, and immobilized metal-ion chromatography was used to purify the 5-HT, receptor. First, the chromatographic conditions were optimized according to yield and purity by varying the metal ion immobilized on NTAagarose (Ni2+> Cu2+> C o 2 + xZn2+), and the wash and elution buffers (see Figure 2A). T h e receptor protein was purified to near homogeneity in a single step, as shown by SDS/ PAGE of the isolated fraction (Figure 2B). About 90% of the ligand-binding activity loaded on to the column was recovered in the eluted fraction. Characterization T h e purified detergent-solubilized receptor showed affinities for radioligands and a rank order of potencies of unlabelled receptor ligands that were virtually identical to the native receptor [13]. T h e purity of the final product was corroborated by a specific radioligand binding of 5 nmol per mg of protein. Conclusion T h e SFV and HEK293EBNA transient-expression sytems yielded a very high expression level of functional 5-HT, receptor in a short time, i.e. within days to weeks. T h e expressed receptor featured ligand-binding [13] and channel [16] properties similar to the native receptor. T h e large-scale expression of the 5-HT3 receptor with the SFV system and the facile purification procedure will allow further biophysical characterization of this receptor. Insight into the secondary structure [13] and the mechanism and stoichiometry of the binding of a fluorescently labelled ligand [16,17] has already been obtained. Moreover, the purified receptor has been used to develop an assay combining functional surface immobilization of the receptor with highly sensitive fluorescence detection [18]. W e wish t o thank Dr. T. Wells (Serono Pharmaceutical Research Institute) for his support and B. Brethon and L. Rey for contributing t o the experimental work W e thank Dr. J. Reid for the 5-HT, receptor-expression yeast strain. This work was financially supported by the Swiss National Science Foundation Priority Program


for Biotechnology. project 5002-35 I80 (to H.V.). W e are also grateful to Dr. S. Mokohliso for corrections t o the English.

I Tyers, M. B. ( I 99 I ) Therapie 46,43 I 4 3 5 2 Maricq, A. V., Peterson, A. S., Brake, A. 1.. Myers, R M. and Julius,D. (I 99 I) Science 254, 432437 3 LiljestrCim, P. and Garoff, H. ( I 99 I ) BioTechnology 9, I 3 5 6 1361 4 Blasey, H. D., Brethon, B., Hovius, R., Vogel, M., Tairi. A. P., Lundstrom, K., Rey, L. and Bernard, A. R ( I 999) Cytotechnology,in the press 5 Berglund, P., Smerdou, C., Fleeton, M. N., Tubulekas, I. and Liljestrom, P. ( I 998) Nat. Biotechnol. 16, 562-565 6 Cachianes, G., Ho, C., Weber, R F., Williams, S. R., Goeddel, D. V. and Leung, D. W. ( I 993) BioTechniques 15, 255-259 7 Berglund, P., Sjodberg, M., Garoff, H., Atkins, J. G., Sheahan, B. J. and Liljestrtjm, P. ( I 993) BioTechnology I I , 9 I 6 9 2 0 8 Lundstrom, K., Michel, A,, Blasey, H., Bernard, A. R, Hovius, R. Vogel, V. and Surprenant, A. ( I 997) J. Receptor Signal Transduction Res. 17, I 15- I26 9 Liljestrom, P. and Garoff, H. ( I 99 I ) Cur. Protocols Mol. Biol. 2, 16-22 10 Lundstrom, K., Mills, A,, Buell, G., Allet, E., Adami, N. and Liljestrom, P. ( I 994) Eur. J. Biochem. 244, 9 17-92 I I I Blasey, H. D., Isch, C. and Bernard, A. R ( I 995) Biotechnol. Techniques 9,725-728 12 Berger, J.. Hauber, J., Hauber, R, Geiger, R and Cullen. B. R. ( 1988) Gene 66, I- I0 13 Hovius, R, Tairi, A.-P., Blasey, H., Bernard, A,, Lundstrom, K. and Vogel, H. ( I 998) J. Neurochem. 70,824-834 14 Blasey, H. D., Lundstrom, K., Tate, S. and Bernard, A. R ( I 997) Cytotechnology 24, 65-72 15 Hovius. R, Tairi. A.-P., Blasey. H., Bernard. A,. Lundstrom, K. and Vogel, H. ( I 998) J. Neurochem. 70,824-834 16 Tairi, A.-P., Hovius, R., Pick H., Blasey, H., Bernard, A,, Surprenant, A,. LundstrCim, K. and Vogel, H. ( I 998) Biochemistry 37, l585&15864 17 Wohland, T., Friedrich, K., Hovius, R. and Vogel, H. (I 999) Biochemistry 38,867 1-868 I I8 Schmid, E. L., Tairi, A.-P., Hovius, R and Vogel, H. ( I 998) Anal. Chem. 70, I33 I -I 338 19 Blasey, H. D., Hovius, R, Rey, L.. Vogel, H. and Bernard, A. ( I 999) In Products from Cells, Cells as Products, Kluwer Academic Publishers, Dortrecht, in the press 20 Cachianes, G., Ho, C..Weber, R F., Williams, S. R., Goeddel, D. V. and Leung, D. W. ( I 993). BioTechniques 15, 255-259 2 I Pietri-Rouxel, F., St. John Manning, B., Gros. J.and Strosberg, A. D. (I 998) BioTechniques 25, 24Cb-244 22 Rominger, D. H., Rominger, C. M., Fitzgerald, L. W., Grzanna, R, Largent, B. L. and Zauek R (I 998) J. Pharmacol. Exp. Ther. 286,459468 23 Boie, Y., Sawyer, N., Slipetz, D. M., Metters, K. M. and Abramovitz, M. (I 995) J. Biol. Chem. 270, I 89 I&I 89 I 6

Received 24 June I999

960