Team:Freiburg bioware/Project/purification




In order to produce and study our different protein constructs they had to be expressed in bacteria.  After cloning the individual parts into the pMA vector the complete expression products were then transferred into the pEx vector (see pEx vector ) and transformed into competent E. coli expression strains via heat shock. We used two different strains for the protein expression: E. coli BL21 de3 (Novagen) and E. coli RV308 (Maurer et al., JMB 1980). Both strains are suited for high-level protein expression.

The expression level of proteins can vary in many parameters, so the conditions often need to be optimized after the first try. Some of these expression parameters are: host strain, temperature, time of induction, inducer concentration, harvest time, etc.


After the successful expression of a construct, the protein of interest had to be purified from the cell lysate created using a sonifier or a french press. There are different methods to do this, one of the easiest is to fuse an affinity tag to the protein of interest. We used three different tags: 6x Histidine-tag , Streptavidine tag and a GST-tag in different approaches. To obtain the tagged proteins a base sequence coding for the information for the tag amino acid sequences has to be inserted at the beginning or at the end of the construct. A specific and commercially available binding partner exists for each tag and by exploiting this interaction, the protein of interest can be purified. Hence in most cases, the protein purification starts with the loading of the cell lysate onto a column, which contains a matrix coupled to one specific binding partner, such as Ni-NTA (Qiagen Ni-NTA Superflow) for the His-Tag, StrepTactin (IBA Strep-Tactin Superflow high capacity) for the Strep-tag and Glutathione (GE Glutathione Sepharose High Performance) for the GST-tag. The tagged proteins bind specifically to their binding partner and whereas the other, non-tagged proteins are washed away. The protein of interest can then be eluted with a solution containing a competitive molecule.

Figure: protein purification scheme

Problems can occur if the protein of interest is too fragile or toxic for the cells. To avoid killing the cells with a toxic protein we used the method of exporting the protein into the periplasm of E. coli by adding a signal sequence (of the periplasmic protein DsbA) at the beginning of the target protein so that it will be cotranslationally exported and can’t cause any damage in the cytoplasm. The signal sequence itself gets split off by a protease in the periplasm, hence interference with the protein is avoided.

For running the column we used the AKTA 900 FPLC purifier system from Amersham Pharmaceuticals. The fluorescence pictures were taken with a Zeiss  AxioVision MTB2004 fluorescence microscope with a COLIBRI LED system.

Left picture: Laura preparing the ÄKTA purifier; right picture: Laura and Hannes using the Zeiss AxioVision fluorescence microscope.

Results and Discussion

Our first proteins chosen for expression were the argonaute proteins (Ago) from  Aquifex aeolicus and Thermus thermophilus. The Ago protein was tagged with a polyhistidine tag so it could be purified by using a Ni-NTA column. We transformed the vector with the Ago sequence into both E. coli BL21de3 and E. coli T7 Express (NEB) which is an enhanced derivative of the BL21 strain. SDS gel electrophoresis showed that both strains expressed the Ago protein at a high level. The cells were harvested after 8 hours of growth at 20°C. The cell pellets were resuspended in washing buffer and a proteinase inhibitor cocktail was added before the cell suspension was sonicated and filtered. Then the filtered cell lysate was loaded on the Ni-NTA column, washed and eluted with a buffer containing 250mM imidazole. The elution fractions were subsequently mixed with Laemmli buffer and loaded on a SDS gel and stained with Coomassie.

SDS gel ; argonaute protein from Aquifex aeolicus; lanes: Marker (NEB prestained protein marker broad range), elution fraction 1, elution fraction 2, elution fraction 3, elution fraction 4, elution fraction 5, elution fraction 6, elution fraction 7, elution fraction 8

The size of the Ago protein is about 83 kDa and distinct bands can be seen. Fractions 1-4 were pooled and further purified by size exclusion chromatography with a GE HiPrep 16/60 Sephacryl S-200 HR column.

SDS gel ; argonaute protein from Aquifex aeolicus after Ni-NTA column and size exclusion chromatography; lanes: Marker (NEB prestained protein marker broad range), fraction 1 (after size exclusion), fraction 2, fraction 3, fraction 4, fraction 5, fraction 6, fraction 7, fraction 8, fraction 22

In lane 10 a distinct band is visible and the amount of unwanted proteins is much lower.


Our second project was the purification of a Fok construct. We started with the expression of an inactive Fok construct (His-Flu_A-SplitLi-Fok_i in the pEx vector).  The first Fok_i construct was expressed in BL21 de3 at 24°C. The cells were harvested after 4 hours of growth.

SDS gel; pEx His-Flu_A-SplitLi-Fok_i in BL21de3 ; Lanes: Marker (NEB prestained protein marker broad range), elution fraction 1, elution fraction 2, elution fraction 3, elution fraction 4, elution fraction 5, flow through fraction 1, flow through fraction 3, washing fraction 1 (250mM imidazole), washing fraction 3

The size of the Fok_i construct was about 45 kDa and  distinct bands can be seen in the first lane, especially in lane 2.

The results of this expression were also verified with a Western Blot using antibodies directed against the terminal His-tag and three positive controls were included.

Western Blot; pEx His-Flu_A-SplitLi-Fok_i in BL21de3; Lanes: 1. Marker (NEB prestained protein marker broad range), 2. sample of pool (fractions 1-4), 3. empty lane, 4.-6. 3 positive controls

The hardest step was the expression of a construct containing the active Fok protein (His-Dig-Split-Fok_a in the pEx vector). After several failed attempts to express the Fok_a protein and purifying the protein of interest out of the cell lysate with a Ni-NTA column we tried to tag the protein with a Strep tag (Strep-DigA-Split-Fok_a). We made several attempts with variations in the growth temperature, the time of induction (at higher OD) and the amount of IPTG. However, in all cases the cells stopped to grow about 2 hours after induction. So in the next attempt we added a signal sequence for periplasmic export to the protein (DsbA-His-DigA-Split-Fok_a). By exporting the protein into the periplasm its toxicity can be reduced and the purity of the protein gained by export in the periplasm is usually higher as well. We performed several attempts to purify the Fok_a construct in this setup but did not obtain detectable amounts of protein. Our next attempts were to add a His-tag in addition to the Strep-tag to purify even small amounts of the Fok_a construct because the expression level remained low and the protein still proved toxic for the cells. A second attempt for the purification of small amounts of an active Fok fusion protein was to use a Glutathione column which was loaded with a GST-tagged FosW, which is a specific Fos-binding helix selected by phage display. We then created a construct containing a Fos linked to an active Fok (His-Fos-SplitLi-Fok_a), so that  the GST-linked FosW binds to the Glutathione immobilized  on the column can be bound by FosW.

Since several approaches to detect and to purify active Fok proteins failed, we also added a fluorescent protein (Venus) to the protein construct. We detected a strong fluorescence under the fluorescence microscope after induction of the cells. The protein construct can be purified using a GFP-trap column.

Picture taken with a fluorescence microscope using a YFP filter; Fok_a_Venus expressed in BL21 cells

The Venus protein was fusioned downstream to the protein construct, so the protein construct is expressed in any case when a fluorescence signal of the Venus protein can be seen.

SDS gel; pEx Venus-Fok_a in BL21de3; lanes: 1. Marker (ColorPlus Prestained Protein Marker, Broad Range), 2. clone 1 uninduced, 3. clone 1 induced, 4. clone 1 induced, 5. clone 2 uninduced, 6. clone 2 induced, 7. control (His-FluA-SplitLi-Fok_i + His-DigA-SplitLi-Fok_a) uninduced, 8. control (His-FluA-SplitLi-Fok_i + His-DigA-SplitLi-Fok_a) induced

In the lanes of the induced clones ( lanes 3,4) appears a band at about 50 kDA which could be the Venus-Fok_a construct.

To detect whether the protein construct with the linked GFP is completely expressed or gets degraded we did a western blot with specific anti-GFP antibodies (first antibody: Santa Cruz Biotechnology sc-9996 mouse αGFP; second antibody: Santa Cruz Biotechnology sc-2005 goat αmouse IgG-HRP)

Western Blot; pEx Venus-Fok_a in BL21de3; lanes: 1. Marker (NEB prestained protein marker broad range, 2. clone 1 uninduced, 3. clone 1 induced, clone 1 induced, clone 2 uninduced, clone 2 induced, control (His-FluA-SplitLi-Fok_i + His-DigA-SplitLi-Fok_a) uninduced, control (His-FluA-SplitLi-Fok_i + His-DigA-SplitLi-Fok_a)  induced

All of the induced clones show a strong signal but show many different bands on the membrane. There’s a signal at 49 kDa, which is the size of the Venus-Fok_a complex. The bands indicating proteins of smaller size could be truncated proteins.


Due to the lack of time at the end our assays to purify Fok_a proteins are still under progress and we keep trying to reach successful expression and purification.