MODELLING

Introduction to Dry Lab

Mathematical and molecular dynamic modellings are important for integrating biological data and conducting confirmatory preexperiments. The modelling work of the dry lab can be categorized into four parts:
     1.Modelling of ion concentrations inside and outside the membrane affecting the binding efficiency of heavy metal ions
     2.Modelling for functional validation of lead-binding protein PbRr
     3.Modelling to verify the possibility of mutual cross-linking of OMVs
     4.Modelling of E. coli growth curves and OMV concentrations

Parts of Modelling

Inside or outside?

After determining the general direction of the project, we wanted to determine whether heavy-metal binding proteins should be expressed on the outer side of the bacterial outer membrane or in the gap between the inner and outer membranes, so we modeled the process of transmembrane transport and binding of ions inside and outside the vesicle.
     We used python to simulate the process of transmembrane transport and binding of ions inside and outside the vesicle and calculated the ion concentration at each time step.

Fig. 1 The intra-and extra-membrane ion concentrations for each time step.

Then We used Comsol to graphically simulate the above simulation process, which takes into account transmembrane transport, but does not consider the binding process, and more intuitively simulates the changes in ion concentration inside and outside the vesicle. (Fig. 2)

Fig. 2 Graphical simulation of heavy metal ion concentrations inside and outside the membrane.

It can be seen that the concentration outside the membrane is always higher than the concentration inside the membrane, and the concentration inside the membrane gradually increases over time, but it is always below the concentration outside the membrane. It is not difficult to imagine that the same heavy metal ion-binding protein works more efficiently at higher ion concentrations. Therefore we decided to express the heavy metal ion-binding protein on the outer surface of the outer membrane.

Lead-binding protein

In early September, during a discussion with colleagues from the SJTU-software, we mentioned that we had designed a ClyA-PbRr protein for capturing Pb2+ in solution and wanted to simulate how this protein binds to lead ions. Consequently, they proposed to assist in modeling and analyzing the interaction between this protein and metal ions. After receiving the protein structure predicted by AlphaFold2 and its corresponding sequence from the modeling group, they initially used BLAST for sequence alignment and successfully identified a rich set of homologous sequences. BLAST results showed that the lead ion-binding functional domain of ClyA-PbRr is nearly identical in sequence to that of a Cd(II)/Pb(II)-responsive transcriptional regulator (NCBI Reference Sequence: WP_000405672.1).
     After reviewing literature related to this protein, they hypothesized that the mechanism by which PbRr protein binds to lead ions involves the lead ions being positioned within three parallel alpha helices, with the protein effectively clamping onto the ions³.
     Subsequently, they examined the structure of ClyA-PbRr using PyMol and confirmed the presence of two regions with three alpha helices within the PbRr domain. They then inserted lead ions into these two regions and performed energy minimization optimization and molecular dynamics simulations using AMBER18. The simulation results indicated that the three-helix structure closer to the middle of the ClyA domain could stably accommodate lead ions, while the three helices closer to the end of the domain could not maintain stable positions for the lead ions during the simulation. By zooming in on the middle three-helix region, they identified residue side chains that could potentially interact with the lead ions during the simulation. (Video. 1)

As a result, through dynamic simulations, we obtained a stable structure of ClyA-PbRr bound to lead ions and partially elucidated the mechanism by which it binds to lead ions(Fig. 3). Based on this mechanism, we believe that the PbRr domain can also capture heavy metal ions of a similar size, such as Cd²⁺.

OMV cross-linking

The Tag and Catcherd structures of docking proteins were predicted using the latest Alpahfold2, and we selected the result with the highest score as our predicted protein structure as follows. (Fig. 3 A,B)
     Then we used the HADDOCK tool to model the protein docking, and the model with the highest score was selected as our docking model through iteration and clustering. The model results showed that our model had lower electrostatic potential energy in the docking region and better binding ability between proteins^4,5. (Video 2, Fig. 3 C,D)

Fig. 3 Structural model of cross-linked proteins.

The results show that SpyTag and SpyCatcher can interact with each other, which indicates that our idea of expressing these two proteins on the surface of OMV to achieve OMV cross-linking is established.

Growth curves and OMV concentrations

In order to determine the replacement frequency of the bacterial agent when our product is put into use, we experimentally measured the OD600 value of the bacterial solution⁶ and OMV concentration at each time point⁷, and plotted the growth curve and the OMV concentration changing curve with time.
     We measured the growth concentration of E. coli as a function of time as shown below(Fig. 4):

Fig. 4 Growth concentration of E. coli.

The results of fitting CPH1 growth curve according to the Logistic model are as follows(Fig. 5):

Fig. 5 Growth curve of E. coli.

In addition, we measured and fitted the curve of OMV concentration in this strain over time, and the results are as follows(Fig. 6):

Fig. 6 OMV concentration changing curve with time

References

     [1] Burger H M, Seebach D. Cyclosporin A: Li- and Ca-specific ionophore. Angew Chem Int Ed Engl, 1994, 33: 442 444
     [2] Ozeki E, Kimura S, Imanishi Y. Calcium transport through lipid membrane by diastereomer cyclic octapeptides.Biopolymers, 1989, 34: 1247 1257
     [3]Zheng C, Li Y, Nie L, Qian L, Cai L, Liu J. Transcriptional and functional studies of a Cd(II)/Pb(II)-responsive transcriptional regulator(CmtR) from Acidithiobacillus ferrooxidans ATCC 23270. Curr Microbiol. 2012 Aug;65(2):117-21. doi: 10.1007/s00284-012-0117-4. Epub 2012 May 4. PMID: 22555344; PMCID: PMC3378830.
     [4]R.V. Honorato, P.I. Koukos, B. Jimenez-Garcia, A. Tsaregorodtsev, M. Verlato, A. Giachetti, A. Rosato and A.M.J.J. Bonvin (2021). "Structural biology in the clouds: The WeNMR-EOSC Ecosystem." Frontiers Mol. Biosci., 8, fmolb.2021.729513.
     [5]G.C.P van Zundert, J.P.G.L.M. Rodrigues, M. Trellet, C. Schmitz, P.L. Kastritis, E. Karaca, A.S.J. Melquiond, M. van Dijk, S.J. de Vries and A.M.J.J. Bonvin (2016). "The HADDOCK2.2 webserver: User-friendly integrative modeling of biomolecular complexes." J. Mol. Biol., 428, 720-725 (2015).
     [6]Tuttle AR, Trahan ND, Son MS. Growth and Maintenance of Escherichia coli Laboratory Strains. Curr Protoc. 2021 Jan;1(1):e20. doi: 10.1002/cpz1.20. Erratum in: Curr Protoc. 2022 Aug;2(8):e552. Erratum in: Curr Protoc. 2022 Aug;2(8):e551. PMID: 33484484; PMCID: PMC8006063.
     [7]Zhao X, Zhao R, Nie G. Nanocarriers based on bacterial membrane materials for cancer vaccine delivery. Nat Protoc. 2022 Oct;17(10):2240-2274. doi: 10.1038/s41596-022-00713-7. Epub 2022 Jul 25. PMID: 35879454.