Laboratory-scale production of eukaryotic proteins with appropriate post-translational modification represents a significant barrier. Here is a robust protocol with rapid establishment and turnaround for protein expression using a mammalian expression system. This system supports selective amino acid, selective labeling of proteins and small molecule modulators of glycan composition.
The art of producing recombinant proteins with complex post-translational modifications represents a major challenge for studies of structure and function. The rapid establishment and high recovery from transiently-transfected mammalian cell lines addresses this barrier and is an effective means of expressing proteins that are naturally channeled through the ER and Golgi-mediated secretory pathway. Here is one protocol for protein expression using the human HEK293F and HEK293S cell lines transfected with a mammalian expression vector designed for high protein yields. The applicability of this system is demonstrated using three representative glycoproteins that expressed with yields between 95-120 mg of purified protein recovered per liter of culture. These proteins are the human FcγRIIIa and the rat α2-6 sialyltransferase, ST6GalI, both expressed with an N-terminal GFP fusion, as well as the unmodified human immunoglobulin G1 Fc. This robust system utilizes a serum-free medium that is adaptable for expression of isotopically enriched proteins and carbohydrates for structural studies using mass spectrometry and nuclear magnetic resonance spectroscopy. Furthermore, the composition of the N-glycan can be tuned by adding a small molecule to prevent certain glycan modifications in a manner that does not reduce yield.
Producing high yields of appropriately folded and post-translationally modified human proteins for detailed analysis of structure and function remains a significant challenge. A large number of expression systems are available that produce recombinant proteins with native-like function and behavior. Bacterial expression systems, predominantly Escherichia coli strains, represent the most accessible and commonly used tools in the research arena, due to the simplicity of these expression systems, though yeast, plant, insect and mammalian systems are also described1-4. However, the majority of these systems are incapable of appropriate post-translational modification of the target proteins. A fundamental interest of the Barb and Moremen laboratories is producing eukaryotic proteins with appropriate glycosylation. Many human proteins require appropriate glycosylation for proper function (see5).
The eukaryotic glycosylation machinery is extensive and capable of making a diverse range of modifications, including both asparagine(N)- and serine/threonine(O)-linked complex glycans6. It is estimated that >50% of human proteins are N-glycosylated7. Glycans are essential components of many proteins including therapeutic monoclonal antibodies, erythropoietin, and blood clotting factors like factor IX, to name a few. Though multiple methods exist to prepare appropriately N-glycosylated proteins and range from purely synthetic8-10, to chemoenzymatic11-14 or recovery from engineered recombinant systems15-20, not surprisingly, human expression systems have thus far proven to be the most robust methods for generating human proteins.
Many therapeutic human glycoproteins are produced in recombinant systems using mammalian cells. Systems of note are the Chinese Hamster Ovary (CHO), mouse myeloma (NS0), Baby Hamster Kidney (BHK), Human Embryonic Kidney (HEK-293) and human retinal cell lines that are employed in adhesion or suspension culture for protein production4,21,22. However, mammalian protein expression systems have required the generation of stable cell lines, expensive growth media and substrate assisted transfection procedures23.
Mammalian cell transfection is achieved with the aid of numerous agents including calcium phosphates24,25, cationic polymers (DEAE-dextran, polybrene, polylysine, polyethylimine (PEI)) or positively charged cationic liposomes26-29. PEI is a polycationic, charged, linear or branched polymer (25 kDa)26 that forms a stable complex with DNA and is endocytosed. Upon acidification of the endosome, PEI is thought to swell, leading to the rupture of endosomes and release of the DNA into the cytoplasm26,30.
Until recently, transient transfection in suspension culture was carried by prior DNA/PEI complex formation followed by addition to the cell culture29. However, Würm and coworkers reported a highly efficient protocol optimized for recombinant protein production in HEK293 cells that formed a DNA/PEI complex in situ31,32. This avoided preparation, sterilization of the complex, and buffer exchange into a culture medium. Further optimization by including expression-enhancing plasmids led to significant yield increases33. Herein is a method that builds upon these advances and is broadly applicable. Expression conditions may also be altered to impact the N-glycan composition.
The HEK293S cell line, with a gene deletion that halts N-glycan processing at an intermediate stage, leads to the expression of proteins with uniform N-glycans consisting of 2 N-acetylglucosamine residues plus five mannose residues (Man5GlcNAc2)34,35. These cells lack the N-acetylglucosaminyl transferase I (GntI) gene which is required for downstream N-glycan processing36,37. The use of glycosyltransferase inhibitors including kifunensine, sialic acid analogs and the fucose analog and 2-deoxy-2-fluoro-fucose has similar effects and limits N-glycan processing38-41.
The protocol reported here uses the pGEn2 vector as shown in Figure 142,43, PEI assisted transient transfection into mammalian cells lines (HEK293F or HEK293S cells), and the recovery of high yields of appropriately glycosylated protein. This system is robust and can accommodate various factors including isotope labeling and glycan engineering for the production of large titers of recombinant proteins.
This protocol is sufficient for expression using either HEK293F or HEK293S cells.
1. Cell Establishment
2. Prepare Materials for Transfection
3. Establishing a Transient Transfection
4. Protein Purification
5. Analysis of Protein by SDS-PAGE
6. Glycans Analysis by Mass Spectrometry
7. Protocol Adjustments for Special Scenarios
High-level protein expression and purity
This optimized expression system generated a high yield of glycosylated proteins. A typical pattern is shown in the expression of IgG1-Fc (Figure 1). In this case, Day 0 is the transfection day followed by Day 1 (dilution) and subsequent culture days up to Day 5. Protein expression is analyzed using the soluble expression fraction in the crude medium. A very small amount of protein expression was observed in Day 1 as the culture aliquot was withdrawn 3 hr after culture dilution. This is easier to visualize with GFP-tagged proteins that display a distinct green color in the culture medium. Such an increase was observed for in the expression of the GFP-FcγRIIIa and GFP-ST6GalI from Day 2 to Day 5. No significant increase in the expression was observed between Day 4 and Day 5. At Day 5, the cells were <50% viable and hence culture was harvested to limit proteolysis. Affinity purification using a Protein A column for IgG1 Fc or nickel column for GFP-FcγRIIIa resulted in proteins with high purity (>99%; Figure 2), yield (Table 1) and complete glycosylation (data not shown).
15N or 13C Isotope labeling of proteins and glycans
This system efficiently expressed isotopically-enriched IgG1 Fc according the described protocol. Although a small reduction in protein yield was observed (25-50%), well-dispersed signals were seen in a 2d NMR spectrum that correlates the resonance frequency of 1H atoms and the directly bonded amide 15N atoms (Figure 3). This level of expression permits efficient protein production on the scale required for structure-based studies (typically 2-20 mg of protein) and illustrates the successful incorporation of the labeled isotopes into the protein. The high degree of similarity between this spectrum and published spectra indicate a high degree of protein labeling occurred with minimal scrambling of the 15N labels43.
Producing proteins with different N-glycans
Different glycoforms were produced with the HEK293F and HEK293S cells. Matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) analyses of the enzymatically-released N-glycans showed the expected glycoforms (Figure 4). Proteins expressed from HEK293S cells harbored N-glycans that were of the Man5GlcNAc2 form and contained no fucose (Figure 4). Glycans from HEK293F-expressed material were complex-type, biantennary forms with core fucose and mostly having terminal N-acetylglucosamine (GlcNAc) or small proportion of mono- or di-galactosylated (Gal) forms. Though it is not know what the native N-glycans of ST6GalI and FcγRIIIa are, the glycan profile for IgG1 Fc is highly similar to IgG1 Fc purified from human serum46. The major differences include the degree of modification; IgG1 Fc from human serum shows a higher degree of galactosylation than observed for the HEK293F expression. Addition of 2-deoxy-2-fluoro-l-fucose, an inhibitor of glycan fucosylation, to an expression conducted using the HEK293F cell line showed a dramatic >90% reduction in fucose incorporation (Figure 4).
Figure 1: High yields of recombinant proteins are recovered by expressing from the pGEn2 vector. (A–B) Two expression vectors used in this study were generated from a pIBI30 plasmid that contains an ampR cassette and an E. coli replication origin. (C) Expression level of the secreted IgG1 Fc protein after transient transfection of HEK293F cells. SDS-PAGE analysis shows the accumulation of expressed proteins from Day 0 to Day 5 and is indicated by the arrow. Please click here to view a larger version of this figure.
Figure 2: Recovery of expressed protein from the culture medium. (A) IgG1-Fc, (B) GFP-FcγRIIIa. "Medium" refers to the culture medium after centrifugation; FT = flow through fraction from the purification column. SDS-PAGE samples are prepared in non-reducing conditions without β-mercaptoethanol. Please click here to view a larger version of this figure.
Figure 3: Amino acid selective labeling of IgG1 Fc. 1H-15N heteronuclear single quantum coherence spectrum of [15N-Tyr; 15N-Lys]-labeled IgG1 Fc expressed using HEK293F cells in custom Medium A supplemented with (15N) labeled l-Tyr and l-Lys. Crosspeaks were assigned based on previous reports43,47. Please click here to view a larger version of this figure.
Figure 4: Protein expression in the presence of small molecule modulators of glycan composition. N-glycan profiles of IgG1 Fc expressed in HEK293 cells using different culture conditions. The top spectrum was prepared using glycans isolated from IgG1 Fc expressed in HEK293S cells. The center spectrum reveals the IgG1 Fc glycoforms from material expressed in HEK293F cells and the bottom spectrum was prepared similarly expect cells were grown in the presence of an inhibitor of GDP-fucose biosynthesis, 2-deoxy-2-fluoro-l-fucose. N-glycans are presented as cartoon diagrams following the CFG convention5: N-acetylglucosamine (GlcNAc), blue squares; Mannose (Man), green circles; Fucose (Fuc) red triangles; Galactose (Gal), yellow circles. Please click here to view a larger version of this figure.
Protein | Yield (mg/L) |
IgG1 Fc | 120 |
GFP-FcgRIIIa | 100 |
GFP-ST6GalI | 95 |
Table 1: Yield for different proteins expressed in HEK293F cells.
This protocol illustrates protein expression via the transient transfection of HEK293F or S cells. The optimal transfection conditions established in the Barb and Moremen labs employ a critical combination of cell density and reagent concentrations to achieve high efficiency transfection. Critical considerations when implementing this protocol include: maintaining a stable culture prior to transfection (with consistent culture doubling times); transfection of actively growing cells (achieved by diluting cells to 1 × 106 cells/ml 24 hr prior to transfection) with cell viability greater than 95%; cell density at transfection should be between 2.5-3.0 x 106 live cells/ml with a viability >95% (transfection density) in a culture containing 90% Medium A and 10% Medium B; and, adding DNA prior to PEI addition at 3 μg/ml and 9 μg/ml, respectively31; at 24 hr post-transfection the culture is diluted 1:1 in medium containing 4.4 mM valproic acid to result in a concentration of 2.2 mM valprioc acid in the diluted culture; the production phase growth then continues in a humidified shake flask at 37 °C and 8% CO2 for an additional 4-5 days. The vector described here is optimized and an important feature of the soluble protein expression system32,42.
If a target protein fails to express, in addition to the factors listed above, multiple other variables may contribute to low yields. Ensure the protein coding DNA sequence contains codons optimized for human cells. This can be assessed with multiple online resources. Sterility throughout the procedure is paramount. An actively growing and pure culture of HEK293F cells should appear slightly grainy to the naked eye, and the medium should be mostly clear (particularly at cell densities <1.0 ×106 cells/ml).
Cultures contaminated with bacteria or fungus should be immediately removed to prevent spread. Is it important to maintain a healthy stock culture. This is achieved by inoculating these cultures at a cells density of 0.3 × 106 cells/ml. Lower inoculation densities can slow the growth rate by limiting the supply of various growth factors required for cell growth and division48. As cultures have been passaged for more than 25 or 30 times, a decrease in expression was observed. For this reason, cultures passaged more than 30 times are discarded.
It is possible the protein is being degraded in the expression medium, or the expression timeframe is too long, leading to protein degradation. For these reasons it is recommend to monitor the culture viability and protein expression at each day to determine an optimal expression timeframe. It is important to purify proteins as quickly as possible after harvesting the medium. For some proteins, it is helpful to include protease inhibitors during purification. In total, these adjustments have improved the recovery of proteins in the Barb and Moremen labs42,43.
The transient transfection of HEK293 cells has shown great utility for producing soluble proteins and protein domains that are naturally targeted to the secretory pathway. It is likely that production of cytoplasmic or integral membrane proteins will require further optimization. A primary advantage of this system is the native substrates and machinery for protein production are present and available in the HEK293 cells49. This largely accounts to its advantages over other recombinant protein production methods such as bacteria with no or limited post-translational modifications, or yeast and baculovirus systems with correctly folded but different N-glycoforms incorporated1,50,51.
Furthermore, the protocol described here shows the added capability to include low molecular weight compounds such as labeled amino acids, labeled glucose or small molecules designed to modify the N-glycan composition. The HEK293 cells also prove adept at expression of non-mammalian proteins, including plant enzymes, and supports co-expression of multiple polypeptides for the recovery of protein complexes (data not shown).
The structural and functional characterization of glycoproteins is often hampered by sample production challenges. The protein expression system described here surmounts this drawback by accomplishing post-translational modifications using the natural complement of sophisticated intracellular glycan processing enzymes6. This robust and simple protocol is proven for the transient transfection of suspension human HEK293 cells. This method demonstrated significant glycoprotein yield (>95 mg/L) with three N-glycosylated proteins and can accommodate supplements such as labeled amino acid residues or the small molecule chemical inhibitor 2-deoxy-2-fluoro-l-fucose. The expressed glycoproteins can be further remodeled post-purification to prepare a wide range of defined glycoforms52.
The authors have nothing to disclose.
This work was financially supported by the grants K22AI099165 (AWB), P41GM103390 (KWM) and P01GM107012 (KWM) from the National Institutes of Health, and by funds from the Roy J. Carver Department of Biochemistry, Biophysics & Molecular Biology at Iowa State University. The content of this work is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Biosafety cabinet | NuAire, Inc. | CellGard ES NU-S475-400 | Class II, Type A2 Biological Safety Cabinet |
Incubation shaker | INFORS HT | Multitron Cell | |
Medium A: FreeStyle Expression Medium | Life Technologies | 12338-018 | |
Medium B: ExCell 293 Serum-Free Medium | SIGMA | 14571C | |
125 Erlenmeyer Flask with Vented Cap | Corning Incorporated/Life Sciences | 431143 | |
250 Erlenmeyer Flask with Vented Cap | Corning Incorporated/Life Sciences | 431144 | |
FreeStyle HEK 293F Cells | Life Technologies | R790-07 | |
1 ml Disposable serological pipette | Fisher Scietific | 13-676-10B | |
10 ml Disposable serological pipette | Fisher Scietific | 13-676-10J | |
25 ml Disposable serological pipette | Fisher Scietific | 13-676-10K | |
Pipettor (Pipet-Aid XP) | Drummond Scientific | 161263 | |
Trypan Blue Solution | Thermo Scientific | SV30084.01 | |
Counting slides | Bio-Rad | 145-0011 | |
TC20 Automated Cell Counter | Bio-Rad | 145-0102 | |
Polyethylenimine (PEI) | Polysciences Inc. | 23966 | Prepare stock solution at a concentration of 1 mg/ml in a buffer containing 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) and 150 mM NaCl (pH 7.5). Dissolve PEI completely; sterilize through 0.22 μM syringe filter and store at -20 °C. |
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) | EMD Chemicals Inc. | 7365-45-9 | |
25 mm Syringe Filter, 0.22 μM | Fisher Scietific | 09-719A | |
Trisaminomethane (Tris base) | Fisher Scietific | BP152-1 | |
XL1-Blue | Stratagene | 222249 | |
Trypton | Fisher Scietific | BP1421-2 | |
Yeast extract | Fluka Analytical | 92144 | |
Sodium chloride | BDH chemicals | BDH8014 | |
Plasmid Purification Kit | QIAGEN | 12145 | |
Valproic (VPA) | SIGMA | P4543 | Prepare stock solution of 220 mM in water, sterilize by passage through a sterile 0.22 mm filter and store at -20 °C |
Corning 250 mL Centrifuge Tube | Corning Incorporated/Life Sciences | 430776 | |
Centrifuge | Thermo Scientific | EW-17707-65 | |
Protein A-Sepharose column | SIGMA | P9424 | |
Ni-NTA superflow | QIAGEN | 30430 | |
3-(N-morpholino)propanesulfonic acid (MOPS) | Fisher Scietific | BP308-500 | |
Glycine | Fisher Scietific | BP381-500 | |
10 kDa molecular weight cut-off Amicon® Ultra centrifugal filters | Millipore | UFC901096 | |
Sodium dodecyl sulfate | ALDRICH | L3771 | |
Beta-mercaptoethanol | ALDRICH | M6250 | |
Glycerol | SIGMA | G5516 | |
Precision Plus Protein All Blue Standards | Bio-Rad | 161-0373 | |
Acetic Acid, Glacial | Fisher Scietific | 64-19-7 | |
coomassie brilliant blue | Bio-Rad | 161-0406 | |
MALDI-TOFMS Voyager-DE PRO | Applied Biosystems | ||
15N labeled L-Tyrosine | ALDRICH | 332151 | |
15N labeled L-Lysine | ALDRICH | 592900 | |
Unlabeled L-Phenylalanine | SIGMA-ALDRICH | P2126 | |
13C6-Glucose | ALDRICH | 389374 | |
2-deoxy-2-fluoro-l-fucose | SANTA CRUZ Biotechnology | sc-283123 |