This protocol describes a purification method for recombinant tag-free calcium binding protein S100A12 and its ion-induced oligomers for human monocyte stimulation assays.
In this protocol, we describe a method to purify human calcium-binding protein S100A12 and its ion-induced oligomers from Escherichia coli culture for immune cell stimulations. This protocol is based on a two-step chromatography strategy, which comprises protein pre-purification on an anion-exchange chromatography column and a subsequent polishing step on a hydrophobic-interaction column. This strategy produces S100A12 protein of high purity and yield at manageable costs. For functional assays on immune cells eventual remnant endotoxin contamination requires careful monitoring and further cleaning steps to obtain endotoxin-free protein. The majority of endotoxin contaminations can be excluded by anion-exchange chromatography. To deplete residual contaminations, this protocol describes a removal step with centrifugal filters. Depending on the available ion-strength S100A12 can arrange into different homomultimers. To investigate the relationship between structure and function, this protocol further describes ion-treatment of S100A12 protein followed by chemical crosslinking to stabilize S100A12 oligomers and their subsequent separation by size-exclusion chromatography. Finally, we describe a cell-based assay that confirms the biological activity of the purified protein and confirms LPS-free preparation.
S100A12 is a calcium binding protein which is predominantly produced by human granulocytes. The protein is overexpressed during (systemic) inflammation and its serum levels, particularly in (auto)inflammatory diseases such as systemic juvenile idiopathic arthritis (sJIA), familial Mediterranean fever (FMF) or Kawasaki disease (KD) can inform about disease activity and response to therapy. Depending on pattern recognition receptors (PRRs) such as toll-like receptors (TLRs), the innate immune system can be activated by pathogen-associated molecular patterns (PAMPs) like lipopolysaccharides (LPS) or damage associated molecular patterns (DAMPs; also termed ‘alarmins’). DAMPs are endogenous molecules such as cellular proteins, lipids or nucleic acids1. DAMP-functions are well described for the members of the calgranulin protein family, S100A8/A9 and S100A122, which are also reported to operate as divalent metal ion-chelating antimicrobial peptides3,4,5,6. Depending on the available ion strength S100A12 can, like other members of the S100 family, arrange into different homomultimers and until recently the impact of S100A12-oligomerisation on PRR-interaction, particularly TLR4, was unknown.
The protein’s monomeric form (92 amino acids, 10.2 kDa) consists of two EF-hand helix-loop-helix structures connected by a flexible linker. The C-terminal EF-hand contains the classical Ca2+-binding motif whereas the N-terminal EF-hand exhibits an S100 protein-specific extended loop structure (‘pseudo-EF-hand’) and reveals reduced Ca2+-affinity. Ca2+-binding by S100A12 can induce a major conformational change in the proteins’ C-terminus, which results in exposure of a hydrophobic patch on each monomer and forms the dimerization interface. Thus, under physiological conditions, the smallest quaternary structure formed by S100A12 is a non-covalent dimer (approximately 21 kDa) in which individual monomers are in antiparallel orientation. When arranged as dimer, S100A12 is reported to sequester Zn2+ as well as other divalent metal ions, e.g., Cu2+ with high affinity7. These ions are coordinated at the S100A12 dimer interface by amino acids H15 and D25 of one subunit and H85 as well as H89 of the anti-paralleling other subunit8,9,10. While earlier studies propose that Zn2+-loaded S100A12 may induce the protein’s organization into homo-tetramers (44 kDa) and to result in increased Ca2+-affinity11,12, recent metal titration studies6 suggest Ca2+-binding by S100A12 to increase the protein’s affinity to Zn2+. Once the S100A12 EF-hands are fully occupied by Ca2+, additional Ca2+ is thought to bind between dimers, triggering hexamer formation (approximately 63 kDa). The architecture of the hexameric quarternary structure is clearly different from that of the tetramer. It is proposed that the tetramer interface is disrupted to give rise to new dimer-dimer interfaces which benefits hexamer formation10. S100A12 is almost exclusively expressed by human granulocytes where it constitutes about 5% of all cytosolic protein13. In its DAMP function S100A12 was historically described as agonist of the multi-ligand receptor for advanced glycation end-products (RAGE), then termed extracellular newly identified RAGE-binding protein (EN-RAGE)14. Albeit we earlier reported biochemical S100A12-binding to both RAGE and TLR415, we recently demonstrated human monocytes to respond to S100A12 stimulation in a TLR4-dependent manner16. This requires arrangement of S100A12 into its Ca2+/Zn2+-induced hexameric quarternary structure16.
Here we describe a purification procedure for recombinant human S100A12 and its ion-induced oligomers for immune cell stimulations16,17. This is based on a two-step chromatography strategy, which initially includes an anion-exchange column to isolate and concentrate the protein and remove bulk contaminations (e.g., endotoxins/lipopolysaccharides)18. Ion-exchange chromatography resins separate proteins on the basis of different net surface charges. For acidic proteins like S100A12 (isoelectric point of 5.81), a buffer system with a pH of 8.5 and a strong anion-exchange resin leads to a good separation. Bound proteins were eluted with a high-salt buffer gradient. With an increase of ionic strength negative ions in the elution buffer compete with proteins for charges on the surface of the resin. Proteins individually elute depending on their net charge and in result of that, the buffers described herein allow to isolate and concentrate the overexpressed S100A12 protein. Due to negatively charged groups in lipopolysaccharides, these molecules also bind to anion-exchange resins. However, their higher net charge results in later elution in the applied high-salt gradient. The second step of the purification procedure has been introduced for polishing purposes. This makes use of the calcium binding ability of S100A12 and removes remaining impurities on a hydrophobic-interaction column. Calcium binding of S100A12 leads to a conformational change and an exposure of hydrophobic patches on the surface of the protein. On that condition, S100A12 interacts with the hydrophobic surface of the resin. Upon calcium-chelating by EDTA, this interaction is reversed. In the presence of ions, especially calcium and zinc, S100A12 arranges into homomeric oligomers. To study structure-function relationships of the different oligomers, we stabilized dimeric, tetrameric and hexameric recombinant S100A12 with a chemical crosslinker and separated the complexes on a size-exclusion chromatography column. Finally, to analyze functionality and biological activity of the purified protein and its ion-induced oligomers, the cytokine release of S100A12 and LPS stimulated monocyte can be compared.
Various methods for purifying S100A12 have been described so far. Jackson et al.19, for example, published a protocol with purification via an anion-exchange column and a subsequent size-exclusion chromatography. Purification polishing on a size-exclusion column leads to good results, but―due to for example limited loading volumes―is less flexible in scalability. A different approach, published by Kiss et al.20, describes purification of tagged protein via Ni2+ affinity column as the first purification step, followed by enzymatic cleavage to remove the tag and further purification steps. In contrast to the aforecited studies19,20, the produced protein as described in this protocol is determined for experiments on immune cells. Therefore, remnant endotoxin contamination from bacterial culture is a challenge. Although different approaches for endotoxin removal have been described so far, there is no uniform method that works equally well for any given protein solution21,22.
In summary, our protocol combines the advantages of a tag-free expression in a bacterial system with efficient endotoxin removal and high yield of pure protein.
NOTE: Please refer to Supplemental Table 1 for preparation of buffers and stock solutions.
1. Protein expression in E. coli
2. Protein Purification
3. Detection and Removal of Endotoxin
4. Chemical Crosslinking and Oligomer Separation
5. Functional Testing on Monocytes
Following pre-purification on the AIEX column (Figure 1A-C) and subsequent calcium-dependent HIC (Figure 2A,B), highly pure protein was obtained (Figure 2C). In addition, measurements of endotoxin revealed successful LPS removal. The LPS content following AIEX was measured in a 1:10 dilution above the assay detection limit, i.e., above 500 EU/mL. After the first filtration through a 50 kDa filter unit, the LPS content was reduced to 1 EU/mL. Following concentration with a 3 kDa filter unit and additional filtration through 50 kDa, the measured LPS contamination was 0.08 EU/mL.
As an additional control, human monocytes were stimulated with the produced wildtype protein (Figure 3A,B). Polymyxin B treatment abrogates TNFα release from LPS-stimulated monocytes, which cannot be observed with S100A12. On the other hand, heat-treatment of both LPS and S100A12 abolishes the protein’s capacity to stimulate cells, while this does not affect cellular response to LPS-stimulation.
Protein exposure to different ions results in arrangement of different S100A12 oligomers (Figure 4A). Chemical crosslinking allows to capture defined complexes such as dimers, tetramers, and hexamers as well as transition states (e.g., ‘trimers’, band at approximately 30 kDa). In order to induce a pronounced shift of the oligomer-equilibrium prior to crosslinking, an excess of ions was applied (Figure 4B).
Isolated oligomers in equal molar concentrations (Figure 5A-C) were then used for monocyte stimulation to compare signaling abilities via PRRs. Monocyte-stimulation with hexameric S100A12 resulted in pronounced TNFα release (Figure 6). Remnant cytokine release could be detected from cells stimulated with tetrameric S100A12, while treatment with dimeric protein does not induce TNFα release.
Figure 1: Results of anion-exchange chromatography. (A) A chromatogram with absorbance at 280 nm (A280) and the percentage of elution buffer B (dashed line). Methods blocks are indicated with A = wash unbound sample, B = linear gradient with elution buffer (buffer B), C = wash out with buffer B, and D = re-equilibration in buffer A. (B) Focus on the relevant peaks with fraction tube numbers in red. (C) Selected fractions were analyzed on 15% Coomassie-stained SDS-PAGE. Please click here to view a larger version of this figure.
Figure 2: Results of hydrophobic-interaction chromatography. (A) A chromatogram with absorbance at 280 nm (A280) and the percentage of elution buffer B (dashed line). Methods blocks are indicated with A = wash unbound sample, B = elution with buffer B, and C = re-equilibration in buffer A. (B) Focus on the relevant peaks with fraction tube numbers in red. (C) Analyzed fractions on 15% Coomassie-stained SDS-PAGE. Please click here to view a larger version of this figure.
Figure 3: Primary human monocytes were stimulated at indicated concentrations. LPS (A) or S100A12 (wildtype, B) were left untreated or heat-denaturated (99 °C, 10 min). Both conditions were tested in the presence and absence of polymyxin B. Please click here to view a larger version of this figure.
Figure 4: S100A12 protein was crosslinked with BS3 after incubation in HBS buffer containing 5 mM Ca2+ and indicated Zn2+ concentrations. (A) Increasing Zn2+ concentrations induce arrangement of S100A12 into tetramers and hexamers upon separation on 4−20% Coomassie-stained SDS-PAGE. (B) Representative result of crosslinked oligomers with conditions as used for separation on a size-exclusion column. S100A12 was crosslinked in presence of either 25 mM Ca2+ (lane 1) or 25 mM Ca2+ and 1 mM Zn2+ (lane 2). (S100A12)2 = dimer; (S100A12)4 = tetramer; (S100A12)6 = hexamer. Please click here to view a larger version of this figure.
Figure 5: S100A12 oligomers were separated on a size-exclusion column. (A) Chromatogram of hexamer/tetramer separation after crosslinking in HBS buffer with 25 mM CaCl2 and 1 mM ZnCl2. (B) Chromatogram for tetramer/dimer separation in HBS buffer with 25 mM CaCl2. (C) Example of pooled and concentrated oligomers after separation on a Coomassie-stained 4−15% gradient SDS-PAGE. Lane 1 = dimer; lane 2 = tetramer; lane 3 = hexamer. Please click here to view a larger version of this figure.
Figure 6: Stimulation of monocytes with purified S100A12 oligomers. TNFα-release after 4 h incubation was quantified by ELISA. The data show the mean value from two independent experiments. Please click here to view a larger version of this figure.
Bed Volume (CV) | 75 mL |
Monitor | Absorbance at 280 nM |
Pressure Max | 3 bar |
Column buffer A | 20 mM Tris-HCl, 1 mM EDTA, 1 mM EGTA, pH 8.5 |
Column buffer B | 20 mM Tris-HCl, 1 mM EDTA, 1 mM EGTA, 1 M NaCl, pH 8.5 |
Sample Volume | variable |
Flow Rate | 1−2 mL/min |
Temperature | 4 °C |
Table 1: Detailed information on the applied parameters of anion-exchange chromatography.
Block | Volume | Buffer | Outlet |
Equilibration | 1−2 column volumes (CVs) | A | Waste |
Sample load | n/a | A | Waste |
Wash out unbound sample | 1 CV | A | High volume outlet |
Gradient–Elution | 0−100 % Buffer B in 1 CV | A to B | Fraction collector |
Wash out–Buffer B | 1 CV | B | Waste |
Re-Equilibration | 2 CVs | A | Waste |
Table 2: Detailed information on the used method of anion-exchange chromatography.
Bed Volume (CV) | 125 mL |
Monitor | Absorbance at 280 nM |
Pressure Max | 4 bar |
Column buffer A | 20 mM Tris, 140 mM NaCl, 25 mM CaCl2, pH 7.5 |
Column buffer B | 20 mM Tris, 140 mM NaCl, 50 mM EDTA, pH 7.5 |
Sample Volume | variable |
Flow Rate | 1−2 mL/min |
Temperature | 4 °C |
Table 3: Detailed information on the applied parameters of hydrophobic-interaction chromatography.
Block | Volume | Buffer | Outlet |
Equilibration | 1−2 column volumes (CVs) | A | Waste |
Sample load | n/a | A | Waste |
Wash out unbound sample | 1−2 CVs | A | High volume outlet |
Elution | 100 % Buffer B | B | Fraction collector |
Wash out–Buffer B | 1 CV | B | Waste |
Re-Equilibration | 2 CVs | A | Waste |
Table 4: Detailed information on the used method of hydrophobic-interaction chromatography.
Bed Volume (CV) | 320 mL |
Monitor | Absorbance at 280 nm |
Pressure Max | 3 bar |
Column buffer A | 20 mM Hepes, 140 mM NaCl, pH 7.2 |
Sample Volume | Up to 13 mL |
Flow Rate | 1−1.5 mL/min |
Temperature | 12−15 °C |
Table 5: Detailed information for the applied parameters of size-exclusion chromatography.
Supplemental Table 1: Preparation of buffers and stock solutions. Please click here to download this file.
In this protocol, we describe tag-free bacterial expression of human S100A12 and its purification as well as separation into different ion-induced oligomers for immune cell stimulation. Compared to published literature on S100A12 protein purification8,23,24, the use of high CaCl2 (25 mM) in hydrophobic-interaction chromatography is to our knowledge unique. Several protocols applying concentrations from 1 to 5 mM do produce pure protein, yet we observed a several times higher yield following our approach using 25 mM CaCl2 instead. This might be explained by a hierarchy of protein interaction with the column material:S100A12 can directly bind to the column material but the excess of Ca2+ may also facilitate indirect binding of S100A12-dimers to the already column-bound protein8. Thus, high Ca2+ concentrations may enlarge the surface available for S100A12 purification. Elution (by using a linear gradient) of S100A12 from HIC as one early (indirectly bound S100A12) and one very late peak (column material bound protein) may support this speculation (data not shown).
For production of recombinant S100A12 (as well as other proteins) at high yields and manageable costs, protein expression in E. coli is still the method of choice. However, the inevitable contamination with bacterial endotoxins remains a problem, when proteins are determined for cell culture experiments, particularly in studies involving innate immune cells. To our experience, even commercially available proteins explicitly declared for cell culture use can contain endotoxin contaminations up to 1 EU/µg protein, which can significantly skew assays. Therefore, a complete removal of endotoxins is mandatory. Endotoxin monomers in solution range from molecular weights of 10 to 20 kDa, but they can form micelles and structures with higher molecular weights. The formation of very large structures is, for example, promoted through bivalent ions21,25.
According to our protocol, we verify the endotoxin-free production of S100A12 protein by combining high-sensitivity endotoxin measurements with monocyte stimulation assays. We consider such combination particularly meaningful as a) low-level endotoxin contamination may be difficult to assess depending on the sensitivity of the assay and b) the use of polymyxin B as LPS inhibitor on monocytes may result in difficult to interpret data due to exclusive polymyxin B effects on cells26,27. Polymyxin B as well as other cationic peptides are reported to bind LPS via negatively charged lipid A28. As the solvent exposed surface of S100A12 also contains large negatively charged patches the observed reduction of TNFα-release from S100A12-stimulated human monocytes in presence of polymyxin B (Figure 3B) may be due to a) unspecific direct binding of polymyxin B to S100A12 and/or b) direct effects of polymyxin B on stimulated cells26,27. Due to the known limitations of both the detection of low-level endotoxin contamination as well as unspecific polymyxin B effects, our protocol further contains a heat-inactivation step to clearly distinguish between LPS- and protein-mediated TLR4-signaling.
Use of LPS-free S100A12 for generation and purification of defined ion-induced oligomers is critical and extra attention should be paid to their subsequent purification to avoid eventual re-introduction of endotoxin via buffers or column material and thus further protein-demanding LPS-depletion via endotoxin removal resins.
The relevance of oligomerization for the biological function of proteins can be assessed by different means. In case of S100A12, we used surface plasmon resonance as well as targeted amino acid exchanges at ion-binding sites and―to most precisely define the protein-complex able to bind and signal through TLR4―we employed chemical crosslinking of Ca2+/Zn2+-pulsed recombinant S100A1216. Chemical crosslinking of S100A12 under different ionic conditions snap-freezes a momentary state including several oligomeric forms that are in transition. From ion titration experiments, we defined conditions under which dimeric, tetrameric or hexameric oligomers could be determined as the predominant oligomers16. In addition, previous experiments have shown that an excess of ions is beneficial for comparable, stable crosslinking and subsequent purification, although oligomerization can also be induced at significantly lower ion concentrations. However, purifying these oligomers by size-exclusion chromatography results in good, but not absolute separation. Still, the selective enrichment of oligomers allows for reliable downstream analyses.
In summary, this protocol provides a method for purification of LPS-free human S100A12 or related calcium binding proteins. To fix ion-induced conformational changes, chemical crosslinking and subsequent complex separation by size-exclusion chromatography is a useful tool to understand the relevance of protein oligomerization for downstream biological processes.
The authors have nothing to disclose.
This study was supported by grants from the intramural innovative medical research program of Muenster University medical faculty (KE121201 to C.K.) and the German Research Foundation (DFG, Fo354/3-1 to D.F.).
pET11b vector | Novagen | ||
BL21(DE3) competent E. coli | New England Biolabs | C2527 | |
100 x Non-essential amino acids | Merck | K 0293 | |
25% HCl | Carl Roth | X897.1 | |
4−20% Mini-PROTEAN TGX Protein Gels | BioRad | 4561093 | |
Ampicillin sodium salt | Carl Roth | HP62.1 | |
BS3 (bis(sulfosuccinimidyl)suberate) – 50 mg | ThermoFisher Scientific | 21580 | |
Calciumchlorid Dihydrat | Carl Roth | 5239.1 | |
Coomassie Briliant Blue R250 Destaining Solution | BioRad | 1610438 | |
Coomassie Briliant Blue R250 Staining Solution | BioRad | 1610436 | |
EasySep Human Monocyte Enrichment Kit | Stemcell | 19059 | Magnetic negative cell isolation kit |
EDTA disodium salt dihydrate | Carl Roth | 8043.1 | |
EGTA | Carl Roth | 3054.3 | |
EndoLISA | Hyglos | 609033 | Endotoxin detection assay |
Endotoxin-Free Ultra Pure Water | Sigma-Aldrich | TMS-011-A | Ultrapure water for preparation of endotoxin-free buffers |
EndoTrap red | Hyglos | 321063 | Endotoxin removal resin |
FBS (heat-inactivated) | Gibco | 10270 | |
HBSS, no calcium, no magnesium | ThermoFisher Scientific | 14175053 | |
Hepes | Carl Roth | 9105 | |
Hepes (high quality, endotoxin testet) | Sigma-Aldrich | H4034 | |
hTNF-alpha – OptEia ELISA Set | BD | 555212 | |
IPTG (isopropyl-ß-D-thiogalactopyranosid) | Carl Roth | CN08.1 | |
L-Glutamine (200 mM) | Merck | K 0282 | |
LB-Medium | Carl Roth | X968.1 | |
Lipopolysaccharides from E. coli O55:B5 | Merck | L6529 | |
Pancoll, human | PAN Biotech | P04-60500 | Separation solution (density gradient centrifugation) |
Penicillin/Streptomycin (10.000 U/ml) | Merck | A 2212 | |
Phenyl Sepharose High Performance | GE Healthcare | 17-1082-01 | Resin for hydrophobic interaction chromatography |
Polymyxin B | Invivogen | tlrl-pmb | |
Protease inhibitor tablets | Roche | 11873580001 | |
Q Sepharose Fast Flow | GE Healthcare | 17-0510-01 | Resin for anion-exchange chromatography |
RoboSep buffer | Stemcell | 20104 | Cell separation buffer (section 5.1.4) |
RPMI 1640 Medium | Merck | F 1215 | |
Sodium chloride (NaCl) | Carl Roth | 3957.2 | |
Sodium hydroxide | Carl Roth | P031.1 | |
Tris Base | Carl Roth | 4855.3 | |
Zinc chloride | Carl Roth | T887 | |
Labware | |||
0,45 µm syringe filter | Merck | SLHA033SS | |
14 mL roundbottom tubes | BD | 352059 | |
2 L Erlenmyer flask | Carl Roth | LY98.1 | |
24 well suspension plates | Greiner | 662102 | |
5 L measuring beaker | Carl Roth | CKN3.1 | |
50 mL conical centrifuge tubes | Corning | 430829 | |
50 mL high-speed centrifuge tubes | Eppendorf | 3,01,22,178 | |
Amicon Ultra-15 Centrifugal Filter Unit MWCO 3 kDa | Merck | UFC900324 | |
Amicon Ultra-15 Centrifugal Filter Unit MWCO 50 kDa | Merck | UFC905024 | |
Culture dish (100 mm) | Sarstedt | 83.3902 | |
Dialysis Tubing Closures | Spectrum | 132738 | |
EasySep magnet 'The Big Easy` | Stemcell | 18001 | |
Fraction collector tubes 5 mL | Greiner | 115101 | |
Lumox film, 25 µm, 305 mm x 40 m | Sarstedt | 94,60,77,316 | Film for monocyte culture plates |
Spectra/Por Dialysis Membrane (3.5 kDa) | Spectrum | 132724 | |
Steritop filter unit | Merck | SCGPT01RE | |
Equipment | |||
37 °C Incubator (with shaking) | New Brunswick Scientific | Innova 42 | |
ÄKTA purifier UPC 10 | GE Healthcare | FPLC System | |
Fraction collector | GE Healthcare | Frac-920 | |
Centrifuge (with rotor A-4-81) | Eppendorf | 5810R | |
Fixed angle rotor | Eppendorf | F-34-6-38 | |
Mini Protean Tetra Cell | BioRad | 1658000EDU | |
NanoPhotometer | Implen | P330 | |
Sonicator | Brandelin | UW2070 | |
Fluorescence reader | Tecan | infinite M200PRO | |
pH meter | Knick | 765 |