A Novel FRET Technique to Characterize the Oligomerization State of Protein-Protein Interactions

452a

Tuesday, February 14, 2017

protein, with information of how often each residue are included in binding sites of high scoring docking models. We applied multiple methods of machine learning to discriminate binders and non-binders using those profiles as input. In this poster we present an evaluation of our method by applying it to proteins of Protein Docking Benchmark ver. 5.0, bacterial chemotaxis, and host-virus protein interactions.We further applied our method to analyze proposed hostvirus PPI interaction surfaces by post-docking analysis to compare surfaces of PPIs between the host proteins. 2222-Pos Board B542 In Silico Screening for Chemical Scaffolds as Suitable Natural Inhibitors of Kinesin EG5 Divulges Morelloflavone, a Biflavonoid, as Potential Anticancer Compound Tomisin Happy Ogunwa, Takayuki Miyanishi. School of Fisheries and Environmental Sciences, Nagasaki university, Nagasaki, Japan. Natural products remain a source of chemical scaffold pool for drug design and development. Kinesin Eg5 has emerged as a clinical target for anticancer agents. In our search for potent natural inhibitors of kinesin Eg5, we employed in silico tools to screen selected diverse chemical scaffolds (compounds) that were obtained from medicinal plants. Surprisingly, the molecular interaction analysis for the selected compounds adjudged morelloflavone (a biflavonoid) as a potential ATP-noncompetitive inhibitor of kinesin Eg5 protein which occupied the putative L5/a2/a3 allosteric pocket on the protein. Compared to STLC with binding energy value 10.0 Kcal/mol, morelloflavone displayed binding energy value of 10.2 Kcal/mol and a 90 percent binding site similarity. It is also worth noting that morelloflavone was embedded within the cavity formed by amino acid residues Ile-136, Glu-116, Glu-118, Trp-127, Gly-117, Ala-133, Glu-215, Leu-214, Tyr-211 and hence, displayed a reliable tendency to block the enzymatic catalysis of kinesin Eg5 allosterically. The compound established hydrogen bonds with Glu-118 and Tyr-211 having minimum length ˚ and hydrophobic interactions occurred with alkyl side chain of resiof 2.97A dues Gly-117, Glu-116, Ala-218, Ile-136, Arg-119 and Asp-130 while p-stacking interaction is observed between the aromatic ring of morelloflavone and Arg-119. These interactions anchored morelloflavone into the binding site. The results obtained in this work indicate the strong affinity and inhibitory potential of this compound on kinesin Eg5, hence lending credence to the yet untapped anticancer capacities of morelloflavone. We therefore suggest in vitro and ex vivo evaluation of this compound as anticancer agent targeting kinesin Eg5 protein. 2223-Pos Board B543 Development of Postprocessing Method of Protein-Ligand Docking using Interaction Fingerprint Nobuaki Yasuo1, Masakazu Sekijima1,2. 1 Department of Computer Science, Tokyo Institute of Technology, Tokyo, Japan, 2Advanced Computational Drug Discovery Unit, Tokyo Institute of Technology, Tokyo, Japan. Protein-ligand docking is an important method in Structure-based Drug Discovery [1]. Although many programs have been developed for docking [2], the accuracy is still insufficient due to the difficulty in the scoring function [3]. Interaction fingerprint is one of the solutions, which generate fingerprints of ligands using the interactions between the ligand and the protein. Interaction fingerprints use the information of known compounds so that compounds that have similar interaction to the known active ligands are expected to find through the virtual screening. However, existing interaction fingerprints such as SIFt [4] and SPLIF [5] only assess the existence or the distance of the interactions and do not consider the strength correctly. In this study, we made a new scoring function of protein-ligand docking called SIEVE-Score (Similarity of Interaction Energy VEctor-Score), which can consider the strength of each interaction explicitly. SIEVE-Score is calculated based on the similarity of the interaction energy vector, which is the list of interaction energy between the ligand and each residue of the protein. We also evaluate the accuracy of virtual screening using SIEVE-Score after the docking by Glide [6]. [1] Chiba, S., et al. Identification of potential inhibitors based on compound proposal contest: Tyrosine-protein kinase Yes as a target. Scientific reports 5:17209, 2015. [2] Elizabeth Y., Jessica H., and Paul A. R., Improvements, trends, and new ideas in molecular docking: 2012–2013 in review. Journal of Molecular Recognition, 28(10):581–604, 2015. [3] Yan L., Li H., Zhihai L., and Renxiao W., Comparative assessment of scoring functions on an updated benchmark: 2. evaluation methods and general results. Journal of Chemical Information and Modeling, 54(6):1717–1736, 2014.

[4] Zhan D., Claudio C., and Juswinder S., Structural interaction fingerprint (SIFt): a novel method for analyzing three-dimensional protein-ligand binding interactions. Journal of Medicinal Chemistry, 47(2):337–344, 2004. [5] Da C., and Kireev D., Structural protein–ligand interaction fingerprints (SPLIF) for structure- based virtual screening: Method and benchmark study. Journal of Chemical Information and Modeling, 54(9):2555–2561, 2014. [6] Richard A. F., et al., Glide: a new approach for rapid, accurate docking and scoring. 1. method and assessment of docking accuracy. Journal of Medicinal Chemistry, 47(7):1739–1749, 2004.

Optical Spectroscopy: CD, UV-VIS, Vibrational, Fluorescence I 2224-Pos Board B544 New Thiol-Reactive Eu-Complex for Distance Measurements by LRET Felix Faschinger1, Mirjam Zimmermann1, Guenther Knoer2, Hermann Gruber1. 1 Institute of Biophysics, Johannes Kepler University, Linz, Austria, 2Institute of Inorganic Chemistry, Johannes Kepler University, Linz, Austria. Crystallography and NMR spectroscopy are ideally suited to resolve the 3D structures of biomolecules but the material and time demand for each structure is high. Fluorescence resonance energy transfer (FRET) provides less structural information but is better suited to study conformational changes and structurefunction relationships by screening a large number of mutants or experimental conditions. Moreover, FRET allows for real time monitoring of conformational changes induced by specific ligands. Usually, FRET yields only a crude estimate of the donor-acceptor distance, due to the fact that the relative orientation of donor and acceptor are rarely known. Luminescence resonance energy transfer (LRET) is much better suited for distance measurements because the orientation factor (and thus the Forster distance) are known, and because energy transfer is measured by a change of lifetime, rather than of signal intensity. In LRET-experiments the ideal donors are highly stable Eu/Tb-complexes, with a single lifetime that is not influenced by attaching the complex to a biomolecule. Several terpyridine-based Eu-complexes described in literature have promising properties concerning uniform lifetimes after protein labeling but all described complexes have rather long linkers which prevent accurate distance measurements. In this study, a new terpyridine-based Eu-complex with maleimide very close to the metal ion center was synthesized and found to have ideal properties for distance measurement by LRET: After linking to the single cysteine of BSA, the complex showed a quantum yield of 30%, a single lifetime of 1.2 ms, comparable to the best known Eu-complexes, and the lifetime was unaffected by phosphate or EDTA. In conclusion, this new Eu complex appears ideally suited for reliable measurement of intra- or intermolecular distances. The work was supported by grant W0125 of the Austrian science fund (FWF) 2225-Pos Board B545 A Novel FRET Technique to Characterize the Oligomerization State of Protein-Protein Interactions Philipp J. Heckmeier, Mark G. Teese, Dieter Langosch. Lehrstuhl f€ur Chemie der Biopolymere, Technische Universit€at M€unchen, Freising, Germany. Protein-protein interactions are the fundamental driving force of numerous cellular processes and cell signaling pathways. Characterizing whether proteins interact as dimers, trimers, or higher oligomers is essential to understanding these interactions. Several microscopy and advanced imaging techniques relying on Fo¨rster resonance energy transfer (FRET) between identical fluorophores (homo-FRET) have been developed to estimate protein stoichiometry. The increased FRET in oligomers is detected by measuring depolarization or emission time. Homo-FRET methods have a strong advantage in requiring only a single fluorophore, greatly simplifying sample preparation in comparison to conventional hetero-FRET methods. However, most homo-FRET methods require sophisticated imaging equipment, and both theoretical models and applications have been restricted to the study of membrane-bound proteins. Using a simple bulk homo-FRET and laser photobleaching approach, we demonstrate the feasibility of characterizing the oligomerization state of an interacting protein in-vitro. To simulate oligomers in a proof of concept, we constructed an extensive repertoire of fusion proteins with 1-6 consecutive green fluorescent protein (GFP) domains. We show how the resulting homo-FRET (measurable via steady-state anisotropy or fluorescence polarization) is proportional to the oligomerization state of proximal GFP domains. For the first time, this is demonstrated with soluble proteins. In both membrane and soluble proteins, oligomerization increases FRET and therefore anisotropy. However for soluble proteins oligomerization also slows fluorophore rotation, leading to a size-dependent decrease in anisotropy. Through gradual photobleaching of

Tuesday, February 14, 2017 fluorophores these two effects can be distinguished, and the oligomerization state of a labeled protein of interest can be estimated. We therefore show how the theoretical framework developed for membrane proteins needs to be adjusted to account for this additional degree of freedom in soluble proteins. Overall, bulk homo-FRET and laser photobleaching is a promising method to determine the oligomerization state of a protein of interest, which can have a low concentration (0.1-0.5 mM) and needs only a single fluorescent label. The method requires only a photometer or microplate reader capable of measuring steady-state anisotropy. 2226-Pos Board B546 Optimizing a Time-Resolved Spectrometer for All Time Scales Christian Litwinski, Sebastian Tannert, Manoel Veiga, Felix Koberling, Marcus Sackrow, Michael Wahl, Olaf Schulz, Marcelle Koenig, Rainer Erdmann. PicoQuant GmbH, Berlin, Germany. Time-resolved fluorescence spectroscopy is a spectroscopist’s most valuable tool for the investigation of excited state dynamics in molecules, complexes, or semi-conductors. In recent years, the study of luminescence properties has gained in popularity in many scientific fields, including Chemistry, Biology, Physics, as well as in Life, Material or Environmental Sciences. The investigations to be carried out in each of these fields impose different requirements. On one side, monitoring dynamic processes in the excited state necessitates high time resolution that can be achieved by fast pulsed lasers and detectors along with appropriate time-correlated single photon counting (TCSPC) units and small monochromators. On the other hand, high spectral resolution is desirable for fluorophore characterization, requiring detectors with high quantum efficiencies, flash lamps for phosphorescence measurements and large monochromators. Up to now, spectrometers have been usually targeted towards either one of these two specifications. Spectrometers equipped with hybrid detectors, versatile TCSPC cards with optional longer time ranges, and pulsed lasers capable of working in a burst mode can offer an combined solution, covering most of the demands of either high time or spectral resolution. We will demonstrate the performance of such a spectrometer in terms of its time resolution, the ability to measure long decays and record time-gated spectra using laser drivers with burst capabilities. This type of instrument is of great value for analytical facilities in research centers, as it offers a wide range of possible spectroscopic applications in a single, easy to use instrument. 2227-Pos Board B547 Hyperspectral Measurements Allow Separation of FRET Signals from Non-Uniform Background Fluorescence Savannah J. West1, Chase Hoffman2, Naga S. Annamdevula2, Kenny T. Trinh3, Thomas C. Rich2, Silas J. Leavesley3. 1 Biomedical Sciences, University of South Alabama, Mobile, AL, USA, 2 Pharmacology, University of South Alabama, Mobile, AL, USA, 3Chemical and Biomolecular Engineering, University of South Alabama, Mobile, AL, USA. In recent years Fo¨rster resonance energy transfer (FRET) has become a standard imaging approach to gain insight into localized biochemical processes within cells. These processes include changes in protein colocalization, second messenger concentration, and activation of effector proteins such as protein kinase A (PKA). However, there are several limitations in FRET measurements that are not often considered. In recent years our group has used hyperspectral imaging approaches to address a subset of these limitations, including the inherently low signal-to-noise ratio of standard two and three filter set FRET measurements. Here we present data demonstrating that hyperspectral imaging approaches can be used to correct for non-uniform background fluorescence in low intensity FRET measurements. We utilized the FRET-based cAMP probe H188. The H188 probe contains a cAMP binding domain between donor (Turquoise) and acceptor (Venus) fluorescent proteins. The probe was transfected into pulmonary microvascular endothelial cells plated on glass coverslips. Hyperspectral image stacks were acquired using a Nikon A1R inverted confocal microscope. Changes in cAMP were triggered by addition of 0.1 mM isoproterenol (a beta adrenergic receptor agonist) 0.1 mM PGE1 (a prostanoid receptor agonist) or 50 mM forskolin (an adenylyl cyclase activator). Data were analyzed using Nikon Elements software and custom MATLAB scripts. Results demonstrate that non-uniform background fluorescence associated with glass coverslips can contaminate FRET measurements from weak fluorescence signals. Linear unmixing approaches were able to separate the abundances of background, donor and acceptor fluorescence signals. Thus, results from this study demonstrate that hyperspectral unmixing approaches can readily separate nonuniform background fluorescence from other fluorescence signatures, allowing for more accurate quantification of FRET efficiency. This work was supported by NIH P01HL066299, NIH S10RR027535, NIH T32HL076125, AHA

453a

16PRE27130004, USA SURF Program, and the Abraham Mitchell Cancer Research Fund. 2228-Pos Board B548 Fluorescent Proteins for Super-Resolution Microscopy Karin Nienhaus1, Gerd U. Nienhaus1,2. 1 Institute of Applied Physics, Karlsruhe Institute of Technology, Karlsruhe, Germany, 2University of Illinois at Urbana-Champaign, Urbana, IL, USA. Super-resolution fluorescence microscopy is the method of choice to monitor cellular and subcellular biological processes in live cells. Among the different fluorescent labels presently available, fluorescent proteins (FPs) of the GFP family have the key advantage of being genetically encodable. Localizationbased super-resolution microscopy approaches require photoactivatable FPs (PA-FPs) that will change their spectral properties upon irradiation with light of a particular wavelength. To be able to distinguish individual, activated fluorophores from the background and to localize them with high precision, a high photon yield in the activated state and a high dynamic range, i.e., the contrast ratio between the fluorescence of the activated (on) and deactivated (off) states, are essential. In stimulated emission depletion (STED) super-resolution microscopy, the sample is raster-scanned by a tightly focused excitation beam followed by a red-shifted, donut-shaped depletion beam. Any FP used for STED must be exquisitely photostable because it has to go through multiple excitation-depletion cycles while the sample is scanned near its location. Moreover, it must be excitable and de-excitable by the lasers that are typically installed in commercial STED microscopes. Therefore, far-red emitting FPs are preferred. We have selected the green-to-red photoconvertible mEosFPthermo and the far-red emitting mGarnet as templates for targeted protein engineering. Considering that FPs are all very similar and share the same scaffold, an obvious strategy was to identify specific amino acid residues that elicit certain properties to one FP and introduce the corresponding amino acid in the other FP variant by using site-directed mutagenesis. As we will show, such simple rational engineering approaches often do not meet with success, which clearly shows that our current understanding of the physics of proteins is far from being complete. 2229-Pos Board B549 Wide Scale Investigation of Protein Interactions by Automation of Fluorescent Polarization and Fluctuation Analysis Tuan A. Nguyen, Grace H. Taumoefolau, Youngchan Kim, Henry L. Puhl, Steven S. Vogel. NIAAA, NIH, Rockville, MD, USA. Monitoring changes in molecular conformation is essential for studying protein interactions within and between complexes. Methods such as FRET or FCS reveal limited aspects of these changes, which in many cases is insufficient for interpretation. Fluorescent Polarization and Fluctuation Analysis (FPFA), a time-correlated single-photon counting technique that combines homoFRET and FCS was developed to address this problem. In FPFA changes in protein complex mass and/or shape, the number of fluorescent subunits per complex, as well as subunit proximity (1 – 10 nm), is simultaneously detected. This multimodal approach was validated with series of 6 fluorescent oligomers composed of between 1 and 6 concatenated Venus molecules, and successfully employed to investigate structural dynamics of calcium/calmodulin-dependent protein kinase II (CaMKII) – a multimeric protein kinase that is enriched in synaptic spines and dendrites that is involved in memory and synaptic modulation, as well as microtubule-associated protein 1A/1B-light chain 3 (LC3) – a soluble protein that is a key component of autophagy, a homeostasis process responsible for the turnover of cellular components. Here we have extended FPFA method by developing a robotic microscope that can measure up to 96 samples automatically. This FPFA automation enables wide-scale biophysical analysis of protein-protein interactions, and hence facilitates the identification of drugs that target protein complexes as potential therapeutic sites for diseases of aggregation and abnormal protein-protein interactions. 2230-Pos Board B550 Comparison Study on Fluorescence Quenching Ability of DNA Wrapped Single- and Multi-Walled Carbon Nanotubes Shusuke Oura, Katsuki Izumi, Kazuo Umemura. Tokyo University of Science, Shinjuku-ku, Japan. In terms of superior physical properties of carbon nanotubes (CNTs), CNT has been intensively studied for various applications including biosensors and drug delivery. For medical applications, hybridization of single-stranded DNA (ssDNA) and CNT (ssDNA-CNTs) is key to utilize them multiply because DNA can bind to various biomolecules. For example, Li et al. used ssDNACNTs for nucleic acid detection by applying the reaction of fluorescent molecules and CNTs. It is known that, due to electron transfer, fluorescence