Results
I was taught that the way of progress was neither swift nor easy.
~ Marie Skłodowska-Curie
During the project, a wide variety of experiments were conducted, yielding numerous results. These results were thoroughly analyzed, and conclusions were drawn based on the findings. On this page, you will find detailed information for each experiment, including the aim of the experiment, the method used, the results, and the conclusions derived from the data. May our findings guide your own experiments!
Achievements
BMV Isolation
We successfully isolated BMVs from both E. coli BL21 and M. smegmatis mc² 155, as can be seen in the Cryogenic Transmission Electron Microscopy (CryoTEM) visualizations of Figure 1 and Figure 2. We also succeeded in increasing the stability of the BMVs by introducing a HEPES buffer as a storage medium.
Genetic Engineering
With the help of antibody staining, we were able to identify the fusion protein using Fluorescence-Activated Cell Sorting (FACS), consisting of the membrane protein OmpA(N21) fused to GFP on E. coli BL21 cells. This proves successful membrane functionalization of BL21 cells via genetic engineering. The FACS result can be seen in Figure 4. The interpretation of the result can be found here.
Post-Insertion
For the post-insertion method, we successfully expressed and purified the CysSpycatcher and Spytag-GFP proteins from BL21 cells. Using SDS-PAGE, we confirmed the formation of a covalent bond between CysSpycatcher and Spytag-GFP (Figure 4). In a subsequent experiment, we attempted to functionalize a DSPE-maleimide (DSPE-mal) micellar solution with CysSpycatcher, preparing it for future combination with BMVs for efficient functionalization. While preliminary results suggest successful micelle formation and functionalization with Spycatcher, further optimization is required to improve homogeneity, consistency and stability.
Outlook
During the lab work and after the lab data analysis, we came up with some future experiments that could guide the wet lab of our project into the right direction. A detailed explanation of this lab outlook can be found on the Project Description page.
BMV Isolation
Background
For BMV isolation, two protocols were used. The first protocol focused on isolating BMVs from E. coli BL21 cells. The second protocol, an improved version of the first protocol, focused on isolating BMVs from M. smegmatis mc² 155. Below, schematic representations of the protocols can be found.
Although our project focuses on M. smegmatis, we decided to perform some of the lab work with E. coli. The reason for this is both the faster growth of E. coli and the large quantity of literature research on E. coli. These aspects could increase accuracy and progression of the lab work.
Click to extend
BMV isolation - E. coli BL21
E. coli BL21 cells were grown in Tryptic Soy Broth growth medium to stimulate BMV formation. After that, the BMVs were successfully isolated via multiple isolation and concentration steps. The isolated BMVs were analysed with both DLS and CryoTEM. With these results, an improved protocol was made to improve specific parameters like homogeneity, vesicle size and protein aggregate filtering.
Aim:
Stimulate BL21 cells to produce BMVs and isolation of these formed BMVs.
Methods:
E. coli BL21 cells are capable of producing BMVs by
themselves. By filtering the bacterial culture with centrifugation
and Amicon filters, bacterial cells and debris can be removed from
the sample. This should result in a concentrated sample containing
BMVs. The protocol can be found
here.
The concentrated sample was analyzed with DLS and CryoTEM. For DLS,
The sample was diluted in a 1:2 ratio with 30 µL sample and 30 µL
milli-Q water. The same sample was measured one day and five days
after completing the protocol. This was done to monitor the
stability of the sample. The expected size range of
E. coli BMVs is 50 – 200 nm
. The parameters used to analyze the DLS data were Polydispersity
Index (PdI), number mean and intensity mean.
PdI is a dimensionless number ranging from 0 to 1, where 0 represents a monodisperse system, meaning all particles are of the same size . A higher values indicate greater size heterogeneity. Number mean is the average particle size calculated by counting the number of particles in each size range. It emphasizes the size of smaller particles because smaller particles are typically more numerous. Intensity mean is the average particle size based on the intensity of scattered light, which tends to emphasize larger particles because the intensity of light scattering increases with the sixth power of the particle radius.
Results:
As seen in , the PdI increases significantly over 5 days. This suggests that the sample becomes more heterogenous over time, which could be due to aggregates being formed or vesicle fusion. This increase in PdI corresponds with the behavior of an unstable sample. As seen in , the number mean is around 18 nm. This size range is unexpected for E. coli BMVs, suggesting that the majority of the sample does not consist of BMVs. Meanwhile, the number mean seems to be stable over 5 days. shows the intensity mean over 5 days. The intensity mean is significantly higher than the number mean, suggesting that the sample contains some large particles. Next to that, the intensity mean of the sample increases significantly, suggesting an increase in larger particles within the sample. This confirms the assumption that the sample is not stable over time, as suggested with the increased PdI.
To confirm the hypothesis that the sample most likely does not contain E. coli BMVs, a CryoTEM measurement was done.
Due to an error with the CryoTEM, the zoom capacity was limited to a maximum of 500 nm. As we would expect the diameter of most particles to be around 18 nm, it is difficult to analyze these results. But by zooming in by hand, it is still possible to identify some potential BMVs.
The sample contains some protein aggregates and debris, but most abundant are the vesicles with a lamellar structure. These vesicles show a wide range of sizes. In the largest vesicles visible in the red squares (Figure 13, Figure 14 and Figure 15), the lamellar structure can be observed. Although the overall size of the vesicles is very small, this CryoTEM visualization suggests that the isolation of BMVs may have been successful for BL21 cells.
Conclusion:
The CryoTEM results show that the sample contains particles that
closely resemble vesicles. The size of these vesicles in the
sample differs and without zooming in correctly it is difficult to
distinguish between vesicles and protein aggregates. Next to that,
the size of the vesicles is smaller than expected, confirmed by
the number mean and CryoTEM.
The PdI and intensity mean show that the particles are not stable
over time.
The results suggest that the performed protocol can be improved. The sample must be filtered more thoroughly to remove debris and increase homogeneity. The isolated vesicles must be stored in a stable buffer to retain their structual integrity.
BMV Isolation - M. smegmatis
An improved protocol with respect to the BL21 BMV isolation was performed on M. smegmatis. The protocol can be found here. A new growth medium, Sauton's minimal medium, was used to promote BMV production of M. smegmatis. More filtration and concentration steps were introduced by (ultra)centrifugation and by using 30kDa Amicon filters. After isolation, DLS measurements were performed on the sample to identify the homogeneity, size and stability of the particles. The sample was stable for at least seven days, an increase compared to the old protocol. Next to that, the number mean was within the expected size range for M. smegmatis BMVs . The homogeneity, represented by the PdI, was roughly the same as the old protocol.
Aim:
Stimulate M. smegmatis cells to produce BMVs and evaluate which isolation steps are necessary in an improved protocol.
Methods:
Literature research was performed to find an improved protocol for
isolation of BMVs. We looked for an isolation protocol that is
compatible to both E. coli and M. smegmatis for
comparison with respect to the old protocol and flexibility in
negative controls.
The improved protocol uses a minimal medium, two more filtration
steps via (ultra)centrifugation and the use of a smaller Amicon
filter of 30 kDa. The sample is, after isolation, stored in a 50 mM
HEPES buffer. These improvements should, in theory, result in a more
homogenous, stable sample.
To identify the effect of an ultracentrifugation step for purification, the concentrated material is divided into two samples. The first sample, 65K, is concentrated with ultracentrifugation at 65.000 x g, while the second sample, 100K, is concentrated with ultracentrifugation at 100.000 x g.
Results:
The DLS measurement, visible in
,
and
, starts at day 3, which is 3 days after finishing the isolation
protocol with the 65.000 x g ultracentrifugation step. The DLS
measurement for the 100K sample starts at day 5, as this protocol
took 2 days longer due to transportation to another lab. The PdI
of the 65K sample is stable for at least 7 days around 0.17, as
can be seen in
. The PdI of the 100K sample seems to fluctuate more, increasing
and decreasing over time. For both samples, the PdI is within the
acceptable range (<0.3) for a homogenous sample, for at least 7
days.
At day 10, the sample is not stable anymore, as the PdI increases
significantly.
shows that the number mean for both samples is within the expected size range for BMVs (50-250 nm). On top of that, the number mean of the 65K sample does not fluctuate significantly in 10 days.
The intensity mean of both samples does not fluctuate
significantly for at least 7 days, as can be seen in
. On day 4, the intensity mean of the 65K sample seems to be
higher than expected, with a relatively large error margin, which
may indicate a potential measurement error.
The intensity mean increases significantly on day 10 for both
samples, suggesting the increase of particle size within the
sample. This could be due to aggregates being formed or fusion of
vesicles.
Conclusion:
The DLS measurements suggests that the 65K sample is more stable than the 100K sample, given by the PdI. The number mean and intensity mean do not differ significantly between both samples. This suggests that a 65.000 x g ultracentrifugation step is preferred over a 100.000 x g step.
Compared to the sample in the BL21 protocol, the stability of the this sample improved when stored in HEPES buffer. It is therefore recommended to use a HEPES buffer for BMV storage.
Since the results are promising, the protocol will be repeated and both DLS and CryoTEM measurements will be performed.
BMV Isolation - 65.000 x g - M. smegmatis
Since the DLS measurements after the improved BMV isolation protocol for M. smegmatis showed promising results, the protocol was repeated and, next to a DLS measurement, a CryoTEM measurement was performed to visualize the sample. The DLS measurement showed little to no difference in number mean and intensity mean, but the PdI was significantly smaller.
The CryoTEM showed BMVs, although a lower quantity than obtained after the BL21 BMV isolation. The size of the BMVs was more consistent. The sample still contained protein aggregates and some of the BMVs were ruptured.
Aim:
Perform the protocol to isolate BMVs from M. smegmatis and analyse the sample with both DLS and CryoTEM.
Methods:
Within the BMV isolation protocol for M. smegmatis, 65.000 x g ultracentrifugation was chosen as this yielded better results than 100.000 x g ultracentrifugation. Apart from that, the protocol was not altered and replicated from previous experiments.
Results:
The PdI, visible in , is lower than in the previous experiment, around 0.10. This suggests that the sample is very homogenous. For at least 7 days, the PdI stays stable, suggesting good stability of the sample. In , the number mean is very stable with little to no variation in size. The same goes for the intensity mean in . These results suggest that the sample shows the same behavior as the previous sample, if not better. Therefore, CryoTEM is performed to visualize the sample.
Within the red squares of , BMVs are visible. These vesicles that are visible in CryoTEM are in the same size range as the number mean measured by DLS. Overall, the amount of BMVs seems to be lower than in the CryoTEM results of BL21 BMV isolation, but the BMVs themselves are more consistent in size. shows lamillar structures, but they seem to have ruptured. This could be due to the high forces on the vesicles during ultracentrifugation or osmotic stress. In , protein aggregates are visible. It can therefore be concluded that with this protocol, we did not filter all the protein aggregates in the sample.
Overall Conclusion:
In previous experiment, the BMV isolation of BL21 was successful, shown by the CryoTEM results. But the BMVs are heterogenous and their average size is smaller than expected. By storing the BMVs in a HEPES buffer, which was done in the improved protocol, the stability of the BMVs increased, resulting in the possibility for long-term storage. The extra filtration and isolation steps in the improved protocol resulted in a more homogenous BMV sample, but the high forces, due to (ultra)centrifugation, most likely caused the presence ruptured vesicles within the population of BMVs.
It is difficult to compare the two CryoTEM visualizations of (E. coli BL21 and M. smegmatis) with each other, as two variables were introduced: change in species and change in protocol. It is therefore adviced to redo both experiments, in which both species are analysed in parallel. This enhances the reliability of result comparisons, leading to more informed decisions regarding the optimal BMV isolation protocol.
Genetic Engineering
Background
Two strategies for the functionalization of the bacterial membrane vesicles were investigated in the laboratory. The first strategy to be discussed is the genetic engineering strategy, where fusion proteins are used.
In the genetic engineering method is visualized. Literature research was performed to find proteins that are known to reside in the membrane of the vesicles of mycobacteria (M. smegmatis, M. tuberculosis and M. bovis BCG), that are suitable for the attachment of model-antigen GFP at the periplasmic face of the membrane. The following selection of such proteins was made: OmpAtb, porin MspA, FtsQ, and RbsC. The structure of these proteins was analyzed to identify certain domains or signal sequences that can be used for the expression of a fusion protein with GFP at its C-terminal end. It is important that the membrane proteins retain their functional domains, to enable effective display of the GFP on the outside of the BMV. Therefore, the membrane proteins that are fused to the antigen are either:
For M. smegmatis several modelling tools were used to design relevant fusion protein constructs, like DeepTMHMM, Signal6.0 and Alphafold. More information can be found on our Modeling page. This resulted in a selection of seven fusion proteins, labeled 3a-3g, described below:
Apart from the fusion proteins in M. smegmatis, two fusion proteins were designed for E. coli. This has been done in parallel to be able to develop a proof of concept for E. coli that is comparable for M. smegmatis. The fusion proteins are based on outer membrane protein A (OmpA), a membrane protein that is native to E. coli. The two fusion proteins are:
An overview of the inserts, plasmids and hosts can be found in Table 1. Inserts 1a and 1b are used for expression in E. coli and 3a-3h for expression in M. smegmatis. Inserts 2a and 2b are used for the post-insertion method, that will be employed as an alternative for the genetic engineering method. More information about 2a and 2b can be found in the post-insertion section.
| ID | Protein inserts | ID (parts registry) | Plasmid | Host |
|---|---|---|---|---|
| 1a | OmpA(N21) + GFP | BBa_K5403012 | pET28a | E. coli |
| 1b | OmpA + GFP | BBa_K5403013 | pET28a | E. coli |
| 2a | CysSpycatcher + 6H | BBa_K5403005 | pET28a | E. coli |
| 2b | Spytag + GFP + 6H | BBa_K5403021 | pET28a | E. coli |
| 3a | OmpAtb(N72) + GFP | BBa_K5403014 | pCHERRY3 | M. smegmatis |
| 3b | OmpAtb(N72) + OmpA + GFP | BBa_K5403015 | pCHERRY3 | M. smegmatis |
| 3c | Porin(N50) + GFP | BBa_K5403016 | pCHERRY3 | M. smegmatis |
| 3d | Porin (C-truncated) + GFP | BBa_K5403017 | pCHERRY3 | M. smegmatis |
| 3e | Porin + GFP | BBa_K5403018 | pCHERRY3 | M. smegmatis |
| 3f | RbsC (C-truncated) + GFP | BBa_K5403019 | pCHERRY3 | M. smegmatis |
| 3g | FtsQ + GFP | BBa_K5403020 | pCHERRY3 | M. smegmatis |
| 3h | GFP | BBa_E0040 | pCHERRY3 | M. smegmatis |
Cloning for E. coli proteins
The genes encoding the fusion proteins OmpA(N21)-GFP and OmpA-GFP that are used for the genetic engineering method were cloned into the pET28a vector by linearization and Gibson assembly. CysSpycatcher and Spytag-GFP, that are later used for the post-insertion method were cloned using the same approach.
Aim:
Linearization and Gibson assembly of fusion proteins 1a, 1b, 2a and 2b into the pET28a plasmid.
Methods:
Vector linearization was achieved through PCR, followed by DpnI digestion and a PCR purification step. The concentration was then determined using NanoDrop. Subsequently, Gibson Assembly was performed, and the resulting samples were visualized on an agarose gel to check whether the correct plasmid was assembled.
Results:
and show the agarose gels for the E. coli inserts. Afterwards, the plasmids were sent to Azenta for sequencing.
Conclusion:
In and , all the plasmids are visible at their expected molecular weight. The Azenta sequencing results confirmed this observation. So, we succesfully assembled our plasmid.
FACS - OmpA(N21) + GFP
The OmpA(N21) membrane protein, fused to GFP, is expressed in BL21 cells by induction of IPTG. The sample is stained with a secondary antibody conjugated to a dye, to increase the fluorescent signal of the GFP. The fluorescent signal is analysed with FACS.
Aim:
Analyze the presence of GFP, bound to membrane protein OmpA(N21), on E. coli BL21 cells.
Methods:
To analyze the expression of both membrane proteins, a large culture was made, consisting of 2YT medium with BL21 cells transformed with a pET28a recombinant plasmid. The culture was grown until a OD600 of ~0.6 was reached. At that point, IPTG was added to the culture to induce protein expression and the large culture was incubated at a lower temperature of 18 ° overnight. The protocol can be found here.
In the first setup of this experiment, the IPTG incubated large culture was directly analyzed with FACS. These FACS results showed little to no GFP signal. The experiment was repeated, but still no signal was visible. To overcome this issue, an antibody staining was introduced to increase the signal.
Antibody staining was done with a primary antibody for GFP and a secondary antibody conjugated to a Cyanine3 dye. Several negative controls were incorporated into the experiment:
Results:
To analyze FACS results, the data has to be gated by hand. By gating, unwanted data that would interfere with the analysis is excluded. This reduces noise and improves data accuracy. Gating is done based on the forward and side scatter of the sample. In , the gating is shown.
To improve accuracy, doublet discrimination is used. Doublets are events where two cells pass through the flow cytometer's laser at the same time or very closely together, producing an artifact in the data. These doublets can be mistakenly counted as a single event, leading to inaccurate results. This discrimination is done by comparing pulse height (FSC-H) and pulse area (FSC-A) in .
When both gating and doublet discrimination have been completed, the quartiles lines that seperate the quartiles can be determined. It is adviced that, in a negative control, maximum 1% of our signal falls in Q1 and Q2. shows the correct separation.
Now that the FACS data is completely filtered, the settings for the gate, doublet discrimination and placement of the quartile lines can be repeated on all samples for OmpA(N21).
shows the signal of the OmpA(N21) sample without IPTG and with no antibody staining. It is expected that the signal will not be visible in Q1 and Q2, which is confirmed by the FACS result. shows the signal of the OmpA(N21) sample with IPTG, but without antibody staining. The signal will therefore only come from the protein-bound GFP. As can be seen in the Figure, around 99% of the signal is visible in Q4 and Q3. This suggests that the GFP itself is not sufficient enough for FACS analysis and the antibody staining is necessary. is the signal of the OmpA(N21) sample with IPTG and only secondary antibody. As only around 1.5% of the signal is present in Q1 and Q2, it can be concluded that no staining occured. Therefore, it can be concluded that little to no non-specific secondary antibody staining occured.
For the OmpA(N21) sample with IPTG and antibody staining, around 9% of the signal is within the color spectrum of the dye (Q1) in . This is larger than the sample without antibody staining.
When looking at , around 4% of the signal is present within the color spectrum of the dye. This is not expected, as no IPTG is added to this sample, meaning that the protein of interest should not be expressed. As the negative control has proven that non-specific staining does not occur, another explanation must be found. One could be leaky expression.
Leaky expression in protein expression refers to the
unintended, low-level production of a protein even when the
gene or promotor that controls its expression is supposed to
be repressed or inactive
. In other words, the regulatory system designed to tightly
control protein expression is not fully effective, allowing
for some leakage of expression in the absence of the intended
inducer or under non-inducing conditions. The fact that the
signal is present at a lower quantity than in the IPTG induced
sampe confirms this hypothesis.
There is no clear explanation for why the signal intensity in the non-induced sample is higher compared to the induced sample.
Conclusion:
Although some leaky expression occurs in the non-induced sample, a clear staining signal can be seen in the induced OmpA(N21) sample. As non-specific staining is ruled out, it can be concluded that the fusion protein consisting of OmpA(N21) bound to GFP is correctly expressed on the membrane of E. coli BL21 cells.
FACS - OmpA + GFP
The OmpA membrane protein, fused to GFP, is expressed in BL21 cells by induction of IPTG. The sample is stained with a secondary antibody conjugated to a dye, to increase the fluorescent signal of the GFP. The fluorescent signal is analysed with FACS.
Aim:
Analyze the presence of GFP, bound to membrane protein OmpA(N21), on E. coli BL21 cells.
Methods:
The sample of OmpA bound to GFP is prepared and analyzed in the same way as the OmpA(N21) sample: antibody staining and analysis with FACS. The protocol can be found here.
Results:
The preproccesing of the FACS results is done in the same way as for the OmpA(N21) sample. First, the results are gated correctly. After this, doublet discrimination is introduced. Lastly, the placement of the quartile lines are determined.
shows that, with IPTG induction, the GFP signal itself is not strong enough for FACS analysis. This is a result that would also be expected for the sample without IPTG induction, but the pattern of this sample in is different from the sample with IPTG. An explanation for this could be that IPTG induction results in misfolding or degradation of OmpA. The expression of the OmpA could also potentially lead to toxicity in the cell, preventing GFP expression from being detectable.
shows the negative control sample for non-specific binding of the secondary antibody. When non-specific staining does not occur, no signal is expected in Q1. But it is visible that 14% of the total signal is in Q1, suggesting a Cyanine3 staining. This suggests that the secondary antibody not only binds to the primary antibody, but also somewhere else on the BL21 cell membrane.
When looking at , around 27% of the sample shows antibody staining. The intensity of this signal closely resembles that of the secondary antibody control group. It is therefore difficult to conclude if this is staining from the antibody complex binding to GFP or non-specific binding of the secondary antibody.
, the sample without IPTG, shows a different signal pattern. Although the total amount of signal is lower than in , the signal intensity is higher. The presence of signal could be due to leaky expression, but there is no clear explanation for why the signal intensity is much higher compared to the induced sample.
Conclusion:
The antibody staining of OmpA + GFP shows non-specific secondary antibody binding. This non-specific staining makes it difficult to confirm the presence of OmpA + GFP on the membrane of the BL21 cells, especially because the signal of the induced sample shows a similar pattern as the secondary antibody negative control. On top of that, the non-induced sample shows a higher intensity than the induced sample, which cannot be explained by leaky expression alone.
Due to these uncertainties, it is adviced to redo the exact experiment to check if the result is the same. If this results in the same FACS data, a different combination of primary and secondary antibody can be chosen to hopefully reduce non-specific secondary antibody binding.
Next to the FACS experiment, it is recommended to perform other
experimental set-ups to identify the presence of the OmpA fusion
protein. These could include visualization techniques such as
super-resolution microscopy (SRM) and immunofluorescence
microscopy (IF), or alternative labelling method like cell surface
biotinylation.
Next to that, the induction system itself can be tested by using
qPCR for mRNA levels or Western blot for protein levels.
Cloning for M. smegmatis proteins
We attempted to linearize the pCHERRY3 plasmid for fusion proteins (3a) up to (3h) for M. smegmatis . The first strategy included a linearization followed with a Gibson assembly. When this kept failing, we altered our strategy to a restriction-ligation.
Gibson Assembly
Aim:
Linearization of fusion proteins (3a) up to (3h) into the pCHERRY3 plasmid.
Methods:
Vector linearization was achieved through PCR, followed by DpnI digestion and a PCR purification step. The concentration was then determined using NanoDrop. Subsequently, Gibson Assembly was performed, and the resulting samples were visualized on an agarose gel to check whether the correct plasmid was assembled.
Results:
One step to ease the linearization of a plasmid is to first digest the plasmid with restriction enzymes. In the case of pCHERRY3, BamHI, PstI and HindIII were used as seen in . show a linearization attempt that was deemed unsuccesful, since the band of the plasmid is ~500bp instead of the wanted ~6000bp.
Conclusion:
We were unable to get the correct plasmid through Gibson assembly, even with the correct digested pCHERRY3. A more in depth description about what we tried can be read on our engineering page and notebook page. Therefore the strategy was changed from Gibson assembly to restriction-ligation cloning.
Restriction-ligation
Aim:
To insert the fusion protein (3c) PorinN50-GFP and the protein (3h) GFP into pCHERRY3 for expression
Methods:First, a plasmid digestion was performed on pCHERRY3 with the restriction enzymes BamHI and HindIII, as seen in . The restriction of the gBlocks was performed only on fusion proteins (3c) and (3h) due to limited materials and the unavailability of the correct primers needed to amplify gBlocks via PCR. Secondly, a ligation was performed on the digested pCHERRY3, for fusion protein 3c and protein 3h. We then put the sample on gel to evaluate whether it succeeded and sent the sample to Azenta for sequencing.
Results:
Conclusion
The gel showed promising results, indicating that restriction-ligation succeeded. As seen in the biggest band was around ~6000bp, which could indicate our constructed plasmid. However, the agarose gel showed contamination in the sample as well. The Azenta results unfortunately confirmed that the sequence was not correct. We plan to assemble the plasmids for the expression of the fusion proteins in M. smegmatis in the future (recombinant plasmids 3a-3g), by restriction-ligation cloning, but with a higher insert concentration to improve the efficiency of the process. This way, we might be able to successfully create our recombinant plasmids and can use them for expression in M. smegmatis.Post-Insertion
Background
Two strategies for the functionalization of the bacterial membrane vesicles were investigated in the laboratory. The second strategy to be discussed is the post-insertion strategy.
The post-insertion method, in contrast to the genetic engineering method, requires isolation of BMVs before functionalization. After isolation, BMVs will be combined with micelles functionalized with Spycatcher proteins, allowing the Spycatcher to integrate in the membrane of the BMV. Unlike the genetic engineering approach—where GFP was fused as a model antigen before introducing modularity—in this method, Spycatcher is directly incorporated into the BMVs. This Spycatcher functionalization of our BMV establishes a ‘plug-and-play’ system for antigen integration. It enables the development of off-the-shelf, personalized vaccines, particularly for cancer treatment, where the desired antigens can be easily ‘clicked’ onto the BMVs.
In order to assemble micelles with Spycatcher on the surface, the amino acid sequence of Spycatcher was adjusted such that a unique cysteine residue exists at its N-terminus ((2a) CysSpycatcher). The cysteine is used to attach the phospholipid DSPE-MAL by maleimide coupling (Click Chemistry). The Spytag is fused to GFP to verify functionalization. An overview of the protein inserts, plasmids and hosts for obtaining the proteins necessary for this method can be found in Table 2.
| ID | Protein inserts | ID (parts registry) | Plasmid | Host |
|---|---|---|---|---|
| 2a | CysSpycatcher + 6H | BBa_K5403005 | pET28a | E. coli |
| 2b | Spytag + GFP + 6H | BBa_K5403021 | pET28a | E. coli |
Protein expression and purification in E. coli
We purified CysSpycatcher and Spytag-GFP proteins for use in the post-insertion method, which will allow functionalization of micelles or bacterial membrane vesicles with Spycatcher for subsequent antigen binding through Spytag.
Aim:
Express and purify proteins that will be employed for the post-insertion method.
Methods:
The cloning of the plasmids for the expression of CysSpycatcher and GFP-Spytag was performed in a pET28a vector by Gibson assembly. The assembled plasmids were multiplied in E. coli TOP10 cells and miniprepped. CysSpycatcher and Spytag-GFP were expressed in BL21 cells, followed by IMAC-purification. The resulting samples were visualized on an SDS-PAGE gel.
Results:
Conclusion:
It can be seen from the SDS-PAGE gel that the purification was successful, resulting in a band at 9.9 kDa for CysSpycatcher and a band at 29.6 kDa for Spytag-GFP.
CysSpycatcher coupling with Spytag-GFP
We coupled Spytag-GFP and CysSpycatcher and ran them on an SDS-PAGE gel to verify the successful conjugation of the two proteins, ensuring that the functional domains of Spytag and Spycatcher interact correctly, forming a stable covalent bond.
Aim:
Check if the recombinantly engineered CysSpycatcher and Spytag-GFP are able to form a covalent isopeptide bond.
Methods:
To couple CysSpycatcher with Spytag-GFP, we prepared the proteins in a 1:1 molar ratio. The two components were mixed in a reaction buffer (e.g., PBS or Tris buffer, pH 7.4) and incubated at room temperature to allow the covalent bond formation between Spycatcher and Spytag.
Results:
The SDS-PAGE analysis in reveals several key observations. In the lanes containing CysSpycatcher and Spytag-GFP, the expected bands for each individual protein are present. CysSpycatcher appears as a band around 10 kDa, while Spytag-GFP shows up at approximately 30 kDa. In the lanes where the two proteins were mixed, a prominent new band appears around 40 kDa, consistent with the expected molecular weight of the CysSpycatcher-Spytag-GFP conjugate.
However, there is an unexpected excess of unbound Spytag-GFP in the conjugate lanes, with almost no detectable free CysSpycatcher remaining, despite mixing in a 1:1 molar ratio. This suggests that the CysSpycatcher is fully utilized in binding Spytag-GFP, but for some reason, the Spytag-GFP is in excess. One possible explanation for this observation is inaccuracies in protein quantification or loss of CysSpycatcher during sample preparation.
Additionally, a band is observed around 75 kDa, which could represent one CysSpycatcher protein coupled to two Spytag-GFP proteins. This may indicate that CysSpycatcher is binding more than one Spytag-GFP molecule, possibly due to an incorrect molar ratio during mixing or due to multimerization of proteins. This suggests that optimizing the ratio and reaction conditions could improve the specificity and efficiency of the coupling, preventing the formation of higher molecular weight complexes.
Conclusion:
Although the coupling reaction and conditions need further optimization, our results demonstrate that CysSpycatcher and Spytag-GFP successfully combine, as evidenced by the appearance of the expected conjugate band on the SDS-PAGE gel. The presence of higher molecular weight complexes and excess Spytag-GFP suggests room for improvement in the reaction setup, but the fundamental Spycatcher-Spytag interaction has been clearly shown.
Micelle formation
We explored the use of DSPE-Mal to form micelles and tested whether CysSpycatcher could be successfully attached. We tried to confirm this attachment using Spytag-GFP. While the results showed some signs of successful functionalization, both the micelle formation and protein attachment were inconsistent, highlighting the need for optimization in the protocol to achieve more reliable and stable micelle formation and functionalization.
Aim:
Make micelles that are functionalized with Spycatcher as preparation for the post-insertion method.
Methods:
We prepared micelles using only DSPE-Mal due to the unavailability of DSPE, deviating from the typical DSPE ratio of 4:1. To form the micelles, DSPE-Mal was dissolved in methylene chloride, and a lipid film was created by evaporating the solvent under a nitrogen stream. The lipid film was then hydrated with an appropriate buffer and sonicated to produce micelles.
For functionalization, CysSpycatcher was added to the micellar solution and incubated, allowing the CysSpycatcher to covalently attach to the DSPE-Mal micelles. The micellar solution was then filtered using a 50 kDa centrifugal filter to remove any unbound CysSpycatcher. Next, the filtered DSPE-Mal-CysSpycatcher micelles were incubated with Spytag-GFP to facilitate covalent binding between Spycatcher and Spytag. This solution was filtered again with a 50 kDa centrifugal filter to eliminate any unbound Spytag-GFP. To verify successful functionalization, the DSPE-Mal-CysSpycatcher + Spytag-GFP micelles were compared with a negative control, where DSPE-Mal micelles were directly incubated with Spytag-GFP and filtered using a 50 kDa centrifugal filter to remove unbound Spytag-GFP.
The structural properties of the DSPE-Mal micelles and the unfiltered/filtered DSPE-Mal-CysSpycatcher micelles were evaluated using DLS to assess their size distribution. Additionally, the filtered solution containing DSPE-Mal-CysSpycatcher + Spytag-GFP was compared to the DSPE-Mal + Spytag-GFP negative control using fluorescence analysis to confirm successful functionalization and binding of Spytag-GFP to the CysSpycatcher-modified micelles.
Results:
Table 3 shows the PdI (Polydispersity Index) and Number Mean (d.nm) for the DSPE-mal micelles and DSPE-mal-CysSpycatcher micelles, both filtered and unfiltered.
| Sample Name | PdI | Number Mean (d.nm) |
|---|---|---|
| micelles DSPE-mal 1 | 0.4 | 167.1 |
| micelles DSPE-mal 2 | 0.434 | 45.29 |
| micelles DSPE-mal 3 | 0.592 | 51.66 |
| micelles DSPE-mal + CysSpycatcher unfiltered 1 | 0.367 | 42.12 |
| micelles DSPE-mal + CysSpycatcher unfiltered 2 | 0.401 | 58.95 |
| micelles DSPE-mal + CysSpycatcher unfiltered 3 | 0.426 | 60.85 |
| micelles DSPE-mal + CysSpycatcher filtered 1 | 0.262 | 18.68 |
| micelles DSPE-mal + CysSpycatcher filtered 2 | 0.254 | 62.38 |
| micelles DSPE-mal + CysSpycatcher filtered 3 | 0.269 | 25.6 |
The DLS data reveal significant variability in the size distribution and polydispersity of both the DSPE-mal micelles and the DSPE-mal-CysSpycatcher micelles, with PdI values ranging from 0.4 to 0.592 for DSPE-mal micelles and 0.254 to 0.426 for DSPE-mal-CysSpycatcher micelles, indicating that all the solutions are highly heterogeneous. While filtration somewhat improves homogeneity, as evidenced by a reduction in PdI values, the number mean size still varies greatly across the samples (e.g., from 18.68 nm to 167.1 nm), suggesting inconsistent micelle formation. This inconsistency highlights the need for the inclusion of DSPE in the formulation, as it plays a crucial role in stabilizing the micelles. Additionally, the protocol for micelle formation requires further optimization to produce uniform and stable particles.
Fluorescence measurements were taken to assess whether Spytag-GFP successfully bound to the DSPE-mal micelles functionalized with CysSpycatcher. The fluorescence signal was measured for three samples: a blank (3 RFU), DSPE-mal-CysSpycatcher + Spytag-GFP (50,465 RFU), and DSPE-mal + Spytag-GFP (27,103 RFU).
The significantly higher fluorescence signal from the DSPE-mal-CysSpycatcher + Spytag-GFP sample (50,465 RFU) compared to the DSPE-mal + Spytag-GFP (27,103 RFU) and the blank (3 RFU) suggests that the micelles were successfully functionalized with CysSpycatcher, allowing the covalent binding of Spytag-GFP. This increase in fluorescence likely indicates the successful attachment of Spytag-GFP to the CysSpycatcher-functionalized micelles.
However, it is unusual that a substantial fluorescence signal remains in the DSPE-mal + Spytag-GFP sample, even after filtering with a 50 kDa centrifugal filter. Since Spytag-GFP is a small protein (~30 kDa), it should have been removed during filtration if it were unbound. This persistent fluorescence suggests that Spytag-GFP might be adhering to the micelles non-specifically, even in the absence of CysSpyCatcher. This could be due to hydrophobic interactions or other surface characteristics of the micelles that cause the Spytag-GFP to stick, rather than being fully removed by filtration.
Conclusion:
The DLS and fluorescence data suggest potential for forming DSPE-mal micelles functionalized with Spycatcher, but micelle formation was inconsistent and showed high variability in size and polydispersity. Non-specific binding of Spytag-GFP was also observed. Overall, this exploratory protocol holds promise but requires significant optimization at each step—micelle formation, functionalization, and filtration—to achieve consistent and reliable results for future applications.
Images created with BioRender.com