Bengt Svensson is a Research Associate in the group led by Professor David Thomas (MSI Fellow; Biochemistry, Molecular Biology, and Biophysics). Dr. Thomas specializes in muscle research. Dr. Svensson came to the University as a post-doc in 1999. He worked in the labs of Bridget Barry (BMBB) and David Ferguson (Medicinal Chemistry) before joining the Thomas lab (a.k.a. the Minnesota Muscle Lab) nearly six years ago. He has been using MSI’s HPC and laboratory resources while with the Thomas lab, and also used MSI laboratories during his time in the other groups. His poster was selected as a finalist in the poster competition at the 2013 Research Exhibition. The title of the poster was, “Simulations of a Fluorescent Probe Attached to SERCA.” Dr. Svensson sat down with MSI recently to talk about his research and this poster.
MSI: You’ve been doing similar research to this poster for several years now. You did an article for us in our Research Bulletin, back in 2011.
Bengt Svensson: Yes, this is part of the same research project.
MSI: Why don’t you tell me about what you’re doing in this poster?
BS: Researchers in the David Thomas lab do spectroscopy studies on muscle-related proteins. In the research project presented on the poster the protein studied is SERCA, the Sarco/Endoplasmic Reticulum Ca2+-ATPase. SERCA is the protein that regulates the muscle by making the muscle relax after contraction. It does that by pumping calcium ions in to the Sarco/Endoplasmic Reticulum, lowering the calcium concentration around the muscle fibers.
MSI: So what you’re doing is creating models of how this process works.
BS: Yes, because we have data from spectroscopy measurements that can be interpreted in a structural biology context. We can get information about amount of motion, distances, and relative orientation, but we don’t really know what those mean unless we can create an atomic model that’s actually going to match the experimental data. The experimental results come from samples where we have attached artificial probes to these proteins, in this case fluorescence labels. From simulations we intend to find out the conformation and motion of those probes. Because if the probes are moving really fast, then their locations are going to be averaged out to a central location, but if they’re moving really slow, then every single possible conformation, or position, is going to be important when we’re trying to interpret a distance measurement from spectroscopy. We are aiming to determine whether specific measurements match an atomic model or not.
So simply expressed, this research project, as presented on this poster, is about modeling how the fluorescent probe is behaving while it’s attached to the protein.
MSI: On the poster, you have images of the protein molecule.
BS: Those images are based on an X-ray crystal structure of the SERCA protein. We have a collaborator in Canada, Howard Young [a co-author on the poster], who has crystallized this protein with one of our probes on it. From that crystal structure, however we only know the conformation promoted by crystallization conditions, and crystal contacts but that’s not going to be the only conformation under natural physiological conditions. Because atoms, protein side chains, and protein regions are expected to be moving around quite a bit. This crystal structure does provide an extremely useful snapshot, which I use as the starting point for simulations or different kinds of computational analysis.
The next thing we did was to perform time-resolved fluorescence anisotropy experiments. Fluorescent molecules absorb light and then emit light at a longer wave length. Fluorescence anisotropy experiments measure the depolarization of fluorescence and give us informations about probe mobility. It tells us both the rate and the amplitude of the motion of the fluorescent probe. From anisotropy experiments we would, in the ideal case, expect to see an exponential decay starting from the theoretical maximum value. We noticed that, in our experiments, the curves on the plots don’t start at that maximum. That means that the light detected is already de-polarized to some extent before we can detect it using our fluorescence instrument.
MSI: So that’s telling you that the probe is moving around?
BS: It tells me that something has happened in the probe itself. Bond stretches - vibrational reorganization have occurred causing the fluorescent light to be emitted at an angle different from what was absorbed by the probe. We have to find out exactly what that angle is, because that’s going to be important when we compare the experimental results with data obtained from molecular dynamics simulations. An approach was used that incorporate quantum chemistry calculations of the excited state of the fluorescence probe and varying the angle between the absorption and emission in the analysis of the molecular dynamics simulation data. We were able to determine the information needed to quite accurately predict the experimental anisotropy results from the molecular dynamics simulations.
MSI: You’re using molecular dynamics simulations. Do you use the supercomputers at all, or do you use the labs?
BS: For these molecular dynamics simulations, since I just simulate the small region of the protein around the probe, it doesn’t make sense to use the supercomputers. These simulations, using implicit solvent and a small number of atoms, only scale up to two or four cores. I mostly run my simulations on the lab queue system, the BSCL lab queue. I do have a couple of computers in my office that I use if I’m just going to submit one or two jobs, but for this project it’s important to have enough data for statistical analysis. I have to run multiple simulations, because averaging several data sets is needed to get interpretable results. For this project I have run 18 simulations for 160 ns each; and that is quite a lot of computer simulation time.
MSI: Is there anything else we should know about this project?
BS: What is presented on this poster are the initial steps for us to be able to go on to further studies determining important structural conformations of SERCA based on results from fluorescence spectroscopy. To make sure we could simulate the probe attached to the protein accurately, we compared data from simulations with fluorescence anisotropy measurements. We believe that we have got the needed parameters right, and we can now use our simulations to interpret results from other fluorescence experiments, like FRET. FRET (Fluorescence Resonance Energy Transfer) experiments are used to measure distances between two fluorescence probes that are attached to different parts of the protein. We have several of these experiments going on in the Thomas lab already
Current hypotheses proposed from available crystal structures suggest that some distances between different parts of the protein will change when the protein is preforming its function. One of our approaches is to use specific FRET distances to verify whether a proposed structural model matches distances measured in a specific biochemical state. The structural models can give us a steady-state snapshot of one conformation, but a working protein is not just flipping between static states. Proteins are always flexible to some degree, and that protein dynamics is really important for a lot of protein functions.
For example, as shown in our recent publication (Pallikuth et al., 2013), we observed that in the biochemical high calcium state, the FRET was higher than in the low calcium state, indicating a shorter distance between the cytoplasmic protein domains. In each of these biochemical states, there were multiple structural states observed. Similar conclusions were made from all-atom molecular dynamics simulation of SERCA (Espinoza-Fonseca, & Thomas, 2011). We can actually quantify how much protein dynamics is observed in each individual state. By using a combination of simulation and experiments, we have discovered new insights into how SERCA works during its catalysis.
MSI: You anticipated my next question - this is the sort of research that’s expanding our fundamental knowledge. It’s not drug design.
BS: The parts I’m involved in are not directed to any product or any cures. Of course, once we have a better understanding of how things work, that’s going to help us with those things, too.
MSI: I know there are high blood pressure medications, calcium channel blockers - are they based on the same sort of thing, this system?
BS: Not really. Those calcium channel blockers affect a different system, the calcium channels that are part of the neurochemical signaling that causes contraction. There are some drugs that affect the function of SERCA. For example beta blocker and stimulator drugs which affect SERCA by affecting another protein that is responsible for its regulation. I believe that I described the regulation of SERCA activity by the small protein phospholamban in the previous MSI Research Bulletin article. Phospholamban is affected by beta adrenergic stimulation, which causes phospholamban to be phosphorylated, and less inhibitory of SERCA. That means, it makes SERCA better at relaxing the muscle, reducing stress on the heart. The Thomas lab is collaborating with a company called Celladon and with scientists at the Icahn School of Medicine at Mount Sinai (New York) on a project that targets phospholamban for gene therapy for heart failure.
SERCA is regulating calcium levels in non-muscle cells too, and may be of interest for several other medical reasons. For example, inhibition of SERCA will cause endoplasmic reticulum stress and trigger cell death. SERCA inhibitors could be useful in cancer treatment because of this. I don’t think there are any current uses because of the toxicity of the inhibitors on healthy cells. If one could develop a safe localized delivery into cancer cells, it could work.
The Thomas group has had numerous publications in recent years. A sample of publications that used MSI resources are shown below. MSI PIs are shown in bold text.
- “Phosphorylated Phospholamban Stabilizes a Compact Conformation of the Cardiac Calcium-ATPase,” S Pallikuth, D Blackwell, Z Hu, Z Hou, DT Zieman, B Svensson, DD Thomas, SL Robia, Biophysical Journal, 105:1812-1821 (2013)
- “Nucleotide Activation of the Ca-ATPase,” JM Autry, JE Rubin, B Svensson, J Li, DD Thomas, Journal of Biological Chemistry, 286(46): 39070-39082, DOI:10.1074/jbc.M112.404434 (2012)
- “TRAIL-Induced Death Receptor 5 Networks are Highly Organized,” CC Valley, JD Perlmutter, AK Lewis, AR Braun, CB Karim, DD Thomas, JR Brody, J. R., JN Sachs, Journal of Biological Chemistry, 287(25): 21265–21278 (2012)
- “Atomic-level Characterization of the Activation Mechanism of SERCA by Calcium,” LM Espinoza-Fonseca, DD Thomas, PLoS One, 6:e26936 (2011)
- “Large-scale Opening of Utrophin’s Tandem CH Domains Upon Actin Binding, by an Induced-Fit Mechanism,” AY Lin, E Prochniewicz, ZM James, B Svensson, DD Thomas, Proceedings of the National Academy of Sciences of the USA, 108:12729–12733 (2011)
- “A Continuous Fluorescence Displacement Assay for BioA: An Enzyme Involved in Biotin Biosynthesis,” DJ Wilson, C Shi, BP Duckworth, JM Muretta, U Manjunatha, YY Sham, DD Thomas, CC Aldrich, Analytical Biochemistry (2011)
- “Atomic-level characterization of the activation mechanism of SERCA by calcium.” LM, Espinoza-Fonseca, DD Thomas. PLoS One, 6:e26936 (2011).
- “Phosphorylation-Induced Structural Changes in Smooth Muscle Myosin Regulatory Light Chain,” D Kast, LM Espinoza-Fonseca, C Yi, DD Thomas, Proceedings of the National Academy of Sciences of the USA, 107:8207 (2010)
posted on February 19, 2014