Production of eukaryotic glycoproteins for structural and functional studies using Expi293F cells

Abstract

Milligram quantities of pure proteins are required for structural, functional, and pharmaceutical screening studies. These requirements can be challenging for a majority of important therapeutic targets that are secreted glycoproteins, receptors, membrane proteins, or large cytosolic complexes. Here we present a protocol for producing and purifying large amounts of secreted glycoproteins using the mammalian cell-based Expi293F system via large-scale transient transfection. This system can be easily adapted for the production of membrane proteins and large cytosolic complexes. The method can be utilized to quickly evaluate numerous expression constructs to identify optimal expressers. The use of mammalian cells ensures proper postranslational modifications, including disulfide bonds and glycosylation, that can be important for accurate functional studies. In addition, minor modifications can be introduced to produce labeled or deglycosylated proteins for structural studies by X-ray crystallography, NMR, or cryo-electron microscopy.

Basic Protocol 1: Production of milligram quantities of plasmid DNA for large-scale transient transfection

Basic Protocol 2: Large-scale culture and transient transfection of Expi293F cells

Basic Protocol 3: Purification of hexahistidine-tagged proteins from media

INTRODUCTION

Comprehensive investigations of protein structure, function, and pharmacologic screening require milligram amounts of pure proteins. For mammalian proteins that contain a signal peptide and are endogenously expressed in the secretory pathway, this task can be challenging. Such proteins usually contain post-translational modifications such as disulfide bonds and N- and O-linked glycosylation that can be difficult to replicate in most common recombinant protein expression systems. For example, bacterial expression systems are fast and inexpensive, but most extracellular mammalian proteins expressed in them are insoluble and improperly folded, requiring further refolding, purification, and evaluation steps to produce properly folded proteins with intact disulfide bonds (Nelson et al., 2014). Even in cases where mammalian extracellular proteins can be refolded from bacterial expressions, these proteins are not glycosylated, which can be critical for proper function. Eukaryotic yeast (Baghban et al., 2018; Daly & Hearn, 2005) and insect cells expression systems (Altmann et al., 1999) can produce proteins containing disulfide bonds and glycosylation. However, the glycosylation patterns differ significantly from mammalian cells, yielding proteins with non-homogeneous glycosylation or containing glycans not usually enriched in mammalian proteins, leading to spurious functional observations (Geisler et al., 2015; Harrison & Jarvis, 2006). For these and other reasons, protein expression systems based on mammalian cell lines have emerged as the preferred method to produce eukaryotic glycoproteins for structural and functional studies (Kober et al., 2015).

The most common mammalian cell lines utilized for recombinant eukaryotic glycoprotein production are based on the human embryonic kidney 293T cell line (HEK293T), as they could be grown in serum-free media, making it easier to purify secreted proteins. The first generation utilized adherent HEK293T cells (Aricescu et al., 2006). These were of limited utility due to the limited cell density per volume that can be achieved with adherent cells. This limitation was overcome by the next generation of systems that utilized HEK293T cells that had been adapted to growing in suspension, such as FreeStyle 293-F cells (Kober et al., 2015). Although the increase in cell density led to increased protein production per culture volume, these cells required careful monitoring of cell density (not to go above 1–2 × 10⁶ cells/mL) and media glucose levels in order to ensure consistent protein production yields. In recent years, the Expi293F system has been introduced, which produces up to 6 times more protein when compared to the FreeStyle 293-F system. The Expi293F cells are more robust, as they can be grown to higher densities (5–6 × 10⁶ cells/mL) and transfected at higher densities (2–3 × 10⁶ cells/mL) without impacting protein production.

Due to speed and the ability to easily produce and test new expression constructs, large scale transient transfection is the optimal method used to introduce expression plasmids into Expi293F cells. Other methods, such as stable integration (which can take several weeks) or viral transduction using BacMam (Dukkipati et al., 2008) (which can take about a month to produce virus) are slow and tedious, which can also limit the ability to try various expression constructs. With large scale transient transfection, it is possible to use gene cloning to produce milligrams of pure protein within 2 weeks, and also allows one to design and trial numerous expression constructs for optimization (Figure 1). One downside to this method is the use of large amounts of transfection reagents, which can be cost prohibitive. In the past, we have used a modified polyethyleneimine (PEI)(Kober et al., 2015) which is cheap, but has a somewhat low transfection efficiency and is mildly cytotoxic. The transfection reagent ExpiFectamine 293 boasts the highest transfection efficiency, but is quite expensive. For our purposes, we have struck a middle ground using the less expensive Hype293 transfection reagent, which has a high transfection efficiency, is not cytotoxic, and consistently produces large quantities of secreted proteins.

Figure 1. Flow schematic for testing and identifying optimal expression constructs.

A) Domain schematic of human CLCA1. Domains labeled as follows: CAT: matrix metalloprotease like catalytic domain; CYS: cysteine-rich domain; VWA: von Willebrand A domain; FnIII: fibronectin III domain; SS: signal sequence. B) Protein of interest (POI), in this case various CLCA constructs, are cloned into a suitable high copy number plasmid. Minipreps of each construct provide suitable amounts of plasmid DNA for sequencing and test expressions. Test transfections can be carried out in 6 well plates (2 mL culture per well), allowed to express for 72 hours, then samples of supernatant can be analyzed by SDS-PAGE/western blot to evaluate secretion levels and identify optimal constructs for large-scale expression.

In this protocol, we outline large-scale transient transfection of Expi293F cells to produce milligram quantities of secreted glycoproteins. This method is easily adaptable to producing membrane proteins, soluble cytosolic proteins, or large complexes, and can be adapted for specialized structural biology purposes such as producing deglycosylated proteins or selenomethionine labeled proteins for X-ray crystallography. The secretion of proteins into a serum-free, chemically-defined medium allows rapid and facile purification for structural, biophysical, and functional studies.

Basic Protocol 1: Production of milligram quantities of plasmid DNA for large-scale transient transfection

Milligram quantities of pure plasmid DNA are required for large scale transient transfection. In this basic protocol we describe how this can be achieved using high efficiency E. Coli through cloning, transformation, and purification. A MaxiPrep typically produces enough DNA to transfect 1–2 liters whereas GigaPrep provides enough plasmid DNA for several liters.

• NEB 10-beta Competent E. coli (New England Biolabs, cat. no. C3019H)

• pHL-sec plasmid (Addgene, plasmid no. 99845)

• pHL-avitag3 plasmid (Addgene, plasmid no. 99847)

• pHL-FcHis plasmid (Addgene, plasmid no. 99846)

• GenElute HP Select Plasmid Gigaprep Kit (Millipore Sigma, cat. no. NA0800)

• Maxiprep kit (Qiagen, cat. no. 12162)

• Lennox L Broth (LB) (Research Products International, cat. no. L24066)

• Carbenicillin, disodium salt (GoldBio, cat. no. 4800-94-6)

• Quick CIP (New England Bio Labs, cat no. M0525S)

• QIAquick Gel Extraction kit (QIAgen, cat no. 28704)

• Quick Ligation Kit (New England Bio Labs, cat. no. M2200S)

• Temperature-controlled upright or chest shaker

• 2 L or 500 ml glass Erlenmeyer shaker flasks, clean and sterile

• 1L or 500 ml centrifuge tubes

• Large-volume floor refrigerated centrifuge (e.g., Beckman Coulter Avanti J-25)

• Thermocycler for standard PCR

Protocol

1. Clone the protein of interest into a high copy number mammalian expression vector using restriction enzyme site cloning.

   1. Use PCR to amplify gene of interest using appropriate primers containing desired restriction enzyme ends
   2. Restriction enzyme digest on plasmid inserts and vector using appropriate restriction enzymes; dephosphorylate the cut vector using Quick CIP
   3. Purify the inserts and vector using the QIAquick Gel Extraction kit
      Gel purification often results in low yield DNA returns
   4. Ligate insert into vector using Quick Ligation Kit according to protocol

   We typically use pHLsec vector with a built in 6His-tag and a strong secretion signal for optimal results.

2. Transform the plasmid into competent cells

   1. Add 20 μl of NEB 10-beta competent E. Coli cells to 1 μg plasmid DNA and incubate on ice for 30 min
   2. Heat shock at 42° for 30–45 sec, then incubate on ice for 2 min
   3. Add 300 μl of microbial growth medium (SOC) and incubate at 37° C for 45 mins, shaking at 220 rpm
   4. Place cells on agar plate with antibiotic selection
      If using pHLsec vector, use 100 μg/mL carbenicillin

3. Select surviving colonies to inoculate 4 × 1L of Lennox L Broth (LB) media supplemented with 100 μg/mL of appropriate antibiotic O/N at 37 °C shaking at 220 rpm

   If a smaller amount of plasmid DNA is sufficient, one can use a Qiagen Maxiprep kit with 250 mL culture and obtain 1–1.5 mg of plasmid DNA for most high-copy number plasmids.

4. Purify DNA from LB culture using GenElute HP Select Plasmid Gigaprep Kit

   GenElute HP Select Plasmid Gigaprep kit produces roughly 10–50 milligrams of plasmid DNA; use kit according to protocol

   1. Elute DNA with endotoxin-free water
   2. Aliquot the purified plasmid at the desired quantity for storage at −20 °C
      We usually prepare 300 μg aliquots for frozen storage, as one aliquot is sufficient to transfect one 300 mL culture, our typical per-flask volume for large-scale transient transfection. Aliquot volumes vary based on sample concentration but typically range between 200–100 μL

Basic Protocol 2: Large-scale culture and transient transfection of Expi293F cells

In this basic protocol we describe large scale culture of mammalian cells in suspension and their use to produce microgram to milligram quantities of proteins containing accurate post translational modifications.

• Expi293F cells (Thermo Fisher Scientific, cat. no. A14527)

• Expi293F GnTI- cells (Thermo Fisher Scientific, cat. no. A39240) optional

• Expi293 Expression Medium (Life Technologies, cat. no. A1435101)

• Expi293 Met(−) Expression Medium (Life Technologies, cat. no. A4096701)

• Trypan blue solution, 0.4% (Thermo Fisher Scientific, cat. no. 15250061) optional

• Opti-MEM reduced serum medium (Life Technologies, cat. no. 31985–070)

• HYPE-293 Transfection Kit (Oz Biosciences, cat. no. HY29330)

• Penicillin Streptomycin 100X (Pen Strep) (Life Technologies, cat. no. 15070–063)

• Antibiotic/Antimycotic (100X) (Life Technologies, cat. no. 15240062)

• Erlenmeyer shaker flask (1000 mL, baffled, vented, sterile) (Greiner, cat. no. 679514)

• Erlenmeyer shaker flask (250 mL, baffled, vented, sterile) (Greiner, cat. no. 679512)

• 15-ml conical centrifuge tubes

• Laboratory water bath

• Automated cell counter

• Tissue culture incubator (set at 37 deg C and 8% CO2)

• CO2 Resistant Shakers (ThermoFisher Scientific, cat. no. 88881103)

• Sticky Mat for CO2 resistant shaker (ThermoFisher Scientific, cat. no. 88881126)

• Serological pipettes and pipette controller

• Tissue culture hood

Protocol

1. Supplement 1 L Expi293 media with 10mL Antibiotic/Antimycotic (100X) (A/A)

   We typically use A/A, which contains Pen/Strep and Amphotericin B to prevent both bacterial and fungal infection, but one can alternatively use Pen/Strep alone if they are concerned about protein yield. We have not noticed any decrease in protein yield when using A/A.

2. Culture Expi293F cells from frozen in 30mL Expi293 Expression Media in 125mL Erlenmeyer shaker flask with vented caps at 37 °C with 8% CO₂ while shaking (120 rpm) in a standard tissue culture incubator until density is about 2×10⁶/mL (roughly 4–6 days).

3. Transfer cells to 1L Erlenmeyer shaker flask with vented caps and dilute to 150mL using Expi293 Expression Media. Allow cells to grow until density reaches about 2×10⁶/mL and then dilute to 300mL.

   Expi293F cell densities can go to about 6×10⁶ per mL. Cells can be split to additional flask and cultured between 0.5–62×10⁶/mL. Expi293F cells typically grow at a rate where cell densities double daily.

4. Dilute cells to 1–1.5×10⁶/mL density using Expi293 Expression Media one day before transfection.

   Cell density on the day of transfection should be about 2–3×10⁶ cells/mL.

5. Calculate volume of DNA required for 1 μg plasmid per ml of cell culture. Under sterile conditions, dilute DNA in 5 mL Opti-MEM reduced-serum medium.

   For example, we typically use 1L vented Erlenmeyer shaker flasks containing 300 mL of cell culture at 2–3×10⁶ cells/per mL, requiring 300 μg of plasmid DNA.

6. Calculate volume of HYPE-293 transfection reagent required to obtain a 1:1.5 DNA:transfection reagent ratio. Dilute in a separate 5 mL of Opti-MEM.

   For the example above, this would require 450 μL of HYPE-293 for 300 μg of plasmid DNA.

7. Add transfection reagent:Opti-MEM solution to the plasmid:Opti-MEM solution. Incubate for 15 minutes at room temperature so that transfection reagent:DNA complexes can form.

   The 15 minute incubation time can be a critical variable. Allowing to go for longer than 20 minutes can impact protein yields.

8. Add transfection reagent:plasmid solution to the cell culture in a drop wise fashion while gently swirling cells. Tightly cap cultures while in tissue culture hood. Place cultures back in incubator on shaker platform and resume shaking. Allow cells to express protein for 3–7 days post transfection.

   To increase protein expression yield, one can add the appropriate volume of 100X B293 reagent, which is included in HYPE-293 transfection kits. The optimal number of days for maximum protein yield will be protein dependent and can be monitored by taking samples each day and analyzing/comparing using SDS-PAGE/western blot.

Basic Protocol 3: Purification of hexahistidine-tagged proteins from media

In this basic protocol we purify the proteins expressed through large scale culture and transfection that are secreted into the expression media. As an example, we discuss the purification of the human calcium activated chloride channel regulator 1 von Willebrand A domain (CLCA1 VWA). This protein (CLCA1 residues 302–476) was cloned into the pHLsec vector and contains the encoded C-terminal 6-histidine tag for purification.

• Ni-NTA superflow (Qiagen, cat. no. 30430)

• 10x Ni-NTA binding buffer (1.5 M NaCl, 0.5 M K2HPO4, 0.1 M Tris pH 8.5, 50 mM imidazole)

• Wash buffer (300 mM NaCl, 50 mM K2HPO4, 20 mM imidazole pH 8)

• Elution buffer (300 mM NaCl, 50 mM K2HPO4, 250 mM imidazole pH 8)

• Pierce High Capacity Endotoxin Removal Spin Columns (ThermoFisher Scientific, cat. no. 88277)

• Cell culture from basic protocol 2 step 7

• Amicon Ultra-15 centrifugal concentrator (MilliporeSigma, cat. no. UFC901024)

• Large-volume floor refrigerated centrifuge (e.g., Beckman Coulter Avanti J-25)

• 1L or 500 ml centrifuge bottles

• Ring stand and clamp

• Econo-column (Bio-Rad, cat. No. 7374253 or 7374156)

Protocol Steps.

1. Decant cell culture from protocol 2 step 8 into centrifuge tubes and centrifuge for 5 minutes at 3,000 rpm to pellet cells. Collect supernatant, which will contain the secreted CLCA1 VWA protein, and centrifuge for 10 minutes at 7000 rpm to further clarify the supernatant. Decant into storage bottle.

   The second centrifuge step serves to remove small particulates that could clog the column. This step is done as an alternative to filtering the supernatant, which can sometimes incur loss of protein.

2. Add 10% volume of 10x Ni-NTA binding buffer (e.g. 30mL per 300mL media).

   Optionally, one can concentrate the supernatant at this step using a large volume concentrator. This increases the 6-His tagged protein concentration in the supernatant, which leads to a higher binding efficiency to the NiNTA resin, increasing the yield of this purification step.

3. Prepare a gravity column by adding about 2mL of Ni-NTA superflow slurry and equilibrate with 10 column volumes (i.e. 20mL) of 1x Ni-NTA binding buffer. Take caution to not allow the resin to go completely dry.

   The binding capacity of Qiagen Ni-NTA resin is about 50 mg of 6-His tagged protein per 1 mL of packed resin. One can increase protein yield using a larger volume of resin, but this comes with the tradeoff of increasing the non-specific binding of contaminants. Chill buffer and protein on ice before preparing gravity column, and keep protein and collected flow through on ice. Beads can be stripped and recharged for multiple uses.

4. Flow the supernatant over the resin and collect flow through. Pour flow through over the column once more and collect.

   As an option, one can also bind the protein to the resin by batch binding. In this process, equilibrated Ni-NTA beads are slurried with the 6-His-tagged-protein containing-supernatant, and gently agitated for 2 hours- overnight, with the slurry then poured into an empty gravity column. This can improve 6-His protein binding yields which might be limited due to slow binding kinetics of some 6-His tagged proteins to Ni-NTA.

5. Wash with 10 column volumes of wash buffer and collect.

6. Elute the CLCA1 VWA protein in 5 column volumes of elution buffer.

   Let the buffer sit on the column for 3 minutes before allowing it to flow through. This will optimize the removal of 6-His tagged protein by the imidazole in the elution buffer.

7. Optional use-dependent step: If the protein is going to used for structural or quantitative studies, it can be further purified using other chromatography techniques, such as size exclusion (SEC) or ion exchange chromatography, as needed. For example, to prepare the CLCA1 VWA protein for crystallization, the eluted protein in elution buffer is concentrated to around 0.5 mL volume using an Amicon Ultra-15 centrifugal filter (10,000 MW cutoff). This protein is then purified using and FPLC sytem (e.g., Akta Pure with s75 increase 10/300 GL column- example chromatogram shown in Fig. 2C). Colum is run using

8. Optional use-dependent step: Endotoxin removal: If the purified protein will be used for in vivo or in vitro functional studies, the sample should be assayed for potential endotoxin contamination. Even though the proteins are expressed in mammalian cells and mostly utilize new sterile plasticware, endotoxin can still be picked from lab glassware if one regularly works with bacteria. Endotoxin can be removed using High Capacity Endotoxin Removal Spin Columns kit according to protocol. Some protein loss can be incurred, but protein recovery should be >85%.

9. Optional use-dependent step: Deglycosylation:

   1. To the eluant from step 6, add 1/10 step 6 eluant volume of 500 mM Na-Citrate pH 5.5.
   2. Add 20 μL of EndoHf (1 × 10⁶ U/mL). Incubate at room temperature for 2 hours.
   3. Prepare the Amylose Resin by washing 3x in phosphate buffered saline (500 ul) (PBS) or final storage buffer.
   4. Mix the resin solution to the purified protein sample from 9a for 1 hour at 4 °C. Spin 5 min at 1000 × g to pellet beads and collect the supernatant.
      Note: The enzyme works optimally at 37 °C, which may cause the concentrated protein to aggregate. Extend the room temperature incubation, if deglycosylation, assessed by SDS-PAGE or immunoblotting, is incomplete. The enzyme does not have activity at 4 °C.

10. Quantitation and storage. Typical protein yield for CLCA1 VWA after NiNTA and SEC is about 3 mg of pure protein per 1 L of cell culture. Protein can used fresh or concentrated to 1–10 mg/mL, aliquoted, frozen and stored at −80C.

Figure 2. Flow scheme for secreted glycoprotein production.

A) For biophysical and functional studies, proteins are expressed in Expi293F cells and purified from the media using NiNTA chromatography, followed by another chromatography step, often size exclusion chromatography (SEC). These proteins can then be used for biophysical studies. If proteins are to be used in functional cellular or in vivo assays, then endotoxin should be removed. B) For structural studies, proteins are expressed in GnTI- as in A). Glycans are then removed using EndoHf. The treated protein can then be crystallized for crystallography or used in cryo-EM studies. C) Representative SEC chromatogram for CLCA1 VWA protein produced with Expi293F cells and analysis of purity by SDS-PAGE.

COMMENTARY

Background Information

While genetics-based studies have identified protein targets of interest in health and disease, follow-up studies often require the production of milligram quantities of pure proteins for structural, binding, functional, and pharmacological screening studies. Many human proteins require post-translational modifications, such as glycosylation, disulfide bond formation, and chaperones, for proper folding and function. For secreted glycoproteins, there are a number of possible expression systems, each with inherent advantages and disadvantages (Table 1). In this protocol, we present a method for producing secreted glycoproteins by large-scale transient transfection of mammalian Expi293F cells. This system allows for a number of advantages. For one, these cells have the required cellular machinery to ensure proper folding and posttranslational modification. Second, these cells are grown in suspension and can be cultured to relatively high densities, allowing for more protein production in a given volume. This can be especially relevant for membrane proteins, where production is limited by the amount of cellular membrane available. Third, transient transfection allows a superior advantage in terms of speed. Production of a stable expressing cell line or production of virus can take over a month. For large scale transient transfection, protein production can occur within the timeframe it takes to clone a new construct and purify μg quantities.

Table 1.

Comparison of protein expression systems

System	Relative cost	Relative speed	Relative ease	Pros/Cons
E. coli	Cheap	Fast	Easy	Pro: Fast and cheap production of soluble or refolded proteins Pro: Easy to incorporate labels for specialized structural biology experiments (e.g., SeMet or spin labels). Con: Native glycosylation and disulfide bond formation are not easily achieved
Mammalian cells (transient transfection)	Moderate	Moderate	Moderate	Pro: Native glycosylation and disulfide bonds Pro: Easy to test multiple expression constructs for optimization Con: Moderately expensive and uses some special equipment
Mammalian cells (generation of stable expressing cell line)	Moderate	Slow	Moderate	Pro: Native glycosylation and disulfide bonds Pro: Stable expression of target protein integrated, so no need for large amounts of expensive transfection reagents Con: Slow to produce stable expressing cell line Con: Cannot test multiple expression constructs
Yeast	Cheap	Moderate	Moderate	Pro: Cheap and supports glycosylation and disulfide bond formation Con: Glycans differ from mammalian cells and can impact functional studies Con: Need to electroporate for transfection
Insect cells (baculovirus)	Moderate	Slow	Moderate	Pro: glycosylation and disulfide bonds Con: Slow; Need to make virus (about 4 weeks) Con: Glycans differ from mammalian cells and can impact functional studies

Problem	Possible cause	Solution
Low protein expression in cells	Disrupted or non-existent Kozac consensus sequence	Analyze cloning strategy to ensure existence of optimal Kozac sequence preceding/including the initiating codon.
Low protein expression in cells	Poor transfection efficiency	Evaluate transfection efficiency by transfecting/expressing eGFP alone or in combination with the protein of interest at 1/10 of the total plasmid DNA transfected. If cells have been passaged many times, you may need to discard and restart using a low-passage frozen stock of Expi293F cells.
Protein not detected secreted into media or secretion is low	Disrupted signal sequence.	Analyze cloning strategy, ensure intact signal sequence exists using bioinformatic analysis (SignalP, etc.). Alternatively, utilize another signal sequence (e.g., HA) or use the signal sequence endogenous to the protein.
Protein not detected secreted into media or secretion is low	Construct boundaries do not yield properly folded or unstable protein	Screen expression and secretion of constructs with variable boundaries based on predicted structural domain boundaries.
Protein not detected secreted into media or secretion is low	Protein may bind strongly to cells.	Wash cells with a high salt buffer (0.5 – 1.0 M NaCl) to release protein from potential electrostatic retention tp cells.
Most of 6-His tagged protein appears in Ni-NTA column flow-through	Media pH not adjusted properly.	Measure pH of meda after addition of the 10X binding buffer. Adjust pH to 8.0 −8.5 with NaOH if needed.
Most of 6-His tagged protein appears in Ni-NTA column flow-through	Low concentration of 6-His tagged protein in media	Use large-volume concentrator to concentrate 6-His tagged protein in supernatant prior to Ni-NTA column. Batch binding can also be used to increase yield.
Little of the 6-His tagged protein elutes from the Ni-NTA column and substantial amounts of it remain on the column.	Protein precipitates when concentrated on Ni-NTA resin.	Adjust the elution buffer (e.g., add 10% glycerol, etc.) to prevent precipitation on column.

An advantage of large-scale transient transfection of mammalian cells is that it can be adapted for various types of protein production and uses. While the method presented here focuses mainly on production of secreted glycoprotein ectodomains, the same system can be used to express cytosolic or membrane proteins (Yeliseev, 2021). In addition, it can also be used to produce multi-protein complexes by either co-transfection of multiple plasmids, utilization of multicistronic vectors, or a combination of both (Portolano et al., 2014). In addition, the system is readily adaptable for specialized protein production for structural biology. For example, attached glycans can often be inhibitory to protein crystallization for X-ray crystallography. For these purposes, glycosylated proteins can be expressed in the engineered Expi293F GnTi- cells. These cells lack N-acetylglucosaminyltransferase I (GnTI) activity, producing proteins that contain immature high-mannose glycans that can later be removed with EndoH (or EndoHf). This produces proteins containing a single N-linked glycan that is more amenable to crystallization. In addition, there is Expi293 Met(−) expression medium available that can be supplemented with modified methionine adducts to label proteins with either selenomethionine for X-ray crystallography experiments or ¹³C methylmethionine for NMR experiments (for overview, see Figure 2).

Critical Parameters and Troubleshooting

The large-scale transient transfection method can be costly as it uses large amounts of transfection reagent and plasmid DNA. For this reason, it is important to utilize an optimal expression vector. The major factors to consider are transfection efficiency, promoter strength, and copy number. We typically use pHLsec (Aricescu et al., 2006), as it is a very high-copy number plasmid. It is a relatively small vector (around 5000 bp), so it transfects efficiently with variable-length inserts. It also utilizes the very strong chicken β-actin promoter. In addition, it contains other useful elements, such as a strong signal sequence for producing secreted proteins, a strong Kozac consensus sequence for high protein expression, and a C-terminal 6-Histidine tag for purification. There are also other related plasmids for easy sub-cloning by restriction enzyme cut-and-paste into vectors to produce Fc-fusion proteins (pHL-FcHis) or C-terminal biotinylation tags (pHL-avitag3). A minor disadvantage of pHLsec that it does not encode a selection marker, so that if one identifies an optimally expressing construct that would be used extensively, one would need to subclone into a different plasmid (e.g., pcDNA3.1) that encodes a selection marker for the purposes of making a stable expressing cell line in order to avoid repeated use of large amounts of transfection reagent.

The speed afforded by transient transfection allows for the quick design and testing of numerous constructs to identify optimal expressing constructs or mutants prior to large-scale production. Expressions are typically tested using 2 mL cultures of Expi293F cells in 6-well tissue culture plates (Fig. 1). These test expressions need only 2 μg of plasmid DNA. For high-copy number plasmids, a miniprep of a colony after cloning provides a sufficient amount for sequencing and test expressions, making it fast and easy to identify optimal expressing constructs which can then be used on larger scales. Once an optimized expression construct has been identified, if that protein will be used for extensive long-term studies, one could opt to then clone into a BacMam system or plasmid to create a stable expressing cell line in order to save on repetitive use of large amounts of expensive transfection reagent.

For optimal transfection efficiency and protein production, Expi293F cells can be passaged up to 30 times, but may need to be discarded after that. Transfection efficiency can be monitored by transfecting and expressing eGFP. In order to diagnose possible challenges with expression, secretion, and purification, one should collect samples (e.g., cells after expression, media supernatant, Ni-NTA flow through, wash, elute, and beads) throughout the expression and purification process for later analysis by SDS-PAGE/western blot.