Final Report Summary - D_IN_VIVO (A general law describing the diffusion of membrane proteins in vivo based on single molecule tracking of membrane proteins in Escherichia coli.)
Quantitative biology a field that aims to assign actual numbers to the key cellular processes, such as reaction rates and diffusion coefficients, in vivo. In this project we proposed an experimental set-up to measure the diffusion of membrane proteins in live Escherichia coli cells using single molecule tracking (SMT). By performing SMT of membrane proteins systematically increasing in size (radius in the membrane) a general model of protein diffusion in the membranes was to be achieved. There are different models describing diffusion of membrane proteins but the most widely accepted one is the Saffman-Delbrueck (the SD model). The model has been shown to hold true in vitro and the question at hand was whether the model also holds in vivo. By addressing the individual objectives of the proposal we anticipated to answer this question. The objectives (O) included:
O1 : To establish relationship between the D and the number of transmembrane helices in vivo.
O2: To determine whether the diffusion of membrane proteins in vivo is normal or anomalous.
O3: To probe whether the diffusion of membrane proteins is homogenous throughout the cell.
O4: To determine the contribution of soluble parts of the membrane proteins to their diffusion coefficient.
A number of genetic constructs has been made, where the membrane proteins were C-terminally tagged with a photoconvertible fluorescent protein mEos2. The constructs included three different membrane proteins and their tandem repeats spanning a range between 6 and 36 transmembrane helices and they were already mentioned in the mid term report.
Subsequently, measurements were made to acquire single molecule tracking data to calculate mean square displacements and two technical problems have arisen: the first one concerning the Escherichia coli cell geometry and the second one concerning the blinking behaviour of the fluorescent protein mEos2. The first problem was solved as described in the mid term report. For the second problem it was decided it will be easier to acquire data if a different fluorescent protein – PamCherry will be used. A decision has been made to make new constructs using PAmCherry instead of mEos2.
Shortly after submitting the mid term report we have become aware that a group at VU in Amsterdam is working on a nearly identical project. Eventually their work has been published (“MreB-Dependent Organization of the E. coli Cytoplasmic Membrane Controls Membrane Protein Diffusion” Oswald et al. Biophys J (2016) 110:1139-1149). In that study the authors have determined the relationship between the diffusion coefficient of membrane proteins of increasing size and their diffusion coefficient in Escherichia coli using single molecule tracking. The diffusion vs. membrane protein size relationship was according to the Saffman Delbrueck relationship, as we have anticipated in our proposal. Since this paper relied on the same model organism, used the same experimental method for measurement and nearly an identical experimental set up, we have decided not to continue with the original experiments described in the proposal. The motivation behind this decision was that this other publication compromised the novelty of our eventual findings. Instead we have decided it will be most efficient to devote the remaining time to the collaborative project with Imperial College London. This collaborative project also addresses properties of the Saffan Delbrueck model in vivo.
The SD model has been shown to also hold true for membranes of E.coli by Oswald et al (see above). In this model the diffusion of membrane proteins scales logarithmical with the size they occupy in the membrane. Another key factor determining the rate of diffusion is the viscosity of the membranes. If we aim at using the Saffman-Delbrueck model to predict/quantitatively estimate the diffusion of membrane proteins (O1, see above), it is crucial to know the viscosity value of E.coli plasma membrane. The viscosity of the plasma membrane of live Escherichia coli cells has not been determined before. In the collaborative project with the group of Dr. Marina Kuimova at Imperial College London we have used molecular rotors - probes her group developed - to determine the viscosity of live E.coli plasma membranes, spheroplasts and vesicles made of E.coli lipid extracts. The measurements were made using FLIM microscopy on single cells.
We have measured that the viscosity of the E.coli plasma membrane is 950 cP and making it very viscous environment as compared to eukaryotic cells or model membrane systems. It also suggests a high degree of lipid ordering within the liquidphase membrane. The viscosity value matched well with the estimate obtained using the Saffman Delbrueck model (Oswald et al) of 1000-1200 cP. This is the most significant finding of this project. Together with the study of Oswald et al we have now parameters that describe quantitatively the diffusion of membrane proteins in E.coli.
Our results were published in an article in Biophysical Journal (“Measuring the Viscosity of the Escherichia coli Plasma Membrane Using Molecular Rotors.” By J.T. Mika et al)
O1 : To establish relationship between the D and the number of transmembrane helices in vivo.
O2: To determine whether the diffusion of membrane proteins in vivo is normal or anomalous.
O3: To probe whether the diffusion of membrane proteins is homogenous throughout the cell.
O4: To determine the contribution of soluble parts of the membrane proteins to their diffusion coefficient.
A number of genetic constructs has been made, where the membrane proteins were C-terminally tagged with a photoconvertible fluorescent protein mEos2. The constructs included three different membrane proteins and their tandem repeats spanning a range between 6 and 36 transmembrane helices and they were already mentioned in the mid term report.
Subsequently, measurements were made to acquire single molecule tracking data to calculate mean square displacements and two technical problems have arisen: the first one concerning the Escherichia coli cell geometry and the second one concerning the blinking behaviour of the fluorescent protein mEos2. The first problem was solved as described in the mid term report. For the second problem it was decided it will be easier to acquire data if a different fluorescent protein – PamCherry will be used. A decision has been made to make new constructs using PAmCherry instead of mEos2.
Shortly after submitting the mid term report we have become aware that a group at VU in Amsterdam is working on a nearly identical project. Eventually their work has been published (“MreB-Dependent Organization of the E. coli Cytoplasmic Membrane Controls Membrane Protein Diffusion” Oswald et al. Biophys J (2016) 110:1139-1149). In that study the authors have determined the relationship between the diffusion coefficient of membrane proteins of increasing size and their diffusion coefficient in Escherichia coli using single molecule tracking. The diffusion vs. membrane protein size relationship was according to the Saffman Delbrueck relationship, as we have anticipated in our proposal. Since this paper relied on the same model organism, used the same experimental method for measurement and nearly an identical experimental set up, we have decided not to continue with the original experiments described in the proposal. The motivation behind this decision was that this other publication compromised the novelty of our eventual findings. Instead we have decided it will be most efficient to devote the remaining time to the collaborative project with Imperial College London. This collaborative project also addresses properties of the Saffan Delbrueck model in vivo.
The SD model has been shown to also hold true for membranes of E.coli by Oswald et al (see above). In this model the diffusion of membrane proteins scales logarithmical with the size they occupy in the membrane. Another key factor determining the rate of diffusion is the viscosity of the membranes. If we aim at using the Saffman-Delbrueck model to predict/quantitatively estimate the diffusion of membrane proteins (O1, see above), it is crucial to know the viscosity value of E.coli plasma membrane. The viscosity of the plasma membrane of live Escherichia coli cells has not been determined before. In the collaborative project with the group of Dr. Marina Kuimova at Imperial College London we have used molecular rotors - probes her group developed - to determine the viscosity of live E.coli plasma membranes, spheroplasts and vesicles made of E.coli lipid extracts. The measurements were made using FLIM microscopy on single cells.
We have measured that the viscosity of the E.coli plasma membrane is 950 cP and making it very viscous environment as compared to eukaryotic cells or model membrane systems. It also suggests a high degree of lipid ordering within the liquidphase membrane. The viscosity value matched well with the estimate obtained using the Saffman Delbrueck model (Oswald et al) of 1000-1200 cP. This is the most significant finding of this project. Together with the study of Oswald et al we have now parameters that describe quantitatively the diffusion of membrane proteins in E.coli.
Our results were published in an article in Biophysical Journal (“Measuring the Viscosity of the Escherichia coli Plasma Membrane Using Molecular Rotors.” By J.T. Mika et al)