Results: Cell-free Expression

We use cell-free expression in the DiaCHIP for the direct production of protein microarrays from a DNA microarray template. This process can be refered to as a molecular copy. To enable antibody detection with this protein microarray, the layer of antigens on the surface have to be as dense as possible. Thus, the expression efficiency of the cell-free system has to be optimized in order to produce a sufficient amount of target proteins within a timespan that is reasonable for preparation of the DiaCHIP in the suggested applications.

One of the key factors influencing expression efficiency is the design of the expression vector. Therefore, we started with the design of a vector based on pSB1C3 that would allow successful cell-free expression. Additionally, we received an expression vector containing a GFP coding sequence, an HA and a double His-tag kindly provided by AG Roth, as an external standard that had already performed well in a cell-free expression system. The third vector we used for our experiments was the pBESTluc^TM encoding a luciferase for performing the luciferase assay.

However, the most important thing of course is the expression system itself.
We established our own cell-free expression system, the "DiaMIX" based on an E. coli lysate. Additionally we obtained a commercially available expression kit.
We also provide the protocol in order to give future iGEM Teams the possibility to produce their cell-free expression mix in a low-budget version themselves.

First Successful Expression

Figure 1: Cell-free expression of GFP. Fluorescing GFP could be detected via fluorescence microscopy after cell-free expression (A). Picture taken of the negative control (B). For detailed reaction performance see our labjournal.

Before using our DiaMIX in the microfluidic chamber, we performed initial experiments in vitro in tubes and, at a later time, in well-plates. In the beginning, we were not able to express anything. This was due to several factors, like the pH value of our mix. It has to be adjusted in the narrow physiological range (between pH 7 and 8). Measurement of pH is difficult as the reaction volume is in the range of 15 µl and we did not want to waste it on pH-paper. Furthermore, the pH value is crucial for the solubility of chemicals for the premix. For pH setting, HEPES-KOH and L-Glutamic acid were used. After adjustment, further parameters were adapted and the source of errors narrowed down. Being the prime error sources, the preparations of both lysate and premix were repeated. The gained experience took care of many errors.

In the end, the reaction premix itself proved to be our greatest source for errors and was prepared once more. Following this, we could record some basal expression indicating that we could now further optimize our system.

Variation of magnesium concentration

Figure 2: Impact of Mg(OAc)₂ addition during cell-free expression of Luciferase. The regular feeding every 20 min with 100mM MgOAc₂ shows an increase of expression to half of that of the commercial kit.

At first, we investigated the influence of magnesium acetate (Mg(OAc)₂) on our reaction. Kim et al.¹⁾ describe an enhancement of cell-free expression by regular addition of small amounts of Mg. We performed expression experiments where reactions were fed with Mg(OAc)₂ while others were not. The repeated addition of 100 nmol of magnesium every 20 minutes during the reaction was found to increase expression levels and was therefore implemented for future experiments. Feeding was performed by pipetting 0.5 μl of a 10 μM magnesium acetate solution into the reaction tube without letting it cool down. Positive effects towards expression were immediately visible in a subsequently conducted luciferase assay (see figure 2). However, the actual impact of feeding varied a lot in the experiments, even among reactions with the same components.

Figure 3: Effect of different Mg(OAc)₂ concentrations (12-16 mM) on cell-free expression. Cell-free expression of pBESTluc^TM and subsequent validation via luciferase assay. Experiment was performed in triplicates.

Supplying a reaction lasting for at least two hours with additional chemicals every 20 minutes is quite uncomfortable, so we varied the initial concentration of Mg(OAc)₂ in the reaction mix. From personal discussions with a leading scientist in the field of cell-free expression, we were adviced to adjust our mix to a concentration of 15-18 mM Mg²⁺. Therefore, we expressed pBESTluc^TM with our DiaMIX in 50 µl reactions at concentrations between 12-16 mM to compare the expression via luciferase assay. The result is shown in figure 3 and reveals the optimal initial concentration of Mg(OAc)₂ to be at 12 mM.

Variation of DNA Concentration

Figure 4: Effects of DNA concentration on cell-free expression. Results were obtained for concentrations of 0.02 µg/µl, 0.04 µg/µl, 0.06 µg/µl, 0.08 µg/µl, 0.1 µg/µl and a negative control without DNA. Validation was performed via luciferase assay and analyzed in a microplate reader.

After establishing the ideal concentration of Mg(OAc)₂ in our reactions, we further investigated the optimal concentration of DNA template for increased expression. Again, pBESTluc^TM was expressed to allow the analysis via luciferase assay in a microplate reader. Triplicates for reactions of 50 µl volumes were set up with amounts of DNA ranging from 0.02 µg to 0.1 µg, as well as a negative control without DNA. A luciferase assay was performed as described in protocols (figure 4+5).

Figure 5: Effects of DNA concentration on cell-free expression (Bar graph of emission maxima). Results for concentrations of 0.02 µg/µl, 0.04 µg/µl, 0.06 µg/µl, 0.08 µg/µl, 0.1 µg/µl and the negative control (water). Validation was performed via luciferase assay and analyzed in a microplate reader.

The optimal amount of DNA established for further experiments was 2 - 3 µg. To make subsequent reactions more cost-efficient, a concentration of 0.04 µg/µl of DNA was used.

Comparing the DiaMIX with a Commercial Kit

Figure 6: Cell-free GFP expression over time. Comparison of the commercial kit and the DiaMIX was performed by expressing the HA-GFP-His vector kindly provided by AG Roth. Relative fluorescence was measured every minute over 2 hours. As negative control, the fluorescence of the respective mix without adding DNA was recorded. The experiment was performed in triplicates and the values normalized to air. A propagation of uncertainty was performed.

Having optimized our self-prepared, low-budget cell-free expression mix, we compared it to a commercially available kit and were thereby able to make a statement about the functionality of our DiaMIX. As already shown on the main results page we set up an experiment using the HA-GFP-His vector kindly provided by AG Roth. Furthermore, a vector that carried a gene for GFP, an N-terminal 10xHis and a Spy-tag on the other side was used. This vector was designed by us, especially for this application. Non-coding areas surrounding the GFP gene were optimized for cell free expression. As seen in figure 6, the DiaMIX shows a higher fluorecence compared to the commercial kit - even with the higher basal fluorescence taken into account (This is an effect of the components in the premix). The commercial kit shows an initial drop in fluorescence. This was observed during every measurement and seems to be an fundamental characteristic of the mix that we attributed to bleaching effects.

Figure 7: Cell-free GFP expression over time. Comparison of the commercial kit with the DiaMIX was performed using the self-synthesized Spy-GFP-10xHis. Relative fluorescence was measured every minute over 2 hours. As a negative control, the fluorescence of the respective mix without adding DNA was recorded. The experiment was performed in triplicates and the values normalized to air. A propagation of uncertainty was performed.

Figure 7 shows the performance of our self-designed vector with the DiaMIX and the commercial mix. As can be observed, the DiaMIX still performs better than the commercial kit. However, the difference is not as obvious as before. This indicates that the DNA does not yet posses an optimal structure for an effective performance in a cell-free system. This could be due to differences in the 5' upstream sequence which was demonstrated to have great influence on the initiation of transcription. Furthermore, tags in the beginning of the coding sequence also enhance the expression level as shown by Haberstock, S. et al. ²⁾
Our self-produced DiaMIX performed about as good as the commercial kit as it is shown in figure 7. Compared to the background, an expression time of two hours resulted in a 2-fold increase of relative fluorescence in both systems. The background fluorescence was estimated by the negative control.

Step by Step Validation

Validation of correct expression via Western blot

Figure 8: Western blot of different GFPs, using our own cell-free expressed expression mix and a commercial one.

After fluorescence analysis, we wanted to validate the expressed protein by performing a western blot. Of each finished cell-free expression reaction, we took 25 µl, performed a SDS-PAGE, blotted it onto a PVDF membrane and incubated it with anti-GFP antibodies. Detection was carried out with secondary HRP-labeled antibodies and ECL solution. The blot shows bands at the right height of about 30 kDa, with GFP being 28 kDa of size. This proves correct protein expression in our system. The Spy-GFP-His is slightly bigger, due to the SpyTag. As expected, no expression is visible in both negative controls. Our mix contains more lysate and therefore shows some smearing. The bands of the commercial kit are much lighter, almost invisible, correlating to its comparably weak expression level (see figure 6). The purified GFP smeared over a small range, but covers the range of our expressed proteins.

iRIf Measurement of Cell-free Expressed GFP

Figure 9: iRIf measurement of cell-free expressed and purified GFP. Anti-GFP antibody specifically bound to both GFPs as seen by an increase of relative iRIf signal.

Encouraged by all the promising results that were obtained, we decided to take the next step and broaden the verification methods of our cell-free expressed proteins by using the iRIf technology. We validated cell-free expressed GFP by pipetting a small amount of the reaction directly onto a glass slide with different surfaces. We checked for specific and unspecific binding. We were able to detect cell-free expressed (DiaMIX) GFP as seen in figure 9. Detailed results for the iRIf experiments performed with cell-free expressed proteins can be found here.

On-slide Cell-free Expression

Figure 10: Microarray scanner picture at 635 nm of cell-free expression on slide. DiaMIX with and without DNA template was spotted and incubated directly on a Ni-NTA slide. Spots containing DNA template show a good signal. The lower left spot showing a half-moon structure had no complete contact with the slide, as spots were incubated within a thick PDMS spotting mask.

To get one step closer to our final approach, we expressed His-GFP directly on Ni-NTA coated iRIf-slides. Therefore, we pipetted 15 µL of our DiaMIX containing the His-GFP DNA template and the DiaMIX without DNA as a negative control on a slide. The slides were incubated for 3 h at 37°C and additionally overnight at 4°C. We performed an iRIf measurement where biotinylated anti-GFP was flushed over the slide, followed by Cy5-labeled streptavidin. In the iRIf, the signal was really weak. Nonetheless, the strep-Cy5 bound to the biotinylated anti-GFP and therefore enabled a detection via a microarray scanner optimized for the detection of Cy-5. The result of the microarray scan at 635 nm is shown in figure 10. We were able to detect small amounts of cell-free expressed His-GFP that diffused to the glass surface and bound to the Ni-NTA with the microarray scanner. The weak binding shows that this part of our systems still needs some optimization.

In-chamber Expression of GFP

The last step towards a final application in the DiaCHIP is the performance of a cell-free expression reaction between the silicone and the glass slide of the microfluidic chamber. Immobilized, linear DNA on the PDMS slide should be transcribed and translated into proteins. Due to time constraints, we were not able to optimize this process. Nonetheless, our initial results can be found on this results page!.

Cell-free Expression of Disease-Specific Antigens

Sadly, we could not show binding of the respective antibodies/serum with the iRIf measurement device. This could be due to folding problems of the antigen during cell-free expression. Moreover, the concentration of expressed C. tetani antigen might not have been high enough to be detected by iRIf measurements.