Use of anionic denaturing detergents to purify insoluble proteins after overexpression
© Schlager et al.; licensee BioMed Central Ltd. 2012
Received: 17 October 2012
Accepted: 4 December 2012
Published: 11 December 2012
Skip to main content
© Schlager et al.; licensee BioMed Central Ltd. 2012
Received: 17 October 2012
Accepted: 4 December 2012
Published: 11 December 2012
Many proteins form insoluble protein aggregates, called “inclusion bodies”, when overexpressed in E. coli. This is the biggest obstacle in biotechnology. Ever since the reversible denaturation of proteins by chaotropic agents such as urea or guanidinium hydrochloride had been shown, these compounds were predominantly used to dissolve inclusion bodies. Other denaturants exist but have received much less attention in protein purification. While the anionic, denaturing detergent sodiumdodecylsulphate (SDS) is used extensively in analytical SDS-PAGE, it has rarely been used in preparative purification.
Here we present a simple and versatile method to purify insoluble, hexahistidine-tagged proteins under denaturing conditions. It is based on dissolution of overexpressing bacterial cells in a buffer containing sodiumdodecylsulfate (SDS) and whole-lysate denaturation of proteins. The excess of detergent is removed by cooling and centrifugation prior to affinity purification. Host- and overexpressed proteins do not co-precipitate with SDS and the residual concentration of detergent is compatible with affinity purification on Ni/NTA resin. We show that SDS can be replaced with another ionic detergent, Sarkosyl, during purification. Key advantages over denaturing purification in urea or guanidinium are speed, ease of use, low cost of denaturant and the compatibility of buffers with automated FPLC.
Ionic, denaturing detergents are useful in breaking the solubility barrier, a major obstacle in biotechnology. The method we present yields detergent-denatured protein. Methods to refold proteins from a detergent denatured state are known and therefore we propose that the procedure presented herein will be of general application in biotechnology.
The purification of natively folded protein from heterologous expression systems is cumbersome and in some cases impossible because of the formation of insoluble protein aggregates called inclusion bodies . Factorial screens of expression conditions or refolding assays can yield soluble proteins in some cases and can increase yield in most cases but many proteins remain resistant . Affinity purification under denaturing conditions followed by renaturation can yield natively folded protein and is a viable alternative.
Anfinsen first demonstrated the reversible denaturation of proteins in solutions of urea in the 1960ies . It has since been the method of choice to denature native proteins and also aberrant protein aggregates. The denaturation of proteins by urea yields protein in a random coil state, i.e. no secondary structure elements are favoured over any other conformation . The interactions of SDS with proteins have been intensively studied since the 1940ies [5–8]. Denaturation of polypeptides by SDS is a multi-step process that starts with the interaction of the negatively charged sulphate group with oppositely charged, basic, amino acid side chains . The hydrophobic tails of SDS molecules then become buried in the hydrophobic core of proteins and start to disrupt the native structure [10–12]. Finally a large fraction of the polypeptide chain, independent of its conformation in the native state, adopts an alpha-helical conformation and is surrounded by a micelle of SDS molecules [9, 13]. The length of this mixed protein-detergent micelle is roughly proportional to the molecular weight of the polypeptide .
Thus the denaturation of polypeptides by urea and SDS are different at the mechanistic level and yield two different results: a random-coil structure in urea and a largely alpha-helical conformation in SDS [4, 6].
Hexahistidine-tagged proteins can be purified under denaturing conditions using chaotropic concentrations of urea or guanidinium hydrochloride . The high-affinity binding of the hexahistidine tag to Ni/NTA resin is based on the co-operative co-ordination of a nickel cation by two histidine side chains . It is independent of the peptide conformation but requires close physical proximity (reviewed in ).
We wondered whether the histidine side chains of an SDS denatured protein, which are buried inside a mixed detegent-protein micelle, would be accessible for binding to Ni/NTA and whether thus detergent denatured proteins could be purfied by immobilized metal ion affinity chromatography (IMAC).
Phusion Polymerase was from New England Biolabs (Ipswich, USA). Cloning vector pET151/D-TOPO and BL21 cells were from Life Technologies (Zug, Switzerland). All chemicals used in this study were laboratory-grade. IPTG and SDS were from Carl Roth GmBH (Karlsruhe, Germany). DTT, imidazole and carbenicillin were from Applichem (Darmstadt, Germany). N-Lauroylsarkosine (Sarkosyl) was from Sigma-Aldrich (Steinheim, Germany) and NHS-activated Sepharose was from GE Healthcare Biosciences (Uppsala, Sweden). Purification was done on 5 ml HisTrap columns from GE Healthcare Biosciences (Uppsala, Sweden). Precast SDS-PAGE gels were from LucernaChem (Luzern, Switzerland) and were stained with Coomassie Brilliant Blue as described elsewhere .
Expression constructs and details of inserts
Insert Protein Domain
Flybase Protein ID and AA range
MW of fusion protein in kDa
Tor kinase catalytic
TSC2 DUF 3384
foxo winged helix
Tor kinase DUF 3385
Pi3K21B N-term SH2
Expression constructs were heat-shock transformed ino BL 21 (DE3) STAR cells, plated on LB plates containing carbenicillin (50 μg/ml) and incubated o/n at 37°C. The next day several colonies were collected and used to inoculate 300 ml cultures of LB medium containing carbenicillin. The cultures were grown at 37°C until the OD 600 was 0.5 and then induced by addition of IPTG to a final concentration of 0.5 mM. Cultures were then grown over night at 30°C.
PCL (lysis buffer) contained 8 mM Na2HPO4, 286 mM NaCl, 1.4 mM KH2PO4, 2.6 mM KCl and 1% SDS (w/v) at pH 7.4. PCW (wash and equilibration buffer) contained 8 mM Na2HPO4, 286 mM NaCl, 1.4 mM KH2PO4, 2.6 mM KCl and 0.1% Sarkosyl (w/v) at pH 7.4. PCE (elution buffer) contained 8 mM Na2HPO4, 286 mM NaCl, 1.4 mM KH2PO4, 2.6 mM KCl, 500 mM imidazole and 0.1% Sarkosyl (w/v) at pH 7.4.
The cultures were harvested in GS3 rotor tubes by centrifugation at 4°C for 12 minutes at 6000 rpm. The pellet was resuspended in 30 ml PCL buffer, supplemented with DTT to 1 mM, and sonicated with a Bandelin Sonoplus HD2070 sonicator, set to 80% cycle and 40% power using a MS73 probe-tip (Bandelin electronic, Berlin, Germany). Samples were sonicated at room temperature twice for 2 minutes each. The lysates were transferred to SS34 tubes and placed in an ice-water mixture and incubated for 30 minutes. The chilled lysates were then centrifuged in a SS34 rotor at 13 krpm for 20 minutes at 4°C. The cleared supernatant was poured off and filtered through a 0.45 μm syringe filter before applying it to affinity purification. Ni/NTA affinity purification was performed on an AKTA Xpress FPLC system using 5 ml HisTrap HP columns and standard purification templates (GE Healthcare Biosciences Uppsala, Sweden). Columns were equilibrated with PCW buffer, the lysate loaded and the columns washed until the absorption of post-column flowthrough returned to base levels. Bound proteins were eluted with a 100 ml linear gradient of buffers PCL and PCE, from 0 to 50% buffer PCE, i.e. from 0 to 250 mM imidazole in 20 column volumes. Weakly bound contaminating proteins typically eluted at 40 mM imidazole, the peak maximum of hexahistidine tagged proteins was between 80 and 150 mM imidazole.
Figures 1 and 2 present the purification of seventeen different fusion proteins of sizes between 13 and 54 kDa, detailed in Table 1. These proteins were previously shown to be insoluble after lysis of induced cells in standard native lysis buffers (PBS containing 0.5% Triton X-100 or 0.3% Sarkosyl, data not shown). They are domains of proteins involved in the Drosophila Insulin and Tor Kinase signaling network and represent various non-homologous structures . The purification of one of these protein domains, the REM domain of rictor, encoded by construct 173, is shown in detail in Figure 1. For the other sixteen proteins only the final, purified protein, is shown in Figure 2.
Summary of the purification procedure
Growth, induction and harvest of induced cells
Overexpression of target protein in E. coli
As described elsewhere 
Resuspension in 1% SDS in PBS
Lysis of cells
Works fine in 1/20 to 1/10 of the original culture volume. Other buffers also work (50 mM Tris/Cl pH 7.0, 100 mM NaCl and 1% SDS)
Solubilisation of proteins from inclusion bodies
Until solution turns clear. Often faster than 2 minutes.
Incubation on ice for 30 minutes
Precipitation of SDS
Precipitation of SDS apparent after 5 minutes. 1 h incubations on ice possible.
Removal of precipitated SDS
SS34 rotor, 13 krpm, 20 minutes, 4°C
Ni/NTA affinity purification
Capture and washing of hexahistidine tagged target protein
See M&M for buffer compositions and details.
Elution in 0.1% Sarkosyl
Elution in a dialyzable detergent that is compatible with refolding 
Other detergents could be used but were not tested.
We wanted to test whether hexahistidine-tagged proteins can be IMAC-purified after SDS denaturation. In our first experiments we included 34 mM (1% w/v) of SDS in all buffers, i.e. in the lysis, wash and elution buffers commonly used for affinity purification. Cell pellets were rapidly lysed by sonication and binding to Ni/NTA resin occurred, however most of the hexahistidine-fusion protein was found in the flow-through, indicating poor binding (data not shown). Upon searching the literature we realized that there was no primary literature concerning the binding of hexahistidine-tagged proteins to Ni/NTA in solutions containing SDS. One manufacturer of Ni/NTA resins publishes a handbook that indicates that no more than 0.3% SDS should be included in the buffers  but these are empirical values that are not based on published experiments (QIAGEN AG, Basel Switzerland, pers. comm.).
In 1988 Suzuki and Terada published that SDS can be selectively removed from solutions containing BSA by cooling . We hypothesized that the same principle might be applied to whole cell lysates and that it should yield a solution with lower concentrations of SDS that would be compatible with high affinity binding to Ni/NTA. We modified the buffer compositions such that only the buffer used for initial lysis contained high concentrations of SDS (34 mM, 1% w/v) and included a low concentration of Sarkosyl (3 mM, 0.1% w/v) in the wash and elution buffers. After lysis by sonication the lysate was cooled in an ice/water bath for 20 minutes and precipitated SDS removed by centrifugation. The residual concentration of SDS was compatible with high affinity binding to NI/NTA. We applied this purification protocol to seventeen different proteins (detailed in Results) and consistently achieved high-affinity binding and elution profiles that indicated sensitivity to imidazole concentration, i.e. contaminating proteins eluted at lower imidazole concentrations than the tagged target protein.
We would like to mention that we did not specifically test whether the dodecylsulphate anion (DS) of SDS precipitates as the sodium salt (SDS) or with another monovalent cation such as potassium (KDS), which is also present in the buffer .
Detergents display complex phase diagrams in aqueous solutions . Many detergents separate into a distinct phase when temperature or salt concentrations change. A landmark paper showed that integral membrane proteins remain in the detergent phase after temperature induced phase separation of Triton X-114 and that therefore phase separation can be used as a tool in protein purification . Nowadays phase separation of detergents, sometimes called “cloud point extraction”, is frequently used to purify membrane proteins .
We would like to point out that our method is similar to cloud point extraction only in as far as the detergent is used to initially solubilize the proteins of interest, in our case from inclusion bodies, and that phase separation is induced experimentally. However, it critically differs from cloud point extraction in that the proteins do not co-partition into the detergent phase after phase separation. Instead the surplus of unbound detergent is removed from the solution to allow subsequent affinity purification.
Curiously as early as in 1944 it was shown that SDS can be selectively removed from solutions containing proteins by cold precipitation with barium chloride . Suzuki and Terada showed that SDS can be removed from solutions containing BSA by cooling , however, we are not aware of a publication that would combine this principle with the dissolution and denaturation of inclusion bodies and subsequent IMAC.
The buffer components used in this method are compatible with automated chromatography and allow high throughput purification of target proteins on a suitable purification platform. One key advantage over purification in urea or guanidinium is that SDS can be used at comparably low concentrations (34 mM versus 6–8 M) and does not tend to crystallize in valves and pumps of FPLC chromatographs .
Finally, we wish to add that protocols for the refolding of proteins from a detergent denatured state are known. One relies on replacing SDS with urea and subsequent removal of urea . The other is based on cyclodextrine mediated stripping of detergent molecules from the protein . We therefore think our method can be of application in the purification and refolding of recalcitrant proteins.
Ionic, denaturing detergents are useful reagents in the solubilization and purification of proteins from inclusion bodies and can be used to replace the more commonly used reagent urea.
We thank Dirk Linke and Ari Helenius for discussions and pointing out important references. B.S. was supported by an EMBO Long Term Post Doctoral Fellowship.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.