|Home | About | Journals | Submit | Contact Us | Français|
X-ray crystal structures of human membrane proteins, while potentially being of extremely high impact, are highly underrepresented relative to those of prokaryotic membrane proteins. One key reason for this is that human membrane proteins can be difficult to express at a level, and at a quality, suitable for structural studies. This protocol describes the methods that we utilize to overexpress human membrane proteins from clonal HEK293S GnTI- cells, and was recently used in our 2.1 Å X-ray crystal structure determination of human RhCG. Upon identification of highly expressing cell lines, suspension cell cultures are scaled-up in a facile manner either using spinner flasks or cellbag bioreactors, resulting in a final purified yield of ~0.5 mg of membrane protein per liter of medium. The protocol described here is reliable and cost-effective, can be used to express proteins that would otherwise be toxic to mammalian cells, and can be completed in 8–10 weeks.
The overexpression of human membrane proteins, in both quantities and qualities amenable for structural studies, is often an immense challenge for numerous reasons. At a fundamental level, the biosynthesis of both prokaryotic and eukaryotic membrane proteins requires two events to occur1,2. Firstly, the newly synthesized polypeptide chain must be inserted into the membrane via recognition by the evolutionarily conserved Sec translocon3,4. This prerequisite for Sec translocon engagement can be seen as placing a lower ceiling on membrane protein biosynthesis, relative to soluble protein biosynthesis, given that the Sec translocon can become saturated when membrane proteins are overexpressed5,6. The manner by which the Sec translocon can recognize heterologous sequences can lower this ceiling even further7, in particular in those cases of expressing human membrane proteins in the most commonly used system for protein expression, Escherichia coli8. Secondly, the inserted polypeptide must laterally traverse into the lipid bilayer where folding into the correct three-dimensional structure predominately occurs1,2,9. Given that the composition of lipid bilayers differ substantially between humans and other species10,11, and given the pronounced effects that lipids and sterols have on membrane protein structure and function12–14, it is apparent that lipid composition is a key variable in determining the yields of functional membrane proteins that are produced.
In addition to these two fundamental principles of membrane protein biosynthesis, eukaryotic membrane proteins often require additional post-translational modifications not found in prokaryotes. N-linked glycosylation of eukaryotic membrane proteins, for instance, can be of critical importance for the folding of membrane proteins possessing large extracellular domains, given that these glycans serve as ligands for the ER-resident molecular chaperones calnexin and calreticulin15. Nevertheless, N-linked glycosylation is often seen as an impediment towards structure determination via X-ray crystallography, given that these glycans are often very large, heterogeneous, and conformationally flexible16. Indeed, considerable effort has been placed into restricting the heterogeneity of these N-linked glycans, via mutagenesis, glycosidases and/or inhibitors, for the purpose of structural studies17–20.
The use of mammalian cell expression systems is becoming increasingly popular for the overexpression of mammalian membrane proteins for structural studies21,22, likely owing to their near-native translocation machinery, lipid milieu, and post-translational modifications. A suspension adapted Human Embryonic Kidney 293 (HEK293S) cell line lacking N-acetylglucosaminyltransferase I (GnTI) has been successfully employed to overexpress a wide variety of mammalian membrane proteins23–25, including that of human RhCG whose X-ray crystal structure was recently reported in our laboratory21. Developed by Gobind Khorana and colleagues to overexpress large quantities of functional rhodopsin20, the HEK293S GnTI- system possesses features that make it amenable for structural studies. Firstly the lack of GnTI restricts N-linked glycans to a homogeneous Man5-GlcNac2 structure, which greatly facilitates their enzymatic removal via endo- and exoglycosidases. Secondly, the system utilizes a tetracyclin inducible promoter, allowing for protein expression to be induced once high-density cell cultures are established (discussed further in Induction of membrane protein expression). As compared to other lower eukaryotic expression systems that have been used to express mammalian membrane proteins for structural studies, such as baculovirus infected SF9 insect cells26, expression from HEK293S GnTI- cells is typically more labor intensive and less cost-effective. Nevertheless, given that the folding and proper trafficking of mammalian membrane proteins can depend critically on the translocation machinery, lipid milieu, and post-translational modification present in the expression system14, the use of mammalian cell expression systems likely ensures the greatest probability of producing properly folded and functional human membrane proteins.
In this protocol, we describe the methods that we used to overexpress quantities of pure, homogeneous and stable human RhCG from stably transfected and clonal HEK293S GnTI-cells, for use in crystallization. Expression trials of human RhCG were performed using Escherichia coli, Saccharomyces cerevisiae, and HEK293S GnTI- cells, however, only in the case of HEK293S GnTI- cells was RhCG expressed at high levels sufficient for structural studies. In addition, three human Rh proteins were transiently expressed in HEK293S GnTI- cells, with RhCG possessing the highest level of solubilized membrane protein (Figure 1). The protocol described here is of use to those that require long-term overexpression of a human membrane protein, for example in antigen production, drug discovery, biochemical characterization, or structural studies, and assumes that suitable expression constructs have already been generated and screened (discussed further in Experimental Design). A summary of the various steps of the protocol are discussed in detail below.
There are two ways by which a transfected transgene can be overexpressed in HEK293S cells, either transiently or stably. In transient expression, a sufficient quantity of plasmid DNA is transfected into cell cultures of varying sizes (mL to L scales), after which overexpression is either immediate or induced, dependent on the promoter type of the plasmid. Continued passage of transiently transfected cells results in a dilution of cells that are transfected, therefore, each overexpression trial must be preceded by transfection. Large scale transient expression, typically involving cell cultures in the 1–10 L range, is becoming increasingly popular given the flexibility of the system and given its potential for high-throughput overexpression trials17,19,27.
Stable expression, on the other hand, occurs from a transgene that has been stably integrated into the transfected cell’s genome. Stably transfected cell lines are generated by introducing Geneticin to the cell culture, over a period of 2–3 weeks, which selects for cells that possess stably integrated plasmid DNA (possessing the transgene and Geneticin resistance). Clones of stably transfected cell lines can then be generated and scaled up for transgene overexpression. The generation of clonal, stably transfected HEK293S cell lines is initially slower and more technically challenging relative to large scale transient expression. Nevertheless, once a clonal cell line is generated, long-term overexpression from stably transfected cells can be much more facile and consistent compared to transient expression given that the purification of large quantities of plasmid DNA, followed by separate large-scale transfections, is not required.
We have explored overexpression of human RhCG from both transiently and stably transfected HEK293S cell lines, using a pACMV-tetO expression vector in both cases, and have found that expression levels are significantly higher in clonal, stably transfected HEK293S cells. Nevertheless, in those cases where shorter time-scales and/or higher throughput are favored over longer-term and consistent expression levels, cell lines and expression vectors designed for transient expression can be explored17,28.
Stably transfected HEK293S cells are generated by selecting for successful integration of the neo gene using the antibiotic Geneticin, following a workflow as shown in Figure 2. Selection occurs in a 10 cm2 tissue culture plate, where it takes ~2–3 weeks for foci of Geneticin resistant colonies to appear, after which the resistant foci are clonally expanded over a period of an additional ~2–3 weeks. While we used Geneticin for selection, other drugs and resistance gene combinations can be used, for instance puromycin and puromycin acetyltransferase, respectively.
The level of transgene expression for any given stably transfected, and clonal HEK293S cell line can vary substantially. Integration of plasmid DNA into the HEK293S genome is random, therefore, the level of transgene expression will be determined in large part by position effects. In addition, given that the site of recombination within the plasmid sequence is random, no transgene expression will be observed if the site of recombination disrupts either the coding sequence or the promoter of the transgene. We have found that expanding 24 clonal cell lines, and screening these clones for expression via western blotting (discussed below in Assessment of membrane protein expression levels from clonal cell lines), is most efficient to find the highest expressing clones. Nevertheless, in those cases where screening 24 clonal cell lines is insufficient to identify highly expressing clones, or in those cases where a large number of clonal cell lines are to be generated in parallel, higher throughput cloning and screening strategies could be explored. For example, higher throughput cloning into 96 well plates, via fluorescent-activated cell sorting (FACS) or limited dilution plating, followed by fluorescence or ELISA quantification of expression levels, can be used to facilitate cloning and screening for highly expressing mammalian cell lines29.
Small scale detergent solubilizations, using both β-octylglucoside (OG) and β-dodecylmaltoside (DDM), are performed on all clonally expanded cell lines. The use of two or more detergents at this stage is optional, as detergent solubilization screening can be performed further downstream upon identification of the highest expressing clone. DDM should be used for small scale solubilizations if a single detergent is to be assessed, given its greater ability to solubilize membrane proteins compared to OG30.
The solubilized material from each clonal cell line is assessed for membrane protein expression via anti-fusion tag western blotting. Comparison of the “before spin” (BS) sample, which reveals total membrane protein expression levels, with the “after spin” (AS) sample, which reveals successfully solubilized protein expression levels, indicates the degree of solubilization. Caution should be exercised in those cases where solubilization only occurs in fos-choline type detergents, given their potential to denature membrane proteins31. Although the western blotting protocol used for assessing membrane protein expression levels is not quantitative, relative expression levels between clones can be assessed, allowing for the identification of clones that express high levels of detergent-soluble membrane proteins. It is the amount of detergent-soluble membrane protein (i.e. AS) produced, and not the total amount of membrane protein (i.e. BS) produced, that defines the useable expression level of a particular cell line.
Once a highly expressing, clonal cell line has been identified, medium (~1–3 L spinner flask) and/or large (~10 L cellbag bioreactor) suspension cultures are established for overexpression trials. The HEK293S GnTI- cells used during the course of this work are adaptable for suspension growth20, allowing for the establishment of higher densities of cell cultures relative to adherent cultures. This adaptation from an adherent monolayer to suspension cultures is brought about solely by changing the medium; the suspension medium lacks the calcium required for forming an adherent monolayer. In practice, adherent HEK293S GnTI- cells do not take long to be adapted to grow as a suspension culture in DMEM supplemented with serum20,32. Nevertheless, the growth of HEK293 cells in suspension is not always facile and, dependent on the type of cells and media used, can require that the cells be adapted for suspension growth over a period of several weeks28. The suspension medium used in this manuscript is a cost effective alternative to commercially available, serum-free formulations designed for suspension cell cultures and, in addition to a lack of calcium, includes Iron-supplemented Bovine Calf Serum (BCS), and the non-ionic detergent Pluronic F-68, with Primatone RL/UF added later as a supplement. Iron-supplemented BCS is a cost effective alternative to Fetal Bovine Serum (FBS), while Primatone RL/UF increases the viability and density of mammalian cells grown in suspension33. Pluorinic F-68 has been shown to both reduce cell adherence and to protect cells against high levels of shear stress introduced by sparging and stirring in the suspension culture34, in a mechanism thought to involve the non-ionic detergent coating the plasma membranes of the suspension cultures34. This suspension medium permits us to grow HEK293S GnTI- cells as a suspension culture, without requiring a lengthy adaptation period from an adherent culture.
The scale of cell culture that is established depends largely on what is known of the membrane protein being expressed. At the initial stages of the structure determination of human RhCG, 1–2 L spinner flasks were grown, induced, and harvested on a periodic basis. These 1–2 L spinner flask cultures were used to determine the optimal parameters for isolating pure, homogeneous and stable RhCG; parameters which included detergent type, pH, glycerol and protease concentrations and incubation periods (for fusion tag removal). Once optimal purification conditions were determined, larger 10 L cellbag, or occasionally 3 × 3 L spinner flask, suspension cultures were established. While cellbag bioreactors are expected to establish higher densities of suspension cultures relative to spinner flasks, owing to more efficient oxygen transfer into the medium, the observed densities of RhCG HEK293S cells in cellbag and spinner flask cultures were comparable (~1–1.5 × 106 cells/mL). We have not, however, explored if increasing or decreasing the rate of air/oxygen sparging into the cellbag can increase cell densities. Given the comparable cell densities observed for RhCG HEK293S in cellbags and spinner flasks, we have also established large scale (3 × 3L) spinner flask cultures of RhCG HEK293S, which are more facile and more cost efficient compared to cellbag cell cultures.
In the HEK293S GnTI- system all membrane proteins are expressed from a tetracycline-inducible promoter. This is an advantage in the overexpression of membrane proteins in particular, since the intrinsic function of certain membrane proteins (e.g. channels, GPCRs, etc.) can be cytotoxic to the cell, thereby preventing high density cell cultures from being established14,32,35,36. For example, it has been shown that stable cell lines constitutively expressing the serotonin transporter (SERT) can only be generated in the presence of SERT inhibitors, possibly owing to the fact that the intrinsic functions of SERT (serotonin transport and channel-like activity) severely stress the cell, in the absence of SERT inhibitors14. The use of a tetracycline-inducible promoter, therefore, delays the expression of potentially toxic membrane proteins until high-density cell cultures are established.
In the protocol described here, membrane protein expression is induced by adding an appropriate amount of doxycycline (a tetracycline antibiotic) to the suspension culture, either grown in spinner flasks or cellbags. For RhCG, we have found that induction with doxycycline at a cell density of ~1.0 × 106 cells/mL, followed by harvesting the cells 36 hours later, is optimal for RhCG expression, nevertheless, these parameters should be empirically determined for each particular membrane protein to be overexpressed. Sodium butyrate is also added to the suspension cultures at the time of induction, given its ability to increase protein expression levels from mammalian cells37, however it alone does not induce expression.
Solubilization of RhCG from whole HEK293S cells was performed subsequent to harvesting the medium- and large-scale suspension cultures. We have found that whole cell solubilizations are faster to perform and can result in higher yields of extracted RhCG compared with solubilization from a membrane preparation. Solubilization of membrane proteins from a membrane preparation, however, should be explored if the final purity of membrane protein after whole cell solubilization is insufficient.
A two-step purification scheme was utilized for the purification of human RhCG. Firstly, FLAG purification was performed, with either OG and DDM, from whole cell solubilized RhCG. This step will depend on the type of transgene that is integrated into the clonal cell; in the case of RhCG, the expression vector (pACMV-tetO20) was modified to possess an N-terminal FLAG and a C-terminal His tag, potentially allowing for two types of affinity purification to be performed. The FLAG purified RhCG was more pure and of better yield, as compared to the immobilized metal affinity chromatography (IMAC) purified RhCG, therefore we did not explore IMAC purification in detail. Secondly, size exclusion chromatography was performed, which resulted in the purification of RhCG to near homogeneity. Size exclusion chromatography is an excellent metric for both the purity of a membrane protein (pure proteins will possess discrete and near-monodisperse major peaks) and its stability (stable proteins will possess few changes in chromatographic profile over multiple days). Purified membrane proteins that are both near-monodisperse and stable, such as was found with human RhCG, are excellent candidates for crystallization trials.
Prior to undertaking the experiments described in this protocol, it is important to screen potential constructs for both expression quality and quantity. The use of fluorescence size exclusion chromatography (FSEC) at this stage to assess protein quality and quantity from transiently transfected HEK293 cells is therefore suggested, in that this will help to ensure that only suitable expression constructs are selected for stable cell line generation38. Constructs suitable for structural studies will possess single, monodisperse FSEC profiles of high magnitude38. An alternative approach is to perform anti-fusion tag western blotting of transiently transfected HEK293S GnTI- cells, in a similar manner as described in this protocol (see steps 26–34). Constructs that possess weak signal on a western blot could then be redesigned, or possibly codon optimized, depending on the importance of the particular membrane protein to the laboratory. The screening of constructs via FSEC or western blotting is also important in that, in our experience, it is not possible to predict a priori if a particular membrane protein class will express well in HEK293S GnTI- cells, as the highest expressing clonal cell lines produced in our laboratory, to date, are an ammonia transporter (RhCG), a cation transporter, a class B GPCR, and a protease.
While the entire procedure described here takes approximately 8–10 weeks to perform, the procedure can be paused at week 6 by freezing the clonal cell line. It is important to ensure that personnel are available to maintain the HEK293S GnTI- cells up until week 6, as clonal cell line generation can fail if the procedure is not followed and the cells are unnecessarily stressed (e.g. by failing to change the medium promptly, allowing the cells to become over confluent, etc.). This requires approximately 0.5–3 hours of cell culture work per day, 1–4 days per week, depending on the stage of the procedure. Similarly, once HEK293S GnTI- cells are harvested from spinner flask or WAVE cellbag suspension cultures, the whole cell membrane solubilization and purification steps should be performed immediately, given that the membrane protein of interest may be unstable; this most closely follows the procedure performed for the solubilization, purification and crystallization of human RhCG. Nevertheless, the procedure can be modified following the cell harvest stage to create an additional pause point by preparing a membrane fraction and freezing these membranes at −80°C39, for future solubilization and purification.
To 1 L of DMEM high glucose, add 10 mL of Penicillin- Streptomycin (100X) and 100 mL of Iron-Supplemented BCS. Filter-sterilize inside a biological safety cabinet using a 0.2 μm syringe filter. Store at 4°C for up to 1 year. ▲CRITICAL The addition of serum to DMEM is suggested for all cell culture work as, in our experience, HEK293S GnTI- cells do not passage well in DMEM alone. Iron-Supplemented BCS is added to both DMEM and Suspension medium, as it is a cost-effective alternative to FBS, suitable for most cell culture work (see DMEM (10% v/v FBS)for exception).
To 1 L of DMEM high glucose, add 10 mL of Penicillin- Streptomycin (100X) and 100 mL of FBS. Filter-sterilize inside a biological safety cabinet using a 0.2 μm syringe filter. Store at 4°C for up to 1 month. ▲CRITICAL Medium that is used during the drug selection process and foci expansion contain FBS, in place of Iron-supplemented BCS, in order to better ensure cell viability.
To 1 L of DMEM high glucose, add 10 mL of Penicillin- Streptomycin (100X), 100 mL of FBS, 40 mL of Geneticin, and 1 mL of Blasticidin S HCl. Filter-sterilize inside a biological safety cabinet using a 0.2 μm syringe filter. Store at 4°C for up to 1 month. ▲CRITICAL Medium that is used during the drug selection process and foci expansion contain FBS, in place of Iron-supplemented BCS, in order to ensure better cell viability.
To 1 L of DMEM high glucose without calcium salts, add 10 mL of Penicillin- Streptomycin (100X), 10 mL of pluronic, 0.3 g of Primatone RL/UF, 100 mL of BCS, and 3.7 g of sodium bicarbonate. Filter-sterilize inside a biological safety cabinet using a 0.2 μm syringe filter. Store at 4°C for up to 1 month. !CAUTION Serum-free mediums formulated for the growth of HEK293S in suspension, such as 293 SFM II (Invitrogen, cat. no. 11686-029), may also be used, however they are typically five-fold more expensive as compared to Suspension Medium.
To 7 L of DMEM high glucose without calcium salts, add 70 mL of Penicillin- Streptomycin (100X), 700 mL of BCS, 70 mL of pluronic, 2.1 g Primatone Rl/UF and 25.9 g sodium bicarbonate. Prepare prior to use and store at 4°C.
Dissolve 50 mg of Blasticidin S HCl in 10 mL of autoclaved water. Filter-sterilize inside a biological safety cabinet using a 0.2 μm syringe filter and store 1 mL aliquots at −20°C. !CAUTION Blasticidin is toxic. Always wear gloves, a mask, a laboratory coat, and safety glasses. Prepare the solution inside a biological safety cabinet. Aliquot in small volumes suitable for one time use, and store at −20°C for up to 6–8 weeks.
To prepare 20 mg/mL of doxycycline solution, dissolve 100 mg of doxycycline hyclate in 5 mL of autoclaved water. Filter-sterilize inside a biological safety cabinet using a 0.2 μm syringe filter and store 1 mL aliquots at −20°C for up to 6–8 weeks. !CAUTION Doxycycline should not be exposed to direct sunlight.
Dissolve 27.5 g of sodium butyrate with autoclaved water to a final volume of 500 mL and filter-sterilize using 0.2 μm syringe filter inside the biological safety cabinet. Store at room temperature (20–25°C) for up to 6–8 weeks.
Dissolve 3g of Primatone RL/UF in 30 mL of autoclaved water and filter-sterilize inside the biological safety cabinet using a 0.2μm syringe filter inside the biological safety cabinet. Prepare immediately before use.
Dissolve 100 mL of glucose in 500 mL of autoclaved water and filter-sterilize inside the biological safety cabinet using a 0.2μm syringe filter inside the biological safety cabinet. Store at 4°C for up to 2–3 months.
Dissolve 50 mL of pluronic in 500 mL of autoclaved water and filter-sterilize using a 0.2μm filter inside the biological safety cabinet. Store at 4°C for up to 6–8 weeks.
40 mM Tris pH 7.4, 200 mM NaCl, 20% v/v glycerol. Immediately before use, add 1 mM PMSF and one complete EDTA-free protease inhibitor cocktail tablet. Chill to 4°C, and discard any unused buffer.
20 mM Tris pH 7.4, 100 mM NaCl, 10% v/v glycerol. Immediately before use, add 1 mM PMSF and one complete EDTA-free protease inhibitor cocktail tablet. Chill to 4°C, and discard any unused buffer.
20 mM Tris pH 7.4, 100 mM NaCl, 10% v/v glycerol, 40 mM OG (or 0.5 mM DDM). Store at 4°C.
Periodically split into 10 cm2 tissue culture plates using DMEM. In a biosafety cabinet, gently wash the cells with 5 mL of D-PBS and aspirate off the medium. Add 1 mL of trypsin, and incubate at room temperature for 1 minute. Gently shake the tissue culture plate to facilitate dislodging of the cells from the plate. Add 5 mL of DMEM, resuspend the cells, and pellet at low speed for 5 minutes. Discard the DMEM, add 1 mL of DMEM to resuspend the cells, and add half of the cells to a new 10 cm2 tissue culture plate containing 10 mL of DMEM. For all cell culture work, HEK293S cells are maintained in a humidified incubator set to 5% CO2/95% air and 37°C. ▲CRITICAL STEP HEK293 cells, as with mammalian cells in general, are very easily contaminated. Therefore, it is of utmost importance that sterile tissue culture techniques are followed at all times. All surfaces (i.e. biosafety cabinet, microscopes, etc.) should be cleaned with 70% v/v ethanol prior to use. Gloves (cleaned with 70% v/v ethanol) and lab coats must be worn at all times.
Clean thoroughly, by separating the different parts of the flask. Add 10% v/v glacial acetic acid, and stir overnight at room temperature. Next day, remove the 10% v/v glacial acetic acid and rinse the spinner flask very well to remove any trace of acid. Fill the spinner flask with distilled water and perform two liquid autoclaves for 30 minutes each, and then a final dry autoclave for 30 minutes. Allow the flasks to cool to room temperature before use.
Assemble the WAVE bioreactor in a tissue culture room, following the manufacturer’s protocols.
Troubleshooting advice can be found in Table 1.
The entire protocol, starting from transfection (step 1) to size-exclusion chromatography purification (step 78) takes approximately 8–10 weeks to complete. Once a stable cell line is generated and characterized, however, 1–2 L spinner flask cultures can be generated in ~2 weeks time, while 10 L cellbag cell cultures can be generated in ~3 weeks time. Multiple spinner flask and/or cellbag cell cultures can be grown in parallel, given adequate access to the appropriate equipment (e.g. incubators, WAVE bioreactors, etc.). The timing for each stage of the Procedure is summarized below and in Figure 2.
We have overexpressed human RhCG, in addition to other human membrane proteins, in HEK293S GnTI- cells using the Procedure described here. For each membrane protein that was expressed, the progression through the protocol was very similar (Figure 2). Approximately 2–3 weeks post drug selection, single foci of cells are typically observed for at least one plating dilution (Figure 3), and clonal cell lines can be generated 2–3 weeks after this. The number of clonal cell lines generated can vary, however, as occasionally only <10–12 individual colonies in total are observed. The amount of membrane protein produced from each clone will likely vary significantly, necessitating the use of western blotting to assess expression levels (Figure 4). While human RhCG was well solubilized using OG, in our experience DDM is a preferred detergent given its greater ability to solubilize membrane proteins from HEK293S cells. For those human membrane proteins that cannot be successfully solubilized with DDM (see Table 1), solubilization using harsher detergents such as FC-14 can be attempted, however, caution should be exercised in these cases given the potential for solubilizing an inactive form of the membrane protein31.
A clonal cell line can typically be utilized indefinitely; human RhCG was expressed from a single, clonal cell line over the course of ~ 1 year with no appreciable loss of protein expression. In addition, cell lines can be frozen at any point, and thawed at a later time, again with no appreciable loss of protein expression. In rare cases, membrane protein expression levels from clonal cell lines have been found to decrease upon continued passage; in these cases care must be exercised to thaw fresh cells, and create new frozen stocks, prior to each cell culture scale up.
The use of both spinner flasks and cellbag WAVE bioreactors (Figure 5) for cell culture scale up is recommended, given that their respective scales complement one another. It is to be expected that cellbag cultures should reach higher density, given its more optimal oxygen transfer; however we did not observe this for the RhCG HEK293S cells. Rarely, for certain clonal cell lines, has viability been observed to decrease significantly following induction with doxycycline (see Table 1). In these cases, the cells should be harvested earlier than normal (see Step 63), which will ultimately result in less membrane protein purified per liter of cell culture medium. The optimal conditions for membrane protein stability (i.e. pH, [salt], [glycerol], etc.) can typically be determined using protein purified from a 1 L spinner flask. For human membrane proteins expressed in HEK293S cells, FLAG affinity purification typically results in higher yields and better purity compared with IMAC purification. Final yields of purified membrane proteins from stably transfected, and clonal HEK293S cells are typically up to ~0.5 mg of protein per L of medium.
This work was supported by NIH/NIGMS grants P50 GM73210, U54 GM094625 and R37 GM24485.
AUTHOR CONTRIBUTIONSS.C., F.G., and R.M.S. designed the experiments. S.C. and F.G. performed the experiments. S.C., J.E.P., F.G., and R.M.S. analyzed the data. V.S. and R.M.S. supervised personnel. S.C., J.E.P., and R.M.S. wrote the paper.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.