In this work, we present proof of principle of a novel direction of protein drug delivery in medicine. Our experimental results show that correctly tailored IBs could be cheap protein cassettes for protein drug delivery that are able to specifically attack pathogen adhesins or receptors of different tissue. Additionally, the tailored IBs are able to carry and release a protein in “programmable” time courses.
Our tailored IBs were prepared via a conjugation reaction using glutaraldehyde. The IBs containing the indicator protein GFP (CBDclos-intein-GFP) were linked to a sialylated and non-sialylated protein, respectively. To test this reaction, the resulting conjugates were tested using our method based on the hemagglutination-inhibition assay . The Sia-dependent hemagglutination is expected to be inhibited by the conjugates containing fetuin but unaffected by the asialofetuin conjugates. However, because the hemagglutination was evoked by the SabA adhesin in the form of IBs, the hydrophobic interactions between the sabIBs and gfpIBs, apparently also between sabIB and gfpIB conjugates, misrepresented the results of our test. Generally, bacterial IBs, as opposed to soluble globular proteins, represent physiological aggregates composed mainly of recombinant proteins in the form of an unfinished tertiary structure; therefore, more hydrophobic acids are exposed on the water-protein interface, thereby allowing themselves IBs formation and clustering IBs to large clusters by hydrophobic interactions. In other words, the ratio of the burial of hydrophobic residues from water balances the aggregate stability and enzyme activity in IBs. Ideally in our fusion constructs, the CBDclos module provides the maximal hydrophobic interaction among fusion protein molecules and forms maximally stable aggregates, and the cleaving module, together with GFP, form the maximal globule stage and provide the maximal activity. However, the hydrophobic interactions between the sabIBs and gfpIBs were not desired; therefore, we decided to change the amino acid used for glutaraldehyde inactivation in the preparation of the gfpIB conjugates. Glutaraldehyde reacts very well with various amino acids ; thus, they are used as deactivators of un-reacted glutaraldehyde [24, 25]. For our conjugation reaction, glycine was the first choice for the glutaraldehyde inactivation because it is a small neutral amino acid without any significant effects on the intramolecular hydrophobic interactions . After observing the unwanted hydrophobic interactions between the sabIBs and conjugates, we applied an excess of lysine to eliminate the un-reacted glutaraldehyde. Lysine is largely polar , and treatment with it “polarised” the surface of the IBs and successfully eliminated the hydrophobic interactions between the sabIBs and gfpIB conjugates as well as between the gfpIBs and plastic surfaces. After applying this approach, our test confirmed the specificity of the Sia-dependent interactions between erythrocytes, sabIBs and conjugates. As expected on the basis of on our previously published results [20, 27], the fetuin-containing conjugates modified the level of hemagglutination, whereas the asialofetuin-containing conjugates maintained the hemagglutination level similar to control wells.
The specificity of the interactions was also confirmed by our in vitro model. We selected the pathogenic bacterium Helicobacter pylori and its SabA adhesin as a model for our study. The SabA adhesin has been well-studied in the pathogenesis of H. pylori. Because it recognises Sias, it binds the sialylated antigens on the inflamed gastric epithelium and on RBC in gastric mucosal blood vessels [3, 4]. SabA aggregated to form IBs is substituting pathogen cells in our model. Under a fluorescent confocal microscope, it was confirmed that only the tailored IBs containing fetuin strongly interacted with the pathogen cells (represented by sabIBs) bound to the Sias on the erythrocyte surface. The conjugates containing asialofetuin did not show the same interactions under the same conditions. The conjugates used for this experiment were compressed into a pellet due to centrifugation and resuspension of the pellet with a pipette resulted in the formation of smaller clusters. These clusters were used for the microscope reaction due to better visibility. However, the basic concept is the application of the IBs as “nanopills” for biomedicine. Several works have been already published, demonstrating that IBs are mechanically stable enough to tolerate the ultrasonication needed to obtain IBs with a median spherical diameter of 200–500 nm [28, 29]. Thus, conjugates could be very easily broken into nanoparticles via sonication. Our results provide convincing evidence that tailored IBs are potentially able to specifically target adherent pathogen cells on human tissues. Erythrocytes in our in vitro model underwent some morphological changes compared to control erythrocytes in neutral phosphate buffer. RBC are very susceptible to the environment changes, and the transformation of their shape might be caused by various aspects . We suggest that morphological changes are a consequence of binding of sabIBs to the Sias on the cell surface. However, this has no significant impact on our results, as this shape modification was observed in a model of the interaction of RBC-sabIBs with gfpIB-fetuin as well as with gfpIB-asialofetuin.
The other part of this study focused on demonstrating the drug release from gfpIBs. The therapeutic protein is represented by GFP, the release of which was tested under two pH conditions. Neutral pH (pH 7.0) was chosen as the primary condition under which also specific pathogen recognition in vitro was demonstrated. However, the pH values in the human body vary widely. Because a model adhesin for our recognition study was SabA from H. pylori, a bacterium colonising stomach, we decided to test also protein release under acidic pH conditions (pH 2.5). The GFP remains fluorescently active when aggregated into the IBs, as was demonstrated by fusion to the VP1 capsid protein of the foot-and-mouth disease virus . We used the CBDclos system that initiates the physiological aggregation of GFP and also maintains protein activity. As a cleaving module, we used the S. aureus sortase A (SrtA) and Ssp DNAB intein from Synechocystis sp. SrtA has already proven to be an efficient tag for the purification of recombinant proteins when fused to their N-termini [21, 31] but, until now, was not applied to the protein release from IBs. Our results show that the SrtA protease is capable of effectively releasing the protein from IBs at a neutral pH in a relatively short period of time. At an acidic pH, such as in the stomach, the progress of GFP release was markedly slower compared to that observed at neutral pH. Nevertheless, the very slow (several weeks lasting) release of the proteins was observed in the aggregates containing intein as the cleaving module. The Ssp DNAB intein was tested as a part of the potential cleavable self-aggregating tag, indicating the aggregation of the target proteins to form IBs and the subsequent intein-cleavage and release of soluble protein. However, the cleavage efficiency of this intein was marked as insufficient; thus, it was not tested further . Although its cleavage activity is not suitable for protein purification, this intein might be beneficial for biomedical purposes, as its sustained, long-lasting drug release might be desirable in some cases, such as in the treatment of chronic diseases . In contrast, the treatment of some infections requires an instantaneous high dose of a drug , indicating the sequential drug release over several minutes to hours. And this can be achieved by the protease cleavage activity. Nowadays, various mechanisms of controlled drug release are described and their application depends on the way of administration as well as the type of infection. However, novel concept of nanocarriers enabling the drug targeting to specific cells is considered as very beneficial .