Team:Freiburg/Binding
Results - Binding
Figure 1: Depiction of a wild type and an anti-GFP nanobody-displaying spore of Bacillus subtilis after GFP staining.
The modified spores displaying the anti-GFP nanobodies as binding moiety are able to bind GFP while the wild type spores remain unstained.
The modified spores displaying the anti-GFP nanobodies as binding moiety are able to bind GFP while the wild type spores remain unstained.
The ability of a chassis to specifically bind to a targeted structure is the crucial part of targeted drug delivery. We examined anti-GFP nanobody-presenting Bacillus subtilis spores as a suitable carrier. The binding of nanobody spores to GFP was tested by incubation of wild type and anti-GFP nanobody spores with purified GFP and subsequent flow cytometry analysis. We were able to show that the transformed spores were functional, since they bound a significantly higher amount of GFP than wild type spores and can thus be explored for their use as a carrier for specific targeting of structures.
Introduction
The goal of targeted drug delivery is to specifically reach the site of interest and therefore minimize the systemic dispersal of drugs throughout the body. Hence, specific binding structures are needed. Marker proteins, which are characteristic for many diseases are either secreted as soluble proteins or expressed on the surface of affected cells. Such surface proteins can be used for direct binding.
We modified Bacillus subtilis spores as a carrier for nanobodies to bind to a desired molecule. Nanobodies are single chain antibodies found in camelids1. We can benefit from their high specificity, simplicity and producibility in bacteria.1 Those advantages lead to affordability and due to their small size, nanobodies can be variously expressed on our modified Bacillus subtilis spores, resulting in multivalency of the nanobody-targeting molecule interaction.
In the case of ulcerative colitis the amount of carcinoembryonic antigen (CEA) increases during flares in the intestinal epithelium2 and hereby marks the affected tissue. Ideally, anti-CEA nanobody would bind to the diseased tissue, which express high amounts of CEA.
Using GFP and our modified spores incorporating the anti-GFP nanobodies, our system is the first step towards the development of a spore that would carry an anti-CEA nanobody and thus bind to the CEA expressing tissue.
We modified Bacillus subtilis spores as a carrier for nanobodies to bind to a desired molecule. Nanobodies are single chain antibodies found in camelids1. We can benefit from their high specificity, simplicity and producibility in bacteria.1 Those advantages lead to affordability and due to their small size, nanobodies can be variously expressed on our modified Bacillus subtilis spores, resulting in multivalency of the nanobody-targeting molecule interaction.
In the case of ulcerative colitis the amount of carcinoembryonic antigen (CEA) increases during flares in the intestinal epithelium2 and hereby marks the affected tissue. Ideally, anti-CEA nanobody would bind to the diseased tissue, which express high amounts of CEA.
Using GFP and our modified spores incorporating the anti-GFP nanobodies, our system is the first step towards the development of a spore that would carry an anti-CEA nanobody and thus bind to the CEA expressing tissue.
Results
To determine the functionality of the displayed nanobodies on the spores we first verified the localization of the nanobody on the surface of the spore. We assembled an integration vector made up of the fusion constructs BBa_K2114001 (containing the nanobody gene fused to the spore coat gene CotZ, the linker aHelix and a hemagglutinin epitope tag (HA tag)), BBa_K2114002 (containing the nanobody gene fused to the spore coat gene CotZ, the linker G4S and a HA tag) and BBa_K2114017 (containing the nanobody gene fused to the spore coat gene CgeA, the linker aHelix and a HA tag) driven by the PCotYZ-RBS promoter (BBa_K2114000) and transformed them into Bacillus subtilis. After selection the resulting spores are to display the nanobody on their surface.
To test display of the nanobody-HA-tag constructs on the spore surface, spores of wild type and construct-containing Bacillus subtilis were generated and stained with Alexa 647-conjugated anti-HA antibodies. After washing, the Alexa 647 fluorescence intensity was quantified using flow cytometry. (Figure 2) As expected, the wild type spores did not bind any anti-HA antibodies, whereas between 0.75 and 4.1% of the nanobody-HA-tag-spores were able to bind to the anti-HA antibody. As a control, all spores without incubation with the antibody, were Alexa 647 negative.
Thus, we verified the localization of the fusion protein on the surface of the spores. (Figure 2) Interestingly, only a small fraction of our modified spores seemed to display the nanobody-HA-tag-construct on their surface. The reason for that is unknown.
To test display of the nanobody-HA-tag constructs on the spore surface, spores of wild type and construct-containing Bacillus subtilis were generated and stained with Alexa 647-conjugated anti-HA antibodies. After washing, the Alexa 647 fluorescence intensity was quantified using flow cytometry. (Figure 2) As expected, the wild type spores did not bind any anti-HA antibodies, whereas between 0.75 and 4.1% of the nanobody-HA-tag-spores were able to bind to the anti-HA antibody. As a control, all spores without incubation with the antibody, were Alexa 647 negative.
Thus, we verified the localization of the fusion protein on the surface of the spores. (Figure 2) Interestingly, only a small fraction of our modified spores seemed to display the nanobody-HA-tag-construct on their surface. The reason for that is unknown.
Spores generated from wild type Bacillus subtilis and Bacillus subtilis transformed with the parts BBa_K2114001 (containing the coat gene CotZ and an a-helical linker), BBa_K2114002 (containing the coat gene CotZ and a G4S linker) and BBa_K2114017 (containing the coat gene CgeA and an a-helical linker) driven by the promoter PCotYZ-RBS were stained with anti-HA antibodies conjugated to the fluorophore Alexa 647. After washing, spores were analysed by flow cytometry. The forward scatter (size of the spores) and the fluorescence intensity of the Alexa 647 channel are displayed.
GFP can specifically bind to the anti-GFP-nanobody spores
To test the functionality of the anti-GFP nanobody-expressing spores i.e. whether these spores could bind to GFP, further flow cytometric analyses were performed. To this end, we incubated wild type and anti-GFP nanobody-presenting Bacillus subtilis spores (containing the constructs BBa_K2114001, BBa_K2114002 and BBa_K2114017) with purified GFP and analysed the samples after washing by flow cytometry. As displayed in the scatter plots, we were able to detect a considerable shift of the populations of anti-GFP nanobody presenting spores towards higher GFP intensities. (Figure 3)To get an accurate quantification, we set a GFP+ gate by comparing treated and untreated wild type and nanobody spores. The spores found in the GFP+ gate successfully bound to GFP, since very few or no events were counted in this gate when wild type spores were used (left panels) or when the treatment with GFP was omitted (untreated, upper panel). After GFP staining we could only detect 2.6 to 4.0% of the nanobody-presenting spores in the GFP+ gate. These small percentages of GFP+ cells are hardly visible in the histograms. However, the majority of the nanobody spores (approx. 95%) show a shift towards higher GFP intensities compared to the wild type spores. This might indicate that low levels of the nanobody-HA constructs are available to bind to the anti-HA tag antibody in these spores.
Interestingly, the spores containing the construct BBa_K2114017 (containing the coat gene CgeA) showed a smaller shift of the curve towards higher GFP intensities in comparison to the spores such as BBa_K2114001 (containing the coat gene CotZ and the linker aHelix) and BBa_K2114002 (containing the coat gene CotZ and the linker G4S). We therefore continued further measurements using BBa_K2114001 and BBa_K2114002 and discarded BBa_K2114017.
When staining the spores with anti-HA antibodies conjugated to the fluorophore Alexa647, we could detect two distinct populations. (Figure 2)
The smaller population of spores found in the Alexa647+ gate are the spores that express the nanobody (fused to a HA tag) on the surface. If those are the spores where the HA tag is accessible, they should also be the spores that express the nanobody on the surface. Therefore, we expected to see those populations also in the GFP staining experiment. We did see GFP+ nanobody spores (Figure 3), however the GFP+ populations were not clearly separated from the GFPlow or GFP- population.
Thus, we performed further experiments to optimize the binding of GFP to the nanobody presenting spores.
Spores were generated from wild type Bacillus subtilis or Bacillus subtilis transformed with the parts BBa_K2114001 (containing the coat gene CotZ and an a-helical linker), BBa_K2114002 (containing the coat gene CotZ and a G4S linker) and BBa_K2114017 (containing the coat gene CgeA and an a-helical linker) driven by the promoter PCotYZ-RBS, displaying anti-GFP nanobodies on their surface.
The spores were incubated with purified GFP and analysed by flow cytometry. The forward scatter (size of the spores) and the fluorescence intensity of the GFP channel (upper panels) or the GFP fluorescence intensity (lower panels) are displayed.
To reduce detectable background signals and improve the readout, we used bovine serum albumin (BSA) as a blocking protein in a second experiment.
We incubated wild type and anti-GFP nanobody presenting spores (containing the constructs BBa_K2114001 and BBa_K2114002) in BSA, washed off the remaining BSA and subsequently incubated the spores with GFP. We analysed the samples by flow cytometry. As seen in figure 4 the two populations were again identifiable in the scatter plots.
We incubated wild type and anti-GFP nanobody presenting spores (containing the constructs BBa_K2114001 and BBa_K2114002) in BSA, washed off the remaining BSA and subsequently incubated the spores with GFP. We analysed the samples by flow cytometry. As seen in figure 4 the two populations were again identifiable in the scatter plots.
Spores were generated from wild type Bacillus subtilis or Bacillus subtilis transformed with the parts BBa_K2114001 (containing the coat gene CotZ and an a-helical linker), BBa_K2114002 (containing the coat gene CotZ and a G4S linker) and BBa_K2114017 (containing the coat gene CgeA and an a-helical linker) driven by the promoter PCotYZ-RBS, displaying anti-GFP nanobodies on their surface.
The spores were blocked with bovine serum albumin (BSA) , washed and subsequently incubated with purified GFP and analysed by flow cytometry. The forward scatter (size of the spores) and the fluorescence intensity of the GFP channel (upper panels) or the GFP fluorescence intensity (lower panels) are displayed.
To achieve a clearer separation of the populations, we performed a third experiment. We blocked wild type and anti-GFP nanobody presenting spores (containing the constructs BBa_K2114001 and BBa_K2114002) in BSA and subsequently incubated them in the same BSA solution but containing GFP. We analysed the samples by flow cytometry.
After staining with GFP, the scatter plots now show that the two populations were better separated. (Figure 5)
As seen in the histograms, the GFP intensity is considerably higher for the transformed nanobody spores than for the wild type spores. The peaks of the plots for the anti-GFP nanobody presenting spores are shifted further towards higher GFP intensities.
After staining with GFP, the scatter plots now show that the two populations were better separated. (Figure 5)
As seen in the histograms, the GFP intensity is considerably higher for the transformed nanobody spores than for the wild type spores. The peaks of the plots for the anti-GFP nanobody presenting spores are shifted further towards higher GFP intensities.
Spores were generated from wild type B. subtilis or Bacillus subtilis transformed with the parts BBa_K2114001 (containing the coat gene CotZ and an a-helical linker), BBa_K2114002 (containing the coat gene CotZ and a G4S linker) and BBa_K2114017 (containing the coat gene CgeA and an a-helical linker) driven by the promoter PCotYZ-RBS, displaying anti-GFP nanobodies on their surface. The spores were blocked with bovine serum albumin (BSA) and subsequently incubated with a BSA/GFP solution and analysed by flow cytometry.
The forward scatter (size of the spores) and the fluorescence intensity of the GFP channel (upper panels) or the GFP fluorescence intensity (lower panels) are displayed.
To statistically validate that the anti-GFP nanobody-presenting spores bind a higher amount of GFP than the wild type spores, we plotted the percentages of spores found in the GFP+ gate in a bar chart. (Figure 6A)
We hereby compared untreated and GFP-stained spores of the same spore type to validate that the nanobody-presenting spores bind more GFP than the untreated spores. We showed that the difference between treated and untreated wild type spores was not significant, however, the differences between the treated and untreated nanobody spores was highly significant (p<0.05). This clearly shows that some of the nanobody-containing spores were functional, since they could bind to GFP.
In addition, we wanted to see whether the two constructs display a difference among each other. To be able to directly compare the two samples, we calculated the mean fluorescence intensity (MFI) of the spores found in GFP+ and the MFI of unstained spores. We divided the MFI of the spores found in GFP+ by the MFI of unstained spores of the same samples and plotted the resulting values in a bar chart diagramm. (Figure 6B) We could not detect a significant difference between BBa_K2114001 and BBa_K2114002.
Together these experiments show that the anti-GFP nanobody-containing spores were functional, and that the spores containing the coat gene CotZ, an a-helical linker and the nanobody were equally good in binding to GFP as the spores containing the same construct but an G4S linker instead.
We hereby compared untreated and GFP-stained spores of the same spore type to validate that the nanobody-presenting spores bind more GFP than the untreated spores. We showed that the difference between treated and untreated wild type spores was not significant, however, the differences between the treated and untreated nanobody spores was highly significant (p<0.05). This clearly shows that some of the nanobody-containing spores were functional, since they could bind to GFP.
In addition, we wanted to see whether the two constructs display a difference among each other. To be able to directly compare the two samples, we calculated the mean fluorescence intensity (MFI) of the spores found in GFP+ and the MFI of unstained spores. We divided the MFI of the spores found in GFP+ by the MFI of unstained spores of the same samples and plotted the resulting values in a bar chart diagramm. (Figure 6B) We could not detect a significant difference between BBa_K2114001 and BBa_K2114002.
Together these experiments show that the anti-GFP nanobody-containing spores were functional, and that the spores containing the coat gene CotZ, an a-helical linker and the nanobody were equally good in binding to GFP as the spores containing the same construct but an G4S linker instead.
(A) Bacillus subtilis wild type spores and Bacillus subtilis spores displaying anti-GFP nanobodies (BBa_K2114001 (containing the coat gene CotZ and an a-helical linker) and BBa_K2114002 (containing the coat gene CotZ and a G4S linker)) were incubated in BSA and GFP and analysed by flow cytometry. The percentages of spores found in the set gate GFP+ were calculated and plotted. (B) The mean fluorescence intensity (MFI) of the GFP binding spores containing the constructs BBa_K2114001 and BBa_K2114002 found in the GFP+ gate was calculated. The MFI of the unstained spores was calculated. The MFI of GFP+ spores was divided by the MFI of unstained spores and plotted.
Adhesion of the anti-GFP-nanobody spores to GFP-coated glass slides
Having shown that our nanobody-modified spores are functional, in that they can bind to GFP, we wanted to further test whether these spores can adhere to GFP-coated glass slides. Hereby, the glass slides represent the solid surface (or in the case of ulcerative colitis the colon epithelium) and the coated GFP represents the disease specific surface marker (CEA) that we want to target.The anti-GFP nanobody-presenting spores should be able to specifically adhere to the GFP-coated slides. We used fluorescent microscopy to screen the surface for bound spores.
As a first step we established an even coating of GFP. We used the silane (3-Glycidyloxypropyl)trimethoxysilane (GOPTS) as coating substance. GOPTS is commonly used to immobilize amine containing substances ( e.g. GFP) on glass slides. We then spotted purified GFP on the GOPTS activated glass slides.
We blocked any unoccupied binding sites with bovine serum albumin (BSA) to avoid unspecific binding which could hamper our results. (Figure 7)
(A) The schematic shows how glass slides were coated with (3-Glycidyloxypropyl)trimethoxysilane (GOPTS) to allow the immobilization of purified GFP. Unoccupied binding sites were blocked with bovine serum albumin (BSA) to avoid unspecific binding. (B) The resulting GFP spots were analysed using a fluorescence microscope.
After successfully binding GFP on the glass slides, we further analysed the GFP via fluorescent microscope and visualized the spots by creating 3D-images of the bound GFP. (Figure 8)
GFP was spotted on a glass slide. The resulting GFP intensity was plotted on the Z-axis. The GFP spot can be seen as a plateau in the middle of the graph.
We performed various experiments, in order to validate the adhesion of the nanobody expressing spores to the GFP spot.
To verify if our anti-GFP nanobody-presenting spores are able to adhere to a surface, we incubated it (in solution) on the GFP spots. By washing off the remaining unbound spores, we confirmed that only specific adhesion is later detected via fluorescent microscopy.
Unfortunately, we were not able to detect the binding of any spores on the glass slides.
To verify if our anti-GFP nanobody-presenting spores are able to adhere to a surface, we incubated it (in solution) on the GFP spots. By washing off the remaining unbound spores, we confirmed that only specific adhesion is later detected via fluorescent microscopy.
Unfortunately, we were not able to detect the binding of any spores on the glass slides.
Discussion
We verified the expression of the anti-GFP nanobodies on the spores and the functionality of these spores by flow cytometry followed by statistical analyses.We found that two types of spores displaying anti-GFP nanobodies fused to the spore coat protein CotZ were functional, since both spores were able to specifically bind to GFP. In sharp contrast, wild type spores treated with GFP could not bind GFP. Spores not treated with GFP served as a negative control. The only difference between the two types of functional spores was the linker between the coat protein CotZ and the nanobody. Thus, the type of linker did not influence the functionality of the spores. These results were statistically significant. In contrast to the flow cytometric analyses, we could not detect nanobody-expressing spores being bound on GFP-coated glass slides. One possible explanation could be that the amount of spores that present the nanobodies was not large enough, as we showed in flow cytometry that only 2.6% to 6.0% of the spores actually express the nanobody in a functional form on the surface. Another possible explanation could be that the sensitivity of detecting the spores was not high enough.
When staining with specific antibodies or with GFP, we could detect two distinguishable populations in scatter plots of flow cytometric analyses. Some spores were functional and could bind to GFP, whereas other spores could not. One explanation could be that the spores were in distinct stages of sporulation, with different accessibility of the heterologous proteins that are displayed on the surface.
Further approaches to improve the percentage of spores that can bind to GFP could be to test different stages of sporulation and use the stage where most of the spores would be functional. In addition, the amount of anti-GFP nanobodies being able to bind to GFP could be increased by including the fusion of the constructs to different spore coat proteins. Also, the inclusion of longer linkers to increase the distance of nanobody and spore to improve the accessibility of the nanobody for targeted structures could be tested.
In conclusion, we clearly show that the anti-GFP nanobody-presenting spores were functional and, due to their enormous versatility and modifiability we suggest that our Nanocillus represents a promising approach for targeted drug delivery.
References
1. S. Muyldermans, T.N. Baral, 2008, Camelid immunoglobulins and nanobody technology, Volume 128, Issues 1–3, pp 178–183
2. Pavelic ZP, Pavelic L, Expression of carcinoembryonic antigen in ulcerative colitis, tubular adenomas and hyperplastic polyps: correlations with the degree of dysplasia, 1991 Sep-Oct;11(5):1671-5
1. S. Muyldermans, T.N. Baral, 2008, Camelid immunoglobulins and nanobody technology, Volume 128, Issues 1–3, pp 178–183
2. Pavelic ZP, Pavelic L, Expression of carcinoembryonic antigen in ulcerative colitis, tubular adenomas and hyperplastic polyps: correlations with the degree of dysplasia, 1991 Sep-Oct;11(5):1671-5
