Results

Repebody characterization

Here we present the results characterizing the expression and binding of the original, unmodified mCherry-binding Repebody BBa_25X0LA8K and our engineered derivative of it, BBa_25IQCJ3B, the ssDNA optimized mCherry-binding Repebody. Various different methods and experimental setups were used to analyze their behaviours with emphasis being put on reliable soluble expression in E. coli BL21 (DE3) cells and binding specificity.

Expression

Aim

Determine optimal IPTG concentration and temperature for expression of the ssDNA optimized mCherry-binding Repebody, compare it to expression of unmodified mCherry-binding Repebody using E. coli BL21 (DE3) cells. The Repebody was cloned into the plasmid backbones pET303 and pCDF-DuetIso.

Experimental design

The Repebody variants were cloned into the pET303 vector with a C-terminal 6xHis-Tag and transformed into E. coli BL21(DE3) cells. The cultures were incubated using 200 mL LB-medium in a 1 L flask containing ampicillin (100 µg/mL) at 37°C to an OD₆₀₀ between 0.5 and 0.6 Four different concentrations of IPTG (0.1 mM, 0.4 mM, 0.7 mM, 1.0 mM) were tested at two different temperatures (18°C for 24 h, 25°C for 16 h) in 20 mL of culture flasks. The final OD₆₀₀ was measured of all cultures and used as an indicator for total cell amount. This OD₆₀₀ was then used to normalize the amount of cells taken out for all samples, in order to measure the expression per cell, and not per culture volume. After induction, cell pellets were lysed by sonication. Soluble and insoluble protein fractions were separated by centrifugation, and analyzed by SDS-PAGE and Western blotting with an anti-His antibody for detection of the Repebody and anti-RNA polymerase beta-subunit antibody as a loading control. For quantification, band intensities were analyzed with FIJI to evaluate optimal conditions.

Results

We detected a protein band corresponding to the expected molecular weight of approximately 30 kDa under all expression conditions on the Western blot (Figure 1). An antibody against the beta-subunit of RNA polymerase was used as a loading control because it recognizes the product of a bacterial housekeeping gene. Notably, the majority of the expressed ssDNA-Repebody was present in the supernatant fraction (Figure 1), which showed a stronger signal than the pelleted fraction (Figure 3). These results suggest that the ssDNA-Repebody is expressed as a soluble protein and predominantly accumulated in the cytosol.

We performed semi-quantitative densitometric analysis of the band intensities using FIJI software to evaluate protein amounts. The band intensities of the soluble fractions were normalized to the RNA-polymerase loading control to determine the expression condition with the highest yield (Figure 2). Analysing expression for 18°C / 24 h, the expression per cell and per culture volume decreases with increased induction concentrations (Figures 2 and 3). Comparing the average protein amounts, 18°C yielded higher expression per cell on average, but when evaluating the expression per culture volume, i.e. not normalized to final OD₆₀₀, 25°C expression resulted in higher protein amounts.

To obtain the protein amount per culture and not per cell, these values (Figure 2) must be multiplied by their respective final OD₆₀₀.

This constitutes the following values:

These are the values most relevant for expressing large amounts of proteins, with expression per cell being relevant for characterization, but expression per culture volume being the practically important factor.

Comparison of ssDNA-Repebody to unmodified-Repebody

To assess whether the ssDNA modification affected protein expression, the ssDNA-Repebody and unmodified-Repebody were expressed under identical conditions in E. coli BL21(DE3) cells using the pET303 vector. Inductions were performed with IPTG concentrations ranging from 0.1 to 1.0 mM at either 18°C for 24 hours or 25°C for 16 hours. Soluble fractions obtained after cell lysis and centrifugation were analyzed by SDS–PAGE and Western blotting using an anti-His antibody for Repebody detection.

Both constructs produced a protein band at approximately 30 kDa, consistent with the expected molecular weight of the Repebody (Figure 5). Comparing the different IPTG concentrations and the two incubation temperatures revealed no significant differences in expression pattern and signal intensities between ssDNA-Repebody and unmodified-Repebody in the soluble fraction. The pellet fractions were not evaluated. This indicates that the ssDNA modification did not substantially alter the expression efficiency or solubility under the tested conditions.

Expression in pCDF-DuetIso backbone

In order to evaluate the compatibility of the pCDF-DuetIso backbone for ssDNA-Repebody expression, the construct was subcloned into this vector and expressed in E. coli BL21(DE3) cells. The pCDFDuetIso plasmid was specifically designed to function with the mutation system used to generate binding diversity variants. Expression tests were conducted under the same conditions as for the pET303 system, with varying IPTG concentrations (0.1–1.0 mM) and incubation temperatures (18°C for 24 hours and 25°C for 16 hours).

Western blot analysis of the soluble fractions revealed a clear band at approximately 30 kDa under all tested induction conditions (Figure 6), confirming consistent expression of the ssDNA-Repebody from the pCDF-DuetIso construct. Densitometric quantification of band intensities using FIJI software indicated a minor increase in expression levels with higher IPTG concentrations at both induction temperatures (Figure 7).

Analysis of the insoluble fractions revealed the presence of the Repebody at the expected molecular weight, indicating that a small portion of the protein was present in aggregated form or inclusion bodies (Figure 8). However, the soluble fraction displayed a much stronger signal, demonstrating that the majority of the ssDNA-Repebody was produced as a soluble cytosolic protein.

Together these results demonstrate the successful expression of the ssDNA-Repebody within the pCDFDuetIso system and validate its functionality as an alternative expression platform for further experiments.

Overall, western blotting confirmed consistent cytosolic expression of both Repebody variants in pET303 vector and also ssDNA-Repebody in pCDFDuetIso vector, allowing further characterization and protein purification.

Conclusion

The results demonstrate successful optimization of Repebody expression in E. coli BL21 (DE3) for both variants in the pET303 vector, as well as for the ssDNA-Repebody in pCDFDuetIso vector. No major differences in protein expression or solubility were observed at 18°C or 25°C or at the various tested IPTG concentrations. Regarding the values used for expression of large culture volume, the ssDNA-Repebody in pET303 (Figure 1) was initially evaluated based on rough band size, resulting in picking 25°C / 0.4 mM as the optimal condition. Later, thorough investigation via densitometric (FIJI) evaluation shows minor differences in protein amounts, revealing that 25°C / 0.1 mM should have optimally been picked (Figure 3).

Western blot analyses confirmed that the Repebody was predominantly present in the soluble fraction, which indicates that it was properly folded and that aggregation was reduced under these conditions. A comparison with the unmodified Repebody revealed no significant differences in soluble expression levels, suggesting that the ssDNA modification does not impair protein expression or solubility. These findings confirms that the engineered modification is compatible with the E. coli expression system. Subcloning of the ssDNA-Repebody into the pCDFDuetIso vector, showing consistent expression as soluble protein, further validated the expression flexibility of this construct across plasmid backbones.

Binding

To test the binding of both the ssDNA-Repebody and the unmodified-Repebody to its target mCherry, we performed pulldown assays and gel filtration chromatography. Emphasis was also placed on testing the specificity of our newly designed ssDNA-Repebody, by investigating its interactions with EGFP and mTagBFP2. Since these proteins share structural similarities to mCherry, the absence of binding by the Repebody variants serves as a strong indicator of specificity. Previously, the unmodified-Repebody revealed an almost equal affinity toward mCherry and mOrange [1], motivating us to test the affinity toward other beta-barrel-structured fluorescent proteins as well.

Co-expression pulldown

Aim

Next, we focused on demonstrating the binding of both Repebody variants to mCherry following co-expression in the same cell, and test whether they can also bind EGFP.

Experimental design

To test the interaction, we designed four expression plasmids based on the pCDFDuet1 backbone. Each carries two coding sequences:

• plG-113: ssDNA-Repebody-6xHis + mCherry-HA
• plG-114: ssDNA-Repebody-6xHis + EGFP-HA
• plG-116: unmodified-Repebody-6xHis + mCherry-HA
• plG-117: unmodified-Repebody-6xHis + EGFP-HA

All constructs were expressed under the optimized conditions previously identified (Figure 1, 25 °C, 0.4 mM IPTG, 16 h). After harvest by centrifugation, the cells were lysed by sonication, and the soluble fraction was applied to a Ni-NTA spin column to capture the His-tagged Repebody. If binding occurred, co-expressed mCherry-HA should be retained and detectable in the elution fractions by western blotting using an anti-HA antibody. EGFP-HA served as a non-binding control.

Results

The HA blot (Figure 9) revealed the presence of mCherry-HA in the elution fraction when co-expressed with both Repebody variants. In contrast, no HA signal was detected when EGFP-HA was co-expressed.

Conclusion

The Western blot supports the specific interaction between Repebody and mCherry (Figure 9).

As the proteins were co-expressed in the same cell and loaded directly onto the column after the soluble fraction was separated from the cell debris pellet, it can be concluded that interaction between mCherry-HA and Repebody occurs directly in cytoplasm.

Pulldown with normalized EGFP and mCherry amounts

Concentration measurement

Aim

Measure the concentration of mCherry-HA and EGFP-HA in their respective soluble fractions after cell lysis.

Experimental design

To compare the affinity of the Repebody to mCherry and/versus EGFP via a binding assay, the Repebody must be exposed to the same amount of molecules for both potential binding partners. Otherwise, no quantitative evaluation about the relative affinity, and thus the specificity, can be made.

As a result, the following plasmids were used in subsequent pulldown experiments:

• pIG25-108: pET303_ssDNA-Repebody-6xHis
• pIG25-110: pET303_mCherry-HA
• pIG25-111: pET303_EGFP-HA
• pIG25-15: pET303_mCherry-TEV-10xHis
• pIG25-01: pET303_EGFP-TEV-10xHis

E. coli BL21 (DE3) cells were used for expression, the soluble fractions were obtained via sonication and centrifugation. For pIG25-15 and pIG25-01, the bacteria were cultivated in 800 ml total culture volume. After sonication and centrifugation, soluble fractions were purified via immobilized metal affinity chromatography (IMAC) and gel filtration (this purification was done by our secondary PI, Dr. Salavei, as was every gel filtration shown by us). With large volumes of highly pure samples, the concentration was measured via nanodrop, and the fluorescence for each concentration in a dilution series was measured via a plate reader (Figures 10 and 11). This approach provides a specific range of fluorescence intensity, in which we can determine the concentration of fluorescing proteins with reasonable accuracy.

Results

Conclusion

Using the linear (regression) equation and the fluorescence values of the soluble fractions of mCherry-HA and EGFP-HA, we obtain the following concentrations of fluorescent proteins in cell lysate supernatant:

mCherry-HA: 2.79 mg/ml

EGFP-HA: 2.85 mg/ml

As these values differ by only ~2%, and mCherry and EGFP have almost the same molecular weight (~28 kDa), for future applications we assumed that the same lysate volume can be used to achieve nearly the same amount of the protein.

Pulldown

Aim

Quantitatively compare the affinities of the ssDNA-Repebody binding to mCherry and EGFP, respectively.

Experimental design

This experiment will be using the concentrations of mCherry-HA and EGFP-HA to attempt a quantitative evaluation of the relative binding affinity of the Repebody. The basic experimental setup is comparable to the pulldown with co-expressed samples. Here, the three proteins used for this binding assay were expressed individually in E. coli BL21 (DE3) cells with the plasmids pIG25-110, pIG25-111 and pIG25-108. In total, eight different input conditions were prepared to test binding across a range of target protein concentrations. The inputs were adjusted to give a final volume of 120 µL with PBS. Inputs 1-6 were designed to have increasing amounts of protein for the Repebody to bind to, whereas inputs 7 and 8 served as negative controls to assess the background binding of HA-tagged proteins to the Ni-NTA resin.

Results

Western blotting was used to analyze the elution and flow-through fraction for His- and HA-tagged proteins, respectively (Figure 12). Anti-His staining confirmed the presence of the Repebody in the elution fractions across all samples.

Conclusion

The anti-HA staining of the elution fraction revealed the presence of mCherry-HA when co-loaded onto the column with the Repebody, indicating interaction between the two proteins (Figure 12). Some background binding of HA-tagged proteins to the Ni-NTA column was observed (visible in control lanes 7 and 8). But the signal intensity of lanes containing both Repebody and mCherry-HA (inputs 1–3) was substantially stronger than in the negative control (input 7). Interestingly, input 1, containing five times less mCherry-HA than input 7, still yielded a stronger signal. This suggests that the detected mCherry-HA was retained through interaction with the Repebody rather than non-specific binding to the resin.

EGFP-HA was only detected in the elution fractions at higher input volumes (inputs 5 and 6), which indicates weaker binding compared to mCherry-HA (Figure 12). The observation that input 6 produced a slightly stronger band than input 8 (despite equal loading volumes) suggests that EGFP may also bind to the Repebody, albeit less efficiently than mCherry.

The anti-HA western blotting of the flow-through fractions revealed residual mCherry-HA and EGFP-HA that did not bind to the Repebody or the Ni-NTA resin. As expected, the intensity of the flow-through bands increased with higher input volumes of each protein, reflecting the proportion of unbound protein passing through the column. This observation confirms that not all available mCherry-HA or EGFP-HA molecules interact with the Repebody under the tested conditions. Specifically, the stronger retention of mCherry-HA in the elutions corresponded with its reduced presence in the flow-through compared to EGFP-HA. This is consistent with the higher affinity binding of the Repebody to mCherry.

The clarity of this result is compromised by the fact that there is background binding of the column to HA-tagged proteins, but the thickness of the bands from inputs 1-3 compared to the band from input 7 confirms that the presence of mCherry-HA in the elution fraction can not solely be explained by unspecific binding to the Ni-NTA column. This is especially true when considering that input 1 had five times less mCherry-HA molecules compared to input 7, yet produced a thicker band, indicating that the presence of mCherry-HA in the elution fraction is largely attributed to the Repebody binding mCherry-HA.

EGFP-HA only produces bands on the western blot when higher volumes are loaded, indicating the Repebody binds mCherry with higher affinity. It is unlikely that the Repebody does not bind to EGFP at all, as the EGFP-HA band from input 6 is bigger than for input 8, despite the fact they had the same input volume. This affinity toward both can be explained by the structural similarity of the two proteins, as both exhibit the beta-barrel structure characteristic for fluorescent proteins, while their differences in fluorescence mostly come from differences in the unexposed fluorophore.

Anti-His blotting of the elution fractions confirmed the presence of the Repebody, and consequently the background interaction of mCherry-HA and EGFP-HA with the affinity gel.

Gel filtration

Aim

Analyze binding and specificity of the Repebody variants via gel filtrations.

Experimental design

Size exclusion chromatography, also called gel filtration, separates proteins natively by hydrodynamic size using a porous material. Unlike gel electrophoresis using agarose or polyacrylamide, the matrix is composed in such a way that its pores exclude larger proteins, causing them to elute faster. The continuous flow out of the column is monitored by a UV-VIS detector. Most proteins containing amino acids tyrosine, tryptophan or cystine could be detected at 280 nm, fluorescent proteins could be detected additionally at their excitation maximum.

To analyze the binding between two proteins, each protein was first applied to the column individually to determine its elution volume. Subsequently, a mixture of both proteins was loaded onto the column. If the complex elutes at an earlier volume, this indicates an increased hydrodynamic size and therefore an interaction between the two proteins. If the elution volume remains unchanged, no interaction is observed.

The proteins analyzed were the ssDNA-Repebody, the unmodified-Repebody, and also mCherry-TEV-10xHis, EGFP-TEV-10xHis and mTagBFP2-10xHis. These were purified via immobilized metal affinity chromatography (IMAC) and gel filtration, yielding large amounts of highly pure proteins. (The purification via IMAC of these samples was performed by our secondary PI, Dr Salavei, as was every gel filtration shown).

Results

Conclusion

At 587 nm, where absorbance specifically corresponds to mCherry (red trace), the elution profile confirms complex formation. The observed shift in mCherry’s elution volume upon addition of either Repebody demonstrates an increase in hydrodynamic size due to binding. Notably, no residual mCherry peak is visible at the original elution volume (~17 mL), suggesting that all mCherry molecules were bound by Repebody under these conditions. Regarding EGFP and mTagBFP2, both their elution profiles remained unchanged upon addition of either Repebody variant (Figure 13). These results clearly demonstrate that the Repebody binds specifically to mCherry under the conditions tested, as no complex formation was observed with two other fluorescent proteins. Both Repebody variants show similar affinity and specificity in the binding to mCherry, EGFP and mTagBFP2.

Far western blot

Aim

The standard Western Blot relies on antibodies to specifically detect proteins after they are immobilized on a membrane, such as PVDF or Nitrocellulose. A primary antibody binds either the target protein directly, or an epitope tag (such as His or HA) that is fused to the target. While antibodies provide high specificity and robustness in this method, the production is time-consuming and costly. It also relies on the use of animals, which requires ethical considerations.

Our project aims to explore Repebodies as an alternative to antibodies. Repebodies can be produced recombinantly in E. coli, far easier and also without the use of animals.

In order to test the potential of the Repebody to replace standard antibodies for methods like Western blotting, we performed a Far Western blot (Figure 14).

Having established the binding of the Repebody to mCherry, this modification to the standard western blot was intended as a proof-of-concept for using Repebodies in protein detection workflows. In our case, we use Repebodies instead of a primary antibody for target recognition. As our Repebody is labeled with a His-tag, it can then be detected with an anti-His antibody, which in turn are recognized by a standard anti-mouse-HRP-antibody, fused to a horseradish peroxidase for imaging via chemiluminescence. This detection sandwich, consisting of Repebody that binds the target protein, an anti-His antibody that binds the Repebody, and a secondary antibody that binds the anti-His antibody, serves as a proof of concept for the potential applications of the Repebody in a staple method of molecular biology, the western blot.

Experimental design

The target proteins immobilized on the membrane were mCherry-HA and EGFP-HA from clarified lysates. The lysates were prepared following expression in E. coli BL21 (DE3) cells harboring the plasmids pIG25-110 and pIG25-111, respectively. After expression, the cells were sonicated and centrifuged to obtain clarified lysates.

The ssDNA-Repebody was expressed in E. coli BL21 (DE3) cells using the plasmid pIG25-108 under optimized conditions (Figure 1; 25 °C, 0.4 mM IPTG, 16 h). The protein was purified via Ni-NTA spin column chromatography, and the elution fraction was used to prepare the blotting solution (PBS containing 3% BSA, 0.1% Tween 20, and 0.1% sodium azide) at a final protein concentration of approximately 1.1 mg/mL.

Results

Experiment 1

Despite the presence of some nonspecific bands most likely from endogenous proteins (Figure 15), the first far-western blot represents a promising initial step toward employing the Repebody as an antibody-like reagent in standard western blotting. Notably, mCherry-HA produced a detectable signal, indicating that the Repebody can recognize its target in a blotting format.

Experiment 2

To validate the results, the experiment was repeated with Ponceau S staining prior membrane blocking and incubation with Repebody and antibodies, and the Repebody blotting solution was loaded in the last lane as a negative control.

Increasing the Tween-20 concentration in Repebody blotting solution improved binding specificity of the Repebody. However, a substantial loss of native proteins during transfer reduced the detectable signal (Figure 16). The appearance of signal for both mCherry-HA and EGFP-HA in the negative controls (Figures 17 and 18) limited the interpretability of these results. Although unintended, this observation may reflect structural similarities between mCherry and EGFP that promote weak cross-reactivity.

Experiment 3

The experiment was repeated with reduced transfer time and purified Repebody was used instead of blotting solution for the negative control blotting.

This further improved experimental reliability. Negative controls behaved as expected, validating that the Repebody functions effectively as a blotting reagent (Figures 19 and 20). The increased signal observed for native samples following a shorter transfer time (reduced by half) suggests improved transfer efficiency. Interestingly, Repebody binding was specific under denaturing conditions but not under native ones—contrary to results from other assays performed under native conditions (pulldown and gel filtration), which indicated no interaction with EGFP.

Experiment 4

The transfer time was reduced further (by an additional 33%), however even samples diluted 1:4 showed a detectable signal under native conditions and a strong signal when denatured (Figures 21 and 22). With all controls functioning properly, this final experiment supports the reproducibility of the previous results and benefits from the use of samples with known concentrations (Figures 10 and 11), allowing for semi-quantitative estimation of Repebody affinities toward mCherry and EGFP.

Conclusion and Outlook

The far-western blot experiments demonstrate that a single Ni-NTA purification step produces a Repebody preparation suitable for use as a functional blotting reagent when expressed in E. coli BL21 (DE3) cells. The purified Repebody remained stable and active at least 1.5 weeks across multiple applications. Interestingly, native EGFP was detected despite pulldown and gel-filtration assays showing little to no interaction, suggesting that the extended incubation time used in the far-western blot (3 h) may permit accumulation of weak or transient binding. While the Repebody displays high affinity for mCherry, weak residual affinity for EGFP could account for this observation.

Future work should include quantitative determination of Repebody affinities toward mCherry, EGFP, and mTagBFP2 using kinetic methods such as fluorescence polarization or biolayer interferometry. Engineering approaches could further enhance applicability, for example by fusing the Repebody to an antibody constant region recognized by existing secondary antibodies or directly to horseradish peroxidase to enable single-step detection.

Overall, ssDNA optimized and unmodified mCherry-binding Repebody variants were successfully expressed and purified in soluble form from E. coli. Pulldown, gel-filtration, and far-western blot assays consistently verified specific target recognition, highlighting the potential of Repebodies as versatile, antibody-mimetic protein binders with broad biotechnological relevance.

Retrons

pCDFDuetIso

Aim

In order to allow us to efficiently mutate a gene of interest (GOI) using our retron system, we had to ensure that the GOI would always be replicated in the same direction during DNA replication. This is because the retron-generated ssDNA only binds to the lagging strand of the replication fork. To stabilize the direction our GOI is replicated in, we decided to move one of the multiple cloning sites (MCS) of pCDFDuetIso adjacent to the origin of replication (ORI).
Since we interfered directly with the expression machinery, it is natural to pose the question about how the plasmid’s expression characteristics were influenced by the changes made to the backbone. Keeping in mind that expression conditions are heavily dependent upon the genes being expressed, we wanted to set up a model in order to test the severity of the changes in expression rate, in absolute terms as well as in relative terms between the two operons of the duet plasmid, as well as determining the general trend of the changes.

Experimental design

Both pCDFDuetIso and pCDFDuet1 share a similar structure, featuring two MCS regions, each with their own lac-T7 promoters. Besides the difference of location, as described above, there is only one terminator in pCDFDuet1, since both MCS are adjacent, whereas each MCS in pCDFDuetIso has its own terminator. In order to determine how the expression rates of the multiple cloning sites between pCDFDuet1 and pCDFDuetIso differ, we cloned HA-tagged EGFP and His-tagged mCherry as transcriptional reporters into each site using the same inserts and enzymes between both plasmids for each fluorescent protein. For this experiment, we traced the fluorescence intensities of both proteins across time. We assumed that the amount of fluorescent protein is directly proportional to the fluorescence emission intensity.
In two test runs, we induced cultures of BL21 cells containing the pCDFDuetIso-EGFP-mCherry and pCDFDuet1-EGFP-mCherry constructs by adding IPTG to a concentration of either 0.1, 0.4, 0.7, 1.0 millimoles per liter. Furthermore, negative controls were prepared by growing untransformed BL21 cells, as well as transformed cells without IPTG. The expression was run for six hours, and the cultures were sampled every hour to retrieve the cultures’ OD600 values, their EGFP fluorescence emission intensity, and mCherry fluorescence intensity. The measurable values we wanted to determine were expression throughput, and the ratio between the expression rates of both multiple cloning sites.

Results

Results from inducing pCDFDuet1 and pCDFDuetIso using IPTG, measured by hourly sampling and fluorescence analysis for EGFP (ex: 467 nm, em: 507 nm) and mCherry (ex: 570 nm, 610 nm), as well as the optical density (abs: 600 nm) of each sample. Multiple different expression conditions are rendered here, the concentration of IPTG for each sample is listed in the graph’s legend.

The measurement results from the induction cultures were plotted, and are displayed in Figures 23-25. In graph A, It can be seen that there is a consistent increase in emission and therefore protein concentration, with most samples generating EGFP at a similar rate, with a relatively large datapoint spread. Similarly to the values from the EGFP measurement, there is a relatively large spread in the measurement values for the mCherry fluorescence intensity displayed in graph B. The fluorescence values from pCDFDuet1 samples seem to be greater on average than those from pCDFDuetIso, but this cannot be said with confidence due to the aforementioned value spread. Finally, the quotient between the fluorescence intensities for EGFP and mCherry was calculated for all measurements, the values of which are shown in graph C. While there appears to be an initially large spread in the fluorescence ratios (t < 4), the ratios converge on two values as time progresses (t => 4), with the exception of the DuetIso 0.1 mM sample. From the normalized mCherry and EGFP fluorescence we can see that for each fluorescent reporter both pCDFDuet1 and pCDFDuetIso are able to produce similar amounts of protein. However, the low accuracy of this measurement prevents us from making any strong statements about the exact differences for each reporter, and it is also difficult to identify an optimal concentration of IPTG for the expression.
On the other hand, the ratio of EGFP and mCherry emission strength, appears to converge between different induction conditions, and remains relatively stable over time. The ratios of EGFP and mCherry emission intensity for either pCDFDuet1-EGFP-mCherry and pCDFDuetIso-EGFP-mCherry approach different values. This indicates that the expression strengths of the promoters have been influenced by converting pCDFDuet1 to pCDFDuetIso. However, as the quotient between the emission ratios of the two plasmids is not particularly large, having a measured value around 1.4, and we found that this was reproducible when we re-ran the experiment. Therefore, it can be concluded with some confidence that the expression strength ratio is slightly affected by using pCDFDuetIso instead of pCDFDuet1, but that this effect is not particularly strong.

Conclusion

From these results, while it is certainly possible to draw some rough conclusions about our experiment, we hesitate in making a confident statement about the exact changes when comparing expression characteristics in pCDFDuet1 with those in pCDFDuetIso. This, however, still achieves the behavior we set out to elucidate, as we were primarily interested in a general trend instead of a concrete numeric result. This is because the optimal expression conditions of a plasmid depend strongly on the genes of interest being expressed, therefore these results would be difficult to apply to other pCDFDuetIso-based plasmid. Nonetheless, we were able to provide evidence for a functional expression system in pCDFDuetIso.
Another experiment, which would have helped to further characterize the expression behavior of pCDFDuetIso’s promoters compared to those in pCDFDuet1, would have been to swap the location of the fluorescent proteins used in this experiment. That way, we would have been able to make a more quantitative statement about the promoter’s expression rates, but as with the experiment described above, we unfortunately lacked the time to perform further experiments.

qPCR

Quantitative PCR (qPCR) is an extension of conventional PCR that enables monitoring of double-stranded DNA (dsDNA) amplification by using a dye that binds to double-stranded DNA. This enables a qPCR machine to read the increase in dsDNA amount every PCR cycle. The main output value that is used to analyze qPCR results, the Ct (cycle threshold) value, is inversely proportional to the logarithm of the DNA template concentration. Under ideal conditions, amplification efficiency approaches a theoretical doubling (2^ⁿ) per cycle, which would be readable as the slope of the graph of Ct values plotted against the log of the concentration.

Aim

Our goal was to determine whether quantitative PCR (qPCR) could serve as a reliable method to detect retron-mediated genome editing events by measuring the mutation-induced disruption of primer binding sites in plasmid DNA. Specifically, we aimed to test if retron-produced single-stranded DNA (ssDNA) donors could generate targeted mutations within plasmid DNA at detectable frequencies.

Experimental design

To study the retron-mediated mutation rate in plasmid DNA, we designed two plasmids, derived from pMS366 [21]: pMS366-RibojRet, containing the RiboJ ribozyme fused to a retron (BBa_252J7V7J), and pMS-366-Ret, containing the same retron (BBa_25RSWBY3), without RiboJ. The construct RiboJ-Ret design includes the RiboJ (BBa_K4216041) self-cleaving ribozyme, which we hypothesized could positively impact mutation efficiency.

Both plasmid express their insert downstream of the single-stranded annealing protein CspRecT. The retron produces ssDNA designed to modify a primer binding sequence located adjacent to the origin of replication on pMS366. The RiboJ-Ret design includes the RiboJ (BBa_K4216041) self-cleaving ribozyme, which we hypothesized could positively impact mutation efficiency.

To monitor potential editing events, two qPCR primer pairs were designed:

• Target primer pair (q:mut): One primer overlaps the potential mutation site within the retron’s ssDNA target region and the reverse primer binds 140 bp upstream.
• Reference primer pair (q:ref): Binds to an unmodified region of the plasmid, serving as normalization control to confirm plasmid presence.

If retron-mediated integration occurred, it would disrupt the q:mut primer binding site, preventing efficient amplification. This disruption would result in a higher cycle threshold (Ct) value for q:mut relative to q:ref. The magnitude of this Ct shift would allow estimation of the proportion of edited plasmid in the sample.

The two retron constructs, pMS366-Ret and pMS366-RibojRet, were used to compare mutation outcomes.

Sub-Experiment 1: Annealing temperature optimization

Aim

First, we needed to test our primer pairs against a temperature gradient, aiming to find the optimal annealing temperature. At his temperature, the Ct values of both primer pairs q:mut and q:ref should be as close as possible.

Experimental Design

To identify the optimal annealing temperature for the q:mut and q:ref primer pairs, a qPCR thermal gradient experiment was performed. The temperature gradient block on the qPCR instrument enabled us to test of a range of annealing temperatures simultaneous in a single run. Identical reactions containing fixed primer concentrations and the same template DNA (pMS366) were prepared for each primer pair.

The tested annealing temperature range spanned approximately 55°C to 70°C, covering temperatures above and below the predicted melting temperature (Tm) of the primers. Ct values were recorded for each temperature, and amplification specificity was assessed by melt curve analysis and agarose gel electrophoresis.

Results

We only approximated the optimal annealing temperature at an unusually high 67˚C, where the Ct values of the two primer pairs were close, but still not equal. Melt curve analysis and gel electrophoresis confirmed the specificity of primer binding.

Sub-Experiment 2: Standard curve generation

Aim

Because qPCR Ct values depend on template concentration and reaction conditions, we generated standard curves for quantitative analysis.

Experimental Design

To establish a qPCR standard curve based on a bacterial culture containing plasmid pMS366, an overnight culture was grown and serially diluted in sterile medium to create a range of bacterial concentrations. The dilutions spanned several orders of magnitude to cover a broad dynamic range of template abundance. Each dilution was used directly as a qPCR template. The Ct values obtained from these serial dilutions were plotted against the logarithm of the dilution factor.
Purified pMS366 plasmid DNA was serially diluted across a range from 10 ng to 0.001 ng per 10 µl reaction, with three technical replicates per dilution. Ct values were plotted against the logarithm of DNA concentration to determine amplification efficiency and sensitivity.

Results

As Ct values are highly sensitive to initial conditions and are largely arbitrary, we needed to establish standard curves using bacterial culture, containing pMS366 and purified pMS366 plasmid DNA, respectively. The mean Ct values of the technical replicates were plotted against the logarithmic concentration of the serial dilution of the sample.
Initially, it was our intention to use diluted pMS366 overnight culture as a template. However, this approach produced inconsistent and variable data (Figure 26).
Miniprepped DNA provided consistent results (Figure 27).

Conclusion

We decided to only use miniprepped DNA for the qPCR experiments, because those yielded the most reliable results. Subsequent to the first standard curve, serial dilution of 2ⁿ was preferred and for each sample a technical quadruplicate was used (see protocols).

Sub-Experiment 3: Induction of retron system

Aim

We aimed to use qPCR to detect retron-mediated plasmid editing. Successful integration of mutation inducing ssDNAs at this site was expected to disrupt primer binding, leading to a relative increase in Ct values for the target primer pair (q:mut) compared to a reference region (q:ref). This proportion can be used to calculate the editing rate of the plasmid.

Experimental Design

Our pMS366 based constructs have the pBAD promoter, they express retron RNA, reverse transcriptase, and CspRecT when induced with arabinose. These genes are necessary for retron-based recombination. The retron ncRNA serves as a template for reverse transcription into ssDNA by the reverse transcriptase. CspRecT increases the annealing of the retron-produced ssDNA during replication.

We prepared samples for each plasmid design (pMS366 and pMS366-RibojRet) and induced them according to induction protocol. 20 µL of the 50 mL pre-culture was diluted with 20 mL of LB media containing chloramphenicol (25 µg/ml) and 0.2% arabinose. After 24 hours, this process was repeated to estimate the number of replications by doing a 1:1000 dilution twice. The liquid cultures were miniprepped after 48 hours of induction.

Results

We compared the resulting mean Ct values of the diluted miniprepped pMS366-RibojRet plasmid DNA produced in the bMS.346 with the linear regression from the standard curves (Figure 28). If analyzed with the pMS366 linear regression, a larger gap (mean Ct value) between the q:mut and q:ref primers could be misinterpreted as retron activity. Therefore, we created a standard curve for each retron plasmid. For each plasmid, a specific qPCR standard curve was constructed using purified DNA. Figure 29 shows the creation of individual standard curves for each plasmid to avoid false positives.

Initially, we initially assumed that plasmids with and without retron inserts would amplify identically. However, this led to a misleading Ct gap between q:mut and q:ref that superficially resembled an editing signal despite no true mutation (Figures 28, 29). This observation revealed that structural differences, including retron DNA sequences, could inherently shift Ct values. Even uninduced samples displayed differences between q:mut and q:ref that could be explained by differences in molar mass, suggesting that the observed variations reflected plasmid structure rather than actual mutations. We also compared induced plasmids to the serial dilution of the same plasmids before induction.

Conclusion

We incorrectly assumed that plasmids with and without retron inserts would amplify identically. This resulted in a misleading Ct gap between q:mut and q:ref, that superficially resembling an editing signal, despite there being no true mutation (Figures 28, 30).

This revealed that structural differences, including retron DNA sequences could, inherently shift Ct values. Even uninduced samples displayed differences between q:mut and q:ref could not be explained by differences in molar mass, suggesting that the observed variations reflected interference with plasmid structure rather than actual mutations.

We also compared induced plasmids to the serial dilution of the same plasmids before induction.

At least one sample per run was sent in for sequencing. The results showed an identical sequence with the negative control, without any uncertainty in mutation sites. These results suggest either low mutation rates or complete failure of mutagenesis, which is consistent with the ambiguous qPCR results.

Sub-Experiment 4: Single colony PCR

Aim

We decided that despite qPCR being viable to detect higher mutation rates, it is not reliable enough to detect low mutation rates. Instead, we aim to test the mutation, by screening induced bacteria by PCR. We will use the same primers, that anneal to specific unmodified plasmid sequences, that we used for qPCR experiments. The absence of bands on the resulting agarose gel would suggest a successful mutation, which then would have to have been confirmed by sequencing.

Experimental Design

We followed a standard colony PCR protocol, using the target primer pair (q:mut), and the reference primer pair (q:ref), to avoid having false positives results. After 48 hours of induction and growth, we isolated and processed the plasmid DNA for colony PCR.

Results

None of the more than 20 screened colonies exhibited detectable retron-induced mutations, as indicated by a PCR band of about 140 bp that was produced by primer binding (Figure 34).

Conclusion

The experimentally determined PCR-efficiencies of the q:mut primer pair range around 86%, while the efficiency of amplification using the q:ref primer pair remained at around 100%.

• pcr efficiency of pms366 with qR primer pair: 111 +/- 1.3 %
• pcr efficiency of pms366 with qT primer pair: 86 +/- 0.5 %
• pcr efficiency of pms366-Ret with qR primer pair: 109+/- 1.8%
• pcr efficiency of pms366-Ret with qT primer pair: 92 +/- 2.4 %
• pcr efficiency of pms366-RibojRet with qR primer pair: 94 +/- 1.7 %
• pcr efficiency of pms366-RibojRet with qT primer pair: 80 +/- 0.8 %

This presents a sub-par experimental setup as the analysis of qPCR results benefits greatly from close to equal PCR-efficiencies between the primer pairs.

Combined with qPCR data, these findings suggest that our retron operon did not produce any observable mutations under the tested conditions. Therefore, we need to redesign the retron operon in order to make it functional. In the absence of a verified mutation, we considered several possible explanation, as discussed with Alejandro Gonzáles-Delgado:

1. The retron sequence was inserted downstream of CspRecT within a single pBAD-controlled operon. This arrangement may have disrupted proper retron RNA folding, resulting in inefficient transcriptional termination efficiency, or it may have created structural hindrances that interfered with reverse transcription. Such interference could prevent the reverse transcriptase (RT) from synthesizing the ssDNA strand.
2. The presence of an unused rpoB-targeting retron donor present in pMS366 could compete with reverse transcriptase activity or interfere with retron complex assembly.

Sub-Experiment 5: Integration into the Repebody

Aim

Our objective was to assess whether retron-mediated integration could occur within the variable leucine rich repeats (LRRVs) of the ssDNA-optimized-mCherry-binding-Repebody (BBa_25IQCJ3B) gene.

Experimental Design

Instead of measuring Ct shifts by qPCR, we opted for a qualitative assay, without determining the efficiency, as a robust proof of concept. We tested for integration by PCR amplification of the Repebody region following induction. The goal was to detect sequence insertions generated by retron activity at specific repeats.

We co-expressed two distinct retron constructs BBa_25DCBAZ7,BBa_254TXZJ8, in E. coli bMS.346. Each construct was designed to create a specific mutation in the LRRVs of the ssDNA-optimized-Repebody caused by ssDNA insertion into the Repebody gene. The Repebody itself was transformed into E. coli bMS.346 in the plasmid pCDFDuetIso, ensuring the positioning of the Repebody on the lagging strand during plasmid replication. This increases efficiency of retron integration.

After co-transforming the retron construct with the Repebody in E. coli bMS.346, the cells were induced with arabinose and the plasmids were miniprepped. PCR of the purified DNA was carried out with a set of primers. Four reference primers were used to amplify each of the unmodified variable repeats, serving as controls for unedited regions. In addition, two mutation-specific primer pairs were employed to amplify potential integration products. Successful retron-mediated insertion at one or both target sites would produce distinct PCR band patterns, reflecting modified repeats. The mutation-specific primers would then also bind to these modified repeats.

Results

Plasmid DNA was analysed before induction showing clear single bands per sample (Figure 34). This result represents an inactive mutation system, as the retrons do not mutate the Repebody sequence at the primer binding site.

After induction both integration 1 and integration 2 show only unspecific primer binding also seen pre-induction. It was expected to see multiple bands, which would suggest that the retrons mutated the Repebody sequence enabling the primers to anneal. The absence of bands suggest that the mutation of the Repebody sequence did not function as intended.

Final Conclusion

Although we did not obtain conclusive results, we know the next steps to troubleshoot our designs. First, we will determine and minimize the interference of the RpoB donor on the expression efficiency of the retrons. Then, we will analyze the ability of the retrons to induce rifampicin resistance by mutating the RNA polymerase. We will analyze successful mutations by plating the cells onto LB plates with chloramphenicol and rifampicin. The presence of many resistant colonies would indicate that our induction protocols are functional. Designs without the RpoB donor and our desired retron donor will be screened as before. We will also assess a design with our desired retron donor upstream of the reverse transcriptase. Additionally, we will test a design with extended homology regions of 27 to 30 bp that introduces fewer substitution mutations.

BACTH

Early into the research about bacterial two hybrid systems, we started focusing on the BACTH system. The BACTH system is based on the adenylate cyclase of the pathogenic bacterium Bordetella pertussis. Normally, this enzyme catalyzes the formation of cAMP from ATP in human cells targeted by the bacterium [1]. However, this adenylate cyclase can be split into three parts, a haemolytic domain, which is not used in the BACTH system, the active site (T25 subunit) and a calmodulin binding site (T18 subunit). If the active site and the calmodulin binding site are brought in spatial proximity, the activity of the enzyme is restored. In the BACTH system, this is achieved by fusing interacting proteins to each subunit. If the subunits are able to interact, a bacterium without an endogenous adenylate cyclase can survive on minimal media with added maltose or lactose [2] [3]. Additionally it is then able to cleave X-Gal resulting in a blue color of the colony. Further information can be found in the BACTH background section of our description page.

Qualitative BACTH assay using the unmodified-Repebody and the ssDNA-optimized-Repebody as well as mCherry as the target

Aim

We aimed to test whether the BACTH system is suitable for detecting interactions of a Repebody and its specific binding partner. Successful detection of the interaction between our Repebodies and mCherry would serve as a proof of concept. We aimed to find out if the size of the Repebody or the position of its binding site would hinder the interaction of the two adenylate cyclase subunits, thereby rendering the BACTH system non-functional. These factors also dictated our decisions in plasmid design.

Experimental design

We tested both the unmodified-mCherry-binding-Repebody (BBa_25X0LA8K) and the ssDNA-optimized-mCherry-binding-Repebody (BBa_25IQCJ3B) against mCherry as the target and eGFP as a negative control. We chose eGFP because of its similar size and structure to mCherry, while there is no interaction reported between the Repebody and eGFP [1]. Due to time constraints, we decided to fuse the Repebodies exclusively to the T25 subunit (BBa_25H9D35L, BBa_255QHYDE, BBa_25L4RPGJ, BBa_25QRQ2FK, BBa_25FKOX5Q, BBa_2540IZIE, BBa_252LEBS1, BBa_25HK2H4O), while mCherry and eGFP were fused to the T18 subunit (BBa_25EO6BEC, BBa_25TAHASE, BBa_259SSWC8, BBa_253I2YTC, BBa_2584YQ8N, BBa_25TTPMRP, BBa_25V2YRYF, BBa_25H5OOE1). We did not expect different results when fusing the Repebodies to the T18 subunit and mCherry or eGFP to the T25 subunit. Nevertheless, these combinations could also be tested, to find the best one. The combinations we created can be found in the protocol of the BACTH assay.

For the assay we followed the steps disclosed in Dr. Bouveret’s review [3], a modified version provided by Dr. Wallner and the protocol of the Euromedex kit [4]: After co-transforming the plasmids into E. coli BTH101, we set up overnight cultures of multiple colonies and induced them with IPTG (0.5 mM). After approximately 18 hours of incubation at 30° C and shaking, 2 µl of the overnight culture were dropped onto M63/Maltose plates with X-Gal (40 µg/ml) and IPTG (0.5 mM). The samples were put on the same plates as their respective controls to ensure easier comparison between them. The plates were photographed after 24 and 48 hours of incubation at 30° C.

Results

After 24 hours of incubation, the positive control showed a clear and intense blue colour. Nine out of sixteen Repebody-mCherry combinations developed a light blue colour, while all Repebody-eGFP combinations as well as the other negative controls remained translucent. The blue colonies indicate the successful reconstitution of the adenylate cyclase activity and therefore, a functional interaction between the Repebody fused to T25 and target mCherry fused to T18 subunit. After 48 hours the blue colouring of the positive control and the Repebody–mCherry pairs intensified, while the controls remained translucent. This time-dependent intensification further supports the reliability of the interaction signal.

Figure 36 presents a representative selection of four Repebody-mCherry constructs alongside controls to visualize these patterns. We also prepared a master control plate that included all individual control conditions except for the co-transformation of Repebody and eGFP constructs.

The results of the BACTH assay are summarized in Table 1:

Table 1: Results of the BACTH assays. pKNT25-unmodified-Repebody (N-unmodified-Repebody) pKT25-unmodified-Repebody (C-unmodified-Repebody), pKNT25-ssDNA-optimized-Repebody (N-ssDNA-Repebdy) and pKT25-ssDNA-optimized-Repebody (C-ssDNA-Repebdy), were co-transformed with either pUT18-mCherry (N-mCherry), pUT18-EGFP (N-EGFP), pUT18C-mCherry (C-mCherry) or pUT18C-EGFP (C-EGFP). The same was also done with a (GGS)3 linker. ‘/’ indicates the result of this assay was inconclusive. It is likely two samples have been mixed up.

	N-mCherry	N-EGFP	C-mCherry	C-EGFP
N-unmodified-Repebody	-	-	-	-
C-unmodified-Repebody	-	-	+	-
N-ssDNA-Repebody	+	-	/	-
C-ssDNA-Repebody	-	-	+	-
(GGS)3	N-mCherry	N-EGFP	C-mCherry	C-EGFP
N-unmodified-Repebody	+	-	+	-
C-unmodified-Repebody	-	-	-	-
N-ssDNA-Repebody	+	-	+	-
C-ssDNA-Repebody	-	-	+	-

Table 2: Results of the BACTH assays of the general controls. pKNT25-unmodified-Repebdy (N-unmodified-Repebdy) pKT25-unmodified-Repebdy (C-unmodified-Repebdy), pKNT25-ssDNA-optimized-Repebdy (N-ssDNA-Repebdy), pKT25-ssDNA-optimized-Repebdy (C-ssDNA-Repebdy), pUT18-mCherry (N-mCherry), pUT18-EGFP (N-EGFP), pUT18C-mCherry (C-mCherry) and pUT18C-EGFP (C-EGFP), were co-transformed with pUT18, pKNT25 or pKNT25. A positive control with pKT25-zip and pUT18C-zip was made. ‘/’ suggests that no assay was performed with this combination.

General Controls	pKNT25	pKT25	pUT18	pUT18C-zip
N-unmodified-Repebody	/	/	-	/
C-unmodified-Repebody	/	/	-	/
N-ssDNA-Repebody	/	/	-	/
C-ssDNA-Repebody	/	/	-	/
N-mCherry	-	/	/	/
C-mCherry	-	/	/	/
N-EGFP	/	-	/	/
C-EGFP	/	-	/	/
pKT25-zip	/	/	/	+
pKNT25	/	/	-	/

Conclusion

The results suggest that the combination of Repebody and mCherry fusion proteins could reconstitute adenylate cyclase activity, as seen in blue colonies. Such reconstitution was not observed with Repebody and eGFP pairs. Interestingly, there is no clear trend indicating which combination yields better results. Incorporation of (GGS)3 linker does also not appear to affect the interaction efficiency. It is possible that the increased flexibility of the fusion proteins through a linker does not facilitate the interaction between the subunits, or that the (GGS)3 linker lacks the necessary length to influence the interaction capability. Both Repebodies appear to be functional, yet the ssDNA-optimized-Repebody seems to work better than the unmodified one, as more assays using it yield positive results (see Table 1).

Surprisingly, certain combinations yield positive results with the ssDNA-optimized-Repebody, but produce negative results with the unmodified variant. These combinations require more investigation and experiments to determine the reason for our results. Possible causes might be varying expression of the constructs, protein folding efficiency, systematic error in the experiment or other unknown issues.

Although this experiment cannot prove a general applicability of the BACTH system for the screening of vast Repebody libraries, since we only tested small volumes and the bacteria in our culture expressed the same Repebody, it nevertheless lays a foundation for further development and optimization efforts to address challenges such as steric hindrance from more complex protein targets.

While these findings demonstrate that the BACTH system can successfully detect Repebody–target interactions, it was important to confirm that the negative results truly reflected a lack of binding, rather than potential issues with protein expression. To address this, we performed a dot blot assay to verify the expression of all fusion proteins and to validate the robustness of the BACTH system within our toolkit.

BACTH dot blot

Aim

The results from the BACTH assay suggest that the system is capable of selecting for Repebodies that bind a desired protein of interest. To strengthen this finding, we aimed to confirm these results by verifying the expression of all parts by a dot blot. This was crucial to ensure that negative BACTH results were not caused by a lack of protein expression. We especially wanted to show that eGFP as the negative control was expressed, as that could also be the reason for these construct combinations yielding negative results. A dot blot functions similar to the western blot, but without separating the proteins by size first.

Experimental design

We decided to use two construct combinations for the dot blot that had been shown to function in the BACTH system beforehand. We selected the combinations of pKT25-unmodified-Repebody-6xHis + pUT18C-mCherry-HA/pUT18C-eGFP-HA and pKT25-ssDNA-optimized-Repebody-6xHis + pUT18C-mCherry-HA/pUT18C-eGFP-HA. The pKT25 + pUT18C were used as a negative control and mCherry with an His- or HA-tag as a positive control. Overnight cultures (50 mL) from the same colonies used in the BACTH assay were lysed. The cell lysate was spotted onto two membranes. One membrane was probed with anti-His antibodies, the other with anti-HA antibodies. Detection was performed with an HRP-linked secondary antibody.

Results

Signals were detected for all samples on both anti-HA and anti-His blots (Figure 37), except for the negative control, which confirmed the expression of the target and Repebody fusion proteins. However, the combinations containing eGFP appeared weaker compared to their mCherry counterparts. As expected, negative controls remained undetectable, thus confirming the specificity of the antibody signals and indicating no background expression.

Discussion:

The close to non visible signal observed in the EGFP samples is consistent with the known regulatory mechanism of the lac promoter, which is activated by cAMP produced upon adenylate cyclase reconstitution. Strong interactions, such as those between Repebody and mCherry, lead to higher cAMP levels, enhanced lac promoter activity, and consequently greater protein expression. Conversely, weaker or absent interactions, as seen in Repebody–EGFP pairs, result in lower cAMP production and reduced expression levels. This feedback can explain why the EGFP controls produce weaker signals despite confirmed protein expression. As expected the negative control without any tagged protein remains unmarked by the antibodies. Ideally we would have wanted to repeat that experiment also with construct combinations, that yielded a negative result in the BACTH essay and with decreased protein concentration of the positive control, to make the EGFP controls more visible.

Conclusion

The weaker signal observed in the eGFP samples is consistent with the known regulatory mechanism of the lac promoter, which is activated by cAMP produced upon adenylate cyclase reconstitution. Strong interactions, such as those between Repebody and mCherry, lead to higher cAMP levels, enhanced lac promoter activity, and consequently greater protein expression. Conversely, weaker or absent interactions, as seen in Repebody–eGFP pairs, result in lower cAMP production and reduced expression levels. This feedback can explain why the eGFP controls produce weaker signals despite confirmed protein expression. As expected the negative control without any tagged protein remains unmarked by the antibodies. Ideally we would have wanted to repeat that experiment also with construct combinations, that yielded a negative result in the BACTH essay and with decreased protein concentration of the positive control, to make the eGFP controls more visible.

BACTH serial dilutions

Aim

After demonstrating that the BACTH system is able to detect the interaction between a Repebody and its target protein in a culture, where all bacteria express these two proteins, we wanted to see whether the system could also select single bacteria with a functional BACTH system from a larger culture of non-functional cells.

Experimental design

We used the same construct combination and cultures as in the dot blot experiment. The OD 600 of every culture was measured and then diluted to OD600 = 1. Then the Repebody-mCherry combinations were serially diluted until a ratio of 1:1,000,000 (Repebody-mCherry : Repebody-eGFP) was reached. The same was done with the positive BACTH control, which was diluted in the negative control. In addition to these dilutions, an undiluted sample and a negative control were included. All samples were then plated on M63/maltose plates containing ampicillin and kanamycin. These plates were kept at 30°C for around 60 hours. The plates were then replicated onto LB plates with X-Gal and IPTG, using autoclaved cotton velvet and a custom made rig. The LB plates were incubated at 30° C overnight. We expected a lot of blue colonies to appear on the plates with the undiluted functional Repebody-mCherry combination, with less blue colonies with increasing dilution. We anticipated not seeing any white colonies, since bacteria without a functional BACTH system would starve on the minimal media plates, due to their inability to use maltose as an energy source.

Results

The results of these experiments, shown in Figure 38, were inconclusive. Around two-thirds of the plates show blue colonies. Unexpectedly, many plates also had white colonies. Furthermore, some negative control plates showed blue colonies despite lacking positive interaction constructs.

Conclusion

It was expected that the LB plates with the undiluted Repebody-mCherry construct combination would be covered by blue colonies. These numbers should decrease the more diluted the samples were. White colonies should not appear. On the negative control plate no colonies should have grown. Based on previous BACTH assay data, we know that the Repebody-mCherry constructs used in this experiment reliably interact, producing blue colonies. We also know that the Repebody-eGFP constructs do not interact and do not result in blue colonies. The bacteria with these colonies should not be able to survive on the M63/maltose plates and henceforth there should be no white colonies on the LB plates. While unlikely, the appearance of white colonies on the LB plates and blue colonies on the negative control plates could be explained by contamination, introduced in the plate replication step. Contamination could either originate from contaminated velvet, although it was autoclaved, or from the M63/maltose plates. The M63 plates were incubated at 30° C for almost 60 hours, which could have led to growth of contaminating bacteria. It would require contamination by two bacterial strains, to cause the results we got. One of these would have to be unable to cleave X-Gal using a &beta-galactosidase, thereby remaining white on the X-Gal containing media, while the other strain should not suffer from that problem. Additionally both strains would have to be resistant against ampicillin and kanamycin. Another possibility is that the concept of the experiment is flawed, although we don’t see a reason to assume this.

Further categorisation of the unexpected colonies, like replating them on fresh M63/maltose and LB plates with antibiotics, IPTG and X-Gal, and repetition of the experiment would be necessary to resolve these issues. However, due to time constraints, these additional experiments were not conducted. Consequently, we cannot conclusively demonstrate that the BACTH system can isolate single bacterial cells with functional protein interactions from a mixed culture. Further testing is required to obtain reliable results.

Targeted protein degradation

In this section of our project, we aim to test the functional capabilities of the Repebody-degrader to bind and target mCherry-fusion protein to lysosome for degradation. We decided to focus on the autophagy-lysosomal pathway by bringing the POI(Protein of interest) into close proximity with p62 via a multifunctional autophagy adaptor protein.

p62, also known as Sequestosome 1 (SQSTM1), is a crucial autophagy cargo receptor that mediates selective autophagy by sequestering ubiquitinated cargo and delivering it for degradation. This protein is fundamental in initiating the process. It facilitates the formation of the phagophore, the initial structure that ultimately becomes the autophagosome. Following by engulfment, the autophagosome fuses with the lysosome to form the autolysosome, where the cargo is degraded. (Figure 39). [5, 6, 7]. p62-induced degradation encompasses a wide range of selective autophagy processes, including aggrephagy, mitophagy, xenophagy, ER-phagy, perophagy, lipophagy, and fluidophagy. In our project, we envisioned our degrader construct utilizing A1E, a recently developed anti-p62 nanobody, to bind to p62 for targeted protein degradation. [8].

To further investigate the efficacy range of our Repebody-degrader construct, we decided to quantify the degradation by implementing the mito-QC reporter system into our design. The mito-QC reporter system utilizes mCherry-GFP tag fused to targeted proteins to form the fusion protein. During autophagy, the fusion protein is exposed in the acidic environment. This results in the pH-sensitive GFP signal being quenched or degraded, while the pH-stable mCherry signal remains fluorescent. Therefore, by using different method such as confocal microscopy with FIJI or flow cytometry to analyze the ratio of mCherry to GFP fluorescent intensity, one can effectively quantify the degradation [9]. We then design two distinct substrate types of mCherry-fusion proteins. Respectively, they are EGFP-mCherry fusion protein as the soluble cytosolic protein and mCherry-EGFP-Fis1 that anchored to the mitochondria. In both fusion proteins, the mito-QC reporter system could be used to quantify degradation.

Quantification of degradation

Mitophagy-Degradation of mitochondria

To quantify mitophagy, the selective lysosomal degradation of mitochondria by the autophagy pathway, we design the construct consisting of a EGFP-mCherry tag fused to Fis1. This would allow the fusion protein to localize to the outer mitochondrial membrane (OMM) via the targeting sequence-Fis1. By analyzing the ratio of mCherry to EGFP fluorescent intensity, we can effectively quantify the degradation and successful lysosomal delivery of mitochondria.

Degradation of Soluble Target

To test the degrader’s ability to selectively degrade soluble targets, we used a soluble mCherry-EGFP fusion protein. Since this soluble target is also delivered to the lysosome for degradation via the p62-autophagy pathway, we expect the same fluorescence readout as the mito-QC system. In essence, the mCherry to EGFP fluorescent ratio could serve as to quantify lysosomal degradation for the soluble targets.

Design

We employed a multi-faceted approach utilizing different techniques to analyze the efficiency and mechanism of the Repebody-degrader.
First, confocal microscopy was performed using U2OS cells, which were chosen for their large, flat morphology and strong adherence. This allows us to perform high-resolution imaging and precise analysis of fluorescent signal co-localization at the single-cell level. Secondly, flow cytometry using HeLa cells was implemented for high-throughput population analysis, enabling the accurate quantification of transfection efficiency and overall protein degradation across numerous cells. We ensured the data accuracy by gating first on live cells then single cells. Afterwards, we gated on EGFP and mCherry signals to track target protein behavior (e.g., lysosomal delivery) and mTagBFP2 fluorescence to monitor the presence of the degrader construct. Lastly, we performed western blot analysis for assessment of the total expression level of our construct and allowed for the detection of degradation intermediates, as evidenced by the presence of multiple bands.

Target expression via confocal microscopy

Aim

Proof expression of the target constructs EGFP-mCherry, EGFP-mCherry-Fis1 and mCherry-EGFP-Fis1 in mammalian cells. Assess the localization of the mitochondrial targets.

Experimental Setup

We designed the following mitochondrial targets: EGFP-mCherry-Fis1 and mCherry-EGFP-Fis1. We decided to use two constructs with different fluorophore order, as we did not know how the construct order with mCherry’s accessibility could interfere with the ability of the Repebody to bind mCherry. To verify degradation of soluble proteins, we designed the EGFP-mCherry target. All target constructs, including the soluble EGFP-mCherry, were cloned into the pcDNA3.1 backbone.
To verify the expression of these constructs in mammalian cells, we transfected the target constructs into U2OS cells for confocal microscopy imaging and into HeLa cells for flow cytometry.
U2OS cells were seeded onto coverslips previously coated with poly-L-lysine. The targets were then transiently transfected into U2OS cells via PEI transfection. After 36 hours, we fixed the samples with 2% PFA and stained the cells’ nuclei with DAPI. Finally, the samples were imaged using a Nikon C2 confocal microscope running on NIS-Elements as a software. Z-stacks of the blue, green and red channels were taken.
After imaging, an average intensity projection (AIP) was conducted on the z-stacks using ImageJ / Fiji. To make the Figures colorblind friendly, the LUTs were adjusted to change blue to cyan, green to yellow, and red to magenta.

Results

The EGFP-mCherry was successfully expressed in mammalian cells (Figure 40). Compared to EGFP (Figure 41) and mCherry (Figure 42) single transfections, the construct is expressed cytosolically, including some signal in the nucleus.

When observing the Fis1 mitochondrial targets, we compared the localization of mito-TurboGFP to our constructs. We observed that mCherry-EGFP-Fis1 localizes as expected (Figure 43), exhibiting a net-like structure similar to that of mito-TurboGFP (Figure 44). EGFP-mCherry-Fis1 however, did not localize as expected and was only distributed diffusely throughout the cytosol (Figure 45).

Discussion and Conclusion

The soluble target EGFP-mCherry can be expressed and behaves as expected when compared to single transfections of these fluorophores.
The mitochondrial target mCherry-EGFP-Fis1 can be expressed in U2OS cells and was localizing as expected, manifesting a net-like structure similar to that of mito-TurboGFP. On the other hand, EGFP-mCherry-Fis1 was not localizing on the OMM as expected. For this reason, we decided to move forward with the mCherry-EGFP-Fis1 construct as our mitochondrial target.
It is worth mentioning, that while both Fis constructs and mito-TurboGFP are supposed to localize to mitochondria, Fis localizes onto the outer mitochondrial membrane. On the other hand, mito-TurboGFP uses a mitochondrial targeting sequence (MTS) to localize GFP on the inner mitochondrial membrane [10, 11]. It is therefore to be expected for the mitochondrial localization to slightly differ from one another.
Furthermore, the transfection efficiency was very low. This could be due to transient expression with PEI not being efficient enough. Something to further explore in the future could be to generate a cell line stably expressing the targets [12].

Target expression via flow cytometry

Aim

Transfect and assess the expression of the target constructs EGFP-mCherry and mCherry-EGFP-Fis1 in mammalian cells, using Flow cytometry.

Experimental Setup

The properly localizing target constructs EGFP-mCherry and mCherry-EGFP-Fis1 were transiently transfected together with empty pcDNA3.1 in HeLa cells.
At 38 hours pos-transfection, samples were prepared and measured using flow cytometry, as described in the protocol “Cell preparation for FACS and WB”

Results

In both analyzed targets, a distinct continuous diagonal distribution extending from quadrant Q4 (non-fluorescent population) to quadrant Q2 (double-positive population) is observed. This pattern reflects the simultaneous detection of EGFP and mCherry fluorescence within the same cells. As the EGFP signal intensity increases, the mCherry signal shows a proportional rise, indicating that both fluorophores are functional and emitted in a consistent ratio as part of the same fusion protein. The minimal cell populations detected in quadrants Q1 and Q3 further confirm that the two fluorophores are expressed together, with no evidence of independent signal emission.

Discussion and Conclusion

Based on this pattern, we can conclude that our degradation assay, which relies on comparing EGFP to mCherry fluorescence, can be employed using the cloned targets.

Degrader expression via confocal microscopy

Aim

Prove expression of Repebody degrader construct on mammalian cells.

Experimental Setup

We verified that our degrader can be expressed in mammalian cell by transfecting the following constructs:

• Repebody-mTagBFP2-HA-A1E
• Repebody-mTagBFP2-HA

We included the fluorescent protein mTagBFP2 in the constructs to observe expression via mTagBFP2 fluorescence. To compare localization of the degrader constructs, we also conducted single transfections of mTagBFP2 and p62-mCherry, respectively.

A1E binds to p62 [13], confocal imaging of the degrader construct should therefore reveal localization in autophagic puncta, similar to that of p62-mCherry. While the target without A1E should occur evenly distributed throughout the cytosol, as it does not have a target to bind to.
U2OS cells were seeded onto coverslips coated with poly-L-lysine. The constructs were then transiently transfected in U2OS cells using PEI transfection. After 36 hours, we fixed the samples with 2% PFA. DAPI staining was not conducted on the samples containing degrader constructs, since this would interfere with the mTagBFP2 fluorescence. This nuclei staining was only conducted on p62-mCherry samples.
Finally, the samples were imaged using a Nikon C2 confocal microscope running on NIS-Elements as a software. Z-stacks of the blue, green and red channels were taken.
After imaging, an maximum intensity projection (MIP) was conducted on the z-stacks using ImageJ / Fiji, as to better represent the autophagic puncta. To make the Figures colorblind friendly, the LUTs were adjusted to change blue to cyan, green to yellow, and red to magenta.

Results

Confocal microscopy of the Repebody (RB) degrader construct without A1E (RB-mTagBFP2-HA, Figure 47) revealed no specific localization of this construct, similar to the expression of single-transfected soluble mTagBFP2 (Figure 48).

When assessing proper expression and localization of our Repebody degrader (RB-mTagBFP2-HA-A1E), we additionally observed p62-mCherry expression and localization of autophagic puncta (Figure 49).

Finally, we proved that our Repebody degrader construct RB-mTagBFP2-HA-A1E can successfully be expressed in mammalian cells. This construct also manifested a similar localization in autophagic puncta (Figure 50).

In general, the mTagBFP2-containing samples were harder to image, since there was no DAPI staining to easily find the samples on the microscopy slides. With these constructs, the transfection rate was also low, as we were working with transiently transfected constructs using PEI.

Discussion and Conclusion

Our degrader constructs can be expressed in mammalian cells. However, the transfection method needs to be improved, as the transfection rate is very low. Transfection rate could be optimized in the future by generating a stable cell line, in which the expression of the degrader construct can be induced with tetracycline, via a Tet-On-System [7].
p62-mCherry was localizing as expected, and similarly to the data reported in literature [14, 15]. The Repebody-degrader construct containing A1E manifested a similar localization to p62-mCherry. This could indicate that this construct, more specifically A1E, is successfully binding to endogenous p62.
Since the construct is binding to p62, it could also be targeted by the autophagy machinery and be degraded. This could explain why not many mTagBFP2-positive cells could be observed.
We observed the samples 36 hours after PEI transfection, as advised by our stakeholder. But expression might take longer due to the large vector size [16].
Few cells expressing the degrader constructs were found during imaging. However, these results could indicate that the degraders had already been degraded.
Finally, the fixation process can also interfere with the signal of fluorescent proteins. Therefore, we could try exploring live cell imaging for microscopy in the future [17].

Degradation ability via flow cytometry experiments

Aim

Assess potential degradation of soluble target EGFP-mCherry and the mitochondrial-anchored target mCherry-EGFP-Fis1 by Repebody-mediated autophagy under nutrient starvation.

Experimental Setup

The degrader construct Repebody-mTagBFP2-HA-A1E was transfected together with the target constructs EGFP-mCherry and mCherry-EGFP-Fis1 to assess possible degradation. As controls, the Repebody construct lacking the p62-nanobody (Repebody-mTagBFP2-HA) was transfected with both targets. Each target was also transfected with the empty vector pcDNA3.1 as an additional reference.
For each double transfection, a mastermix was prepared and split to transfect two wells of a 6-well plate.
At 36 hours post-transfection, one well of each pair was subjected to starvation treatment by replacing DMEM with Earle’s Balanced Salt Solution (EBSS) starvation medium. Two hours later, the second well of each pair was also switched to EBSS for 0 hours, ensuring both samples were treated the same way.
Following starvation, cells were harvested and each sample was split into two Falcon tubes. One aliquot was prepared for Western blot (WB) analysis, and the other for flow cytometry. The WB and Flow Cytometry samples are therefore identical in terms of transfection, treatment, and timing. (See Protocol FACS/WB: Repebody degradation + EBSS treatment)

Results

Repebody-mTagBFP2-HA exhibits a strong mTagBFP2 fluorescence signal throughout all four samples, indicating successful expression of the construct. In contrast, Repebody-mTagBFP2-HA-A1E displays a markedly reduced mTagBFP2 intensity compared to Repebody-mTagBFP2-HA. Nonetheless, the mTagBFP2 signal of Repebody-mTagBFP2-HA-A1E remains higher than that observed in the target-only control, confirming expression of the construct (Figure 52).
The minor fluctuations seen when comparing samples at 0 and 2 hours of EBSS treatment suggest that starvation did not lead to an altered degradation and thus mTagBFP2 amount.

When comparing fluorescence detection of the targets across EGFP and mCherry channels, a consistently higher number of mCherry-positive cells was observed in all samples. Additionally, the overall fluorescence profiles of EGFP and mCherry followed similar trends. Notably, samples subjected to EBSS-induced starvation for 2 hours exhibited an increased number of fluorescently positive cells in both channels, with the exception of the co-transfected samples Repebody-mTagBFP2-HA + EGFP-mCherry and Repebody-mTagBFP2-HA-A1E + mCherry-EGFP-Fis1.

When analyzing the mCherry/EGFP mean fluorescence intensity ratio (Figure 54), which serves as an indicator of lysosomal targeting through EGFP quenching, as described in the introduction, we observed relatively constant values across all samples.
No significant difference was detected between the target-only control constructs, Repebody-mTagBFP2-HA and Repebody-mTagBFP2-HA-A1E, whose bar chart profiles appeared nearly identical. The 2 hour EBSS starvation treatment also did not appear to affect the fluorescence ratio, as the 0 and 2 hour samples displayed comparable values.

Discussion and Conclusion

We can therefore conclude that neither of the target constructs underwent detectable degradation. However, the consistently low mTagBFP2 emission observed for A1E remains intriguing and suggests that the construct may undergo rapid degradation prior to engaging with its intended binding partners.

Western blot of degradation

Aim

Assess potential degradation of soluble target EGFP-mCherry and the mitochondrial-anchored target mCherry-EGFP-Fis1 by Repebody-mediated autophagy under nutrient starvation.

Experimental Setup

The degrader construct Repebody-mTagBFP2-HA-A1E was transfected together with the target constructs EGFP-mCherry and mCherry-EGFP-Fis1 to assess possible degradation. As controls, the Repebody construct lacking the p62-nanobody (Repebody-mTagBFP2-HA) was transfected with both targets. Each target was also transfected with the empty vector pcDNA3.1 as an additional reference.
For each double transfection, a mastermix was prepared and split to transfect two wells of a 6-well plate.
At 36 hours post-transfection, one well of each pair was subjected to starvation treatment by replacing DMEM with Earle’s Balanced Salt Solution (EBSS) starvation medium. Two hours later, the second well of each pair was also switched to EBSS for 0 hours, ensuring both samples were treated the same way.
Following starvation, cells were harvested and each sample was split into two Falcon tubes. One aliquot was prepared for Western blot (WB) analysis, and the other for flow cytometry. The WB and Flow Cytometry samples are therefore identical in terms of transfection, treatment, and timing. (See Protocol FACS/WB: Repebody degradation + EBSS treatment)

Results

We would expect a band at approximately 72 kDa if the Repebody–degrader construct (Repebody-mTagBFP2-HA-A1E) was fully expressed. However, this band was not detected in the western blot. The multiple bands observed in panel (a.) may represent degradation products of the Repebody–degrader, and their intensity increases after cell starvation in EBSS medium. The degradation of the Repebody–degrader could also explain the absence of the mTagBFP2 signal in the FACS experiments.

The Western blot results suggest non-specific binding of the antibodies, as we failed to detect a clear band corresponding to our Repebody-mTagBFP2-HA constructs at the expected molecular weight of 57 kDa. This is further supported by the slightly higher signal observed on the right side of the blot (b), which should have been devoid of any HA-tagged proteins.
Therefore,in conclusion, although flow cytometry confirmed the presence of the degrader construct via mTagBFP2 fluorescence, detection using the HA-tag failed entirely. A plausible explanation for this is the inaccessibility of the HA epitope in the Repebody-mTagBFP2-HA construct for anti-HA-antibody. Notably, the Repebody-degrader construct (Repebody-mTagBFP2-HA-A1E) includes a linker positioned between mTagBFP2 and the HA tag, which leads us to believe that the multiple bands presented in (Figure 55) for the degrader construct might due to protein degradation and not unspecific binding of antibody.

When looking into western blot for Repebody-degraders and pcDNA probed with Anti-mCherry (Figure 67) and compare the two side (c.) and (d.), there’s a reduction of the EGFP-mCherry target samples under different EBSS treatment time. This by itself can’t claim to be a successful degradation, as simply the increased metabolic burden from expressing the Repebody-degrader could also result in this. To confirm whether this is from the degradation, we have to look into the anti-mCherry staining of the non-degrading negative control, Repebody-mTagBFP2-HA (Figure 68), and see if the same reduction is observed.

The western blot (Figure 59) revealed the same reduction of EGFP-mCherry construct as observed in the previous western blot (Figure 58 (c)). This indicates the absence of signals might not be due to degradation, but rather errors with loading the gel or sample preparation.
To further investigate how degradation looks like when analyzed via western blotting. We transfected the p62-mCherry construct. This construct should be degraded with the presence of p62 as part of the natural degradation machinery.

p62 in theory has a molecular weight of around 50 kDa, but is known to run consistently at 62 kDa, as its name. When fusing to mCherry (around 28 kDa), the construct is expected to run at around 90 kDa. For this transfection, we would assume some level of degradation of the target protein is occurring, as p62 is known to facilitate this process. However, as observed in (Figure 60), the degradation isn’t fast enough to completely break down all the expressed constructs.
After seeing how the potential degradation rate could be, we would therefore expect to see a distinct band corresponding to the full molecular weight of the target construct, even when degradation happened for our Repebody-degrader experiments (Figure 55). Since the full construct was not observed, we would conclude that the multiple bands observed in this Figure are likely not true degradation products but the artifacts caused by low signal strength and non-specific binding of the HA antibody.

Discussion and Conclusion

1. Discrepancy of signals from Repebody-mTagBFP2-HA construct in FACS and WB
   One serious discrepancy exists between our flow cytometry and the western blot data concerning the expression of the Repebody-mTagBFP2-HA construct. While the flow cytometry shows mTagBFP2-positive rate of approximately 35% (Figure 51), indicating the successful protein expression that should be detectable, no obvious signals were detected on the western blot probed with anti-HA antibodies(Figure 56). This conflict is likely due to the lack of a linker between the mTagBFP2 fluorescent protein and the HA tag. We hypothesize that the C-terminal amino acids of mTagBFP2 sterically block the small HA epitope, making it inaccessible to the antibody. This corresponds to the degrader expression western blot we did, as it also showed no clean bands after 60 hours after transfection (Figure 57). This hypothesis is also supported by our literature review, which showed that published constructs using mTagBFP2-HA consistently included at least two intervening amino acids, suggesting that incorporating a short linker is standard practice to ensure the HA tag is functionally exposed. The failure of our blotting procedure is therefore likely attributed to this structural constraint, not a lack of protein expression. [13, 19, 20].
   To address the unexpected behavior of our non-degrading binding control-Repebody-mTagBFP2-HA in the western blot, a better approach would be to mutate the A1E nanobody in the control construct to render it non-functional while preserving the overall size and structure. This modification would create a more comparable negative control than simply omitting A1E.
   
   
2. Several bands observed in Repebody-mTagBFP2-HA-A1E
   Analysis of the Western blot following co-transfection of Repebody-mTagBFP2-HA-A1E and its targets revealed multiple bands. These bands could represent either non-specific antibody binding or successful lysosomal degradation of the construct (Figure 55).
   To distinguish between these possibilities, we compared the result to a blot of the control construct-Repebody-mTagBFP2-HA (lacking the A1E nanobody). The Repebody-mTagBFP2-HA blot showed very low signal for both the HA-tagged protein and the pcDNA3.1 (negative control) (Figure 56). As discussed above, If we assume the HA-tag on Repebody-mTagBFP2-HA is sterically inaccessible, this result would rule out non-specific binding of the anti-HA antibody.
   Further discussion into the Repebody-degrader blot, the negative control with pcDNA3.1 showed much weaker bands (Figure 55, (b)), thus validating the HA signal detecting in Figure 55, (a). This outcome aligns with flow cytometry data, which detected non-zero levels of mTagBFP2 in the same samples, suggesting a signal detectable by Western blot.(Figure 51)
   
   
3. Potential explanation for inefficient degradation
   One potential reasons resulting from the inefficient degradation presented in our western blot might be due to the insufficient affinity of our Repebody to its targets or its folding kinetics in mammalian cell environment is significantly slower than that of the A1E nanobody (the lysosome-targeting domain). This kinetic imbalance could mean that the entire Repebody-A1E construct is degraded after A1E binds p62, but before the Repebody is successfully folded and bound to its intended target, preventing its delivery to the lysosome.
   Another potential reason is the saturation of the endogenous p62. The EBSS treatment was implemented to induce endogenous p62 production, which would then bind to the A1E domain of the degrader. However, the consistent signal observed in the mTagBFP2 channel (indicating degrader levels) suggests that the starvation treatment did not reduce the amount of degrader present (Figure 51). This implies that the construct may have already been saturated with naturally occurring endogenous p62 prior to the onset of the starvation treatment, rendering the EBSS induction ineffective.
   Overall, these experiments were inconclusive due to multiple complex factors. The construct design and experimental setup must be refined in the future to properly assess the degrader’s efficacy.

Outlook and future experiments

To improve our degrader system and experimental design. These adjustments including construct engineering, assay refinement, and alternative induction methods can be made.

Construct Design

Construct and Affinity Optimization

To optimize fluorescent tags and binding domains. We can consider replacing the current fluorescent reporter with mOrange, as previous work suggests it interacts more strongly with this specific Repebody variant than mCherry. Moreover, to make the affinity and folding speeds of the Repebody and A1E domains be balanced for coherent function. We can use a mutation and selection system to generate a library of Repebody variants. Using this library, we can select variants that bind the fluorescent reporter and its folding speed in mammalian cells to compare to the A1E domain.

Improve Target-Degrader Linkage

To ensure a proper ratio and proximity between the target and the degrader, we could encode the following construct on the same plasmid backbone. First, inserting the P2A Site between the coding sequences to ensure a 1:1 stoichiometric ratio of target to degrader upon translation [18]. This immediate proximity should improve degradation efficiency by giving the Repebody more time to attach to the target before the degrader construct is recruited to p62. Alternatively, using an Internal Ribosome Entry Site (IRES) between the two sequences could be leveraged to express the degrader in excess compared to the target, as the IRES site typically favors the translation of one sequence over the other. While not applicable to degrading endogenous proteins, this approach would serve as a crucial proof-of-concept for the system’s core function.

Assay Refinements

Western Blotting Improvements

For future Western blot analysis, we could include staining with anti-mCherry and anti-EGFP antibodies in addition to anti-HA to double-check the functionality of the fusion construct (e.g., confirming the presence of the fluorescent reporter). Also, to provide more reliable data for interpreting true degradation, potentially showing a shift from full-length target to free reporter protein.

Stable Expression and Control

To establish a more consistent and adjustable system, we can create a stable cell line that inserts the degrader construct into the cell line’s genome using an inducible promoter to control its expression. This would eliminate cell-to-cell variability from transient transfection. Additionally, adjusting expression levels by using a weaker promoter to control target expression and a stronger one for the degrader. This could ensure excessive expression of the degraders, which should lead to a more readily detectable change in target levels.

Microscopy and Visualization

One adjustment could be changing the mTagBFP2 reporter. Since mTagBFP2 is difficult to use with the standard nucleus staining dye DAPI, consider using alternative fluorophores for the degrader could ensure there is no spectral overlap with the target reporters to maintain proper flow cytometry detection. Moreover, perform live-imaging instead of fixed-cell microscopy, as cell fixation can quench some fluorescence. This should improve visualization of positive cells.

Lysosomal Function Assay

To definitively test whether low degrader signal is due to rapid turnover, we can treat cells expressing the Repebody constructs with Bafilomycin A1, a drug that inhibits lysosomal acidification and autophagosome-lysosome fusion. We can also monitor fluorescence in the mTagBFP2 channel to help determine if the low emission of the A1E construct is, as hypothesized, attributable to rapid post-expression degradation. Another alternative to induce degradation for the mitochondrial-anchored targets (the Fis1 constructs) could be switching from using EBSS, which induces general, unselective autophagy via nutrient deprivation to treating with Rapamycin. Rapamycin inhibits mTORC1 and is the preferred method for inducing more selective forms of autophagy, such as mitophagy.

References

[1]: Kim et al. (2016). https://doi.org/10.1016/j.bbrc.2016.01.129

[1] Carbonetti, N. H. (2010). Pertussis toxin and adenylate cyclase toxin: Key virulence factors of Bordetella pertussis and cell biology tools. Future Microbiology, 5(3), 455–469. https://doi.org/10.2217/fmb.09.133

[2] Karimova, G., Pidoux, J., Ullmann, A., & Ladant, D. (1998). A bacterial two-hybrid system based on a reconstituted signal transduction pathway. Proceedings of the National Academy of Sciences, 95(10), 5752–5756. https://doi.org/10.1073/pnas.95.10.5752

[3] Battesti, A., & Bouveret, E. (2012). The bacterial two-hybrid system based on adenylate cyclase reconstitution in Escherichia coli. Methods, 58(4), 325–334. https://doi.org/10.1016/j.ymeth.2012.07.018

[4] https://shopresearch.euromedex.com/document/FTPDF/EUK001.pdf

[5] akobi, A.J., Huber, S.T., Mortensen, S.A. et al. Structural basis of p62/SQSTM1 helical filaments and their role in cellular cargo uptake. Nat Commun 11, 440 (2020). https://doi.org/10.1038/s41467-020-14343-8

[6] Vargas, J.N.S., Hamasaki, M., Kawabata, T. et al. The mechanisms and roles of selective autophagy in mammals. Nat Rev Mol Cell Biol 24, 167–185 (2023). https://doi.org/10.1038/s41580-022-00542-2

[7] Kumar, A. V., Mills, J., & Lapierre, L. R. (2022). Selective Autophagy Receptor p62/SQSTM1, a Pivotal Player in Stress and Aging. Frontiers in cell and developmental biology, 10, 793328. https://doi.org/10.3389/fcell.2022.793328

[8] Jia, W., Li, W., Li, Z. et al. An all-in-one targeted protein degradation platform guided by degradation condensates-bridging bi-specific nanobodies. Cell Res 34, 389–392 (2024). https://doi.org/10.1038/s41422-024-00942-3

[9] Montava-Garriga, L., Singh, F., Ball, G., & Ganley, I. G. (2020). Semi-automated quantitation of mitophagy in cells and tissues. Mechanisms of ageing and development, 185, 111196. https://doi.org/10.1016/j.mad.2019.111196

[10] Xu, Y., Shen, J., & Ran, Z. (2020). Emerging views of mitophagy in immunity and autoimmune diseases. Autophagy, 16(1), 3–17. https://doi.org/10.1080/15548627.2019.1603547

[11] Ma, X., & Ding, W. X. (2021). A fluorescence imaging based-assay to monitor mitophagy in cultured hepatocytes and mouse liver. Liver research (Beijing, China), 5(1), 16–20. https://doi.org/10.1016/j.livres.2020.12.002

[12] Thiel, Z., Marcellin, D., Manneville, C., Goretzki, B., Egger, L., Maher, R., Siccardi, N., Sansig, G., Torres, L., Probst, A., Müller, C. S., George, N., Vogel, M., Sinterhauf, S., Lavoisier, A., Choi, J.-Y., Forcellino, L., Hauck, P., Be, C., Villard, F., Gutmann, S., Meyer, M., Freuler, F., Patoor, M., Mitchell, G., & Nyfeler, B. (2024). Development of a p62 biodegrader for autophagy targeted degradation [Preprint]. bioRxiv. https://doi.org/10.1101/2024.11.19.624312

[13] Jia, W., Li, W., Li, Z. et al. An all-in-one targeted protein degradation platform guided by degradation condensates-bridging bi-specific nanobodies. Cell Res 34, 389–392 (2024). https://doi.org/10.1038/s41422-024-00942-3

[14] Jakobi, A. J., Huber, S. T., Mortensen, S. A., Schultz, S. W., Palara, A., Kuhm, T., Shrestha, B. K., Lamark, T., Hagen, W. J. H., Wilmanns, M., Johansen, T., Brech, A., & Sachse, C. (2019). Structural basis of p62/SQSTM1 helical filaments, their presence in p62 bodies and role in cargo recognition in the cell [Preprint]. bioRxiv. https://doi.org/10.1101/671792

[15] Feng, X., Sun, D., Li, Y. et al. Local membrane source gathering by p62 body drives autophagosome formation. Nat Commun 14, 7338 (2023). https://doi.org/10.1038/s41467-023-42829-8

[16] Thermo Fisher Scientific. (2025). Plasmid Transfection. Thermo Fisher Scientific. https://www.thermofisher.com/de/de/home/references/gibco-cell-culture-basics/transfection-basics/applications/plasmid-transfection.html

[17] Campbell, B.C., Paez-Segala, M.G., Looger, L.L. et al. Chemically stable fluorescent proteins for advanced microscopy. Nat Methods 19, 1612–1621 (2022). https://doi.org/10.1038/s41592-022-01660-7

[18] Choi et al. (2011). High Cleavage Efficiency of a 2A Peptide Derived from Porcine Teschovirus-1 in Human Cell Lines, Zebrafish and Mice. Public Library of Science https://doi.org/10.1371/journal.pone.0018556

[19] Bartel et al. (2025) Identifying and characterizing a missing peroxin—PEX8—in Arabidopsis thaliana. The Plant Cell, 37(7) https://doi.org/10.1093/plcell/koaf166 https://pubs.acs.org/doi/full/10.1021/acssynbio.2c00058

[20] Mihara et al. (2022) Effects of Hydrophobic Residues on the Intracellular Self-Assembly of De Novo Designed Peptide Tags and Their Orthogonality. ACS Publications, 11(6) https://pubs.acs.org/doi/full/10.1021/acssynbio.2c00058

[21] M.G. Schubert, D.B. Goodman, T.M. Wannier, D. Kaur, F. Farzadfard, T.K. Lu, S.L. Shipman, & G.M. Church, High-throughput functional variant screens via in vivo production of single-stranded DNA, Proc. Natl. Acad. Sci. U.S.A. 118 (18) e2018181118, https://doi.org/10.1073/pnas.2018181118 (2021).