TU Darmstadt

Fine Tuning The Modification Ratio

Introduction

To broaden the modularity and flexibility of our application we have tested different expression procedures. This is very important due to the need of different modification ratios of the Virus-like particles (VLPs). The P22 VLP consists of coat proteins (CP, BBa_K3187017) and scaffold proteins (SP, BBa_K3187021) ^[1]. For our modular Virus-like particle platform (MVP), we used the LPETGG-tagged CP (BBa_K3187000), an altered version of the CP. The LPETGG-tag (BBa_K3187019) was used to modify the surface of the VLPs via the Sortase A7M (BBa_K3187028) after VLP assembly ^[2]. By varying the ratio of CP to CP-LPETGG, the density of LPETGG-tags presented on our particles can be modified. On this basis, the number of proteins of interest fused to the VLPs by the Sortase A7M can be deliberately changed. Consequently, the ratio of CP to CP-LPETGG is essential for the level of modification of our customized VLPs.

To address our idea of modification regulation, we first need to design and characterize a proper expression cassette. Therefore, we constructed a dual expression vector consisting of pT7 sfGFP x pTet mCherry in pTeTW3con2, which encodes the reporter proteins mCherry, controlled by a tetA promoter, and sfGFP, controlled by a T7 promoter (Fig. 1). We chose these proteins for proof-of-concept experiments, as their expression levels are easily measured via fluorescence. During the experimental process we introduced different induction times and inducer concentrations to achieve a wide spectrum of expression levels. The combination of these manifold expression circumstances should then make it possible to identify the fitting expression variables for the desired composition of any Virus-like particle.

Figure 1: Characterization of the two expression sites using mCherry and sfGFP as gene reporters.

The data, which was collected during the process of characterization, was then used in our Tech project. In order to construct a bioreactor for the production of VLPs in a specific composition, the fitting induction strategy is indispensable. By working with an automized bioreactor we wanted to enhance the reproductivity of our MVPs in an upscaling production.

In the future, we are planing to use a dual expression plasmid for the in vivo assembly of our VLPs. This plasmid consists of P22 coat and scaffolding Protein x CP-LPETGG in pTeTW3con2, which encodes the CP and SP on a tetA regulated site, and the CP-LPETGG on a T7 site regulated by a lac operator (Fig. 2). We expect to reach the variation of the modification level by tuning the expression ratio of both expression sites. We will use just a single plasmid for the in vivo assembly. This will simplify our future production of our modular Virus-like particles (MVPs).

Figure 2: Tunable expression system for adjusting the modification ratio of our modular Virus-like particles (MVPs) during the in vivo assembly.

Achievements

Successful characterization of our pTeTw3con2-sfGFP-mCherry construct

Checkbox

Providing a functional induction strategy for varying ratios of sfGFP to mCherry

Checkbox

Establishing a mathematical context of the expression levels in relation to an induction concentration

Checkbox

Creating a basis for the software development of the syringe pump system used for the induction of the bioreactor

Cloning

The cassette pTeTW3con2-ptet-mCherry--sfGFP-pT7 was cloned in two steps via restriction and ligation. First, the mCherry gene was amplified using overhang primers with the recognition sequences of the restriction enzymes EcoRI and SalI. The purified backbone (pTeTw3con2) and PCR product were restricted with the respective enzymes and ligated after gel extraction. Sequencing analysis was carried out to test whether the cloning was positive. Next, the sfGFP gene was amplified using overhang primers with the recognition sequences of NdeI and XhoI. The purified backbone (pTeTw3con2-ptet-mCherry) and PCR product were restricted with the respective enzymes and ligated after gel extraction. The cloning of the final product was checked via sequencing.

Expression Assay

Before the experiment, E. coli BL21 (DE3) were transformed with the desired dual expression plasmid and selected overnight at 37 °C on an agar plate with fitting antibiotic. The plate could be stored up to three days in the cooler.

The day before the measurement, three 5 mL overnight cultures per plasmid variant were inoculated with a single colony. Following the incubation in a rotation shaker the OD₆₀₀ was measured, so that the 1.5 mL of LB media in every well of the 24 well plate could be inoculated at OD₆₀₀ = 0.1. Following, the 24 well plate was incubated at 37 °C and 200 rpm until OD₆₀₀ = 0.4 was reached. The moment this value was reached was defined as 0 h (t=0) and the following incubation period was performed at 30 °C. The wells were induced with 37.5 µL of an inducer depending on the experiment. At 6 h, 100 µL samples were collected from each well and the OD₆₀₀ was measured as well. This process was repeated after the 24 well plated had been incubated overnight (22 h) at 30 °C and 200 rpm.

The 100 µL samples were centrifuged at 11000x g for one minute and the supernatant was discarded. At this point the samples were snap-frozen in liquid nitrogen for storage. Then the cell pellet was resuspended in a 200 mM Na₂HPO₄ (Disodium hydrogen phosphite) solution, so that OD₆₀₀ = 5 was reached. The suspension was mixed 1:1 with 2x Rotiload Buffer. 20 µL of the samples were used to perform a SDS-PAGE at 120 V for 100 min. After the SDS-PAGE, fluorescent images were taken using an Amersham™ Imager 600.

Spectrophotometric Measurement using a TECAN Reader

Before the experiment E. coli BL21 (DE3) were transformed with the desired dual expression plasmid and selected overnight at 37 °C on an agar plate with fitting antibiotic. The plate could be stored up to three days in the cooler.

The day before the measurement, three 5 mL overnight cultures per plasmid variant were inoculated with a single colony. After incubation in a rotation shaker, the OD₆₀₀ was measured, so that the 1.5 mL of LB media in every well of the 24 well plate could be inoculated at OD₆₀₀ = 0.1. Following, the 24 well plate was incubated at 37 °C and 200 rpm for 30 min. Following the incubation the spectrophotometric measurement at 30 °C was started and measured for 1 h without induction. Later, the wells were induced with 37.5 µL of an inducer depending on the experiment. The measurement was ended 6 h after the first induction. At the end of the measurement, 100 µL samples were collected from each well and the OD₆₀₀ was measured as well. This process was repeated after the 24 well plated had been incubated overnight (22 h) at 30 °C and 200 rpm.

The 100 µL samples were centrifuged at 11000 rpm for one minute and the supernatant was discarded. At this point the samples were snap-frozen in liquid nitrogen for storage. The cell pellet was then resuspended in a 200 mM Na₂HPO₄ (Disodium hydrogen phosphite) solution, so that OD₆₀₀ = 5 was reached. The suspension was mixed 1:1 with 2x Rotiload Buffer. 20 µL of the samples were used to perform an SDS-PAGE at 120 V for 100 min. After the SDS-PAGE, fluorescent images were taken using an Amersham™ Imager 600.

Spectrophotometric Measurement using a Spectramax M5e

Before the experiment E. coli BL21 (DE3) were transformed with the desired dual expression plasmid and selected overnight at 37 °C on an agar plate with fitting antibiotic. The plate could be stored up to three days in the cooler.

The day before the measurement, three 5 mL overnight cultures per plasmid variant were inoculated with a single colony. After incubation in a rotation shaker, the OD₆₀₀ was measured, so that the 1.5 mL of LB media in every well in the 96 well plate could be inoculated at OD₆₀₀ = 0.1. Following, the 96 well plate was incubated at 37 °C at 400 rpm for 30 min. After the incubation the spectrophotometric measurement at 30 °C was started and measured for 90 min without induction. Later, the wells were induced with 37.5 µL of an inducer depending on the experiment. The measurement was ended after 5 – 7 h of measuring time.

Fluorescent imaging

Fluorescence images were taken using the Epi-Blue (460 nm), Green (520 nm) and Red (630 nm) light sources in combination with the Cy2 (525 nm), Cy3 (605 nm) and Cy5 (705 nm) emission filters of the Amersham™ Imager 600. The exposure settings were kept the same for every gel. Afterwards, the different light channels were merged to match the shown colors with the wavelength of the emission signal, instead of the excitation wavelength. The images were edited using ImageJ. The brightness and contrast of the images were not modified for the green and red channels, which represent the fluorescence of sfGFP and mCherry. Only the signal of the blue channel, which represents most of the ladder, has been increased in brightness.

Notebook

To find more detailed versions of the method protocols, please have a look at our notebook .

Measuring the expression levels

After finishing the cloning of pT7 sfGFP x pTet mCherry in pTeTW3con2, we started measuring the expression levels for both expression sites by performing spectrophotometric measurements using a Spectramax M5e. Therefore, we used triplicates induced with Isopropyl ß-D-1-thiogalactopyranoside (IPTG) (Fig. 4) or Anhydrotetracycline (AHT) (Fig. 3) in different concentrations. The IPTG concentrations used in this experiment ranged from 0.1 – 1 mM and AHT was used in the range of 0.1 – 0.3 µg/mL. Additionally, one triplicate was left uninduced and another triplicate was left uninoculated and used as a blank for the measurements. We expected the results to show the basic expression characteristics of our both expression sites.

Figure 3: Spectrophotometric measurement of the fluorescences of mCherry (red) and sfGFP (blue) triplicates after inducing with AHT. AHT was induced at 90 minutes. The induction with 0.1 µg/mL is shown in light red and light blue, and the induction with 0.3 µg/mL is shown in dark red and dark blue. View full size image.

Figure 4: Spectrophotometric measurement of the fluorescences of mCherry (red) and sfGFP (blue) triplicates after inducing with IPTG. IPTG was induced at 90 minutes. The induction with 0.1 mM is shown in light red and light blue, and induction with 1 mM is shown in dark red and dark blue. An uninduced sfGFP variant is shown in orange. View full size image.

The measurement showed a strong background expression of the T7 site, represented by the increasing fluorescence signal of sfGFP. A possible cause for this background expression could be a missing terminator following the coding sequence of the chloramphenicol resistance. For that reason, the RNA polymerase may not dissociate from the plasmid. This may lead to transcription of the sfGFP gene that is oriented in tandem to the resistance.

However, this background expression of sfGFP lessened with a rising AHT concentration. This came as quite a surprise, since inducing with different AHT concentrations was supposed to mainly regulate the tetA regulated site. Generally, the data showed that sfGFP fluorescence is clearly dominating. Nevertheless, the data was not sufficient enough to determine the ratio of the two sites to each other due to the differences in maturation and brightness between mCherry and sfGFP. Therefore, we needed to perform further measurements.

Unfortunately, we could not select the fitting settings for monitoring the fluorescence signals of both sfGFP and mCherry in the Spectramax M5e. The final fluorescence signal of sfGFP in IPTG induced triplicates was stronger than the maximal detection point of the instrument while a change in the signal of mCherry was barely detected during the measuring process. For this reason, we changed our method to an expression assay where the fluorescence was detected after a semi denaturized SDS-PAGE via a fluorescence imager. This method has the advantage that the fluorescence signals of sfGFP and mCherry can be detected separately with fitting settings. The different settings were noted for the followed analysis.

In the further process, we wanted to test a new set of plasmids which vary in gene orientation and ribosomal binding site (RBS) on the tetA site. By measuring the expressions levels of these plasmids, we intended to find a variant that fits our demands better. Another advantage over the dual expression plasmid used in this experiment is the existence of a T7 terminator behind the chloramphenicol resistance gene.

Screening for induction strategies

Because we could not generate statistically relevant data by performing assays with the improved plasmids, as described in the last paragraph, we went back to using pTeTW3con2-ptet-mCherry--sfGFP-pT7 for the following assays. Due to the results of the spectrophotometric measurement using the Spectramax M5e, we used another approach to assess the expression data. In particular, we performed semi-denatured SDS-PAGEs and measured the fluorescence of applied proteins. By doing so, we could better distinguish the expression levels of sfGFP and mCherry.

We used two induction strategies for this assay (Fig. 5 and Fig. 6). On the one hand we delayed the induction with IPTG and on the other hand varied its concentration. We tried to find a strategy that compensates the sfGFP background expression in order to generate a comparable ratio of sfGFP to mCherry.

Figure 5: Representation of the relative fluorescence intensities of mCherry (red) and sfGFP (blue) triplicates after an expression time of 6 h. First, a constant concentration of AHT was induced, while the concentration of IPTG was varied (0.1 mM and 0.5 mM). The induction time of IPTG after inducing with AHT was changed as shown in the brackets. View full size image.

The samples collected after 6 h of expression showed a distinct trend. Whenever IPTG was induced during the experimental procedure there was a large excess of sfGFP in comparison to mCherry. The used induction times or concentrations of IPTG only had a small effect on the resulting ratio of sfGFP to mCherry. Solely the variant that was induced with AHT only showed a significant difference: There we observed a 2:1 ratio of sfGFP to mCherry. As expected from the previous experiments, the uninduced control showed the highest sfGFP expression in this series.

Figure 6: Representation of the relative fluorescence intensities of mCherry (red) and sfGFP (blue) after an overnight expression. First, a constant concentration of AHT was induced, while the concentration of IPTG was varied (0.1 mM and 0.5 mM). The induction time of IPTG after inducing with AHT was changed as shown in the brackets. View full size image.

The next samples were collected after continuing the expression overnight. These samples showed the same trend as the ones taken after 6 h of expression. As before, there was only a very small difference between the various induction strategies when IPTG was used. Again, the uninduced control had the largest excess of sfGFP in this series. Unlike the other samples, the "AHT-only" induced variant showed a 1:1 ratio of sfGFP to mCherry.

To conclude, the data showed that the T7 site had, under any induction condition, a much higher activity than the tetA site. The expression was too strong for an effective tuning of the expression levels while IPTG was induced. Surprisingly, the background expression of the T7 site compensated the AHT-induced expression of the tetA site over night and was even stronger at 6 h after induction. Strikingly, this unexpected behavior gave us the opportunity to receive results in the fitting area of ratios. We therefore changed our strategy from dual induction to single induction while still performing dual expression. Fortunately, this change of strategy simplified things a lot. It became possible to control the expression levels of both expression sites with a single inducer, paving the way for VLPs with customized modification ratios.

Tuning the expression ratio

At first, we tried to delay the induction with IPTG after the induction of our bacterial cultures with AHT in order to produce comparable concentrations of sfGFP and mCherry. Our results showed comparable levels of sfGFP to mCherry in the samples where the T7 site was left uninduced. Using this information, we tried inducing with different AHT concentrations to regulate the ratio of mCherry to the background-expressed sfGFP (Fig. 8 and Fig. 9). At the same time we measured the OD₆₀₀ of all samples to find out wheather AHT induction interferes with cell growth (Fig. 7).

Figure 7: OD₆₀₀ measurement of triplicates induced with various concentrations of AHT at minute 60. The induction concentration ranged from 0.1 - 0.3 µg/mL in steps of 0.5 µg/mL. In addition, one triplicate was left uninduced. View full size image.

The OD₆₀₀ measurements showed that the induction of AHT did not affect the bacterial growth of the various triplicates prominently. Nevertheless, the uninduced control had the highest growth rate.

Figure 8: Spectrophotometric measurement of triplicates with mCherry as a reporter. The induction of various AHT concentrations was at minute 60. The induction concentration ranged from 0.1 - 0.3 µg/mL in steps of 0.5 µg/mL. In addition, one triplicate was left uninduced. View full size image.

Figure 9: Spectrophotometric measurement of triplicates with sfGFP as a reporter. The induction of various AHT concentrations was at minute 60. The induction concentration ranged from 0.1 - 0.3 µg/mL in steps of 0.5 µg/mL. In addition, one triplicate was left uninduced. View full size image.

Surprisingly, the spectrophotometric measurements showed a decline in the production of sfGFP with an increasing AHT concentration, while the mCherry production was the same for all tested concentrations. These results were unexpected, but provided us with a way to produce varying ratios of sfGFP to mCherry in dependence of the AHT concentration.

We guess that increased mCherry mRNA levels cause reduced production of sfGFP. Thereby, the protein translation machinery is overloaded with mCherry as well as sfGFP production. Another possibility is an inhibition of sfGFP transcription due to the plasmids' altered secondary structure in the absence of tetR. Both ideas could be tested by cloning two plasmids, one having deleted its ribosomal binding site on the mCherry coding site, and the other having its entire mCherry coding sequence deleted. Inducing these plasmids with AHT could provide enough information to further assess the cause for the decline in the sfGFP production. A reduction of the sfGFP concentration for both plasmids would indicate that the inhibition is based on a change in the plasmids' secondary structure. No change in either of the plasmid would suggest a competitive translation inhibition. Unchanged sfGFP levels in the RBS-deleted plasmid, and decreased levels in the non-mCherry coding plasmid, would imply a competitive transcription inhibition to be the cause.

Figure 10: Shown is an image of a semi denaturated SDS-PAGE of the trplicates shown in Fig. 7 and Fig. 8 after an overnight expression. The upper bands in red represent mCherry and the lower bands in green represent sfGFP. The induction concentration increases from the left side to the right side (0, 1, 1.5, 2, 2.5, 3, 3.5, 4 µg/mL), triplicates were used. For the protein size determination the BlueStar Prestained Protein Ladder was used. The blue channel was increased in brightness for a better visibility of the ladder. View full size image.

The semi denatured cell culture samples, taken 6 hours and the day after induction, prove that we can adjust the ratio of our two fluorescent proteins by modulating the AHT concentration. As can be seen in Fig. 10 and Fig. 11, inducer concentrations of AHT ranging from 0.1 µg/mL to 0.4 µg/mL caused a change in the ratio of mCherry:sfGFP from 1:2 to 2:1 for overnight cultures of E. coli. The ratio seems to approach a maximum of around 2:1 following an increasing inducer concentration. We sadly did not manage to take measurements at higher or lower concentrations in time, leaving this a hypothesis.

Figure 11: Representation of an asymmetric sigmoidal regression (red) of the ratio of mCherry to sfGFP. The data were obtained by various induction concentration of AHT (blue). The samples were taken in triplicates after an overnight expression. The function of the regression is shown in Equation 1. View full size image.

After plotting the collected data, we did a software-based regression by testing different types of functions. The best fitting function with the highest determination coefficient was an asymmetric sigmoidal function, as presented in Equation 1. The determination coefficient of this function is 0.989, which shows the great reliability of this function for further applications. Based on this, the function was then used by the Tech project for developing their software, which can be used to control the settings of the bioreactor.

$$y = {0.4756 + {2.5966 \over (1 + 10^{((0.3047 - x) * 11.33)})^{0.2072}}}$$

Equation 1: Equation of the asymmetric sigmoidal regression with a determination coefficient of 0.989. The variable x represents the induction concentration of AHT in µg/mL and y represents the ratio of mCherry to sfGFP in relative fluorescence units (RFU).

In our opinion, it should be possible to use the vector pTeTW3con2 for the coexpression of CP, SP and CP-LPETGG, while regulating the ratio of CP to CP-LPETGG using the AHT concentration. We then came up with two simple methods to improve our part. First, replacing the T7 promoter with a weaker promoter should make higher ratios of mCherry to sfGFP possible. Secondly, adding a terminator at the 5' UTR of the chloramphenicol resistance should quench the background expression of sfGFP.

The AHT concentration used to produce ratios of proteins other than the ones tested would most likely differ from our results. Nonetheless, our results have proven that defining protein ratios by varying inducer concentrations with our vector is in fact possible and easy. To properly define the protein ratios, a western blot would have been another option. The fluorescence-based ratios we defined in our results only appoximate the real protein concentration ratios, since semi-denaturating the samples presumably has different effects on the integrity of mCherry and sfGFP, consequently resulting in an unpredictable change of the emission intensity coming from each of the fluorescent proteins.

References

↑ Earnshaw, W., Casjens, S., & Harrison, S. C. (1976). Assembly of the Head of Bacteriophage P22: X-ray Diffraction from Heads, Proheads and Related Structures. 387–410. [1]
↑ Patterson, D., Schwarz, B., Avera, J., Western, B., Hicks, M., Krugler, P., … Douglas, T. (2017). Sortase-Mediated Ligation as a Modular Approach for the Covalent Attachment of Proteins to the Exterior of the Bacteriophage P22 Virus-like Particle. [2]

[1] Earnshaw, W., Casjens, S., & Harrison, S. C. (1976). Assembly of the Head of Bacteriophage P22: X-ray Diffraction from Heads, Proheads and Related Structures. 387–410. [1]

[2] Patterson, D., Schwarz, B., Avera, J., Western, B., Hicks, M., Krugler, P., … Douglas, T. (2017). Sortase-Mediated Ligation as a Modular Approach for the Covalent Attachment of Proteins to the Exterior of the Bacteriophage P22 Virus-like Particle. [2]

[1]

[2]

Team:TU Darmstadt/Project/Modification Ratio