Modeling the membrane docking geometry of the synaptotagmin 7 C2A domain through electron paramagnetic resonance

Material Information

Modeling the membrane docking geometry of the synaptotagmin 7 C2A domain through electron paramagnetic resonance
Osterberg, John ( author )
Place of Publication:
Denver, CO
University of Colorado Denver
Publication Date:
Physical Description:
1 elecvtronic file (97 pages). : ;


Subjects / Keywords:
Whitebark pine -- Seeds ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Thesis (M.S.)--University of Colorado Denver. Department of Chemistry
Includes bibliographic references.
Statement of Responsibility:
by John Osterberg.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
904609813 ( OCLC )


This item is only available as the following downloads:

Full Text


MODELING THE MEMBRANE DOCKING GEOMETRY OF THE SYNAPTOTAGMIN 7 C2A DOMAIN THROUGH ELECTRON PARAMAG NETIC RESONANCE by JOHN OSTERBERG B.S. Chemistry, Oregon State University, 2008 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Chemistry 2014




iii This thesis for the Master of Science degree by John Osterberg has been approved for the Department of Chemistry by Jefferson Knight, Chair Hai Lin Scott Reed November 20, 2014


iv Osterberg, John (M.S., Chemistry) Modeling the Membrane Docking Geometry of the Synap totagmin 7 C2A Domain through Electron Paramagnetic Resonance Thesis directed by Professor Jefferson Knight. ABSTRACT Synaptotagmin (Syt) is found ubiquitously throughou t human physiology and plays an important role along with SNARE proteins in vesicle -membrane docking and fusion during Ca2+ induced exocytosis. Syt membrane binding function derives from its two C2 domains, C2A and C2B. Syt isoforms 1 and 7 are kno wn to differ significantly in their Ca2+ sensitivity and membrane activity; in particular, the C2A domain from Syt7 binds Ca2+ and membranes much more tightly than the C2A domai n from Syt1. While the structure and membrane activity of Syt1 have been e xtensively studied for decades, the structural origins of functional differences betwee n Syt7 and Syt1 are unknown. The goal of the present study was to use site-directed spin labeling and electron paramagnetic resonance (EPR) spectroscopy to determine the membr ane docking geometry of the C2A domain from Syt7, for comparison to analogous previ ous studies with Syt1. Site-directed spin labeling was performed from a library of C2A d omain mutants with single cysteine residues inserted at key positions including the Ca2+ binding loops. EPR power saturation measurements were used to calculate dept h parameters for each spin-labeled mutant and applied to a publically accessible solut ion NMR structure to develop a suite of models representing the membrane bound state of Syt7 C2A. The resulting models depict membrane penetration of binding loops 1 and 3 of Syt7 C2A, analogous to Syt1 C2A. However, the position of binding loop 1 appea rs to be, on average, more deeply


v inserted in the membrane than the corresponding bin ding loop of Syt1 C2A, giving a possible mechanistic explanation for the slower dis sociation kinetics of Syt7 C2A relative to Syt1. The form and content of this abstract approved. I recommend its publication Approved: Jefferson Knight


vi ACKNOWLEGEMENTS I would like to thank the Knight lab for the suppo rt and assistance in making this research possible. In particular IÂ’d like to thank Favinn Maynard for her hard work on the stopped flow measurements, and Arthur Boo for h is equally hard work in purifying proteins. IÂ’d like to thank Anette Erbse for all h er time and expertise in teaching me the EPR method. IÂ’d like to thank my parents and famil y for all their love and support in getting me through my continuing education. IÂ’d li ke to thank all my graduate professors for everything theyÂ’ve taught me. And IÂ’d like to thank Jefferson Knight for all of his excellent guidance and mentoring throughout the las t few years.


vii TABLE OF CONTENTS CHAPTER I. INTRODUCTION ...................................... ................................................... ................1 Introduction to Synaptotagmin .................................................. ........................................... 1 Synaptotagmin Structure and Function .............. ................................................... ...1 Synaptotagmin Functioning as a Fusion Clamp for SNA RE Proteins ....................2 Synaptotagmin 1 Characterization and Cooperativity between C2A and C2B Domains ........................................... ................................................... .....................2 Synaptotagmin Bridging Lipid Membranes............. ................................................3 Synaptotagmin Inducing Curvature in Lipid Membranes .......................................4 Synaptotagmin 7 Characterization .................. ................................................... ......4 Introduction to Electron Paramagnetic Resonance .................................................. .......... 5 II. EXPERIMENTAL ...................................... ................................................... ................8 Materials .................................................. ................................................... ............................. 8 Protein Mutagenesis, Expression, Spin-Labeling, and Purification ..............................8 Preparation of Lipid Vesicles ..................... ................................................... ...............10 Measuring EPR Spectra ............................. ................................................... ...............11 Continuous-Wave Power Saturation Measurements ..... ..............................................11 Fluorescein Maleimide Assay to Assess Spin-Labeling Efficiency ............................13 Stopped Flow Kinetic Measurements ................. ................................................... ......13 Modeling Docking Geometry ......................... ................................................... ..........13 III. RESULTS .................................................. ................................................... ......................... 16 Site Selection, Protein Purification, and MTSSL Lab eling ............................................ 16


viii Effects of Spin-Labeling on Syt7 C2A Function ..... ................................................... .20 Orientation and Depth of Penetration by Syt7 C2A .. ..................................................2 5 IV. DISCUSSION ........................................ ................................................... ...................31 Comparing Syt7 and Syt1 Docking Models .................................................. .................... 31 Observations about Mutant Kinetics .................................................. ................................ 32 Potential Sources of Error in the Model .................................................. .......................... 32 Potential Effects of Cooperativity ................................................. .................................... 33 The Existence of Multiple Protein Conformations................................................... ........ 34 Future Experiments .................................................. ................................................... ......... 35 REFERENCES 36


ix LIST OF TABLES TABLE 1. Absorbance/Concentration data for each single c ysteine mutant ..........................17 2. Lipid aggregation data ........................ ................................................... ...............20 3. Stopped flow kinetic rates .................... ................................................... ...............21 4. Depth parameter/accessibility parameter/physica l depth data ...............................26


x LIST OF FIGURES FIGURE 1. Syt7 C2A solution NMR structure with mutated re sidues.....................................16 2. Fluorescein maleimide fluorescent image showing spin-labeling efficiency ........19 3. Doxyl lipid EPR spectra ....................... ................................................... ..............23 4. Overlay of bound/unbound EPR spectra for each S yt7 C2A mutant ....................24 5. Hyperbolic tangent fit ........................ ................................................... .................27 6. Optimized best fit docking model for Syt7 C2A ..................................................2 9 7. Solution NMR state 2 docking model for Syt7 C2A .............................................30


xi LIST OF EQUATIONS EQUATION 1. Power saturation curve fitting equation ....... ................................................... .......12 2. Accessibility parameter equation ............. ................................................... ..........12 3. Depth parameter equation ...................... ................................................... .............12 4. Hyperbolic tangent fit equation ............... ................................................... ...........13 5. Physical depth calculation ........................ ................................................... ..........14


xii LIST OF ABBREVIATIONS Syt Synaptotagmin EPR Electron Paramagnetic Resonance PIP2 phosphatidylinositol 4,5-bisphosphate WT Wild type NiEDDA Ni 2+ ethylenediamine diacetic acid EDTA Ethylenediaminetetraacetic acid POPC 1-Palmitoyl-2-oleoyl-sn-glycero-3phosphocholine POPS 1-Palmitoyl-2-oleoyl-sn-glycero-3phosphoserine SNARE soluble N -ethylmaleimide-sensivite fusion protein attachment receptor DEER double electron-electron resonance FRET Fluorescence Resonance Energy Transfer PH Pleckstrin Homology GRP1 General Receptor of Phosphoinositides 1 PKC Protein Kinase C Alpha NMR Nuclear Magnetic Resonance DMF Dimethylformamide cPLA2 Cytosolic Phospholipases A2 IPTG Isopropyl -D-thiogalactopyranoside SDS-PAGE Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis HEPES (4-(2-hydroxyethyl)-1piperazineethanesulfonic acid MTSSL 1-Oxyl-2,2,5,5-tetramethyl3 -pyrroline-3methylmethanethiosulfonate PDB Protein Data Bank


1 CHAPTER I INTRODUCTION Introduction to Synaptotagmin Synaptotagmin Structure and Function Synaptotagmins (Syt) are a family of proteins chara cterized by their membrane binding C2 domains. Syt proteins contain a transme mbrane helix which anchors a cytoplasmic region consisting of two C2 domains (C2 A and C2B) connected by a short linker. Each C2 domain is composed of two four-str anded -sheets forming a sandwich structure with three flexible binding loop s. There are currently 17 human isoforms of Syt, eight of which show varying degree s of Ca2+ binding affinity. The individual C2 domains of each Syt exhibit a variety of binding affinities including specific targeting of lipid head groups such as pho sphatidylinositol 4,5-bisphosphate (PIP2) and the nonspecific binding of anionic lipids [4 ]. The two C2 domains can differ in both electrostatic properties and structure as w ell. This is demonstrated clearly in Syt1 with the C2A and C2B binding loops coordinating thr ee and two Ca2+ atoms respectively in its membrane bound state [5]. Additionally, whi le some Syts are expressed primarily in specific tissue types such as Syt1 in neuronal t issue, other isoforms such as Syt7 are present in a wide variety of tissues [6]. The pres ence of multiple isoforms of Syt in one cell type would also suggest specific functions for each individual protein. This variability between each isoform is indicative of t he various and complex functions this family of proteins has evolved to fill.


2 Synaptotagmin Functioning as a Fusion Clamp for SNA RE Proteins During exocytosis soluble N-ethylmaleimide-sensiti ve factor attachment protein receptor (SNARE) proteins both on the plasma membra ne and secretory vesicle interact to form the fusion complex. The exact mechanism of exocytosis still remains unclear but one model proposes the v-SNAREs (on the vesicle mem brane) and t-SNAREs (on the target membrane) interact forming a four -helix coiled-coil which “zippers” together bringing both membranes into closer proximity. Syn aptotagmin then interacts with the SNARE complex and acts as a fusion clamp which afte r a Ca2+ signal releases allowing fusion to take place. An alternate hypothesis iden tifies complexin as the fusion clamp and asserts synaptotagmin is the Ca2+ sensor which causes complexin to release acting as a cofactor for inducing fusion [7]. There is gener al agreement, however, of the regulatory role of Ca2+ binding to synaptotagmin during exocytosis. Synaptotagmin 1 Characterization and Cooperativity between C2A and C2B Domains To date the most extensively studied member of the Syt family is Syt1, and studies on this protein have often framed further s tudies into other Syt isoforms. Syt1 acts as a Ca2+ sensor for the fast synchronous release of synapti c vesicles in neuronal cells. This is the fastest known membrane fusion e vent, occurring in the s to ms timescale [8-11]. The C2A domain of Syt1 binds non specifically to anionic phospholipids such as phosphatidylserine (PS), whil e the C2B binds only weakly to PS and instead has a strong affinity for PIP2 [12-16]. The C2B domain affinity for PS in 3:1 PC/PS membranes becomes stronger, however, in a tan dem domain with C2A suggesting a degree of cooperativity between the two C2 domain s [17, 18]. Recent studies involving


3 mutation of the Syt1 linker region to a rigid polyp roline variant, designed to keep the C2A and C2B domains from interacting, have shown di sruption of function similar to the Syt1 KO phenotype [19]. FRET experiments also sugg est a change in conformation may bring C2A and C2B domains of Syt1 closer together d uring Ca2+ signaling events [20]. While there is supporting evidence for interactions between C2A and C2B domains, the exact nature and extent of this interaction is far from understood. Synaptotagmin Bridging Lipid Membranes Due to the flexibility of the linker between C2 dom ains in Syt, it has been hypothesized that the two C2 domains may adopt eith er a cis or trans conformation in relation to one another during fusion events. Diff ering binding properties between the C2 domains such as observed in Syt1 could indicate dif fering target membranes for each C2 domain. Experiments utilizing site directed spin l abeling followed by double electronelectron resonance (DEER) distance measurements of Syt1 C2AB fragments have yielded models requiring the binding loops of C2A a nd C2B to be oriented in opposite directions ( trans ) when membrane bound [21]. Further studies adding the environment sensitive fluorescent probe NBD to the bottom of Sy t1 C2B domains in C2AB fragments (N396C) has identified hydrophobic interactions tak ing place in both the binding loops and bottom face of C2B which could be attributed to direct bridging of Syt1 C2AB across two membranes [22]. This finding is supported by t he identification of two arginine residues of Syt1 C2B which after mutation (R398Q an d R399Q) nearly abolished Ca2+ induced neurotransmitter release as well as reducin g liposome clustering activity [7, 22, 23]. The relative orientation of the two C2 domain s during fusion events remains as one of the unanswered questions in regards to the mecha nistic role of Syt during exocytosis.


4 Synaptotagmin Inducing Curvature in Lipid Membranes The fundamental question of how Syt initiates fusion has been probed by many different studies. Syt in the presence of Ca2+ has been observed to tubulate vesicles containing anionic head groups through induced curv ature [16, 24-26]. This observation of Syt inducing membrane curvature has been echoed in other studies and discussed as a means of lowering the energy barrier for fusion to take place [27]. One study identified a polybasic strand on Syt1 C2B which could interact e lectrostatically with negatively charged PIP2 head groups to cause a kind of ‘ratcheting’ motion which could induce membrane curvature as well as bring two opposing me mbranes closer together [23]. These unanswered questions about Syt1 are important reference points for discussing structure and function of other related Syt protein s. Synaptotagmin 7 Characterization Syt7, although not as well studied as Syt1, perform s many very important functions throughout the human body. In particular Syt7 is involved in the regulation of insulin secretion in pancreatic -cells. Dysregulation of insulin secretion has bee n implicated in the development of Type II diabetes a nd in part efforts to further understand the underlying factors contributing to this disease has lead to research into the structure and function of Syt7 [28-30]. Similar to Syt1 C2A, Syt7 C2A binds nonspecifically to anionic lipid head groups electrostatically with no special affinity for PIP2 beyond normal electrostatic interactions [31]. Similar to Syt1, Syt7 is regulated through Ca2+ signaling. Syt7, however, is much more sensitive to lower conc entrations of Ca2+ with Syt7 C2A Ca1/2 values reported at 1-6 M as compared to 29-33 M for Syt1 [31-36]. The kinetics


5 differ greatly between these two proteins as well, with a 2 fold slower on rate and 60 fold slower off rate when comparing Syt7 C2A to Syt1 C2A [31, 32]. The significantly slower off rate of Syt7 C2A could be attributed in part to a hydrophobic component of the Syt7 C2A docking mecha nism. In the presence of 600 mM NaCl Syt7 C2A still remains docked to lipid memb ranes unlike the primarily electrostatic interactions of Syt1 C2A [31]. Direc t probing of Syt1 C2A structure through EPR has modeled membrane penetration by the 1 and 3 binding loops with loop 3 binding more deeply [3]. The deeper penetration of Syt1 C2A binding loop 3 may be due to membrane insertion of the hydrophobic aromatic r ing of phenylalanine (F235), with binding loop 1 penetrating less deeply due to a met hionine (M173) in the analogue position. The equivalent binding loops of Syt7 C2A both contain phenylalanine at these positions (F229 and F167) which has been proposed t o lead to more hydrophobic interactions and deeper penetration of binding loop 1 [31]. Introduction to Electron Paramagnetic Resonance EPR has been useful in providing a detailed molecu lar picture for membranedocked proteins. Examples of proteins modeled by t his method include the pleckstrin homology (PH) domain of general receptor of phospho inositides 1 (GRP1) and the C2 domains of Syt1, protein kinase C alpha (PKC), and calcium-dependent phospholipase A2 (cPLA2) [1-3, 21, 37]. EPR detects the relaxati on of unpaired electrons excited by incident microwaves within a magnetic field of 8-10 GHz. Incorporation of unpaired electrons into a protein of interest involves a pre viously established approach known as site-directed spin labeling [44]. Single cysteine mutations are introduced at locations of interest on the target protein through site-directe d mutagenesis. A nitroxide spin label


6 such as MTSSL is then attached through formation of a disulfide bond to the mutated protein. Nitroxide spin labels will bind indiscrim inately to solvent exposed cysteine residues so any native cysteines near the solvent-e xposed surface of the protein must first be mutated to avoid labeling at multiple sites. EPR spectra are useful for observing local environm ental interactions with an unpaired electron; however, to model protein behavi or an application of EPR termed continuous wave power saturation is necessary. Pow er saturation relies on the accessibility of a spin label to various paramagnet ic probes. EPR signal amplitude increases linearly with the square root of incident power until a saturation point after which the signal decreases in intensity. The ampli tude at which a sample is halfsaturated (P1/2) is proportional to the relaxation rate of the spi n label. Heisenberg spin exchange between the spin label and paramagnetic pr obes increases the relaxation rate allowing determination of accessibility parameters. Paramagnetic probes with differing concentration gradients such as O2 and NiEDDA allow for the calculation of depth parameters or phi () values. These values can then be fit to a known hyperbolic relationship which, with some known calibration poi nts, gives a physical translational value for each individually tagged location on the protein of interest [38]. These translational values used in conjunction with solut ion NMR or crystal structures allow for development of a membrane docking model with optimi zed geometry to fit the experimental results. In previous EPR studies of Syt C2 domains between 10 and 18 spin labeled mutants were created covering locations all through the 1 and 3 binding loops [1-3]. In the present study depth parameters for 16 separate MTSSL spin labeled mutants,


7 covering key locations in the binding pockets of Sy t7 C2A, have been measured. In our present study depth parameter data is interpreted b oth by an established means of modeling docking geometry based on a previously pub lished solution NMR structure, and by a collaborative effort to further characteri ze Syt7 C2A through molecular dynamics simulations [39].


8 CHAPTER II EXPERIMENTAL Materials All reagents were reagent grade unless otherwise sp ecified. Synthetic lipids: 1Palmitoyl-2-oleoylsn -glycero-3-phosphocholine (phosphatidylcholine, POP C, PC) and 1-Palmitoyl-2-oleoylsn -glycero-3-phosphoserine (phosphatidylserine, POPS, PS) were obtained from Avanti Polar Lipids (Alabaster, AL) i n choloroform. The spin label 1Oxyl-2,2,5,5-tetramethyl-3-pyrroline-3-methylmethanethiosulfonate (MTSSL, R1) was from Toronto Research Chemicals. Fluorescein-5-mal eimide was obtained from AnaSpec Inc. (Fremont, CA). Doxyl lipids: 1-palmi toyl-2-stearoyl-(12-doxyl)-snglycero-3-phophocholine (12 Doxyl PC), 1-palmitoyl2-stearoyl-(10-doxyl)-sn-glycero3-phophocholine (10 Doxyl PC), 1-palmitoyl-2-stearo yl-(7-doxyl)-sn-glycero-3phophocholine (7 Doxyl PC), and 1-palmitoyl-2-stear oyl-(5-doxyl)-sn-glycero-3phophocholine (5 Doxyl PC) were from Avanti Polar L ipids. Protein Mutagenesis, Expression, Spin Labeling, and Purification The initial goal for protein mutagenesis was to cre ate a cysteine-free (cysless) variant of Syt7 C2A, so that unique cysteine residu es could be introduced at desired positions. The single domain Syt7 C2A DNA originat ed from American Type Culture Collection (Syt 7: 11045721) and was cloned previou sly into a customized vector RDX1. All mutants were generated using the QuikChange II XL (Agilent) site directed mutagenesis kit using the manufacturerÂ’s standard p rotocol. The cysless mutant of Syt7 C2A (C260S) was used as the parent DNA for developi ng a library of single cysteine mutants. A total of 17 single cysteine mutants wer e created at residues: 136, 148, 164,


9 167-170, 195, 196, 224, 226, 228, 229, 231, 232, an d 234 of Syt 7 C2A. All mutations were verified by primer-extension sequencing of the full C2A coding sequence. Each single mutant was expressed as a fusion prote in with a GST tag from a plasmid also containing a sequence encoding ampicil lin resistance. After transformation into E. coli BL21 cells, the transformed bacteria were streaked on LB media agar plates (Tryptone 10 g/L, Yeast Extract 5 g/L, Agar 15g/L, 86 mM NaCl, autoclaved 30 min, 100 g/ml ampicillin), and incubated overnight at 37C, an optimal temperature for E. coli growth. Single colonies were picked and added to s terile LB Media (Tryptone 10 g/L, Yeast Extract 5 g/L, 86 mM NaCl, autoclaved 30 min) in shaker tubes and incubated overnight at 37C, with shaking at 250 RPM The b acterial cultures were then used to inoculate 500 mL of sterile 2XYT media (Tryptone 16 g/L, Yeast Extract 10 g/L, 86 mM NaCl) with ampicillin, and incubated at 37C with s haking at 250 RPM. Once bacterial growth reached an OD600 of 0.6, protein expression was induced with IPTG ( 50 M) and incubated while shaking at room temperature, condit ions which favor protein expression over bacterial growth. After 8 hours of protein ex pression the bacteria were centrifuged and supernatant discarded. Cell pellets were then stored at -80C. Purification involved the lysis by probe sonication (Vibra Cells VCX130) of each cell pellet in column wash buffer (50 mM Tris base, 0.5 mM CaCl2, 1 mM MgCl2, 0.5 mM EDTA, 1 mM Benzamidine) with protease inhibitors (1 mg/L each of aprotinin, pepstatin, antipain, and leupeptin) The lysate w as then centrifuged and glutathione sepharose beads added to the supernatant in order t o bind the protein. The glutathione sepharose beads were washed with high salt column w ash buffer (50 mM Tris base, 1 M NaCl, pH 7.5) and MTSSL labeling buffer (20 mM HEP ES, 100 mM KCl, pH 7.7),


10 before addition of the MTSSL tag (37.9 mM in DMF). Spin-labeling occurred at 4C for 1 hr, followed by a final wash in TCB buffer (50 mM Tris-HCl, 150 mM NaCl, 50 M EDTA, pH 7.72) and addition of thrombin. Thrombin cleavage was allowed to proceed overnight at 4C followed by elution with assay buf fer (140 mM KCl, 25 mM HEPES, 15 mM NaCl, 0.5 mM MgCl2, pH 7.4) and spin concentration using 10 kDa Amico n tubes. Protein purity was verified by SDS-PAGE and ultravi olet-visible spectroscopy. Due to the positive charge on Syt 7 C2A DNA contamination is more likely. Any proteins showing a high 260/280 ratio, which is indicative o f DNA contamination, was further treated with bezonase for 1hr at 25C and a new spe ctra run to confirm the removal of DNA contamination. An alternate procedure for spin-labeling was also u sed involving the addition of the MTSSL tag to the concentrated protein after pur ification. Spin labeling was allowed to proceed overnight at 4C. The protein was then spin concentrated with 30 mL of assay buffer to remove excess MTSSL label. Each spin lab eling procedure was tested for labeling efficiency with no significant differences observed. A mixture of both methods was used for preparation of samples for EPR. Preparation of Lipid Vesicles Each lipid preparation was dried under N2, and then placed under vacuum for 2 hrs in order to remove all traces of solvent. Drie d lipids were then hydrated with assay buffer buffer and mixed by vortexing. Small unilam ellar vesicles (SUV) were then generated through probe sonication and allowed to e quilibrate overnight at 25 C. Vesicles for EPR were prepared with a mole ratio of 75:25 POPC:POPS and a total lipid concentration of 80 mM.


11 Measuring EPR Spectra An EPR spectrum was taken for each single cysteine using a Bruker ELEXSYS E500 spectrometer (9.4GHz) with a loop gap resonato r (Medical Advances). Each sample consisted of MTSSL labeled Syt7 C2A in concentratio ns of 30-150 M with either no lipids present for unbound spectra or POPC:POPS lip ids for bound spectra. A high total lipid concentration of 30 mM (15mM accessible), sim ilar to previous EPR studies [1-3], was used in order to minimize the possibility of sp in-spin interactions between the MTSSL tags of bound Syt7 C2A domains. Measurements were performed at 2.0 mW incident power with a minimum of 5 scans of 100 G. EPR spectra were normalized to their second integral representing the total number of spins. This technique has proven effective in previous studies [1, 2] for comparing spectra of varying spin label concentrations particularly when comparing bound an d unbound states. Continuous Wave Power Saturation Measurements Each sample was loaded into a gas-permeable TPX cap illary tube (Medical Advances). Power saturation curves were measured b etween 0.2 mW and 50 mW taking saturation measurements at predefined intervals, ov er 30 G with at least 2 scans. Power saturation curves for each sample were measured und er three separate conditions: (1) atmospheric oxygen (20%), (2) after equilibration u nder a stream of N2 for 15 mins, and (3) after addition of 10 mM of NiEDDA and equilibra tion under N2 for 15 mins. Each accessibility measurement for a single protein was taken on the same day in order to minimize the effect of random variation in sample p reparation and minor instrument fluctuations when calculating the depth parameter. Amplitudes were plotted as a function of microwave power squared and fit in Kaleidagraph to the equation [38,40]:


12 n r (1) where A represents the peak-to-peak amplitude of th e signal, C is a scaling factor, P is the microwave power, P1/2 is the power at which half saturation occurs, and is the measure of homogeneity of saturation. Accessibility parame ters were then calculated as follows [38,40]: (2) where X represents either O2 or NiEDDA paramagnetic species, represents the accessibility parameter for the given species, and Hpp is the average peak to peak line width over the linear region of the power saturatio n curve. Depth parameters were then calculated from the following equation: !"# $ %&''( ) (3) where is the depth parameter. Each depth parameter was m easured at least twice for each mutant and three times if the depth parameters differed by more than 0.4 in value.


13 Fluorescein Maleimide Assay to Assess Spin-Labeling Efficiency Fluorescein maleimide was dissolved in DMF (50 M) and reacted with spin labeled protein (10 M) in Tris buffer (pH 8). The reaction was allowed to proceed for 1 hr at room temperature while protected from light. After reacting, loading dye was added to each sample and SDS-PAGE performed. The unlabel ed cysteine of wild type Syt7C2A was used as a positive control. Fluorescen t images were taken of each gel and the relative intensities of wild type and single cy steine mutant bands were qualitatively compared to determine estimated labeling efficiency The intensities of the fluorescent images were then compared to the coomassie stained gel to ensure equal loading of protein in each lane. Stopped Flow Kinetic Measurements Stopped flow experiments were performed by Dr. Knig ht and Favinn Maynard as outlined previously [31]. Modeling Docking Geometry Optimal geometry for Syt7 C2A was modeled using PyM ol with the “MTSSL Wizard” plug-in using a previously published high r esolution solution NMR structure (PDB code 2D8K) [39]. The position for each spin l abel was initially determined by averaging the coordinates for every allowed conform ation of the MTSSL tag. Coordinates from PyMol were then fit to a hyperboli c tangent function: *+,-. / 012+. n3 (4)


14 where A and D represent the bulk values of in water and hydrocarbon, C sets the inflection point of the curve, B determines the slo pe and, depth represents the distance of the MTSSL spin label from the phosphate plane given by: 012+.456-78n9:;57<=>:;57856-7

15 equations 4 and 5, experimental depth parameters, D oxyl lipid calibration data, and the relevant solution NMR structure for Syt7 C2A.


Site Selection, Protein purification, and MTS The process of s ite directed spin protein through the MTSSL positions used in a previous EPR power saturation study with S strategy allows for a more direct comparison betwee n Syt7 and its well studied isomer Syt1. Spin labeled positions are summarized in or near the Ca2+ binding loops opposite the binding site Single cysteine mutants were purified through gluta thione affinity chromatography of a GST Figure 1 Solution NMR structure of Syt7 C2A with mutations highlighted by yellow spheres. CHAPTER III RESULTS Site Selection, Protein purification, and MTS SL labeling ite directed spin labeling introduces an unpaired electron into the protein through the MTSSL nitroxide spin label. This spin label attaches through a disulfide bond to any accessible cysteine residues on the protein As a result any native cysteine residues must be mutated in order to selectively label at the desired locations of interest cysless (C260S) variant of Syt7 C2A was first generated to act as a base for the creation of a series of single cysteine mutants. The positions selected for site directed spin labeling in the functional cysless variant of Syt7 C2A are a previous EPR power saturation study with S yt1 C2A [3]. This strategy allows for a more direct comparison betwee n Syt7 and its well studied isomer Syt1. Spin labeled positions are summarized in Figure 1 and Table 1 cover binding loops 1 and 3 of Syt7 C2A, with one negative control placed (Q148C). Single cysteine mutants were purified through gluta thione affinity chromatography of a GST linked protein followed by cleavage of the GST with Solution NMR structure of Syt7 C2A with mutations highlighted by yellow 16 labeling introduces an unpaired electron into the label attaches through a any accessible cysteine As a result any native cysteine residues must be mutated out in order to selectively label cysteine residues of interest A cysless (C260S) variant of Syt7 C2A was to act as a base for the creation of a series of single cysteine The positions selected for site directed spin labeling in the functional C2A are a nalogous to yt1 C2A [3]. This strategy allows for a more direct comparison betwee n Syt7 and its well studied isomer cover ing 16 sites on with one negative control placed Single cysteine mutants were purified through gluta thione affinity linked protein followed by cleavage of the GST with


17 thrombin. Purity of the concentrated protein was v erified by SDS-PAGE showing single bands around 16 kDa, the expected size of Syt7 C2A. The level of contamination with nucleic acids in purified protein stocks was measur ed using ultraviolet-visible spectroscopy and absorbance at 260 nm (absorbance o f nucleic acids) and 280 nm (absorbance of the protein) were recorded as summar ized in Table 1. High amounts of nucleic contamination may influence the kinetic rates of membrane binding proteins such as Syt7 C2A. This is a concern for our chosen method of fluorescent resonance energy transfer (FRET) assay which measures kinetic on and off rates to determine any perturbing cysteine mutants. Only samples with a 260/280 ratio of 1.0 or less, indicating only minor levels of DNA contamination, were determined to be acceptable for kinetic studies. EPR measurements a re less sensitive to nucleic acid contamination due to the long time period in which measurements are taken, allowing proteins to be in equilibrium with the large excess of lipid vesicles. Concentrations were calculated from 280 nm measurements using an extinc tion coefficient of 14280 M-1 cm-1 and are also summarized in Table 1. Table 1. Absorbance data and calculated concentration for each Syt7 C2A mutant. Sample Type 260/280 Ratio Concentration(M) Wild Type 0.85 170 C260S (Cysless) 0.91 161 L136R1 0.73 138 Q148R1 0.79 66 A164R1 0.68 438 F167R1 0.78 91 S168R1 0.69 684 G169R1 0.82 363 T170R1 0.75 787 N195R1 0.99 113 L196R1 1.0 166 L224R1 0.64 143 Y226R1 0.95 107 R228R1 0.68 220 F229R1 0.76 242 R231R1 0.70 753 N232R1 0.67 3362 P234R1 0.67 1023


18 In order to characterize the MTSSL spin-labeling ef ficiency a fluorescein maleimide assay was utilized. Fluorescein maleimid e will attach a fluorophore to any unreacted cysteines through Michael addition. Prot eins labeled with the fluorophore were examined by SDS-PAGE. Fluorescent pictures of the gels revealed a qualitative look at labeling efficiency relative to a positive control wild type protein, utilizing the wild type native cysteine residue (C260) for fluoro phore attachment. In each case MTSSL spin labeled mutants fluoresced at roughly ha lf the intensity of the positive control indicating a >50% labeling efficiency as sh own in Figure 2. Both the overnight and 1 hr spin-labeling procedures appeared to achie ve roughly similar levels of labeling efficiency after confirming equal loading of protei ns through comparison with the coomassie stained gel.


Figure 2. Fluorescent image of proteins spin spinlabeling procedure (A164R1, L196R1, and N224R1). Balancing protein conditions for EPR. T he h had a tendency to induce previously in Syt 1 C2B and C2AB domains aggregated samples lead to concentrations of both protein and lipid involved experiment s which could quantify the extent of aggregation. Instead simple qualitative experiments were performed Table 2 represents a series of protein/lipid ratios and the observed aggregation levels i Fluorescent image of proteins spin -labeled through both the procedure (R228R1 and N195R1) and 1 hr spinlabeling (A164R1, L196R1, and N224R1). Balancing protein / lipid ratios was another consideration when setting sample he h igh protein concentrations necessary for strong EPR signals had a tendency to induce lipid aggregation, a phenomenon that has been observed 1 C2B and C2AB domains [24]. The nonhomogenous to a greater variability in repeat EPR measurements. The high of both protein and lipid involved limited the practicality of turbidity s which could quantify the extent of aggregation. Instead simple qualitative performed to determine ideal protein/lipid conditions for EPR samples. represents a series of protein/lipid ratios and the observed aggregation levels i 19 both the overnight labeling procedure lipid ratios was another consideration when setting sample igh protein concentrations necessary for strong EPR signals a phenomenon that has been observed homogenous natures of highly EPR measurements. The high limited the practicality of turbidity s which could quantify the extent of aggregation. Instead simple qualitative to determine ideal protein/lipid conditions for EPR samples. represents a series of protein/lipid ratios and the observed aggregation levels i n


20 multiple samples of the G169R1 mutant. Protein/lip id ratios of roughly 1:300 were identified as the approximate ratio for maximizing protein concentration and minimizing lipid aggregation. This corresponds to 100 M final protein concentration in EPR samples containing 30 mM total lipids. Table 2. Lipid aggregation study data. Protein Concentration (M) Lipid Concentration (mM) Protein/Lipid Ratio Observed Aggregation1 582 13.3 1:22.9 + + + 291 13.3 1:45.7 + + + 58.2 5.32 1:91 + + 145 13.3 1:92 + + 72.8 13.3 1:183 + 116.4 23.9 1:206 + 58.2 13.3 1:229 + 87.3 24.6 1:282 + 53.2 16 1:301 58.2 21.3 1:366 58.2 25.3 1:434 1 Qualitative ranking of protein/lipid aggregation: +++ indicates large amounts of aggregation, ++ indicates moderate amounts of aggre gation, + indicates only minor aggregation, indicates no visible aggregation Effects of Spin Labeling on Syt7 C2A Docking Functi on As has been demonstrated previously most spin label s introduced into membrane targeting regions that are not specific recognition sites, do not disrupt normal protein function [1,41]. Through a protein membrane FRET a ssay we examine the kinetics of each mutant and compare to the wild type and cysles s Syt7 C2A. Rates within a factor of 2.5 in comparison to the C260S (cysless) mutant wer e considered non-perturbing to normal protein function. Summarized in Table 3 are the off and on rates for wild type, C260S, and each single cysteine mutant.


21 Table 3. Measured rate constants for kinetic measurements Sample Type k off (s 1 ) Relative to C260S k obs (s 1 ) Relative to C260S Wild Type 26 1.5 27 0.47 C260S (Cysless) 17 1 57 1 L136R1 19 1.1 35 0.61 Q148R1 20 1.2 27 0.47 A164R1 21 1.2 18 0.32 F167R1 27 1.6 52 0.91 S168R1 34 2.0 29 0.51 G169R1 19 1.1 25 0.44 T170R1 19 1.1 51 0.89 N195R1 33 1.9 65 1.14 L196R1 38 2.2 60 1.05 L224R1 35 2.1 34 0.60 Y226R1 69 4.1 5 0.09 R228R1 24 1.4 73 1.28 F229R1 69 4.1 34 0.60 R231R1 16 0.9 50 0.88 N232R1 23 1.4 34 0.60 P234R1 21 1.2 65 1.14 Nine of the mutants were in good agreement with bo th the off and on rates for C260S. Four proteins: G169R1, L196R1, L224R1, and Q148R1; were beyond a 2 fold difference as compared to C260S but less than 2.5. Y226R1, F229R1, and A164R1 each showed differences in either on or off rates >2.5 f old higher in comparison to C260S. One protein (I235R1) had no observable membrane kin etics which may be due to the interior location of this mutation site causing pro tein misfolding. Mutants with kinetic on or off rates greater than a magnitude of 2.5 differ ence were considered perturbing to the normal function of the Syt7 C2A protein justifying the exclusion of these mutants from the docking model. Previous studies have limited t he included proteins to those within a factor of 2 for Syt1 C2A using a terbium affinity a ssay [3]. Alternately assays measuring the Ca2+ dependence of lipid binding have been used for pro teins such as PKC [1].


22 These previously established methods are difficult to use with Syt7 in part due to how tightly Syt7 binds Ca2+. The FRET assay we use here is suited for studyin g Syt7 kinetics but has more natural variability in its measurement s as compared to methods used for Syt1. In addition, certain proteins were observed to have double exponential components in their rate constants. Notably N232R1 and Y226R1 showed double exponential off rates, with R228R1 and R231R1 showi ng a similar component to their measured on rates. In each of these cases, the fas ter component had a greater amplitude. The origin of the slow, low-amplitude FRET change i s unclear but could arise from either (a) a small population of misfolded proteins, which would not impact EPR measurements significantly due to its low abundance, or (b) slow vesicle aggregation, which would not affect EPR measurements due to the much smaller pro tein-to-lipid ratio used in EPR samples. In addition to kinetic FRET measurements, continuou s wave EPR spectra were taken for the 13 spin labeled mutants included in t he docking model and the four doxyl lipids. EPR spectra, taken for both a bound and un bound state of the protein, are useful in determining changes in the local environment aro und a spin label during membrane docking. Doxyl lipid EPR spectra are displayed in Figure 3. Both bound and unbound spectra were measured for each mutant and compared as represented in Figure 4. Bound spectra included 3:1 POPC:POPS membranes with 30 mM total lipid concentration and excess Ca2+. Unbound spectra were measured with the free prot ein in solution and equivalent Ca2+ concentrations to the bound sample. The presence o f excess lipid and


23 Ca2+ concentrations provide conditions for near complet e docking of Syt7 C2A in the bound sample. Figure 3. Doxyl lipid EPR spectra measured in 27 mM 3:1 PC:P S with 1% doxyl lipids. Each measurement is an average of 20 100 G scans in assay buffer.


nr rr r rrr "r #$ nnrr%&r nr "r &r r !r'()(* '(!' #$ +r rrr "r r,"rrr nnrr%&r 24 "r &r "#((( '(!' +r rrr "r r,"rrr ) nnrr%&r


25 Four of the proteins (R231R1, S168R1, R228R1, G169R 1, and F167R1) exhibited significant signal broadening when comparing the bo und state to the unbound state. Signal broadening generally indicates a local pertu rbation in mobility of the MTSSL tag likely due to contact with the lipid membrane or du e to changes in protein conformational or rotational properties. The location of these mu tations along the binding loops of Syt7 C2A make it likely this broadening is due to insert ion of these binding loops into the plasma membrane interior followed by steric clash w ith lipid acyl chains decreasing the MTSSL tag mobility. The rest of the mutants showed varying degrees of signal broadening including, interestingly, Q148R1 located opposite the binding loops. Orientation and Depth of Penetration by Syt7 C2A EPR power saturation was performed in order to meas ure the depth parameters for each of the 13 spin labeled positions on Syt7 C2A a s well as four doxyl lipids with spin labeled acyl chains at the 5, 7, 10, and 12 locatio ns. Doxyl lipids have a known depth within the lipid membrane and are useful in calibra ting the hyperbolic tangent fit used to translate depth parameters into physical depths [42 ]. Each measurement was repeated at least once with any depth parameters differing by > 0.4 given a third repeat. Depth parameter errors were derived from a 95% confidence interval propagated from the power saturation curve fit done in Kaleidograph (Eq uation 1). A weighted average was then used for repeat depth parameter measurements t o give the final error. Table 4 summarizes the depth parameters along with the O2 and NiEDDA accessibility parameters used to calculate the depth parameters. Positive depth parameters are indicative of more deeply penetrated spin labels wi th depth parameters < -2 being in the


26 bulk aqueous phase (>5 away from the phosphate pla ne) and less significantly impacted by any contacts with the lipid membrane. Table 4. Probe accessibility and depth parameters for doxyl lipids and spin-labeled C2A domains Sample Type E dPi FGHIIJ dPi dPhi Depth ()* L136R1 1.37 0.10 7.39 0.23 -1.59 0.18 -15.0 Q148R1 2.99 0.24 12.66 0.58 -1.38 0.21 -45.1 F167R1 3.19 0.34 2.25 0.31 0.27 0.19 4.4 S168R1 3.32 0.21 1.06 0.25 1.12 0.30 5.9 G169R1 2.18 0.16 3.49 0.26 -0.43 0.17 1.8 T170R1 1.67 0.49 12.01 0.91 -1.85 0.40 -3.3 N195R1 1.86 0.17 11.70 0.59 -1.83 0.20 -10.6 L196R1 2.19 0.31 14.55 0.64 -1.89 0.21 -4.3 L224R1 1.410.23 7.620.42 -1.540.24 -9.1 R228R1 2.42 0.25 0.92 0.14 0.69 0.24 5.9 R231R1 1.55 0.33 2.50 0.24 -0.48 0.29 1.7 N232R1 1.240.24 3.570.35 -0.940.46 -2.2 P234R1 0.96 0.14 4.18 0.23 -1.42 0.23 -13.0 Doxyl 5 4.48 0.17 0.74 0.09 1.73 0.21 8.1** Doxyl 7 5.97 0.52 0.55 0.15 2.42 0.42 10.5** Doxyl 10 7.53 0.36 0.70 0.12 2.42 0.24 14.0** Doxyl 12 9.03 0.62 0.38 0.16 3.13 0.54 16.0* *Calculated depths based on depth parameters: **pre viously published depths [42] Calculated depth parameters for each doxyl lipid w ere in good agreement with previous literature values [42]. From the 13 Syt7 C2A mutants 7 showed depth parameters below -1: L136R1, Q148R1, T170R1, N195R1 L196R1, L224R1, and P234R1; indicating aqueous exposed locations with l ittle to no membrane contact. Three of the positions: R231R1, F167R1, and S168R1; had p ositive depth parameters, indicating membrane penetration potentially deeper than the phosphate plane of the lipid membrane while R228R1 and G169R1 appeared to be in close proximity to the phosphate plane. Depth parameters are more useful in determining relativity; these values will vary on different instruments and under different reaction conditions. In order


27 to positively identify membrane penetrating residue s, depth parameters must be translated into physical depths. This is done through use of a hyperbolic tangent fit (Figure 5) which models the relationship between depth paramet ers and the physical distance from the phosphate plane. Figure 5. Hyperbolic tangent fit of experimentally determine d depth parameters. Open circles indicate doxyl lipids, solid black circles indicate single-cysteine mutants. Fit is based on Equation 4. Error bars are based on the w eighted average over two or three measurements and are listed in Table 4. The initial depth of each MTSSL tag was defined as the average position of all allowed conformations for each spin label. Depth p arameters and average position depths were input into the hyperbolic tangent fit ( Equation 4). Any outliers to the hyperbolic tangent fit were then adjusted to better fit the experimental data through defining MTSSL dihedral angles as described in the Methods section. The average position fit described the experimental data well w ith the exceptions of F167R1, G169R1,


28 and L196R1. For these mutants the 1 and 2 side chain dihedral angles were set to (g+,g+) for F167R1 moving it shallower, (g-,t) for L196 also moving it shallower, and (g+,t) for G169R1 moving it deeper. The optimized bes t fit depths are outlined in Table 4. From these depth values a self-consistent docki ng model is created using the solution structure for Syt7 C2A (Figure 6). Calculated dept hs indicate potential membrane penetration of both binding loop 1 (F167R1, S168R1, G169R1) and binding loop 3 (R228R1, R231R1). The deepest measured residue of binding loop 3 (R228R1) has a depth of 5.5 as compared to the deepest residue o f binding loop 1 (S168R1) at 5.8 . The similar depth of penetration for both of these residues is reflected in the optimized docking model which indicates near equal penetratio n of both binding loops 1 and 3 into the lipid membrane.


Figure 6. Optimized best fit docking model for Syt7 C2A solution structure (PDB code 2D8K). The phenylalanine residues at positions 167 and 229 are shown in green. Protein orientation cor responds to Euler angle rotations of manipulation of the y Protein docking cannot be fully characterized by a single docking model, rather proteinmembrane interactions are fluid with many potential valid conformations. The solution NMR structure, from which we base our Syt7 C2A model, is itself an ensemble of po tential structures. The optimized best fit model i n solution NMR structure. model we repeated the same modeling procedure for t he stat shown in Figure 7. Notable in the state 2 docking model is the shallow er penetration of binding loop 3 and the much shallower penetration o f binding loop 1. Optimized best fit docking model for Syt7 C2A based on the solution structure (PDB code 2D8K). The green line defines the lipid phosphate plane. The phenylalanine residues at positions 167 and 229 are shown in green. Protein responds to Euler angle rotations of x by 17.5 and z manipulation of the y -translational value -20.6 -. Protein docking cannot be fully characterized by a single docking model, rather membrane interactions are fluid with many potential valid conformations. The solution NMR structure, from which we base our Syt7 C2A model, is itself an ensemble tential structures. The optimized best fit model i n Figure 6 is based on the state 1 In order to better visualize uncertainty in the opt imized docking model we repeated the same modeling procedure for t he stat e 2 solution NMR stru Notable in the state 2 docking model is the shallow er penetration of binding loop 3 and the much shallower penetration o f binding loop 1. Residues S168R1 29 based on the state 1 NMR The green line defines the lipid phosphate plane. The phenylalanine residues at positions 167 and 229 are shown in green. Protein by 4.6 and a Protein docking cannot be fully characterized by a single docking model, rather membrane interactions are fluid with many potential valid conformations. The solution NMR structure, from which we base our Syt7 C2A model, is itself an ensemble is based on the state 1 In order to better visualize uncertainty in the opt imized docking 2 solution NMR stru cture Notable in the state 2 docking model is the shallow er penetration of Residues S168R1


and R228R1 are assigned depths of 6.0 and 4.7 r espectively similar to t model. The differences in docking geometry influenced by the dynamic therefore may be less representative Figure 7. Optimized best fit docking model for Syt7 C2A solution structure (PDB code 2D8K). The green line defines the lipid phosphate plane. The phenylalanine residues at positions 167 and 229 are shown in green. Protein orientation corresponds to Euler angle rotations of manipulation of the y and R228R1 are assigned depths of 6.0 and 4.7 r espectively similar to t The differences in docking geometry between each model may be dynamic nature of protein-membrane interactions, however, representative of the error in a single given model Optimized best fit docking model for Syt7 C2A based on the state 2 NMR solution structure (PDB code 2D8K). The green line defines the lipid phosphate plane. The phenylalanine residues at positions 167 and 229 are shown in green. Protein orientation corresponds to Euler angle rotations of x by 23.1 and z by manipulation of the y -translational value -22.2 -. 30 and R228R1 are assigned depths of 6.0 and 4.7 r espectively similar to t he state 1 may be heavily however, and based on the state 2 NMR solution structure (PDB code 2D8K). The green line defines the lipid phosphate plane. The phenylalanine residues at positions 167 and 229 are shown in green. Protein by -6.5 and a


31 CHAPTER IV DISCUSSION Comparing Syt7 and Syt1 Docking Models The extensive body of research detailing the struct ure and function of Syt1 is an important starting point for the discussion of Syt7 Each isoform of Syt has evolved a very precise function suited to its biological envi ronment. The highly electrostatic interactions of Syt1 to its target membrane are wel l suited to the fast release of neurotransmitters. Syt7 is known to have a much hi gher sensitivity to Ca2+ and to bind membranes much more tightly with significantly slow er dissociation kinetics. One potential model for explaining the slower kinetics has been proposed in previous studies by the two state docking model. In this model for Syt7 C2A there is an initial electrostatic association followed by deeper penetr ation of both binding loops into the hydrophobic interior of the lipid membrane. This i s, in part, facilitated by the aromatic rings of phenylalanine at the tips of each binding loop [31]. This mechanism is in agreement with our proposed docking model for Syt7 C2A. A comparison of previously published models for Syt1 C2A and our current propo sed model for Syt7 C2A show near equal penetration of binding loop 3 but much deeper penetration of binding loop 1 for Syt7 C2A [3]. The best fit model for Syt7 C2A plac es three residues in binding loop 1 inside the lipid membrane: G169R1(1.8 ), F167R1(4 .4 ), S168R1(5.9 ), in comparison to one for Syt1 C2A: G174R1(5.3 ) with the nearest residue (G175R1) reported as being in the aqueous phase (-1.1 ) [3] The equal penetration of both binding loops 1 and 3 could account for the reporte d slower dissociation kinetics of Syt7.


32 Observations about Mutant Kinetics The kinetic rate differences between the F229R1 and F167R1 mutants may also give us more insight into the role of each binding loop in Syt7 C2A. Generally the F229 residue has been considered to be essential to the binding function of binding loop 3 due to its hydrophobic phenylalanine residue. The pert urbation of kinetic rates when mutating this residue is unsurprising in this conte xt. It has been hypothesized that F167 plays the same role for binding loop 1, however, ki netic rates were not significantly impacted by mutation to cystein and addition of an MTSSL spin label. The relatively shallow depth parameter measured for F167 might sug gest some loss of function but not significant enough a difference to be unaccounted f or by movement of the spin label. The differing effects of mutating these residues ma y indicate a more essential mechanistic role for binding loop 3 as compared to binding loop 1. Potential Sources of Error in the Model EPR based models are useful in directly visualizing docking geometry, but it is important to realize the limitations of such method s. Many assumptions go into our proposed docking model. The average position metho d for determining spin label depth treats every conformation as equally favorable when some conformations may be preferred over others. Any outliers to the average position hyperbolic tangent fit are then assigned fairly arbitrary dihedral angels. Apart f rom crystallographic data showing a natural tendency of MTSSL side chains to adopt the (g+,g+) conformation [43] all other conformations, when (g+,g+) is sterically hindered, are chosen through an edu cated guess introducing potential bias. Additionally, the solu tion NMR structure which our model is


33 based on is not bound to Ca2+, making any structural changes to Syt7 C2A due to binding of Ca2+ unaccounted for in the experimental design. We at tempt to illustrate part of this error directly through adoption of our state 2 mode l. Even comparing the shallower binding of loop 1 in the state 2 model we still obs erve overall deeper penetration of binding loop 1 in comparison to Syt1 C2A. Potential Effects of Cooperativity Potential cooperativity between the C2A and C2B dom ains is another relevant factor in the overall mechanism of Syt7. A follow up study to the Syt1 C2A EPR docking model reported deeper penetration of both b inding loops 1 and 3 of Syt1 C2A when in tandem with Syt1 C2B (G174R1 6.9 and M17 3R1 7.8 ) [21]. This may indicate an association between the two C2 domains that causes deeper penetration of the C2A domain. A similar cooperative affect between S yt7 C2A and C2B would be unaccounted for in our experiment as we deal only w ith the C2A domain. A possible contributor to the result of Sy1 C2A penetrating de eper, however, may be the induction of curvature into target membranes by the tandem do main. EPR measurements of C2A domains occur under constant levels of membrane cur vature. Alternately tandem domains may introduce negative curvature due to the mechanistic functions of the complete protein. Negative curvature in the lipid membrane would increase the packing of lipid headgroups thereby decreasing collisions b etween any shallow spin labels and the NiEDDA paramagnetic probes. Decreased accessibilit y to NiEDDA would tend to give apparent deeper depth parameter values.


34 The Existence of Multiple Protein Conformations The qualitative observation that the bound spectrum for Q148R1 broadens in comparison to the unbound spectra is interesting. Collaborative molecular dynamics simulations have identified a preference for multip le conformations of the Syt7 C2A domain on a lipid membrane. The ‘standing’ conform ation simulates the protein nearly perpendicular to the lipid bilayer with both loops near equally inserted. The ‘lying down’ conformation has binding loop 3 fully inserted with binding loop 1 only partially inserted, and the protein near parallel with the lipid bilaye r. Our model depicts a ‘standing’ conformation in which the entire protein is mostly perpendicular to the lipid membrane. In this conformation the Q148R1 spin label would re main relatively unhindered in both a bound and unbound states. Differences in Q148R1 bo und and unbound EPR spectra seem to indicate some kind of interaction taking pl ace when in the bound state resulting in a slight broadening of the signal. Potential me mbrane interactions are also suggested by the depth parameter itself, which is less negati ve then would be expected for a spin label so far from the membrane interface. There ar e a few possibilities that may explain this observation. The first derives from the tende ncy of Syt7 C2A to self associate when in the presence of Ca2+. Multiple Syt7 C2A domains forming dimers may res ult in electrostatic contacts being made in regions near t he Q148R1 mutation. Secondly, Syt7 C2A may act as a bridging protein between two lipid membranes. A second membrane interacting electrostatically with Syt7 C2A residue s near Q148R1 would result in broadening of the EPR spectrum and a less negative depth parameter. Finally, Syt7 C2A may adopt a ‘lying down’ conformation with electros tatic interactions of the 4 loop bringing the protein near parallel with the lipid m embrane. Indeed this is reflected by a


35 similar EPR spectra broadening of the mutation on 4 (N195R1). Any of these factors could potentially account for broadening of the bou nd Q148R1 signal. Future Experiments Here we present the first EPR based model for the d ocking geometry of Syt7 C2A. More experimentation is needed to gain a full er understanding of the complex mechanism through which this protein functions. Di rect structural models are an important tool in the characterization of these com plex biological processes. Similar EPR experiments involving Syt7 C2B and tandem Syt7 C2AB proteins could provide a much fuller picture and a more accurate interpretat ion of how this protein functions. The spin-labeling of more sites along surface residues of the 4 sheet may also give more information about potential contacts along this reg ion, supporting the presence of a lying down conformation.


36 REFERENCES 1. Landgraf, K.E., Malmberg, N. J., Falke, J. J. (200 8) Effect of PIP2 on the Membrane Docking Geometry of PKC C2 Domain: An EPR Site-Directed Spin-Labeling and Relaxation Study. Biochemistry 47, 8301-8316. 2. Chen, Huai-Chun, Ziemba, B.P., Landgraf, K.E., Corb in, J.A., Falke, J.J. (2012) Membrane Docking Geometry of GRP1 PH Domain Bound t o a Target Lipid Bilayer: An EPR Site-Directed Spin-Labeling and Re laxation Study. PLoS One 7: e33640. 3. Frazier, A.A., Roller, C.R., Havelka, J.J., Hinderl iter, A., Cafiso, D.S. (2003) Membrane-Bound Orientation and Position of the Syna ptotagmin I C2A Domain by Site-Directed Spin Labeling. Biochemistry 42, 96-105. 4. Hurley, J.H., and Misra, S. (2000) Signaling and Su bcellular Targeting by Membrane-Binding Domains. Annu. Revs. Biophys. Biomol. Struct. 29, 49-79. 5. Perin, M.S. et al. (1990) Phospholipid binding by a synaptic vesicleprotein homologous to the regulatory region of protein kina se C. Nature 345, 260–263. 6. Andrews, N. W., Chakrabarti, S. (2005) There’s mor e to life than neurotransmission: the regulation of exocytosis by synaptotagmin VII. Cell Biology 15, 626-631. 7. Krishnakumar, S.S., Radoff, D.T., Kmmel, D., Girau do, C.G., Li, F., Khandan, L., Baguley, S.W., Coleman, J., Reinisch, K., Pince t, F., Rothman, J.E. (2011) A conformational switch in complexin is required for synaptotagmin to trigger synaptic fusion. Nature Struc. & Mol. Bio. 18, 934-941. 8. Chapman, E.R. (2008) How Does Synaptotagmin Trigge r Neurotransmitter Release ? Annu. Rev. Biochem 77, 615-641. 9. Katz, B. (1969) The Release of Neural Transmitter S ubstances. Springfield, IL: Thomas 10. Llinas, R., Steinberg, I.Z.,Walton, K. (1981) Presy naptic calcium currents in squid giant synapse. Biophys. J. 33:289–321. 11. Almers, W. (1994) How fast can you get? Nature 367:682–83.


37 12. Schiavo, G., Gu, Q.M., Prestwich, G.D., S.llner, T.H., Rothman, J.E. (1996) Calcium-dependent switching of the specificity of p hosphoinositide binding to synaptotagmin. Proc. Natl. Acad. Sci. USA 93:13327–32. 13. Bai, J., Tucker, W.C., Chapman, E.R. (2004) PIP2 increases the speed of response of synaptotagmin and steers its membrane-penetratio n activity toward the plasma membrane. Nat. Struct. Mol. Biol. 11:36–44. 14. Damer, C.K., Creutz, C.E. (1996) Calcium-Dependent Self-Association of Synaptotagmin I. J. Neurochem. 67:1661–68. 15. Damer, C.K., Creutz, C.E. (1994) Synergistic membra ne interactions of the two C2 domains of synaptotagmin. J. Biol. Chem. 269:31115–23. 16. Arac, D., Chen, X., Khant, H.A., Ubach, J., Ludtke, S.J., Kikkawa, M., Johnson, A.E., Chiu, W., Sdhof, T.C., Rizo, J (2006) Clos e membrane-membrane proximity induced by Ca2+-dependent multivalent binding of synaptotagmin-1 t o phospholipids. Nat. Struct. Mol. Biol. 13:209–17. 17. Bai, J.,Wang, P., Chapman, E.R. (2002) C2A activate s a cryptic Ca2+-triggered membrane penetration activity within the C2B domain of synaptotagmin I. Proc. Natl. Acad. Sci. USA 99:1665–70. 18. Hui, E., Bai, J., Chapman, E.R. (2006) Ca2+-Triggered Simultaneous Membrane Penetration of the Tandem C2-Domains of Synaptotagm in I. Biophys. J. 91:1767– 77. 19. Liu, H., Bai, H., Xue, R., Takahashi, H., Edwardson J.M., Chapman, E.R. (2014) Linker mutations reveal the complexity of synaptota gmin 1 action during synaptic transmission. Nature Neuroscience doi: 10.1038/nn.3681. 20. Garcia, R.A., Forde, C.E., Godwin, H.A. Calcium tr iggers an intramolecular association of the C2 domains in synaptotagmin. Proc. Natl. Acad. Sci. U S A 2000;97:5883-5888. [PubMed: 10811903] 21. Herrick, D.Z., Kuo, W., Huang, H., Schwieters, C.D. Ellena, J.F., Cafiso, D.S. (2009) Solution and Membrane-Bound Conformations of the Tandem C2A and C2B Domains of Synaptotagmin 1: Evidence of Bilaye r Bridging. J. Mol. Biol. 390(5): 913-923. doi: 10.1016/j.jmb.2009.06.007. 22. Seven, A.B., Brewer, K.D., Shi, L., Jiang, Q., Rizo J. (2013) Prevalent mechanism of membrane bridging by synaptotagmin-1. PNAS doi: 10.1073/pnas.1310327110.


38 23. Kuo, W., Herrick, D.Z., Cafiso, D.S. (2011) Phosph atidylinositol 4,5 bisphosphate alters Synaptotagmin 1 Membrane Dockin g and Drives Opposing Bilayers Closer Together. Biochemistry 50(13): 2633-2641. doi: 10.1021/bi200049c. 24. Martens, S., Kozlov, M.M., McMahon, H.T. (2007) How Synaptotagmin Promotes Membrane Fusion. Science 316:1205–8. 25. Wu, Y., He, Y., Bai, J., Ji, S.R., Tucker, W.C., et al. (2003) Visualization of synaptotagmin I oligomers assembled onto lipid mono layers. Proc. Natl. Acad. Sci. USA 100:2082–87. 26. Fukuda, M., Katayama, E., Mikoshiba, K. (2002) The Calcium-bindinjg Loops of the Tandem C2 Domains of Synaptotagmin VII cooperat ively Mediate Calciumdependent Oligomerization. J. Biol. Chem. 277:29315–20. 27. Chernomordik, L.V., Kozlov, M.M. (2003) Protein-Lip id Interplay in Fusion and Fission of Biological Membranes. Annu. Rev. Biochem. 72:175–207. 28. Gauthier, B.R., Wollheim, C.B. (2008) Synaptotagmi ns bind calcium to release insulin. Am. J. Physiol. Endocrinol. Metab. 295: E1279-E1286. doi: 10.1152/ajpendo.90568.2008. 29. Wajchenberg, B.L. (2007) Beta-cell failure in diabe tes and preservation by clinical treatment. Endocr Rev 28: 187–218. 30. Weir, G.C., Bonner-Weir, S. (2004) Five stages of e volving beta-cell dysfunction during progression to diabetes. Diabetes 53, Suppl 3: S16–S21. 31. Brandt, D.S., Coffman, M.D., Falke, J.J., Knight, J .D. (2012) Hydrophobic Contributions to the Membrane Docking of Synaptotag min 7 C2A Domain: Mechanistic Contrast between Isoforms 1 and 7. Bio chemistry. doi: 10.1021/bi3007115. 32. Nalefski, E.A., Wisner, M.A., Chen, J.Z., Sprang, S .R., Fukuda, M., Mikoshiba, K., Falke, J.J. (2001) C2 domains from different C a2+ signaling pathways display functional and mechanistic diversity. Biochemistry 40, 3089-3100. 33. Sugita, S., Shin, O.H., Han, W., Lao, Y., Sudhof, T .C. (2002) Synaptotagmins form a hierarchy of exocytotic Ca(2+) sensors with distinct Ca(2+) affinities. EMBO J. 21, 270-280. 34. Nalefski, E.A., Falke, J.J. (2002) Use of fluoresc ence resonance energy transfer to monitor Ca(2+)-triggered membrane docking of C2 domains. Methods Mol. Biol. 172, 295-303.


39 35. Ziemba, B.P., Knight, J.D., Falke, J.J. (2012) Ass embly of membrane-bound protein complexes: detection and analysis by singl e molecule diffusion. Biochemistry 51, 1638-1647. 36. Ubach, J., Lao, Y., Fernandez, I., Arac, D., Sudhof T.C., Rizo, J. (2001) The C2B domain of synaptotagmin I is a Ca2+-binding module. Biochemistry 40, 5854-5860. 37. Herrick, D.Z., Sterbling, S., Rasch, K.A., Hinderli ter, A., Cafiso, D.S. (2006) Position of Synaptotagmin I at the Membrane Interfa ce: Cooperative Interactions of Tandem C2 Domains. Biochemistry 45, 9668-9674. 38. Malmberg, N.J., Falke, J.J. (2005) Use of EPR Powe r Saturation to Analyze the Membrane-Docking Geometries of Peripheral Proteins: Applications to C2 Domains. Annu. Rev. Biophys. Biomol. Struct. 34: 71-90. doi: 10.1146/annurev.biophys.34.040204.144534 39. Nagashima, T., Hayashi, F., Yokoyama, S. (2009) So lution structure of the first C2 domain of synaptotagmin VII. doi: 10.2210/pdb2 d8k/pdb. 40. Altenbach, C., Greenhalgh, D.A., Khorana, H.G., Hub bell, W.L. (1994) A collision gradient method to determine the immersio n depth of nitroxides in lipid bilayers: application to spin-labeled mutants of b acteriorhodopsin. Pro.c Natl. Acad. Sci. U.S.A. 91, 1667-1671. 41. Malmberg, N.J., Van Buskirk, D.R., Falke, J.J. (20 03) Membrane-docking loops of the cPLA2 C2 domain: detailed structural analys is of the protein-membrane interface via site-directed spin-labeling. Biochemistry 42: 13227-13240. 42. Frazier, A.A., Wisner, M.A., Malmberg, N.J., Victor K.G., Fanucci, G.E., Nalefski, E.A., Falke, J.J., Cafiso, D.S. (2002) M embrane Orientation and Position of the C2 Domain from cPLA2 by Site-Direct ed Spin Labeling. Biochemistry 41, 6282-6292. 43. Langen, R., Oh, K.J., Cascio, D., Hubbell, W.L. (2 000) Crystal structures of spin labeled T4 lysozyme mutants: implications of the i nterpretation of EPR spectra in terms of structure. Biochemistry 39, 8396-8405. 44. Altenbach, C., Flitsch, S.L., Khorana, H.G.,Hubbell W.L. (1989) Structural studies on transmembrane proteins. 2. Spin labeling of bacteriorhodopsin mutants at unique cysteines. Biochemistry 28:7806–12. 45. Tombolato, F., Ferrarini, A. (2006) Modeling the Ef fects of Structure and Dynamics of the Nitroxide Side Chain on the ESR Spe ctra of Spin-Labeled Proteins. J. Phys. Chem. B 110(51), 26260-26271.