Citation
Calcium and membrane association for th C2 domains of synaptotagmin 1 and synaptotagmin 7 by molecular dynamics

Material Information

Title:
Calcium and membrane association for th C2 domains of synaptotagmin 1 and synaptotagmin 7 by molecular dynamics
Creator:
Chon, Nara Lee ( author )
Place of Publication:
Denver, Colo.
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file : ;

Subjects

Subjects / Keywords:
Membrane proteins ( lcsh )
Exocytosis ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references.
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Nara Lee Chon.

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:
on10596 ( NOTIS )
1059640231 ( OCLC )
on1059640231

Downloads

This item has the following downloads:


Full Text
CALCIUM AND MEMBRANE ASSOCIATION FOR THE C2 DOMAINS OF
SYNAPTOTAGMIN 1 AND SYNAPTOTAGMIN 7 BY MOLECULAR DYNAMICS
by
NARA LEE CHON B.S., University of Colorado, 2014
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 Program
2018


This thesis for the Master of Science degree by
Nara Lee Chon
has been approved for the
Chemistry Program
by
Hai Lin, Chair Jefferson Knight Michael Crowley
Date: May 12, 2018
n


Chon, Nara Lee (M.S., Chemistry)
Calcium and Membrane Association for the C2 Domains of Synaptotagmin 1 mid Synaptotagmin 7 by Molecular Dynamics
Thesis directed by Professor Hai Lin
ABSTRACT
The C2 domain is an important membrane binding motif in cell signaling pathways found ubiquitously in mammals. Two synaptotagmin (Syt) proteins, Sytl and Syt7, each employ two C2 domains (C2A and C2B) to interact with membranes during exocytosis, acting as Ca -sensors. To explore Ca ion binding and membrane association of the Sytl and Syt7 C2 domains, we performed molecular dynamics simulations on a series of model systems. First, in
the solvated proteins simulations, we found that Ca ions were chelated by the protein oxygen
2+
atoms in the Ca -binding loops (CBLs): the side chain oxygen atoms in the aspartate mid the
2+
serine residues, mid backbone oxygen atoms. However, the number of bound Ca ions mid their
2+
coordination shell compositions varied. Sytl C2A and Syt7 C2A and C2B each bound 3 Ca ions tightly, but Sytl C2B only bound 2 Ca ions tightly and the outermost Ca ion weakly. Water molecules and Cl ions were also recruited from the bulk solution to complete the Ca solvation shells. Two chimeric C2 domains (Syt7:lC2ACH and Sytl:7C2BCH) were also studied as comparisons with the wild-type proteins. Syt7:1C2A is a hybrid of the Syt7 C2A body and the Sytl C2A CBLs, whereas Sytl :7C2Bch the Sytl C2B body and the Syt7 C2B CBLs. Our data suggests that Syt7:1C2A shows similar Ca -binding to the wild-type C2A of Sytl and Syt7. In contrast, the Ca ion binding of Sytl:7C2B resembles Sytl C2B in agreement with
iii


experiment. Second, we generated four of Syt7 C2Ato 3:1 POPC/POPS membrane associated models. F229 in CBL3 inserted deep into the membrane shortly after the beginning of the simulations, on the other hand, F167 in CBL1 entered mid left the membrane constantly. We suspect that this oscillating motion is caused by the competition between the electrostatic attractions (between the polybasic region near the p4 strand and lipid head groups) and hydrophobic interactions (between FI67 and the lipid acyl chains). We conclude that the affinity of the two Syt isoforms for Ca ions is important predictor of the Syt-mediated membrane fusion.
The form mid content of this abstract are approved. I recommend its publication
Approved: Hai Lin
IV


DEDICATION
To my husband Song Mo mid lovely two-year-old daughter Dana for their unconditional love mid solicitude that allow me to write this thesis weekdays and weekends with full of joy. To my family in Seoul mid here in Denver for their spiritual support. To Adam Duster for taking his time to review this thesis mid giving me honest and friendly feedback. To Hai Lin and all members of the Lin Lab for making me fall in love with computational chemistry, mid enjoy my life as a computational chemist.
v


ACKNOWLEDGMENTS
I would like to thank my advisor, Dr. Hai Lin, for his guidance and immense support in preparing this thesis. I am thankful to Dr. Jefferson Knight, Dr. Nathalie Reuter, mid Dr. Arun Anantharam, and their group members who have widened my perspectives by sharing their knowledge and enriching ideas with me.
This work is supported by National Science Foundation (CHE-0952337) and Camille & Henry Dreyfus Foundation (TH-14-028). This work used the Extreme Science mid Engineering Discovery Environment (XSEDE) under grant CHE-140070 and MCB160138, which are supported by National Science Foundation grant number ACI-1053575, mid the National Energy Research Scientific Computing Centre (NERSC) under grant m2495, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. This research is also supported by the University of Colorado Denver Chemistry Departmental Student Research Fellowship.
vi


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION.................................................................1
1.1. Membrane Association of C2 domain-containing Synaptotagmins............1
1.2. Sequence and Structural Comparisons between Sytl mid Syt7..............3
1.3. Unanswered Questions and Aims of the Present Study.....................5
II. COMPUTATIONAL METHODS........................................................7
2.1. Model Preparation.......................................................7
2.1.1. Standalone Membrane or Protein......................................7
2.1.1.1. Standalone Membrane............................................7
2.1.1.2. Standalone C2 Domains..........................................8
2.1.1.2.1. Wild-type Sytl Protein: SytlC2AWT-3Ca2+, SytlC2BWT-2Ca2+, and
S yt 1C 2 BWT- 3 Ca2+......................................13
2.1.1.2.2. Wild-type Syt7 Protein: Syt7C2AWT-2Ca2+, Syt7C2AWT-3Ca2+, and
Syt7C2BWT-3Ca2+............................................16
2.1.1.2.3. Chimeric Syt Proteins: Syt7: lC2AGH-3Ca2+, Sytl:7C2BGH-2Ca2+, and
Syt 1: 7C2BGH-3Ca2+........................................19
2.1.2. Syt7 C2A Protein-Membrane Complex Models...........................20
2.1.2.1. Construction mid Solvation of the Pre-insertion Models........21
2.1.2.2. Construction mid Solvation of the Embedded Models.............23
vii


2.2. Data Analysis
24
2.2.1. Area per Lipid (APL) and Lipid Order Parameter (&ch)..................24
2.2.2. Poisson-Boltzmann Calculation.....................................25
III. RESULTS AND DISCUSSION....................................................26
3.1. Simulations ofthe Standalone 3:1 PC/PS Membrane.......................26
3.2. Simulations ofthe Standalone C2 Domains...............................27
3.2.1. Comparisons of 2-Ca2+ with 3-Ca2+ Models in Each C2WT Domain......28
3.2.2. Comparisons of WT mid CH Domains..................................39
3.3. Simulations of Syt7 C2A Protein-membrane Association..................42
3.3.1. Comparisons of the Pre-insertion with the Embedded Syt7C2AWT-membrane.43
3.3.2. Comparisons of Simulation with Experiment.........................52
3.4. Coordination of the Ca2+ Ions in the Standalone Syt Proteins and Syt7C2AWT-
membrane Complexes....................................................53
IV. CONCLUSION.................................................................58
viii
REFERENCES
59


LIST OF ABBREVIATIONS
a
APL
P
CBL
CHARMM
CHARMM-GUI
CH
CMAP
EPR
HMMM
ScH
MD
NAMD
POPC or PC
POPS or PS
P04
PDB
PQR
RMSD
Syt
Sytl
Syt7
VMD
WT
alpha
area per lipid
beta
2+
Ca -binding loop
chemistry at Harvard macromolecular mechanics
CHARMM, a web-based graphical user interface
(www.charmm-gui.org)
chimeric or chimera
cross-term for the <|>, y (backbone dihedral angle) values or grid-based energy correction maps electron paramagnetic resonance
highly mobile membrane mimetic
lipid order parameter
molecular dynamics
nanoscale molecular dynamics
l-palmitoyl-2-oleoyl-^?7-glycero-3-phosphocholine
l-palmitoyl-2-oleoyl-^?7-glycero-3-phospho-L-serine
phosphate group
protein data bank (www.rcsb.org)
PDB data with atomic charge and radius parameters
root-mean-square deviation
synaptotagmin
synaptotagmin 1
synaptotagmin 7
visual molecular dynamics
wild-type
IX


CHAPTER I
INTRODUCTION
1.1. Membrane Association of C2 domain-containing Synaptotagmins
The C2 domain, first identified in protein kinase C as the second-conserved domain,1 is found to be a key component of many cell signaling proteins such as synaptotagmin (Syt) proteins. The Syt proteins contain two C2 domains, C2A and C2B. Each domain comprises approximately 130 residues. All C2 domains have eight beta strands (P1-8) connected by flexible loops including three calcium(Ca )-binding loops (CBLs, CBL1-3) (Figure 1). With two to four Ca ions bound in the CBLs, synaptotagmin 1 (Sytl) and synaptotagmin 7 (Syt7) bind to membranes containing negatively charged phospholipids such as phosphatidylserine (Figure
2) 4, 5, 6, 7
2+
Sytl mid Syt7, both of which serve as Ca ion sensors, are the two well-studied proteins in the Syt family. They are responsible for exocytosis in synaptic vesicles and in secretory granules of endocrine cells. Despite their structural similarities, these two proteins facilitate vesicle release in markedly different ways. In neurons, Sytl primary responds to the rapid synchronous neurotransmitter release, while Syt7 facilitates the slow asynchronous release by jointly regulating the exocytosis with other Syt isoforms.891011 Both C2A domains in Sytl and Syt7 bind three Ca ions to the CBLs. However, the C2A domain of Syt7 has stronger electrostatic attractions between its polybasic lysine cluster region mid the phospholipid head groups than Sytl.121314 The coordination of the Ca2+ ions by the protein and lipid, mid insertion of hydrophobic residues into the membrane allow Syt7 C2Ato dock deeply to the membrane.15
1


*2 I
A relative role for the C2 domains of Sytl and Syt7 in Ca -triggered neurotransmitter
0 -I- m m m
release is reversed. For Sytl, Ca binding to the C2B domain is more important for synaptic transmission than Ca2+ binding to the C2A domain.16 The isolated Sytl C2B domain can simultaneously bind to two membranes: one membrane through the CBL region at the bottom of the C2B domain, and another membrane through the top face loops in the presence of Ca ions, supporting that the Sytl protein may trigger neurotransmitter release by bringing the synaptic vesicle and plasma membranes together mainly via C2B domains. For Syt7, however, the C2A domain actively militates in favor of fusion opening but the C2B domain is selectively
90 &
essential for neurotransmitter release. ,
Figure 1. The structures of the C2 domains in Syt7 in gray (p4 strand in orange). (A) C2A and (B) C2B domains are with each of the highlighted Ca -binding loops (CBL1-3; red, yellow, and blue spheres). The Ca2+ ions are represented by small green spheres with labels from 1 to 3. The side chains of the critical residues in Syt7 C2A (FI 67 and F229) are only shown as sticks (white,
91
H; cyan, C). Figure modified from Fig. 2 in Ref.
2


o
Figure 2. The chemical structures of (A) l-palmitoyl-2-oleoyl-s-glycero-3-phosphocholine and (B) l-palmitoyl-2-oleoyl-*src-glycero-3-phospho-L-serine. Figure modified from Fig. 1 in Ref.21
1.2. Sequence and Structural Comparisons between Sytl and Syt7
2_|_
The two Syt isoforms, Sytl and Syt7, each contain a set of two Ca -activated C2 domains. Figure 3A displays the sequence alignment for the aforementioned proteins. The Sytl and Syt7 C2A domains can each accommodate three Ca ions in the binding sites. 5 The residues of CBL2 and CBL3 are well-conserved across Syt isoform (Figure 3B). Two residues (arginine and phenylalanine) located in the middle of CBL3 are strictly conserved across these proteins; in particular, F234 in Sytl and F229 in Syt7 play a key role in the protein-binding to the membrane.35 2425 The CBL1 residues in Sytl and Syt7 are less strictly conserved, with M173 in Sytl corresponding to FI67 in Syt7. The CBL region is known to be essential for Ca -dependent phospholipid binding. This binding is mediated by CBL1 and CBL3 primarily through electrostatic attractions and secondarily through hydrophobic interactions. When the protein approaches the membrane surface, the two hydrophobic residues in CBL 1 and CBL3 of Syt7 (FI67 and F229) act as anchors to the hydrophobic lipid tails. However, M173 in Sytl C2A does
3


not exhibit the hydrophobic interactions with the lipid tails as strong as the phenylalanine does. Hence, the binding affinity ofSytl is less than of Syt7.1521 14
Just like C2A, both C2B domains in Sytl and Syt7 have a cluster of the acidic residues
2+
such as the aspartate residue in the CBLs. However, the number of bound Ca ions are different:
26 27
two and three Ca ions bind to Sytl mid Syt7, respectively. A serine residue in CBL3 of
2+
Syt7 C2B (S362) is not present in Sytl C2B domain. Crystallography revealed the third Ca ion
9+
binds to the CBLs in Sytl C2B but at very high Ca ion concentration with coordination of the side chain oxygen atom in the asparagines residue. Moreover, Sytl C2B contains two a-helices: one of these helices is in Syt7 C2B, while the other helix near the C-terminal loop is predicted to be present in only Sytl. The functional differentiation in Sytl mid Syt7 arises in part from subtle sequence variations.
4


C2A CBL1
C2A CBL2
(A)
Sytl WT C2AB 140
syt7 WT CRAB 134
3yl7: ; 1 CH C2A 134
Sytl: : 7 CH (J2B
Sytl WT C2AB 211
3yt7 WT C2AB 235
3yL7: : 1 CH C2A 235
Sytl: : 7 CH C2B 272
Sytl WT C2AB 231
3yt7 WT C2AB 275
Syt 7: : 1 CH C2A
Sytl; ; 7 CH C2B 231
Sytl WT C2AB 352
Sy L7 WT C2AB 341
Syt 7: : 1 CH C2A
Sytl: : 7 CH C2B 352
EKLGKLQYSLDYDFQKHQLLVGIIQAAELPAL3MGGT3DPYVKVELLPDKKKKFBTKVHR?rrLNPVFNEQF FNLGRIQFSVGYNFQESTI .TVKIMKAQE LP AKDFSCTSI3PFVKT YTiIjPCKtCHKLE TKVKEKN LNPHWNE ?F ENLGRIQFSVGYNFQESTLT'/KVMKAQELPALDMGGTSDPFVKIYLLPCKKHKLBTKVHRKTLNPHWNETF
C2A CBL3
TFK-VPYSEIXJGKTLVMAVYDFDRFSKHDIIGEFKVPMNTVDFGI-IVTEEWRDLQSAEKEEQEKLGDICFSL LFEGF PYEKWQRILYLQVLDYDRFSRNDPIGEVSIP LNKVDLT QMQT FWKDLKP SGPSSGS R-GEL LLS 1 LFEGFPYEKWQRVLYLQVLDFDRFSKHDPIGEVSIPLNKVDLTQMQTFWKDLKP (259)
KLGDICFSi
C2B CBL1 C2B_CBL2
RWPTAGKLTv'VILEP.KNIKKUDVGGLSDPWKIHLKQNGKRLKKKKTTIKKNTLMPYYNESFSFEVPFEQ
CYNPSAKSIIVNIIKARNLKflMDIGGTSDPYVKVWLMYKDKRVEXKKTVTMKKMLNPIFNESFAFDIPTEK
RY'/PTAGKLn'VILEAKNLKSMDIGGTSDPYVKIHmQNGKRLKKKKTTIKKRNLNPYYNESFSFEVPFEQ C2B CBL3
IQKVQVWTVLDYDKIGKNDAfGKVFVGYNSTGAELRHWSDMI.A>IPRRPIAQWHTLQVEEEVDAMLAV (419; LRETTIIITVMDKDK1SRNDVIGKIYLSWKSGPGEVKHWKDMIARFROPVAQWHQLKA (403)
IQKVQWVTVLDKDKLSRNDAXGIC/FVGYNSTGAELRHtfSDMLAXPRRPIAQWHTLQVEEEVDAMLAV 1419)
(B)
CBLl region
Sytl
Syt2
Syt9
Syt6
SytlO
Syt5
Syt3
Syt8
Syt4
Syt7
LPALD LPALD LAALDH LPAKD LPAKD-LPAKD-LPAKD LKA-LPAMD LPAKD-I
3
3

i js
GT GT [JGGS FCGS F TGT F|SGT GF |EGT MT 0SGT
CBL2 region
RKTLN
RKTLN
RQTLN
RKTLN
RKTLN
RKTLN
RKTLN
RGTLC
RKTLD
RKNLN
CBL3 region
DRFSKHD
DRFSKHD
DRFSRND
DRFSRHD
DRFSRHD
DRFSRHD
DRFSRHD
KRFSEHE
DRFSRDD
DRFSRND
Figure 3. Sequence alignment and comparison. (A) C2 domains of all proteins employed in this
I m m m m
work, showing each three of Ca -binding loops (CBLs) with red shaded boxes. The critical residues in the CBLl and CBL3 of Syt C2A domains are also highlighted with the red font. For the chimeric (CH) domains, Syt7:1C2A contains the Syt7 C2A sequence, but with the CBLs
/"ITT
replaced by the counterpart in Sytl C2A domain. Sytl:7C2B contains the Sytl C2B sequence, but with the CBLs from Syt7 C2B. (B) The C2A domain sequences of CBLl to CBL3 regions among mammalian Syt isoforms using Clustal Omega. Boxed regions indicate positions at the tip of CBLl and CBL3. Figure taken from Fig. SI in Ref.21
1.3. Unanswered Questions and Aims of the Present Study
Detailed mechanisms of the membrane associations by Sytl and Syt7 are not well understood at the molecular level. Atomistic molecular dynamics (MD) simulations could provide some much needed missing information about the functions of the proteins. There are several questions that we seek to answer:
5


2_|_
(1) How many Ca ions are most likely bound to Syt7 C2A? It has been experimentally
9+
shown that the C2A domain in Syt7 can bind three Ca ions, but the affinity is much lower for the third Ca than the first two Ca ions. The available NMR structure of Syt7 C2A is a Ca2+-free form.31 Because it is essential for Syt7 C2A to bind Ca2+ ions before inserting into the membrane, it is important to determine the most optimal stoichiometry of Ca ion binding.
(2) What are the structural features of Sytl and Syt7 at membrane binding sites? Recently Rao et al. found that Sytl and Syt7 are sorted to different populations of chromaffin granules, and fusion pores of granules harboring Sytl expand more rapidly than pores of granules expressing Syt7. Here we performed wild-type (WT) versus chimera (CH) simulations by generating the Syt7:lC2A and Sytl:7C2B domains as a preliminary study. The Syt7: 1C2Ach denotes that the C2A domain contains the Syt7 C2A body with Sytl C2A CBLs, while Sytl:7C2BCH contains the Sytl C2B body with Syt7 C2B CBLs.
(3) What is the orientation of Syt7 C2A when it docks to the membrane? We will characterize the geometry of the protein-membrane complex. Our results will be compared with the experiments from electron paramagnetic resonance (EPR) depth parameter measurements.34
The results presented in this thesis were from two independent research works: (i) Ca -binding for C2AWT of Sytl and Syt7, and membrane association for Syt7 C2A,2134 mid (ii) Ca2+-binding for C2Ach, C2Bch, and C2BWT of Sytl and Syt7.35 Therefore, the simulation settings in corresponding to the experiments are different from each work.
6


CHAPTER II
COMPUTATIONAL METHODS
2.1. Model Preparation
2.1.1. Standalone Membrane or Protein
2.1.1.1. Standalone Membrane
A phospholipid bilayer model was generated by mixing l-palmitoyl-2-oleoyl-.s'w-glycero-3-phosphocholine (POPC, PC) mid l-palmitoyl-2-oleoyl-,s'?7-glycero-3-phospho-L-serine (POPS, PS) at a molar ratio of 3:1, employing Membrane Builder in CHARMM-GUI.36 The 3:1 PC/PS membrane model is composed of 192 POPC lipids, 64 POPS lipids, 64 Na+ ions as counter ions, and 9219 water molecules were prepared by the CHARMM36 force fields and TIP3P water model. Using NAMD, the model was equilibrated at 310 K and 1 bar for 160 ns under the NpT ensemble with a periodic boundary.40
7


Table 1. Numbers of atoms and dimensions for protein, membrane, and protein-membrane complex models, employed in this work.
Protein Water Lipid Ca2+ Na+ K+ CL Initial cube dimension [A3]
Standalone membrane
3:1 PC/PS 0 27,657 33,856 a 0 64 0 0 92x92x79
Standalone protein WT
SytlC2A 3Ca2+ 2,083 19,839 0 3 0 0 4 80x80x50
SytlC2B 2Ca2+ 2,410 75,657 0 2 0 71 81 101x126x103
SytlC2B 3Ca2+ 2,410 75,651 0 3 0 71 83 101x126x103
Syt7C2A 3Ca2+ 2,181 12,150 0 3 0 0 9 50x65x65
Syt7C2B 3Ca2+ 2,300 75,762 0 3 0 71 89 101x126x103
CH
Sytl:7C2B 2Ca2+ 2,413 75,246 0 2 0 71 82 101x126x96
Sytl:7C2B 3Ca2+ 2,413 75,648 0 3 0 71 84 101x126x103
Syt7:lC2A 3Ca2+ 2,160 75,864 0 3 0 71 78 101x126x103
Protein-membrmie Pre-insertion
Lying-down 2,181 75,951 33,856 a 3 64 0 9 92x94x120
Standing-up 2,181 75,747 33,856 a 3 64 0 9 91x95x118
Embedded
jSt 2,181 65,805 33,454 b 3 64 0 9 92x94x106
2nd 2,181 74,898 33,454 b 3 64 0 9 91x94x118
a256 lipids, including 192 POPC (134 atoms each) mid 64 POPS (127 atoms each).
b 253 lipids, including 189 POPC and 64 POPS (deleted 3 of PC lipids close to protein).
2.1.1.2. Standalone C2 Domains
All structural manipulations were performed using VMD.41 Minimizations and MD simulations were performed using NAMD as detailed below. The Syt proteins were solvated in either water (for WT C2A domain models) or 0.15 M of KC1 solution (for WT C2B domain and all chimera, CH, C2 domain models), described by CHARMM22 force fields with the CMAP correction42,43,44 mid TIP3P38 water model. The number of water molecules and counter ions (K+ or Cf) in the systems varied across models, but all systems were charge neutral (Table 1-2). For
8


the CH Syt C2 domain, homology models were created from amino acid sequences (translated from the DNA sequences through Translate Tool in ExPASy45) using the Swiss-Mo del46 program. The procedure of constructing the Syt protein models is illustrated in Figure 4, and the detailed information will be discussed in the following sections. The minimization mid MD simulations were performed at 298 K mid 1 bar under the NpT ensemble with a periodic boundary. Harmonic restraints were imposed upon the distances between Ca ions mid their coordinating oxygen atoms during the early stage of equilibration; the force constants were 50 kcal mol'1 A'2 for Syt7: lC2AGH-3Ca2+, and 200 kcal mol'1 A'2 for both Sytl :7C2Bch-2 and -3Ca2+ models. The simulation times are listed in Table 3.
9


Table 2. The net charges (in e) of secondary structures and the whole proteins.
Structure Protein C2A C2B
True ID Relative Charge True ID Relative Charge
WT CH WT CH
Overall: Sytl E140-K267 1-128 -2 + 1 K272-V419 1-148 +6 +7
Start-End Syt7 E134-G265 1-132 +3 S266-A403 1-138 + 12
PI Sytl K144-D152 5-13 -1 D275-L280 4-9 -1 -1
Syt7 R138-Y145 5-12 + 1 + 1 E269-277 4-12 -1
P2 Sytl N157-A166 18-27 0 L289-K297 18-26 0 0
Syt7 L152-Q160 19-27 +2 +2 S282-R291 17-26 +2
CBL1 Sytl L168-T176 29-37 -1 -1 K301-D309 30-38 -1
Syt7 L162-T170 29-37 0 A295-D303 30-38 -2
P3 Sytl P179-L186 40-47 + 1 P310-Q318 39-47 + 1 + 1
Syt7 P173-L180 40-47 + 1 + 1 P304-Y312 39-47 + 1
P4 Sytl K192-E194 53-55 0 R322-K327 51-56 +5 +5
Syt7 K186-E188 53-55 0 0 K315-K321 50-56 +4
CBL2 Sytl H198-N203 59-64 +2 +2 K332-P337 61-66 + 1
Syt7 K192-N197 59-64 +3 K326-P331 61-66 +2
P5 Sytl E208-F212 69-73 -1 Y338-F345 67-74 -1 -1
Syt7 E202-F206 69-73 -1 -1 I332-D340 67-75 -2
P6 Sytl T223-D230 84-91 -1 V355-D363 84-92 -1 -1
Syt7 I218-D225 85-92 -1 -1 T349-D357 84-92 -1
CBL3 Sytl F231-H237 92-98 + 1 + 1 Y364-D371 93-100 0
Syt7 Y226-N232 93-99 + 1 K358-D365 93-100 + 1
P? Sytl I239-P246 100-107 0 A372-G379 101-108 + 1 + 1
Syt7 D233-P241 100-108 -2 -2 V366-L372 101-107 + 1
P8 Sytl V255-D261 116-122 -2 I401-T406 130-135 0 0
Syt7 Q251-K255 118-122 + 1 + 1 V395-Q400 130-135 0
10


C2A
CNJ
CS
o
CO
<
C2B
CNJ
CS
o
CO
<
Syt1C2AWT-3Ca2+ B Syt1C2AWT-3Ca2+ E $yt1C2AWT-3Ca2+

\ Syt7C2AWT-0Ca2+ Syt7C2AWT-2Ca2+ E Syt7C2AWT-2Ca2+
I 1 B fD: 3rd Ca2+
1 Syt7C2AWT-3Ca2+ E Syt7C2AWT-3Ca2+
1 I
i' A: 3 Ca^+|
Syt7:1 C2ACH-0Ca2+ B Syt7:1 C2ACH-3Ca2+ E Syt7:1 C2ACH-3Ca2+

r Syt1 C2BWT-2Ca2+ B Syt1 C2BWT-2Ca2+ E Syt1 C2BWT-2Ca2+
i B
Syt1 C2BWT-3Ca2+ E Syt1 C2BWT-3Ca2+
A: 3rd Ca2+
i
i Syt7C2BWT-3Ca2+ B Syt7C2BWT-3Ca2+ E Syt7C2BWT-3Ca2+

Syt1:7C2BCH-0Ca2+ i i B Syt1:7C2BCH-2Ca2+ E Syt1:7C2BCH-2Ca2+
t D: 3rd Ca2+
Syt1:7C2BCH-3Ca2+ E Syt1:7C2BCH-3Ca2+
r- Legend
A: adding B: building D: deleting E: equilibrating : modeling direction template structure
experimental structure
translated amino acid sequence
initial geometry
equilibrated structure
Figure 4. The construction procedure of C2 domain model systems. When building a 3-Ca bound model, the addition of all three or the outermost (3rd) Ca2+ ion(s) from one structure to another is carried out after the superposition (on backbone heavy atoms: N, C, O, and Ca) of the two structures. The MD simulations are performed as listed in Table 3.
11


Table 3. The settings of NAMD in MD simulation for all models.
Equilibration times a (Restraints b)
C2A w/o membrane
SytlC2AWT-3Ca2+ 20ps-20ps-200ps-200ns
Syt7C2AWT-2Ca2+ 20ps-20ps-200ps-200ns
Syt7C2AWT-3Ca2+ 20ps-20ps-200ps-200ns
Syt7:1C2 ACH-3Ca2+ 500ps(50)-500ps(50)- Ins- 100ns
C2B w/o membrane
SytlC2BWT-2Ca2+ 500ps-500ps- Ins- 100ns
SytlC2BWT-3Ca2+ 500ps-500ps- Ins- 100ns
Syt7C2BWT-3Ca2+ 500ps-500ps- Ins- 100ns
Sytl :7C2BGH-2Ca2+ 500ps(200)-500ps(200)- Ins- 100ns
Sytl :7C2BGH-3Ca2+ 500ps(200)-500ps(200)- Ins- 100ns
C2A w/ membrane
Syt7C2A-3Ca2+ 400ps(100, 2000)-400ps(100, 2000)-2ns-500ns
a The equilibration was performed in four stages. In the first two stages, the temperature and pressure controls were turned on sequentially while the protein backbone atoms were frozen and the Ca solvation shells were restrained. Then the backbone was allowed to move freely in the third stage. In the final stage, the MD simulations were performed without any constraint or restraint.
b The harmonic force constants (in kcal mol"1 A"2) of the restraint potentials for the Ca2+-0 distance (see Table 4). For the C2A with membrane simulations, 100 kcal mol"1 A"2 were put upon the Ca2+-0 distances in pre-insertion models and 2000 kcal mol"1 A"2 were in the embedded models.
12


2.1.1.2.1. Wild-type Sytl Proteins: SytlC2AWT-3Ca2+, SytlC2BWT-2Ca2+, and SytlC2BWT-3Ca2+
The experimentally determined structures for the Ca -bound Sytl C2A and C2B domains [Protein Data Bank(PDB) entries 1BYN22 and 1K5W26, respectively] were utilized to build the solvated models. In this Sytl C2A structure, the coordination of three Ca ions is listed in Table 4. The C2B domain of Sytl, however, only have 2 Ca ions in the binding site. Thus, the first Ca2+ ion is coordinated by the oxygen atoms of the D303, D309, D363, and D365 side chains, mid the Y364 backbone; mid the second Ca ion by the D303, D363, D365, and D371 side chains, and the M302 backbone.
13


9+
Table 4. The distances between Ca and their coordinating oxygen atoms in the binding site of 3-Ca2+ bound Syt C2 domain models. Experimental distances given in parenthesis. a
WT C2A Sytl Syt7 CH C2A Syt7:l
Residue Atom Distance [A] Residue Atom Distance [A] Residue Atom Distance [A]
| st Ca2+ D172 OD1 2.6 0.2 (2.81) D166 OD1 2.2 0.1 D166 OD1 2.1 0.1
D172 OD2 2.1 0.1 (2.82) D166 OD2 2.5 0.2 D166 OD2 2.8 0.3
D178 OD2 2.1 0.1 (2.81) D172 OD2 2.1 0.1 D172 OD1 2.1 0.1
D230 OD1 2.1 0.1 (2.81) D225 OD1 2.1 0.1 D225 OD1 2.1 0.1
F231 0 2.3 0.1 (2.81) Y226 O 2.3 0.1 F226 O 2.3 0.1
D232 OD1 2.2 0.1 (2.82) D227 OD1 4.2 0.1 D227 OD1 2.2 0.1
2nd Ca2+ L171 0 2.3 0.1 (2.79) K165 O 2.3 0.1 LI 65 O 4.8 0.2
D172 OD1 2.2 0.1 (2.83) D166 OD1 2.2 0.1 D166 OD1 4.1 0.1
D230 OD1 3.9 0.1 (3.31) D225 OD1 3.9 0.1 D225 OD1 4.0 0.1
D230 OD2 2.1 0.1 (2.84) D225 OD2 2.1 0.1 D225 OD2 2.2 0.1
D232 OD1 2.5 0.2 (2.80) D227 OD1 2.2 0.1 D227 OD1 2.4 0.1
D232 OD2 2.2 0.1 (2.83) D227 OD2 2.5 0.1 D227 OD2 2.2 0.1
D238 OD2 2.1 0.1 (2.82) D233 OD2 2.1 0.1
3rd Ca2+ D232 OD2 6.0 0.2 (2.84) D227 OD2 6.6 0.1 D227 OD2 4.8 0.6
S235 OG 6.9 0.8 (2.84) S230 OG 5.7 0.1 S230 OG 5.3 0.8
K236 0 4.8 0.7 (2.81) R231 O 4.9 0.1 K231 O 4.3 0.3
D238 OD1 2.1 0.1 (2.84) D233 OD1 2.1 0.1 D233 OD1 3.6 0.4
D238 OD2 4.2 0.2 (2.84) D233 OD2 3.7 0.1 D233 OD2 2.1 0.1
WT C2B CH C2B
Sytl Syt7 Sytl: 7
| st Ca2+ D303 OD1 2.2 0.1 (2.59) D297 OD1 2.4 0.1 (2.67) D303 OD1 2.7 0.3
D303 OD2 2.6 0.2 (2.82) D297 OD2 2.2 0.1 (2.43) D303 OD2 2.2 0.1
D309 OD2 2.1 0.1 (2.23) D303 OD2 2.1 0.1 (2.37) D309 OD2 2.1 0.1
D363 OD1 2.1 0.1 (2.81) D357 OD1 2.1 0.1 (2.40) D363 OD1 2.1 0.1
D363 OD2 3.1 0.2(3.31) K358 O 2.3 0.1 (2.35) K364 O 2.3 0.1
Y364 0 2.3 0.1 (2.82) D359 OD1 2.3 0.1 (2.32) D365 OD1 2.2 0.1
D365 OD1 2.2 0.1 (2.80)
2nd Ca2+ M3 02 0 2.3 0.1 (2.82) M296 O 2.3 0.1 (2.32) M3 02 O 2.3 0.1
D303 OD1 4.1 0.1 (2.81) D297 OD1 2.2 0.1 (2.42) D303 OD1 2.2 0.1
D363 OD2 2.1 0.0(2.80) D357 OD2 2.1 0.0(2.30) D363 OD2 2.1 0.1
D365 OD1 2.4 0.1 (2.81) D359 OD1 2.3 0.1 (2.50) D365 OD1 2.5 0.1
D365 OD2 2.2 0.1 (2.81) D359 OD2 2.3 0.1 (2.49) D365 OD2 2.2 0.1
D371 OD2 2.1 0.1 (2.81) D365 OD2 2.1 0.1 (2.32) D371 OD2 2.1 0.1
3rd Ca2+ D365 OD2 4.3 0.2 D359 OD2 2.2 0.1 (2.38) D365 OD2 4.6 0.2
K369 0 4.8 0.5 S362 OG 2.3 0.1 (2.44) S368 OG 5.1 0.6
D371 OD1 2.1 0.0 R363 O 2.2 0.1 (2.32) R369 O 4.1 0.2
D371 OD2 3.5 0.2 D365 OD1 2.1 0.1 (2.45) D371 OD1 2.1 0.1
For Syt7C2Aw l and the CH domains, and the third Ca2+ ion in SytlC2Bw l, the experimental
values are not available.
14


Two 3-Ca2+ models (SytlC2AWT-3Ca2+ and SytlC2BWT-3Ca2+) mid one 2-Ca2+ model (SytlC2BWT-2Ca2+, as mentioned earlier) was generated for Sytl. Among the 20 NMR structures available in the PDB (1BYN), the first was selected for building the SytlC2AWT-3Ca2+ model. Just like Sytl C2A domain, the first model of 20 NMR structures in the PDB (1K5W) for SytlC2BWT-2Ca2+ model was used, followed by solvation and adding ions for charge neutrality. To build the three Ca2+-bound C2B model, SytlC2BWT-2Ca2+ was superimposed with Syt7C2BWT-3Ca2+ (PDB entry 3N5A27), and the third Ca2+ in the Syt7C2BWT-3Ca2+ binding pocket was manually copied to the SytlC2BWT-3Ca2+ model. Note that the C2B domains of Sytl and Syt7 are very similar with a RMSD of only 0.9 A for the backbone heavy (N, C, O, and Ca) atoms (Figure. 5). By construction, the third Ca ion in SytlC2B -3Ca interacts with the side chain oxygen atoms of the D365 mid D371, and the backbone oxygen atom of the K369
(Table 4), which are corresponding oxygen atoms in Syt7C2BWT-3Ca2+. The serine residue that
2+
is expressed in CBL3 of Syt7 C2B is not present in Sytl C2B domain. Therefore, in all 3-Ca models except for SytlC2B -3Ca model, the third Ca ion was chelated by the hydroxyl oxygen atom in the serine along with the carboxyl oxygen atoms in the aspartate and the carbonyl oxygen atoms in the protein backbone.
15


Figure 5. (A) Superposition of the Sytl and Syt7 C2B domains based upon the root-mean-square deviation (RMSD) per residue. As the RMSD values become small to big, the color range goes from blue to red. The last 10 residues (E410-EEVDAMLA-V419; a red a-helix in the figure; green arrow) in Sytl C2B were ignored during the comparison, because Syt7 C2B does not have the corresponding residues. (B) RMSD per residue with respect to the start of the Syt7 C2B domain. The residue numbers in the context of the full-length protein can be obtained by adding 271 and 265 for the C2B domains in Sytl and Syt7, respectively.
2.1.1.2.2. Wild-type Syt7 Proteins: Syt7C2AWT-2Ca2+, Syt7C2AWT-3Ca2+, and Syt7C2BWT-3Ca2+
The solvated Syt7 C2A with Ca ions were acquired from the experimental NMR
O 1 rs.
structures (PDB entry 2D8K ). While the presence of up to three Ca ions is known to stabilize the Syt7 C2A domain, no Ca ions were present in the experimental structure. 5 5 In order to
r\_|_
find the most feasible Ca -bound C2A structures, two models were rendered by augmenting two and three Ca2+ ions, respectively (Syt7C2AWT-2Ca2+ and S)47C2Aw-3Ca2+) using the first model of 20 NMR structures. First, SytlC2AWT-3Ca2+ was aligned with Ca2+-free Syt7 C2A. Then the coordinates of all 3 Ca ions were extracted from Sytl and added to Syt7, resulting in Syt7C2A -3Ca All three Ca ions were coordinated by the oxygen atoms as listed in Table
4. For Syt7C2Awr-2Ca2+, the third Ca2+ ion was simply deleted from Syt7C2Aw-3Ca2+. Figure
16


1A offers an overview of the Syt7 C2A, with two critical residues in CBL1 and CBL3. The
9+
conformations of the CBLs with Ca ions after minimization were presented in Figure 6, showing that some Ca ions in the binding pocket required not only the protein oxygen atoms but also water molecules to form a complete solvation shell (see Section 3.4. Coordination of the Ca2+ Ions in the Standalone Syt Proteins and Syt7C2Axv l mcmbranc Complexes).
A model of Syt7 C2B was prepared in a similar way. The three Ca -bound crystal structure (PDB entry 3N5A, Syt7C2BWT-3Ca2+) was used as an initial structure. The protein was, solvated, mid ions were added for neutralizing the excess charge. The coordination of all three Ca ions was again described in Table 4.
17


(A) C2A domains
(B) C2B domains
Figure 6. Geometries of Ca2+ ions in the binding sites after minimization. (A) C2A domains in gray (SytlC2Awr~3Ca2+, top; Syt7C2Awr~2Ca2+, middle; Syt7C2Awr~3Ca2+, bottom) and (B) C2B domains in purple (SytlC2Bwr~2Ca2+, top; SytlC2Bw-3Ca2+5 middle; Syt7C2Bw-3Ca2+5 bottom). All Ca2+ ions were in green spheres with surrounding residues shown as sticks (N, blue; O, red; C, cyan), and the coordinating oxygen atoms for Ca2+ ions shown as red spheres. Protein H omitted for clarity. Figure modified from Fig. 2 in Ref.21
18


2.1.1.2.3. Chimeric Syt Proteins: Syt7:lC2ACH-3Ca2+, Sytl:7C2BCH-2Ca2+, and Sytl :7C2BCH-3Ca2+
A chimeric (CH) C2A domain with 3 Ca ions in the Ca -binding pocket (Syt7: lC2AGH-3Ca2+) was constructed, where Syt7:1 denotes the combination of Syt7 C2A body, and Sytl C2A Ca2+-binding loops. The homologous model for Syt7: lC2AGH-3Ca2+ was built using the Ca -free NMR experimental structure of Syt7 C2A as a template. The model consists of 126 residues E134-P259, as S260-GPSS-G265 in the template sequence was excluded. These 6 residues were added at the C-terminus of Syt7:1C2A using the molefracture plugin in VMD, which assists the comparisons between Syt7: lC2AGH-3Ca2+ mid Syt7C2AWT-3Ca2+ in the presence of the phospholipid bilayer. Finally, the equilibrated Syt7C2AWT-3Ca2+ structure supplied the coordinates of three Ca ions in the binding site to the CH C2A model.
Two models of CH C2B (Sytl:7C2BGH-2Ca2+ and Sytl:7C2BGH-3Ca2+) were created by mixing the bodies of Sytl C2B and the CBL region of Syt7 C2B. The CH C2B domains were constructed using the Sytl C2B NMR structure as a template. The three Ca ions were copied from the crystal structure of Syt7 C2B and added to the homologous model to generate the Sytl:7C2BGH-3Ca2+ model. By construction, the overall Ca2+-binding geometry in Sytl:7C2BCH-3Ca2+ is similar to that in Syt7C2BWT-3Ca2+, with a one-to-one correspondence between the Ca2+ coordinating residues in the two models. The coordination of three Ca2+ ions was given in Table 4. The Syt 1: 7C2BGH-2Ca2+ model was then constructed from the equilibrated Sytl:7C2BCH-3Ca2+ trajectory after deleting the third Ca2+ ion.
19


2.1.2. Syt7 C2A Protein-Membrane Complex Models
42
All protein-membrane complex models were built using CHARMM and VMD programs, and minimization and MD simulations were performed using NAMD. Two types of protein-membrane relative geometries were studied for Syt7C2AWT: the protein above the membrane (pre-insertion) mid inserted into the membrane (embedded) in Figure 7. The equilibrated 3:1 PC/PS membrane was used for constructing all Syt7C2AWT-membrane complexes. After the equilibrated standalone protein and membrane were merged, the energy minimization was carried out to relax the structures with harmonic restraints on selected atoms (Table 5). After adding water molecules mid counter ions into the systems, the initial protein-membrane structures were minimized again using NAMD for 18,000 steps. MD simulations
were carried out under the NpT ensemble at 298 K and 1 bar with four stages of sequential
2+
equilibration: first two stages for 800 ps with a frozen protein backbone mid restrained Ca solvation shells, then for 2 ns after releasing the restraints on the protein backbone. The lengths of the final stage of equilibration varied from system to system (Table 3).
Table 5. The harmonic restraints on the selected atoms of the protein and membrane in the energy minimizations through CHARMM program for building the initial protein-membrane complex structures.
Force constants (kcal mol"1 A"2) Selected atoms
150 - protein backbone atoms (N, C, O, mid Ca) - water molecules in the standalone membrane model - counter ions (Na+ ions) in the standalone membrane model
100 - membrane atoms located >5 A away from the protein - protein side chain atoms located >10 A away from the membrane
50 - membrane atoms located <5 A from the protein
1 - protein side chain atoms located <10 A from the membrane
20


2.1.2.1. Construction and Solvation of the Pre-insertion Models
The equilibrated structure of Syt7C2AWT-3Ca2+ model was extracted from the trajectory (at t = 6 ns) generated from the 200-ns standalone solvated protein simulation. The water molecules and ions were then deleted. The Ca -bound protein was placed above the equilibrated 3:1 PC/PS membrane in two different orientations, namely the lying-down mid standing-up orientations, in which the tilt angle between the long axis of the C2A domain mid the membrane surface was set to 23 and 36, respectively (Figure 7A). The tilt angle for the protein with respect to the membrane plane is defined in Figure 8A. In the lying-down model, the Syt7 C2A tilt angle is similar to a previously reported Sytl C2A experimental docking geometry.47 The critical residues F167 in CBL1 and F229 in CBL3 of the lying-down model were in the aqueous phase above the membrane, but CBL3 was more closely located to the membrane surface. In contrast to the lying-down model, CBL1 and CBL3 in the standing-up model were similarly positioned from the membrane in order to place the residues FI67 and F229 in the aqueous phase right above the membrane; this required no lipid molecule deletion (Table 1). Then the models were solvated in water, and ions were added to neutralize the excess charge.
21


(A) Syt7C2AWT Pre-insertion: Lying-down (left) and standing-up (right) (B) Svt7C2AWT Embedded
Figure 7. Initial models used for simulations of Syt7C2A -membrane complexes. Two preinsertion models were made: (A, left) lying-down and (A, right) standing-up models have slightly different tilt angles with respect to the membrane plane and initial penetration depths of F229 of CBL3 (23 and -6.9 A for the lying-down model, and 36 and -4.2 A for the standing-up model, respectively). (B) In the embedded models, both CBL1 and CBL3 were inserted into the membrane by 1.8 A and the protein remained approximately perpendicular to the membrane plane throughout the simulation. The protein structures are shown as cartoon in gray ((34 strands in orange, and the critical residues FI67 and F229 for Syt7C2Aw are represented by van der Waals spheres). Lipids are shown as lines, and water molecules and counter ions are hidden for clarity. Color code: N, blue; O, red; C, cyan; H, white; and P, brown. Figure modified from Fig.
3 in Ref.21
(A) l ilt angle
(B) Depth of penetration
Figure 8. (A) The tilt angle (0) of the protein (gray, cartoon) with respect to the membrane surface (green shaded area, M). The plane of the protein (blue shaded area, P) was defined by the three center-of-mass vectors of the outer most (3 sheets, (35 (red), (36 (blue), and (38 (purple). Each normal vector of the membrane surface and the protein was denoted as Uu and nP, respectively. (B) The penetration depth d between the Ca atom (yellow sphere) in the residues FI67 and F229 for S)47C2AW (the conserved residue F452 in Slp4C2AAA given in parentheses) and the average phosphate plane of the membrane (black horizontal line). Lipids in this figure are shown as lines except phosphate atoms as licorice. Water molecules and counter ions are hidden for clarity. Color code: N, blue; O, red; C, cyan; H, white; and P, brown. Figure modified from Fig. 3-4 in Ref.21
22


2.1.2.2. Construction and Solvation of the Embedded Models
Two embedded simulations were made with different initial structures of the equilibrated Syt7 C2A solvated model (at t = 6 and 8 ns). Each protein was shallowly inserted into the membrane, and hence three overlapping POPC lipids were deleted. In addition, water molecules and ions (not Ca2+ ion) which were close to (< 3 A) or contacted directly with the proteins were eliminated. In both embedded models, the Ca for FI67 of CBL1 was placed above the average phosphate plane by 1.4 A while the Ca for F229 of CBF3 was below by 1.8 A on average. Importantly, however, both side chain phenyl rings were embedded in the membrane (Figure 7B). The initial tilt angles were 48 and 35 for the first and second embedded models, respectively. The starting protein conformations in the two embedded models were similar to each other except the amine group in the side chain of K184, which was near the membrane surface in the first model but oriented away from the membrane in the second model (Figure 9).
Figure 9. Two orientations of the residue K184 that observed during the embedded models simulations; when the side chain of K184 was (A) located in the bulk water and (B) interacting with the head groups of lipids. The protein is shown as cartoon in gray, with p4 strand in orange, K184 as ball-and-stick (located in the loop between p3 and p4 strands), and lipids (without H) as licorice. Color code: N, blue; O, red; C, cyan; H, white; and P, brown.) Water and ions are omitted for clarity. During the simulations of Syt7C2AWT, sometimes p4 strand was represented by a loop like in (A) due to its short sequence containing K186-F187-El 88 residues. Figure taken from Fig. S2 in Ref.21
23


2.2. Data Analysis
2.2.1. Area per Lipid (APL) and Lipid Order Parameter (Ach)
The area per lipid (APL) describes the average size taken up by a lipid leaflet area in the membrane. For the 3:1 PC/PS membrane, the APL value was obtained by dividing the average area of one leaflet in the membrane by the total number of lipids in the leaflet (A)jp = 128) as seen in Equation 1:
APL
{xtyt)
tflip
Equation 1
where xt and vt are the dimensions of the membrane in the x-v plane at the instantaneous time t. Only the reference APL values for the pure PC (APLPOpc) and PS (APLPOps) are available from previous experimental48 mid computational49,50 studies (Table 6), which were used to estimate the reference APL for the 3:1 PC/PS mixture membrane with the following expression:
APLPC/PS 0.75 APLpopc T 0.25 APLpgps Equation 2
Table 6. Comparisons of the area per lipid (APL) in A2 between the literature and simulated membranes.
____________Pure____________ 3:1 PC/PS
______________________POPC__________POPS_________________________________
Experiment 64.3 1.3a N/A N/A
Computation 65.5 1.1 b 55 1.0 c 62.9 0.9 di 62.6 0.9 e
a From Ref.48
b From Ref.49
c From Ref.50
d The estimated APL using the values from h and c.
e The calculated APL of the 3:1 PC/PS membrane using the last 120 ns simulation.
24


The order parameter (Sea) is a measure of structural orientation or flexibility of lipids in a membrane. The lipid head groups, which are charged, tend to interact with other head groups, ions, or water molecules, restricting their motions. On the other hand, the motions of the tails occur in a more fluid environment. Thus, the order parameters of different carbons in the membrane give the information about the orientation anisotropy. The value of Sea was evaluated by measuring the spatial orientation of a C-H bond, described by the angle a between the C-H vector and the membrane normal (Eq. 3).
Vh = 2 (3 1)
Equation 3
where (cos2 a) denotes the average cosine square angles between each C-H bond in a methylene group (or a methane group near c A-double bond) and the membrane norm over all lipids.
2.2.2. Poisson-Boltzmann Calculation
The Poisson-Boltzmann (PB) calculations for the protein structures were performed using
51 2b 2b
the Adaptive Poisson-Boltzmann Solver. Two sets (Ca -bound vs. Ca -free) of PB
2_|_
calculations were performed for the 3-Ca bound WT mid CH C2 domains of Syt proteins. The Ca -bound Syt structures for C2 domains were provided by utilizing the last snapshot of the protein standalone simulations, followed by the generation of the PQR (PDB data with the charge and radius parameters) files at a pH of 7.4 with CHARMM22 force field. Finally, the electrostatic potentials were described in 0.15 M of KC1.
25


CHAPTER III
RESULTS AND DISCUSSION 3.1. Simulations of the Standalone 3:1 PC/PS Membrane
The standalone simulation for 3:1 PC/PS membrane was performed in water with the counter ions of 64 Na+ ions for 160 ns in order to calculate the area per lipid (APL) mid order parameter (S'ch) for the membrane (Figure 10). Because the instantaneous APL dropped quickly in the first 10 ns and became stable afterwards, only the data from the last 120-ns of the trajectory was used to calculate the average APL and SCu- The average APL was 62.6 A2, and was similar to the APL of 62.9 A2 estimated by utilizing the literature data, taken from the NMR measurements (Table 6) 48,49''50 The Sch for each C in the palmitoyl mid oleoyl chains (Figure 2) of the membrane model were computed using the same data set as APL calculations. In Figure 10B, Sen generally fell between those for the pure POPC mid pure POPS.52,53 Note that for the oleoyl chain, there are only a few reference *Sch- Nevertheless, the calculated values still agreed with the literature values. Overall, our results of APL and *Sch agreed well with the literature values, suggesting that the simulated membrane structure was well equilibrated after 40 ns.
26


(A) 80
ST
< 75
H
ill
<
55
0 40 80 120 160
Time [ns]
3:1 PC/PS pure POPC pure POPS
(B) 0.3
palmitoyl
S. 0.3
o
Y o
oleoyl
2 4 6 8 10 12 14 16 18
Carbon Atom Index
Figure 10. (A) Area per lipid (APL) as a function of simulation time for the standalone 3:1 PC/PS membrane for 160 ns. (B) Order parameter (Sce) as a function of carbon atom index of the palmitoyl (top) and oleoyl (bottom) chains. The values from the NMR measurements (pure POPC52 in green and POPS53 in blue curves) and this study (black curves) were compared. Only the values ofScH for the palmitoyl chain in a pure POPS membrane were available. Figure modified from Fig. S3 in Ref.21
3.2. Simulations of the Standalone C2 Domains
Both C2 domains of Sytl and Syt7 have the ability to bind to a phospholipid membrane in the presence of Ca2+ ions.545556 The exact number of bound Ca2+ ions to the Ca2+-binding loops (CBLs) are unknown, but up to four Ca2+ ions can bind to the CBLs.233057 The NMR and crystal structures were used for modeling WT C2A and C2B domains for Sytl and Syt7 to study
o_|_ 09 OA 97 91 9+
those different Ca -sensitivities. 5 5 5 Experimentally, Ca -bound structures are available for each domain of Sytl (SytlC2AWT-3Ca2+ and SytlC2BWT-2Ca2+) and C2B domain of Syt7
27


(Syt7C2BWT-3Ca2+), but not for Syt7 C2A. Therefore, the two models for Ca2+-bound Syt7 C2A were created by adding two mid three Ca ions to the Ca -free Syt7 C2 A experimental structure (Syt7C2AWT-2Ca2+ mid Syt7C2AWT-3Ca2+), respectively. Furthermore, one Ca2+ ion was added in SytlC2BWT-2Ca2+ model to build the 3-Ca2+ bound Sytl C2B model (SytlC2BWT-3Ca2+), as described in Computational Methods 2.1.1.2. A series of homologous CH C2 domains (Syt7:lC2AGH-3Ca2+, Sytl:7C2BGH-2Ca2+, mid Sytl:7C2BGH-3Ca2+) was also created to provide additional information on the role of three loops in Ca2+-binding.3126
3.2.1. Comparisons of 2-Ca2+ with 3-Ca2+ Models in Each C2WT Domain
For Syt7C2AWT-2Ca2+ and -3Ca2+, and SytlC2BWT-2Ca2+ and -3Ca2+ models, the initial positions of the first two Ca ions in the CBLs were the same. However, their positions were rearranged during the minimizations, leading to the different ways of coordination with the protein oxygen atoms in the CBLs (Figure 6). In the Syt7C2A -3Ca model, the first Ca ion was coordinated by the side chains of D166, D172, D225, mid D227, the backbone oxygen atom of Y226, mid one water molecule; the second Ca ion by the side chains of D166, D225, D227, and D233, the backbone oxygen atom of K165, mid one water molecule; and the third Ca ion by the side chains of D227, S230, and D233 and the backbone oxygen atom of R231. In the Syt7C2A -2Ca model, however, the second Ca ion moved towards CBL3, so one extra water molecule entered the empty space. The overall conformation of the side chains in CBL3 was very different from that in the 3-Ca2+ model. The first Ca2+ ion was coordinated by the side chains of D166 mid D172, the backbone oxygen atoms of K165 mid Y226, mid two water
28


2 I m m m m
molecules. The second Ca ion was coordinated solely by the side chains of D225, D227, S230, and D233.
The C2B domains of Sytl with 2- and 3-Ca2+ are illustrated in Figure 6B. In the 3-Ca2+
2+
model of Sytl C2B domain, the first Ca ion was coordinated by the side chains of D303, D309,
2+
D363, mid D365, the backbone oxygen atom of Y364, and two water molecules; the second Ca
ion by the side chains of D303, D363, D365, and D371, the backbone oxygen atom of M302, and
2+
two water molecules (note that one of the water molecules here was also shared by the first Ca
2+
ion); and the third Ca ion by the side chains of D365 mid D371, the backbone oxygen atom of K369, mid three water molecules. In this model, the position of the third Ca ion is located somewhere between the binding pocket and the bulk water, as the ion interacts with less protein oxygen atoms (three) than in Sytl C2A (four) or Syt7 C2B (four). The loose binding of the third Ca ion is primarily due to the absence of a serine residue in CBL3 of Sytl C2B (S235 in Sytl C2A, S230 in Syt7 C2A, and S362 in Syt7 C2B), which plays a key role in trapping the third
Ca ion inside the CBLs in Sytl C2A and Syt7 C2AB. The importance of this serine residue
2+
will be covered in Section 3.2.2. Comparisons of WT and CH Domains. In the 2-Ca model, just like in the Syt7C2A -2Ca model, the second Ca ion moved towards CBL3 mid allowed one water molecule to come in by creating more space. Therefore, the first Ca2+ ion was surrounded by the side chains of D303, D309, D363, and D365, the backbone oxygen atom of Y364, mid two water molecules, and the second Ca ion by the side chains of D303, D363,
D365, mid D371, the backbone oxygen atom of M302, and two water molecules. Unlike the Syt7C2AWT-2Ca2+ model, here the solvation shells of the first and second Ca2+ ions do not share the same water molecules.
29


After minimization, a 200-ns trajectory for each of the three C2A domains (SytlC2AWT-3Ca2+, Syt7C2AWT-2Ca2+, mid Syt7C2AWT-3Ca2+) mid a 100-ns trajectory for each of the three C2B domains (SytlC2BWT-2Ca2+, SytlC2BWT-3Ca2+, and Syt7C2BWT-3Ca2+) were generated. The overall conformation of the protein was quite stable, as can be seen from the consistently small root-mean-square deviation (RMSD) with respect to the starting geometry of the protein backbone N, C, O, and Ca as a function of the simulation time (Figure 11A). As listed in Table 7, the average RMSD of the backbone atoms are 1.1 A for SytlC2AWT-3Ca2+, 2.6 A for Syt7C2AWT-2Ca2+, mid 1.8 A for Syt7C2AWT-3Ca2+. The smallest average RMSD value and the fluctuations (0.1) were found in SytlC2AWT-3Ca2+ model, implying that the Sytl C2A model which started from the Ca -bound NMR structure was the most stable during the simulation time. The largest RMSD fluctuations occurred in the 2-Ca model of Syt7 C2A, in which two Ca2+ ions in the CBLs moved substantially compared to the equivalent Ca2+ ions in the 3-Ca2+ model of Syt7 C2A (the respective RMSD values are 3.2 and 1.8 A for the two Ca2+ ions in the 2-Ca2+ model, and 1.0 and 0.5 A in the 3-Ca2+ model). Recall that the NMR structure for Syt7 C2A was Ca2+-free. The addition of multiple Ca2+ ions into the Ca2+-free Syt7 C2A models required the adaptations of the protein, leading to the greater protein movements than observed in other models. In the case of SytlC2B -3Ca model, only the third Ca ion from Syt7C2BWT-3Ca2+ was added to SytlC2BWT-2Ca2+, likely requiring less protein conformational changes. For the C2B domains, the average RMSD of the backbone atoms and the fluctuations over simulation time are substantially smaller (1.4 A for SytlC2BWT-2Ca2+, 1.1 A for SytlC2BWT-3Ca2+, and 0.9 A for Syt7C2BWT-3Ca2+) than for the C2A domains. The RMSD value for the 2-Ca2+ model of SytlC2BWT was larger than for the 3-Ca2+ model, but the difference is minor. Interestingly, the first and second Ca ions in the CBLs of the 2-Ca model
30


2_|_
exhibited smaller movements compared to those of the 3-Ca model (RMSD values for the two Ca2+ ions in the 2-Ca2+ model are 0.4 and 0.5 A, and RMSD values in the 3-Ca2+ model are 0.9 and 0.6 A), in contrast to the 2- and 3-Ca2+ models of Syt7C2AWT. The Ca2+ ions different stabilities in the Syt7C2A and SytlC2B were found in agreement with experiments.
(A) Standalone proteins
Syt7:1CH C2A 3Ca2+
Syt1WT 2Ca2+ Syt7WT 2Ca2+ Syt1:7CH C2B 2Ca2+
- Syt1WT 3Ca2+ - Syt7WT 3Ca2+ - Syt1:7CH C2B 3Ca2+
Q
cn
QC
4
3 2 1 0
0 50 100 150 200
4
3 2 1 0
0 50 100 150 200
C2B domain
JLi, uaJU
tv*#*****'-
(B) Protein-membrane
Syt7C2A Pre-insertion Lying-down Syt7C2A 1st Embedded Standing-up Syt7C2A 2nd Embedded
Figure 11. Root mean square deviation (RMSD) of protein backbone atoms (N, C, O, and Ca) with respect to the first frame of the simulation, as a function of simulation time for all (A) standalone proteins (top, C2A; bottom, C2B) and (B) Syt7 C2A~membrane complexes models. (A) The traces are shown for both C2 domains of SytlWT~3Ca2+ in red and Syt7WT-3Ca2+ in navy. For 2-Ca2+ models for Syt7C2AWT and SytlC2BWT, the traces are represented by light blue and pink, respectively. For the chimeric (CH) domains, Syt7:lC2ACH is shown as green, Sytl:7C2BGH~2Ca2+ as gray, and Sytl:7C2BGH~3Ca2+ as black. The simulation times are different per domain; C2AWT for 200 ns, and C2BWT and all CH for 100 ns. (B)The traces are shown for two Syt7C2A pre-insertion models; lying-down in red and standing-up in yellow, and for two Syt7C2A embedded models; 1st in green and 2nd in navy for 500-ns each simulation.
31


Table 7. Root mean square deviations (RMSD) in A, averaged over the trajectories for selected atoms in the simulations of standalone Sytl mid Syt7 C2A domains. The maximum RMSD values are given in parentheses. a
Backbone b First Ca2+ Second Ca2+ Third Ca2+
Standalone C2Aw 1
Sytl-3Ca2+ 10-ns c 1.0 0.1 (1.4) 0.6 0.2 (1.1) 0.6 0.2 (1.1) 3.8 1.2 (6.1)
200-ns d 1.1 0.1 (1.5) 0.5 0.2 (1.1) 0.5 0.2 (1.3) 4.1 0.3 (6.0)
Syt7-2Ca2+ 10-ns c 1.9 0.5 (2.8) 1.7 0.5 (3.0) 1.2 0.4 (2.4) N/A
200-ns d 2.6 0.2 (3.5) 3.2 0.6 (5.1) 1.8 0.5 (3.6) N/A
Syt7-3Ca2+ 10-ns c 1.6 0.4 (2.3) 0.8 0.2 (1.4) 0.7 0.3 (1.6) 1.6 0.8 (3.6)
200-ns d 1.8 0.1 (2.4) 1.0 0.2 (1.8) 0.5 0.2 (1.2) 2.4 0.4 (4.0)
Standalone C2BWT
Sytl-2Ca2+ 100-ns e 1.4 0.1 (1.8) 0.4 0.2 (1.2) 0.5 0.2 (1.3) N/A
Sytl-3Ca2+ 100-ns e 1.1 0.1 (1.4) 0.9 0.2 (1.7) 0.6 0.2 (1.4) 1.6 0.5 (3.2)
Syt7-3Ca2+ 100-ns e 0.9 0.1 (1.2) 0.5 0.2 (1.0) 0.5 0.2 (1.1) 0.5 0.2 (1.4)
Standalone CH
Syt7:l C2A-3Ca2+ 100-ns e 1.9 0.1 (2.4) 1.1 0.2 (1.7) 1.4 0.2 (2.1) 3.6 0.6 (5.4)
Sytl:7 C2B-2Ca2+ 100-ns e 0.9 0.1 (1.2) 0.6 0.2 (1.3) 0.6 0.2 (1.5) N/A
Sytl:7 C2B-3Ca2+ 100-ns e 1.2 0.1 (1.5) 0.5 0.2 (1.3) 0.6 0.2 (1.3) 2.8 0.3 (3.9)
Syt7C2AWT-membrane
Pre-insertion
Lying-down 500-ns * 1.5 0.1 (2.2) 0.4 0.2 (1.2) 0.4 0.2 (1.4) 2.0 0.4 (4.7)
Stand-up 500-ns f 1.4 0.1 (2.0) 0.4 0.2 (1.0) 0.4 0.2 (1.1) 2.4 0.3 (3.8)
Embed
First 500-ns * 1.1 0.1 (1.6) 0.5 0.2 (1.4) 0.5 0.2 (1.3) 1.1 0.5 (2.8)
Second 500-ns f 1.2 0.2 (1.9) 0.5 0.2 (1.3) 0.7 0.2 (1.6) 1.1 0.3 (2.3)
a Mean S.D. calculated with the initial structure as reference and averaged over the trajectory of 10-ns and 200-ns simulations for the protein standalone models, and 500-ns simulations for the protein-membrane complex models.
b Protein backbone atoms were defined as N, O, C, and Ca.
c Calculated using the last 6 ns of trajectories of 10-ns standalone SytC2AWT simulations.
d Calculated using the last 80 ns of trajectories of 200-ns standalone SytC2AWT simulations.
e Calculated using the last 20 ns of trajectories of 100-ns standalone SytC2BWT and chimeric C2 domain simulations.
^Calculated using the last 400 ns of trajectories of 500-ns standalone Syt7C2A-membrane simulations.
32


The coordination of three Ca2+ ions in the CBLs per C2WT domain closely resembles each
2+
other, as depicted by the distances between the Ca ions and their coordinating oxygen atoms from the domains as a function of simulation time (Figure 12). The first two Ca ions in both C2Awt models were relatively stable in the binding site, maintaining approximately constant distances with the chelating protein oxygen atoms most of the time. Substantial conformational change only occurred early in the simulations. For example, at t = 3 ns, the distance of the first and second Ca2+ ions with the carboxyl oxygen atom of the D227 in Syt7C2AWT (the red curves with markers in Figure 12A) quickly dropped from 6 to 4 A and from 4 to 2 A, respectively. As the oxygen atoms in D227 side chain, which was initially located in the middle of the second mid third Ca ions (Figure 6A bottom panel), went closer to the first and second Ca ions, one water molecule diffused in mid took the oxygens position. The addition of this water molecule initiated the conformational changes of S230 and R231, resulting in the distances of the third Ca ion with the hydroxyl oxygen atom in S230, and with carbonyl oxygen atom in R231 increase from 2 to 6 A and from 2 to 5 A, respectively, at t = 6 ns. Subsequently, up to 5 water
molecules moved near the third Ca ion during the simulation. The average distances between
2+
three Ca ions and their coordinating protein oxygen atoms are provided in Table 4.
These results indicate that the first two Ca2+ ions in the Sytl and Syt7 C2AWT models are strongly enfolded in the binding site, while the third Ca ion binds less tightly. This interaction occurs primarily through interactions with the side chain oxygen atoms of the D238 in Sytl and the D233 in Syt7 and secondarily with the backbone oxygen atoms of the K236 in Sytl and the R231 in Syt7. The structural and functional similarities between two C2AWT domains of Sytl and Syt7 propose that Syt7C2AWT binds three Ca2+ ions in a similar way to that of SytlC2AWT as suggested by a previous NMR measurement (Figure 13 top panel). The RMSD plot for the
33


third Ca2+ in Syt7C2AWT as a function of the 200-ns simulation time shows the stable binding in the CBLs. Therefore, the 3-Ca2+ model for Syt7C2AWT was used for the protein-membrane association simulation in this study.
Next, for the C2BWT models, the first two Ca2+ ions were substantially held by the coordinating oxygen atoms, as evidenced by the relatively consistent distances between the atoms (Figure 12B). The third Ca2+ ion in SytlC2BWT model, however, behaved very differently to that in Syt7C2BWT model. In Syt7C2BWT, the side chain oxygen atom of the S362 clung tightly to the third Ca ion, in collaboration with the side chain oxygen atoms in the D359 and D365, mid the backbone oxygen atom in the R363. As mentioned earlier, this serine is not found
WT 2t WT
in SytlC2B The third Ca^ ion in SytlC2BWi was surrounded by five water molecules from the beginning of the equilibration (three water molecules after minimization, Figure 6B middle panel). For comparison, the third Ca2+ ion in Syt7C2BWT was surrounded by two water molecules (one water molecule after the minimization, Figure 6B bottom panel). Both the previous experiment26 and this computation (Figure 13 bottom panel) suggest that SytlC2BWT exhibits a low affinity for the third Ca2+ ion.
34


(A) C2A domains
<
u
c
rs
w
Q
Syt1WT Syt7WT SytCH
12 12
8 8
1 4
0 0
0 50 100 150 200 0 50 100 150 200 0 50 100 150 200
12 12
8 8
4 il am Jim
R"IP|1
0 0
0 50 100 150 200 0 50 100 150 200 0 50 100 150 200
12
8
4
0
50 100 150 200 0
50 100 150 200
(B) C'2B domains
Time [ns]
Figure 12. Distances between each Ca2+ ion and its coordinating protein oxygen atoms, as a function of time of standalone (A) C2A and (B) C2B domains with three Ca2+ ions. In each column (left, Sytl' ; center, Syt7' ; right, chimeric domains), the top, middle, and bottom panels are for the first, second, and third Ca2+ ions, respectively. The consistent color codes are used for the equivalent residue atoms in the order of SytlC2A-Syt7C2A-Syt7:lCHC2A-SytlC2B-Syt7C2B-Sytl:7chC2B domains: L171-K165-L165-M302-M296-M302 O, black; D172-D166-D166-D303-D297-D303 OD1, red; D172-D166-D166-D303-D297-D303 OD2, green; D178-D172-D172-D309-D303-D309 OD2, gray; D230-D225-D225-D363-D357-D363 OD1, blue; D230-D225-D225-D363-D357-D363 OD2, purple; F231-Y226-F226-Y364-K358-K364 O, black with markers; D232-D227-D227-D365-D359-D365 OD1, red with markers; D232-D227-D227-D365-D359-D365 OD2, green with markers; S235-S230-S230-N/A-S362-S368 OG, gray with markers; K236-R231-K231-K369-R363-R369 O, brown with markers; D238-D233-D233-D371-D365-D371 OD1, blue with markers; D238-D233-D233-D371-D365-D371 OD2, purple with markers.
35


+
CM
CO
O
T3
i
CO
a
co
O'
8
C2B domain Syt1WT

- Nl 1 ibni II E Syt7WT
iil'j lx lUi JliiiiJiukl, UkUuj.iti., Syt7:1CH (top)
iuV'^nW ih..'-Mj r^jpi Syt1:7CH (bottom)
0 50 100 150 200
Time [ns]
0_L m m
Figure 13. Root mean square deviation (RMSD) of the third Ca ion with respect to the first frame of the simulation, as a function of simulation time for all 3-Ca2+ models of WT and CH C2A domains for 200-ns (top) and C2B domain for 100-ns simulations (bottom). The traces are
v / WT WT v y . .
shown for both C2 domains of Sytl in red and Syt7 in green, and chimeric domains in yellow.
Each C2A and C2B domain in Sytl and Syt7 displays generally similar electrostatic
2+
potentials from Poisson-Boltzmann calculations (Figure 14). The Ca binding sites in all domains have negative potentials in the Ca -free structures due to exposed aspartate residues and backbone oxygen atoms in the CBLs (Figure 6). However, this region has a positive potential after three Ca ions are bound on the membrane binding. Moreover, the region from p3-loop-p4 to CBL2 has a positive potential in both of the Ca -free and Ca -bound forms due to a cluster of lysine and arginine residues. The net charge of this region is rather positive (Table 2). The positive potentials of the CBLs in the presence of Ca ions and the p3-4 strands with the
36


adjacent loops promote the anionic membrane-protein association through electrostatic
+ +- 15,23, 58
interactions.
Subtle differences still existed between domains. For the two C2AWT domains, the overall electrostatic potential of Sytl C2A is more negative than that of Syt7 C2A. This is consistent with the protein net charges reported in Table 2: -2 for Sytl and +3 for Syt7 C2AWT. Particularly, the CBL1 region in Sytl has a net charge of -1, while in Syt7 it is charge neutral. This leads to the more negative potentials in CBLs of SytlC2AWT (black arrows in Figure 14). Similarly, Sytl C2B is less positive than Syt7 C2B: the total charges are +6 for SytlC2BWT mid +12 for Syt7C2BWT. However, an a-helix in the C-terminus of SytlC2BWT (but not in Syt7C2BWT) has a net charge of -4, making its entire C2B domain less positive (green arrows in Figure 14). If only from pi to p8 (with interconnecting loops) in those two C2BWT domains are compared, the net charges will be +11 for SytlC2BWT mid +12 for of Syt7C2BWT. Overall, Syt7 C2ABWT domains have more positive potentials than Sytl C2ABWT domains, implying that Syt7 possibly possesses greater electrostatic interactions with anionic lipid membranes in the presence of Ca2+ ions.56,59
37


C2A
C2B
WTSvtl
Side view
Bottom view
Ca2+-bound Ca2+-free
Ca2+-bound Ca2+-free
WT Syt7
Side view
Bottom view
Ca2+-bound Ca2+-free

Ca2+-bound Ca2+-free
CH
Syt7:l &
Side view
Bottom view
Ca2+-bound Ca2+-free
Ca2+-free
Figure 14. Electrostatic potential isosurfaces for Syt C2AB and Slp4 C2A domains. The proteins are shown as a cartoon with [34 strands colored orange, Ca2+ ions as yellow spheres. The potential maps computed with and without the three Ca2+ ions for Syt proteins. Blue for +50 mV and red for -50 mV equipotential contours (assuming 0.15 M KC1 and pH 7.4).
38


3.2.2. Comparisons of WT and CH Domains
One C2A and two C2B chimeric (CH) models were created in 0.15 M KC1 solution. The C2Ach model retained the Syt7 body with Sytl CBLs, while the C2Bch models retained the Sytl body with Syt7 CBLs specifically for the purpose of corresponding with the collaborative experiments. For C2A model, both Sytl mid Syt7 C2A domains require to have three Ca in CBLs, so only a 3-Ca model was built. However, Sytl and Syt7 have two and three Ca ions
2+ CH
in the binding sites of C2B domains. Thus 2- and 3-Ca models were generated for C2B models.
Figure 11A illustrates the overall protein stabilities of the CH domains. The average RMSD values are all smaller than 2 A: 1.9 A for Syt7: lC2AGH-3Ca2+, 0.9 A for Sytl:7C2BCH-2Ca2+, and 1.2 A for Sytl:7C2BGH-3Ca2+ (Table 7). The ~0.6 A of the average RMSD for both the first and second Ca ions in the two Sytl :7C2B models indicates that the major difference was observed for the third Ca ions in Sytl:7C2B -3Ca Furthermore, the third Ca ions in Syt7: lC2AGH-3Ca2+ mid Sytl :7C2BGH-3Ca2+ models have even greater RMSD values than their first two Ca2+ ions (respective RMSD values are 3.6 and 2.8 A for the third Ca2+ ions). The fluctuations of the RMSD for the third Ca ion in C2A however, is doubled in magnitude of 0.6 compared to the fluctuations for the third Ca2+ ion in Sytl :7C2BGH-3Ca2+ (0.3), implying that the third Ca ion binding in C2B is more stable than C2A In Figure 13 top panel, the first
CH
big jump in the C2 A RMSD plot appears at t = 6 ns for the oxygen atoms of S230 side chain
2+
and K231 backbone. Those oxygen atoms moved away from the third Ca ion as two water molecules diffused into the site rapidly (the distances between the third Ca ion and the oxygen atoms doubled from 2 to 4 A). At t = 78 ns, the second jump occurred when one water molecule
39


9+
approached the third Ca ion from bulk, causing the side chain of D227 to slightly move away from the third Ca ion towards the first and second Ca ions. This interaction has been observed in C2Awt as mentioned earlier. The two carboxyl oxygen atoms in D233 closely interacted with the third Ca ion but one of the oxygen atoms began to interact with the second Ca ion after the last 5 ns of the simulation. The overall conformation of CBL3 is somewhat similar to the two 3-Ca2+ C2Awt models. But longer simulations are required to compare the C2A domains between the WT and CH proteins to reduce the statistical uncertainties.
Overall the RMSD for the third Ca ion in the C2B model resembles that for SytlC2BWT, but the ions solvation shell is different in the CH and WT models. In the first snapshot of the C2B equilibration, the third Ca ion was surrounded by the side chain oxygen atoms in D365, S368, and D371, the backbone oxygen atom in R369, mid three water molecules. An early RMSD change occurred at t = 9 ns when the aspartate residue (D365 in Sytl:7C2B ) moved towards the first and second Ca ions and two water molecules diffused into fill the evacuated space. Subsequently, the side chain oxygen atom in S368 moved away from the third
Ca ion, while the backbone oxygen atom in R369 came nearer to the third Ca ion. As the
2+
distances between the third Ca ion and D365 and S368 became longer, more water molecules
entered the binding site, mid the third Ca ion was coordinated by a total of 6 water molecules in
20 ns. After that, two of these water molecules were replaced by Cf ions at t = 20 and 40 ns.
This Cl recruitment was only detected in C2B model, not in C2 A mid we found that the
third Ca2+ ion in SytlC2BWT model also recruited one Cf ion during the simulation. The details
2+
will be discussed more in Section 3.4. Coordination of the Ca Ions in the Standalone Syt Proteins and Syt7C2AWT-membrane Complexes.
40


The side chain of S368 in Sytl:7C2BCH (G368 was in SytlC2BWT instead), showed two
orientations in the simulations (Figure 15). Rotations of the side chain in S368 caused the distances between the third Ca ion and its hydroxyl oxygen atom (OG) to vary. The distances tended to be longer when the hydroxyl group pointed inward from the CBL region, characterized by a gauche(+) conformation with 60 of the %i angle in S368 (Figure 16). As the side chain exhibits the gauche{+) conformation, one more water molecule diffused in between the hydroxyl
0 -I- m m m
oxygen atom and the third Ca ion. But when the serine residue had the tram conformation (180 of the %i angle), the oxygen atom helped to hold the third Ca ion, preventing its escaping from the binding site into the bulk water. These changes impact the RMSD of the third Ca ion (Figure 13 bottom panel).
Figure 15. Two orientations of the residue S368 that observed during the standalone SytlC2BGH-3Ca2+ model when the side chain of S368 faced (A) inward and (B) outward the CBLs (green circles). The orientations of S368 represent the gauche{+) conformation in (A), see Figure 16. The three Ca ions and were in green spheres with surrounding residues in CBL3 (sticks: N, blue; O, red; C, cyan), and the coordinating oxygen atoms (red sphere) and Cl- ions (blue sphere) for the third Ca ions are shown. Protein H is omitted for clarity. The number under the dotted line in (A) and (B) indicates that the distance in A between the third Ca2+ ion and the hydroxyl oxygen atom in the serine residue.
41


0 50 100
Time [ns]
Figure 16. (A) Distances between the side chain oxygen atom (OG) of S368 and the third Ca2+ ion (gray curves with markers) and the backbone nitrogen atom (N) of S368 (blue curves) in Sytl:7C2BCH model. (B) The dihedral xi angles (N-Ca-Cp-OG) for the same S368. The serine residue had mostly trans conformations but had gauche(+) at ~50 ns when the angles of ~60.
3.3. Simulations of Syt7 C2A Protein-membrane Association
The membrane-targeting mechanisms of C2A domains in Syt7 require detection of multiple signals such as Ca2+ ions and target lipids (usually anionic phospholipids).475 2560 The MD simulations of 500 ns were performed for each of the four Syt7C2Awr-membrane complexes (two pre-insertion models and two embedded models). The initial structure of Syt7C2AWT in the pre-insertion standing-up model was used to build the embedded models.
42


3.3.1. Comparisons of the Pre-insertion with the Embedded Svt7C2Axvlmcm branc
In order to characterize the orientation of the membrane-associated Syt7C2AWT, the protein tilt angles and the penetration depths of FI67 mid F229 were calculated as a function of simulation time (Table 8). The tilt angle is defined to be the angle between the membrane plane and the norm of a protein plane containing the three centers-of-mass (COM) of the three outermost p strands (|35, [36, and |38) (Figure 8A). The average tilt angles were 13.5 mid 37.1 for the pre-insertion lying-down mid standing-up models, and 32.8 mid 32.2 for two embedded models, respectively. There were no significant angle differences between each model except the lying-down model. These angles indicated that the lying-down and standing-up proteins were relatively more parallel and perpendicular to the membrane surface, respectively. Strikingly, as the Syt7C2AWT domain started to lay down, it tilted towards the side that contains the [33-4 strands, which are associated with positive electrostatic potentials (see Section 3.2.1 Comparisons of 2-Ca2+ with 3-Ca2+ Models in Each C2WT Domain). We found strong interactions of this polybasic region (formed by the cluster of the lysine residues) with the lipid head group in all simulations. Interestingly, this polybasic region is identified to be essential for the strong Ca -dependent binding of Sytl to PS-lipid containing membrane. The protein orientation was generally inferred from the tilt angle calculation, while the penetration depth calculation was used to describe which residues have significant interactions with membrane.
The depths of the critical residues are defined by the distances between their Ca atom and the average phosphate (PO4) plane of the membrane, as exemplified in Figure 8B for F167 in CBL1 and F229 in CBL3. A positive penetration depth indicates the Ca atom of the residue is present on the lipid face of the average PO4 plane. It is evident that F167 often passed in and out of the membrane interface. This type of motion of FI 67 was more common in the pre-insertion
43


standing-up mid two embedded models (Figure 17A). In the lying-down model, FI67 remained in the aqueous phase all the time following the initial position (initial depth: -15.9 A in the bulk water). For F229, the average penetrated depths are similar in all models: 0.2 A for the lying-down and 2.2 A for the standing-up pre-insertion models and 2.8 A for the first and 2.4 A for the second embedded models, respectively, suggesting that F229 in CBL3 drives the penetration process prior to F167 in CBL1. In Figure 18, the penetration depths for the pre-insertion models of individual residues at various simulation times were plotted for the two pre-insertion models. The overall trend of the protein as a whole approaching the average PO4 was evident when comparing the traces of the first 50-ns (red curves) with those of the last 50-ns (purple curves). However, the actual penetration process is far more complicated. For example, in the lying-down model, the protein moved back to the aqueous phase after 300 ns, while CBL1 and the polybasic region of [33-4 strands moved near the membrane. During the last 100 ns, CBL1 constantly returned to the bulk water while the polybasic region was stuck by the interaction with the lipid head groups. For the standing-up model, the depths for CBL1 (particularly for FI67) mid the polybasic region were similar at 250 ns. After that, CBL1 started to anchor deeper to the membrane surface, while the polybasic region moved back towards the bulk water. In contrast, the regions of CBL3 in both pre-insertion models exhibited steep increase in depth (approaching and binding to membrane) over time (Figure 18).
44


Table 8. The averaged (final 400 ns from 500-ns simulations) tilt angles mid the penetration depths with respect to the membrane surface. The initial values are given in parentheses. The selected atoms are positioned below the average PO4 of the membrane when the depth values are positive (depth of avg. PO4 = 0).
Tilt angle () _____________Depth (A)______
F167 F229
Pre-insertion
Lying-down 13.510.0 (23.1) Ca -9.82.5 (-15.9) 0.22.2 (-6.9)
phenyl* -7.82.7 (-14.1) 3.12.3 (-4.1)
Standing-up 37.19.5 (36.1) Caa -3.13.3 (-7.1) 2.22.0 (-4.2)
phenyl* -1.03.9 (-4.9) 4.92.0 (-3.9)
Embedded
1st model 32.814.4 (48.2) Caa -2.03.0 (-1.3) 2.81.6 (2.3)
2nd model phenyl* 0.93.1 (1.8) 4.81.7 (5.4)
32.29.1 (34.7) Caa -2.62.7 (-1.4) 2.41.8 (1.3)
phenyl* 0.23.2 (1.4) 4.31.9 (4.9)
a The penetration depths were calculated using the Ca position.
* The penetration depths were calculated using the center-of-mass of the side chain phenyl ring.
45


100 200 300 400 500
Q_ -40
0) on
1st Embedded
-20
-40 -30 --20 -10 0 -10
100 200 300 400 500
Pre-insertion: Standing-up
100 200 300 400 500
2nd Embedded
100 200 300 400 500 0
Time [ns]
100 200 300 400 500
Figure 17. The tilt angles (black) and penetration depths of the protein center of mass (gray) and the critical residues (FI67 in red and F229 in blue) for different Syt7C2AWT-membrane models over the 500-ns simulation time. Note that the penetration depth of 0 corresponds to the position of the average P04 plane. Figure taken from Fig. 7 in Ref.21
46


Figure 18. Penetration depths for the lying-down (top) and standing-up (bottom) Syt7C2Awr pre-insertion models, as a function of the protein sequence. Each of 10 traces per plot represents on the average position of the residue for 50 ns increments of the 500 ns trajectory. The rainbow-colored order indicates the simulation time (red for the first 50-ns to purple for the last 50-ns). Three pink regions on the graph correspond to the regions of the CBLs (left to right, CBL1 to 3) and the blue region for the polybasic region of the [34 strand.
47


Interestingly, all simulations depicted large swinging motions of the protein between a more parallel (smaller tilt angle) and a more perpendicular (greater tilt angle) orientation (Figure 17). For two pre-insertion models, even with large swinging motion occurred (Aangle > +40 at t = 170 ns for the lying-down mid t = 70 ns for the standing-up models), the stable binding of F229 maintained. For the two embedded models, the large swinging motions (Aangle > +50) appeared at t = 70 ns for the 1st embedded mid t = 140 ns for the 2nd embedded models but again, F229 stayed inside the membrane. The overall stability of F167 binding to the membrane was different from the 1st to 2nd embedded models. It is likely that the insertion is only partially completed over the 500 ns simulation time.
Tilt angles were correlated with F167 penetration depths (Figure 19). It is less predominant for the tilt angle and the depth of F229, because of the stability of F229 in membrane. Visual inspection of the trajectories revealed that, as the depth of F167 became more positive after 200 ns, the tilt angle increased. The protein orientation in the pre-insertion standing-up model began to resemble the two embedded models. When the depths of F167 approached 0 A with approximately 45 of the tilt angles, the correlation manifested in all but the lying-down model. The polybasic region of the proteins experienced strong electrostatic attractions to the negatively charged head groups in PS lipids. Thus, CBL1 (represented by F167) moved away from the membrane, leading to the more parallel protein orientation. On the other hand, when the protein tilt angle increased, the interactions between F167 in CBL1 and the hydrophobic environment of the membrane also increased. The competition between the electrostatic attractions (the polybasic region near p4 strand with the lipid head groups) and the hydrophobic interactions (FI67 in CBL1 and F229 in CBL3 with the lipid acyl tails) led to seesaw-like movements of Syt7C2AWT domain with F229 as the fulcrum.
48


This type of seesaw-like movement of Syt7C2AWT domain was also observed in the two embedded model simulations. We counted the interacting pair of the heavy atoms in protein mid lipids as a simple way to measure the electrostatic attractions (by positively charged residues) and the hydrophobic interactions (by FI67 mid F229). When the distance was shorter than 5.0 A, the pair was counted (Table 9). The residues in the CBLs (mostly in CBL1 and CBL3) produced most of the lipid contacts, in line with the depth calculations. Two residues at the flexible C-terminal loop, S264 and G265, are treated as exceptions because they belong to the connecting loop between C2A and C2B domains. Both F167 and F229 were surrounded by the hydrophilic residues (D166 and S168 for F167; R228, S230, and R231 for F229). The sums of the lipid head group and acyl chain contacts were 66.8 and 64.0 for F167 and F229, respectively. Among these contacts, about 30 % and 70 % are from the acyl chain carbon atoms for F167 and F229, respectively. More specifically, F229 had the lipid acyl chain contacts of 45.5, which is the greatest among all residues, because the aforementioned residue had the deepest penetration depth (2.6 A). The second greatest 19.9 is found for for FI67. Unlike F229, F167 stayed above the membrane on average, but the big standard deviation (2.0 A) of its penetration depth. This is because FI 67 frequently traveled into mid out of the acyl chain region due to the seesaw-like movements described above.
49


Table 9. The protein residues with the largest numbers of lipid contact, averaged from the two
embedded model simulations. a
Residue Lipid contacts Head group Acyl chain Depth (A)
FI 67 46.9 19.9 -2.3 2.0
R231 39.3 7.4 -1.8 1.3
G265 32.3 0.1 -4.9 1.5
R228 28.1 3.8 -0.3 1.2
S230 20.5 8.6 1.0 1.2
S264 18.8 0.0 -6.5 1.8
F229 18.5 45.5 2.6 1.2
S168 17.8 1.4 -3.5 2.1
K194 14.8 0.1 -8.4 1.6
D166 12.7 0.0 -4.5 1.6
a Numbers of contacts are averaged over the last 400 ns of the simulations of the first and second embedded models. A cutoff of 5.0 A was used in counting the heavy atom contacting pairs; i.e., if the distance between a heavy atom of the protein mid a heavy atom of the lipids was less than 5.0 A, this pair was counted. Entries for residues that average more than three interactions with acyl chains are shown in bold. The penetration depths are measured from the average phosphate plane to the Ca position for a given residue; a positive value indicates that the Ca atom is located below (deeper than) the phosphate plane.
50


F167
F229
Q.
D
0 30 60 90 0 30 60 90
-30 JVPre-insertion: Standing-up ti&p R2 = 0.60 -30 Pre-insertion: Standing-up
-20 -20 R2=0-28
-10 -10
0 0
10 i i 10
0 30 60 90 0 30 60 90
-30 -30
2nd Embedded 2nd Embedded
-20 R2 = 0.47 -20 R2 = 0.00
-10 -10
0 0
10 1 1 10 1 1
0 30 60 90 0 30 60 90
Angle [Deg] Angle [Deg]
Figure 19. The correlation between the penetration depths of F167 (red) and F229 (blue) and tilt angles of Syt7C2AWT associated with the membrane, for each simulation of the indicated protein-membrane complex models. Figure taken from Fig. 8 in Ref.21
51


3.3.2. Comparisons of Simulation with Experiment
Penetration depths for critical residues of Syt7 C2A from the embedded protein-membrane complex simulations were linearly correlated (R2 = 0.8) with electron paramagnetic resonance (EPR) depth parameter measurements34 (Figure 20). This suggests that our docking results are consistent with the experiments, although the experimental depths are on average deeper by roughly 4 A. Note that our calculations were taken from the protein-membrane complex models, while the experimental data were derived from EPR spin-labeling depth measurements employing the NMR structure31 (which is the Ca2+-free form) and a solvated protein structure from our standalone protein simulations. Our calculated insertion depths of CBL1 and CBL3 are also consistent with available computational observation through MD simulations using the highly mobile membrane mimetic (HMMM) model, although the HMMM allows slightly deeper penetration into the membrane relative to all-atomistic membrane model.14
Experimental Depth [A]
O'

Figure 20. Comparisons between experimental and computational membrane penetration depths of selected residues. The computational depths of Ca atoms were averaged over the last 400 ns In the left panel, the i?2 value of the regression line is 0.8. The horizontal error bars indicate experimental statistical errors, and vertical error bars indicate computational statistical standard deviations. The right panel shows the side-by-side comparisons of penetration depths for each residue (red, experimental; blue, computational). Figure modified from Fig. 10 in Ref.21
52


3.4. Coordination of the Ca2+ Ions in the Standalone Syt Proteins and Svt7C2Axvl-
Membrane Complexes
9+
The presence of the Ca ions is essential for the Syt-protein membrane association process. The results in this work suggest three Ca2+ ions bind to both C2AWT domains of Sytl and Syt7, two to SytlC2BWT, and three to Syt7C2BWT, in good agreement with experiments.26,30 To further probe the roles of the Ca ions in the protein binding to the anionic lipid-containing membrane, the integrated coordination number (ICN) was computed with the radial pair distribution function tool in VMD41 (Table 10). The ICN counts the average number of electronegative atoms (protein oxygen, lipid oxygen, or solvent atoms; water oxygen and/or Cf
ion) within the first solvation shell of Ca2+ ion (r < 3.25 A from the ion), providing important
2+
information about the Ca ion coordination.
The literature value for the Ca2+ ion ICN ranges from 6-8 _61>62 63 This agrees well with the ICN across all standalone 3-Ca protein models, which the ICN of 7.9 (6.9 from protein oxygen and 1.0 from water oxygen atoms) mid 7.0 (6.0 from protein oxygen mid 1.0 from water oxygen atoms) for the first and second Ca ions, respectively. A minor variation was found for the second Ca ion in Syt7:1C2A model, which has 5.0 of protein oxygen and 2.0 of water oxygen atoms.
The third Ca ions had distinct ICNs from domain to domain, and from WT to CH. For the solvated C2A domain models, the three domains (SytlC2AWT, Syt7C2AWT, and Syt7:lC2A ) showed different ICN: 7.0, 7.1, and 7.3, respectively, where the water oxygen atoms contributed more to the coordination shell than the protein oxygen atoms. For the solvated C2B domain models, the ICN for Syt7C2BWT was similar to those for the C2A domains (7.0),
53


and the contributions are mostly from the protein oxygen atoms, as expected from the tight
WT CH
binding (Figure 12B). The SytlC2B and Sytl:7C2B each carry one protein oxygen atom, three or four water oxygen atoms, and one or two Cl ions. Therefore, the Ca ion coordination in the CH C2B domain seemed to resemble SytlC2BWT to a great extent, even though the CBLs in C2B came from Syt7C2B This striking findings were in agreement with measurements
35
by Bendahmane et al.
Both SytlC2BWT and Sytl:7C2BCH recruited Cf ions to complete the coordinate shell of the outermost Ca ion (Figure 21). If anion (here Cl ) recruitment is analogous to the anionic lipid membrane binding, these results suggest that the existence of the third Ca ion in the CH C2B model will lead to a larger affinity for the membrane than that in SytlC2BWT.
In the presence of the anionic lipid membrane (3:1 PC/PS), the third Ca ions in the C2Awt domains had different interactions with the coordinating oxygen atoms based upon the initial positions of the proteins. The ICNs were 7.1, 7.0, mid 7.3 for the third Ca ions in the pre-insertion lying-down model, and pre-insertion standing-up model, and the (average) embedded model, respectively. The embedded model had a similar ICN for the third Ca ion to the standalone models. However, the solvation shell composition changed. In the standalone models, 4.0 water molecules participated in the Ca coordination, but as the proteins approached the membrane surface, one lipid oxygen atom replaced the water. In the pre-insertion standing-up model, 3.4 water molecules mid 0.5 lipid oxygen atom were present in the third Ca ions solvation shell, while in the average embedded model, 2.5 water molecules and 1.3 lipid oxygen atoms in the solvation shell. Noticeably, the contribution of the lipid oxygen atoms was larger when the protein associated with membrane. The change in the coordination shell composition of
54


2+ . . 9+
the third Ca ions suggests that the third Ca ion may not only prepare the protein for membrane association by modifying the CBLs electrostatic potential, but also directly participate in C2A-lipid binding. Based on the similarities of these results for the solvated C2A and C2B domains, such interactions between the outermost Ca ion mid membrane lipids may also exist in Sytl C2A and C2B domains.57,60
55


Table 10 Integrated coordination numbers for the first solvation shell (r < 3.25 A) for each Ca2+ ion in the standalone 3-Ca bound protein models and protein-membrane complex models.
Ligand First Ca2+ Second Ca2+ Third Ca2+
SytlC2Awl Protein O 6.9 6.0 3.0
Water O 1.0 1.0 4.0
CL N/A N/A N/A
Sum 7.9 7.0 7.0
Syt7C2AWT Protein O 6.9 6.0 3.3
Water O 1.0 1.0 3.8
CL N/A N/A N/A
Sum 7.9 7.0 7.1
Syt7:lC2ACH Protein O 6.9 5.0 3.3
Water O 1.0 2.0 4.0
CL 0.0 0.0 0.0
Sum 7.9 7.0 7.3
SytlC2BWT Protein O 6.9 6.0 1.2
Water O 1.0 1.0 4.0
CL 0.0 0.0 1.0
Sum 7.9 7.0 6.2
Syt7C2BWT Protein O 6.9 6.0 5.0
Water O 1.0 1.0 2.0
CL 0.0 0.0 0.0
Sum 7.9 7.0 7.0
Sytl:7C2BCH Protein O 6.9 6.0 1.1
Water O 1.0 1.0 3.0
CL 0.0 0.0 2.0
Sum 7.9 7.0 6.1
Pre-insertion: Lying-down Protein O 6.9 6.0 3.1
Water O 1.0 1.0 4.0
Lipid O 0.0 0.0 0.0
Sum 7.9 7.0 7.1
Pre-insertion: Standing-up Protein O 6.9 6.0 3.1
Water O 1.0 1.0 3.4
Lipid O 0.0 0.0 0.5
Sum 7.9 7.0 7.0
1st embedded Protein O 6.9 6.0 3.4
Water O 1.0 1.0 3.0
Lipid O 0.0 0.0 1.3
Sum 7.9 7.0 7.6
2nd embedded Protein O 6.9 6.0 3.6
Water O 1.0 1.0 2.0
Lipid O 0.0 0.0 1.3
Sum 7.9 7.0 6.9
56


(A) Syt1C2BWT
(B) Syt1:7C2BCH
0 50 100 0 50 100
Time [ns]
Time [ns]
_i_
Figure 21. The distances between the third Ca and Cl ions as a function of the 100-ns simulation time, indicating the Cl recruitment by the third Ca ions. The third Ca ions are completely stabilized by (A) one Cl ion in SytlC2BWT and (B) two in Sytl :7C2Bch after ~40 ns.
57


CHAPTER IV
CONCLUSION
In this work, we investigated the C2 domains in Sytl mid Syt7 proteins through MD simulations. The main conclusions are as follows:
i. Syt7C2A preferentially binds three Ca ions in agreement with the experiment, and its third Ca ion is coordinated by a combination of protein residues in the CBLs, water, and lipid if presented. Syt7: 1C2Ach also binds three Ca2+ ions in the CBLs, and its Ca2+ ion affinity somewhat fell somewhat between the SytlC2A mid Syt7C2A models.
ii. The SytlC2BWT domain, which is known to bind two Ca2+ ions, can hold one additional Ca2+ ion in the CBLs. The binding of the third Ca2+ ion to the protein, however, is less firm than that in SytlC2AWT or Syt7C2BWT, as the third Ca2+ ion is coordinated by fewer protein residues and more water molecules. The WT mid CH C2B domains of Sytl recruit Cl ions from the bulk solution if the third Ca ion is bound. In particular, Sytl :7C2B tended to attract more Cl ions. The stronger anion recruitment by Sytl:7C2BCH implies a greater anionic lipid affinity than SytlC2BWT.
iii. When the CBLs with three bound-Ca2+ ions in Syt7C2AWT are sufficiently close to an anionic phospholipid membrane, the protein spontaneously approaches and inserts into the membrane. The orientation of the protein is determined by the competition between the electrostatic attractions mid hydrophobic interactions. Three of the four simulated geometries of the protein-membrane complex (not the lying-down model) are in good agreements with the experimental EPR measurements.34
58


REFERENCES
1. Coussens, L.; Parker, P. J.; Rhee, L.; Yang-Feng, T. L.; Chen, E.; Waterfield, M. D.;
Francke, U.; Ullrich, A. (1986) Multiple, Distinct Forms of Bovine mid Human Protein Kinase C Suggest Diversity in Cellular Signaling Pathways. Science 233, 859-866.
2. Nalefski, E. A.; Wisner, M. A.; Chen, J. Z.; Sprang, S. R.; Fukuda, M.; Mikoshiba, K.; Falke,
9+
J. J. (2001) C2 Domains from Different Ca Signaling Pathways Display Functional and Mechanistic Diversity. Biochemistry 40, 3089-3100.
3. Sutton, R. B.; Davletov, B. A.; Berghuis, A. M.; Sudhof, T. C.; Sprang, S. R. (1995)
2+
Structure of the First C2 Domain of Synaptotagmin I: A Novel Ca /Phospholipid-binding Fold. Cell 80, 929-938.
4. Brose, N.; Petrenko, A. G.; Sudhof, T. C.; Jahn, R. (1992) Synaptotagmin: A Calcium Sensor on the Synaptic Vesicle Surface. Science 256, 1021-1025.
5. Perin, M. S.; Fried, V. A.; Mignery, G. A.; Jahn, R.; Sudhof, T. C. (1990) Phospholipid Binding by a Synaptic Vesicle Protein Homologous to the Regulatory Region of Protein Kinase C. Nature 345, 260-263.
6. Chapman, E. R. (2008) How Does Synaptotagmin Trigger Neurotransmitter Release? Annu. Rev. Biochem. 77, 615-641.
7. Rickman, C.; Craxton, M.; Osborne, S.; Davletov, B. (2004) Comparative Analysis of Tandem C2 Domains from the Mammalian Synaptotagmin Family. Biochemistry 378, 681-686.
8. Bacaj, T.; Wu, D.; Yang, X.; Morishita, W.; Zhou, P.; Xu, W.; Malenka, R. C.; Sudhof, T. C. (2013) Synaptotagmin-1 and Synaptotagmin-7 Trigger Synchronous and Asynchronous Phases of Neurotransmitter Releas. Neuron 80, 947-959.
9. Sudhof, T. C. (2013) Neurotransmitter Release: The Last Millisecond in the Life of a Synaptic Vesicle. Neuron 80 (3), 675-690.
10. Luo, F.; Bacaj, T.; Sudhof, T. C. (2015) Synaptotagmin-7 is Essential for Ca2+-triggered Delayed Asynchronous Release But Not for Ca -Dependent Vesicle Priming in Retinal Ribbon Synapses. J. Neurosci. 35, 11024-11033.
11. Dolai, S.; Xie, L.; Zhu, D.; Liang, T.; Qin, T.; Xie, H.; Kang, Y.; Chapman, E. R.; Gaisano, H. Y. (2016) Synaptotagmin-7 Functions to Replenish Insulin Granules for Exocytosis in Human Islet p-Cells. Diabetes 65, 1962-1976.
12. Takahashi, H.; Shahin, V.; Henderson, R. M.; Takeyasu, K.; Edwards on, J. M. (2010) Interaction of Synaptotagmin with Lipid Bilayers, Analyzed by Single-Molecule Force Spectroscopy. Biophys J. 99 (8), 2550-2558.
59


13. Hui, E.; Bai, J.; Wang, P.; Sugimori, M.; Llinas, R. R.; Chapman, E. R. (2005) Three Distinct Kinetic Groupings of the Synaptotagmin Family: Candidate Sensors for Rapid and Delayed Exocytosis. Proc. Natl. Acad. Sci. U. S. A. 102, 5210-5214.
14. Vermaas, J. V.; Tajkhorshid, E. (2017) Differential Membrane Binding Mechanics of Synaptotagmin Isoforms Observed in Atomic Detail. Biochemistry 56, 281-293.
15. Brandt, D. S.; Coffman, M. D.; Falke, J. J.; Knight, J. D. (2012) Hydrophobic Contributions to the Membrane Docking of Synaptotagmin 7 C2A Domain: Mechanistic Contrast between Isoforms 1 and 7. Biochemistry 51 (39), 7654-7664.
16. Mackler, J. M.; Drummond, J. A.; Loewen, C. A.; Robinson, I. M.; Reist, N. E. (2002) The
2_|_
C2B Ca -binding Motif of Synaptotagmin Is Required for Synaptic Transmission in vivo. Nature 418, 340-344.
17. Rizo, J.; Chen, X.; Arac, D. (2006) Unraveling the Mechanisms of Synaptotagmin and SNARE Function in Neurotransmitter Release. Trends Cell Biol. 16, 339-350.
18. Fernandez-Chacon, R.; Konigstorfer, A.; Gerber, S. H.; Garcia, J.; Matos, M. F. (2001) Synaptotagmin I Functions as a Calcium Regulator of Release Probability. Nature 410, 41-49.
19. Voleti, R.; Tomchick, D. R.; Sudhof, T. C.; Rizo, J. (2017) Exceptionally Tight Membrane-binding May Explain the Key Role of the Synaptotagmin-7 C2 A Domain in Asynchronous Neurotransmitter Release. PNAS114 (40), E8518-E8527.
20. Segovia, M.; Ales, E.; Montes, M. A.; Bonifas, I.; Jemal, I.; Lindau, M.; Maximov, A.; Sudhof, T. C.; Alvarez de Toledo, G. (2010) Push-and-pull Regulation of the Fusion Pore by Synaptotagmin-7. PNAS 107 (44), 19032-19037.
21. Chon, N. L.; Osterberg, J. R.; Henderson, J.; M., K. H.; Reute, N.; Knight, J. D.; Lin, H. (2015) Membrane Docking of the Synaptotagmin 7 C2A Domain: Computation Reveals Interplay between Electrostatic and Hydrophobic Contributions. Biochemistry 54, 5696-5711.
22. Shao, X.; Fernandez, I.; Sudhof, T. C.; Rizo, J. (1998) Solution Structures of the Ca2+-Free and Ca -Bound C2A Domain of Synaptotagmin I: Does Ca Induce a Conformational Change? Biochemistry 37 (46), 16106-16115.
23. Ubach, J.; Zhang, X.; Shao, X.; Sudhof, T. C.; Rizo, J. (1998) Ca2+ Binding to Synaptotagmin: How Many Ca2+ Ions Bind to the Tip of a C2-domain? EMBO J. 17, 3921-3930.
24. Shao, X.; Davletov, B. A.; Sutton, R. B.; Sudhof, T. C.; Rizo, J. (1996) Bipartite Ca2+-Binding Motif in C2 Domains of Synaptotagmin mid Protein Kinase C. Science 273, 248-251.
60


25. Vasquez, J. K.; Chantranuvatana, K.; Giardina, D. T.; Coffman, M. D.; Knight, J. D., Lateral Diffusion of Proteins on Supported Lipid Bilayers: Additive Friction of Synaptotagmin 7 C2AC2B Tandem Domains. Biochemistry 2014, 53, 7904-7913.
26. Fernandez, I.; Arac, D.; Ubach, J.; Gerber, S. FL; Shin, O.; Gao, Y.; Anderson, R. G. W.; Sudhof, T. C.; Rizo, J. (2001) Three-dimensional Structure of the Synaptotagmin 1 C2B-domain: Synaptotagmin 1 as a Phospholipid Binding Machine. Neuron 32, 1057-1069.
27. Xue, M.; Craig, T. K.; Shin, O. FL; Li, L.; Brautigam, C. A.; Tomchick, D. R.; Sudhof, T. C.; Rosenmund, C.; Rizo, J. (2010) Structural and Mutational Analysis of Functional Differentiation between Synaptotagmins-1 mid -7. Plos One 5, 12544.
28. Cheng, Y.; Sequeira, S. M.; Malinina, L.; Tereshko, V.; Sollnerr, T. FL; Patel, D. J. (2004)
2+ 2+
Crystallographic Identification of Ca and Sr Coordination Sites in Synaptotagmin I C2B Domain. Protein Science 13, 2665-2672.
29. Ubach, J.; Garcia, J.; Nittler, M. P.; Sudhof, T. C.; Rizo, J. (1999) Structure of the Janusfaced C2B Domain of Rabphilin. Nat. Cell Biol. 1, 106-112.
30. Maximov, A.; Lao, Y.; Li, H.; Chen, X.; Rizo, J.; Sorensen, J. B.; Sudhof, T. C. (2008) Genetic Analysis of Synaptotagmin-7 Function in Synaptic Vesicle Exocytosis. ProcNatl AcadSci U. S. A. 105, 3986-3991.
31. Nagashima, T.; Hayashi, F.; Yokoyama, S. (2015) Solution Structure of the First C2 Domain of Synaptotagmin VII, unpublished results.
32. Rao, T. C.; Passmore, D. R.; Peleman, A. R.; Das, M.; Chapman, E. R.; Anantharam, A. (2014) Distinct Fusion Properties of Synaptotagmin-1 and Synaptotagmin-7 Bearing Dense Core Granules. Mol. Biol. Cell 25, 2416-2427.
33. Rao, T. C.; Santana Rodriguez, Z.; Bradberry, M. M.; Ranski, A. H.; Dahl, P. J.; Schmidtke, M. W.; Jenkins, P. M.; Axelrod, D.; Chapman, E. R.; Giovannucci, D. R.; Anantharam, A.
(2017) Synaptotagmin Isoforms Confer Distinct Activation Kinetics and Dynamics to Chromaffin Cell Granules. J. Gen. Physiol. 149, 763-780.
34. Osterberg, J. R.; Chon, N. L.; Boo, A.; Maynard, F. A.; Lin, FL; Knight, J. D. (2015) Membrane Docking of the Synaptotagmin 7 C2A Domain: Electron Paramagnetic Resonance Measurements Show Contributions from Two Membrane Binding Loops.
Biochemistry 54, 5684-5695.
35. Bendahmane, M.; Bohannon, K. P.; Rao, T. C.; Schmidtke, M. W.; Bradberry, M. M.;
Abbineni, P.; Chon, N. L.; Tran, S.; Lin, FL; Chapman, E. R.; Knight, J. D.; Anantharam, A.
2+
(2018) The Synaptotagmin-1 C2B Domain Ca -binding Loops Modulate the Rate of Fusion Pore Expansion. Mol. Cell 29, 1-44.
36. Jo, S. K., T.; Iyer, V. G.; Im, W. (2008) CHARMM-GUI: A Web-based Graphical User Interface for CHARMM. J. Comput. Chem. 29, 1859-1865.
61


37. Klauda, J. B.; Venable, R. M.; Freites, J. A.; O'Connor, J. W.; Tobias, D. J.; Mondragon-Ramirez, C.; Vorobyov, I.; MacKerell, A. D.; Pastor, R. W. (2010) Update of the CHARMM All-atom Additive Force Field for Lipids: Validation on Six Lipid Types. J.
Phys. Chem. B 114 (23), 7830-7843.
38. Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. (1983) Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 79 (2), 926-935.
39. Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. (2005) Scalable Molecular Dynamics with NAMD. J.
Comput. Chem. 26(16), 1781-1802.
40. Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, EL; Pedersen, L. G. (1995) A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 103 (19), 8577-8593.
41. Humphrey, W.; Dalke, A.; Schulten, K. (1996) VMD: Visual Molecular Dynamics. J. Molec. Graphics 14 (1), 33-38.
42. Brooks, B. R.; Brooks, C. L., 3rd; Mackerell, A. D., Jr.; Nilsson, L.; Petrella, R. J.; Roux, B.; Won, Y.; Archontis, G.; Bartels, C.; Boresch, S.; Caflisch, A.; Caves, L.; Cui, Q.; Dinner, A. R.; Feig, M.; Fischer, S.; Gao, J.; Hodoscek, M.; Im, W.; Kuczera, K.; Lazaridis, T.; Ma, J.; Ovchinnikov, V.; Paci, E.; Pastor, R. W.; Post, C. B.; Pu, J. Z.; Schaefer, M.; Tidor, B.; Venable, R. M.; Woodcock, H. L.; Wu, X.; Yang, W.; York, D. M.; Karplus, M. (2009) CHARMM: The Biomolecular Simulation Program. J. Comput. Chem. 30 (10), 1545-1614.
43. Mackerell, A. D., Jr.; Bashford, D.; Bellott, M.; Dunbrack, R. L.; Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher, W. E., Ill; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.; Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. (1998) All-Atom Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins. J. Phys. Chem. B 102 (18), 3586-3616.
44. Mackerell, A. D., Jr.; Feig, M.; Brooks, C. L. (2004) 3rd, Extending the Treatment of Backbone Energetics in Protein Force Fields: Limitations of Gas-Phase Quantum Mechanics in Reproducing Protein Conformational Distributions in Molecular Dynamics Simulations. J.
Comput. Chem. 25 (11), 1400-1415.
45. Gasteiger, E.; Gattiker, A.; Hoogland, C.; Ivanyi, I.; Appel, R. D.; Bairoch, A. (2003) ExPASy: The Proteomics Server for In-depth Protein Knowledge mid Analysis. Nucleic Acids Res. 31, 3784-3788.
46. Biasini, M.; Bienert, S.; Waterhouse, A.; Arnold, K.; Studer, G.; Schmidt, T.; Kiefer, F.; Cassarino, T. G.; Bertoni, M.; Bordoli, L.; Schwede, T. (2014) SWISS-MODEL: Modelling Protein Tertiary and Quaternary Structure Using Evolutionary Information. Nucleic Acids Res. 42, W252-W258.
62


47. Frazier, A. A.; Roller, C. R.; Havelka, J. J.; Hinderliter, A.; Cafiso, D. S. (2003) Membrane -bound Orientation and Position of the Synaptotagmin I C2A Domain by Site-Directed Spin Labeling. Biochemistry 42 (1), 96-105.
48. Kucerka, N.; Nieh, M.-P.; Katsaras, J. (2011) Fluid Phase Lipid Areas mid Bilayer Thicknesses of Commonly Used Phosphatidylcholines as a Function of Temperature.
Biochim. Biophys. Acta, Biomembr. 1808, 2761-2771.
49. Bockmann, R. A.; Hac, A.; Heimburg, T.; Grubmiiller, H. (2003) Effect of Sodium Chloride on a Lipid Bilayer. Biophys. J. 85, 1647-1655.
50. Mukhopadhyay, P.; Monticelli, L.; Tieleman, D. P. (2004) Molecular Dynamics Simulation of a Palmitoyl-Oleoyl Phosphatidylserine Bilayer with Na+ Counterions and NaCl.
Biophysical J. 86(3), 1601-1609.
51. Baker, N. A.; Sept, D.; Joseph, S.; Holst, M. J.; McCammon, J. A. (2001) Electrostatics of Nanosystems: Application to Microtubules and the Ribosome. Proc. Natl. Acad. Sci. U. S. A. 98 (18), 10037-10041.
52. Seelig, A.; Seelig, J. (1974) Dynamic Structure of Fatty Acyl Chains in a Phospholipid Bilayer Measured by Deuterium Magnetic Resonance. Biochemistry 13 (23), 4839-4845.
53. Venable, R. M.; Luo, Y.; Gawrisch, K.; Roux, B.; Pastor, R. W. (2013) Simulations of Anionic Lipid Membranes: Development of Interaction-specific Ion Parameters and Validation using NMR Data. J. Phys. Chem. B 117, 10183-10192.
54. Li, C.; Ullrich, B.; Zhang, J. Z.; Anderson, R. G.; Brose, N.; Siidhof, T. C. (1995) Ca2+-dependent and -independent Activities of Neural and Non-neural Synaptotagmins. Nature 375, 594-599.
55. Sugita, S.; Han, W.; Butz, S.; Liu, X.; Fernandez-Chacon, R.; Lao, Y.; Siidhof, T. C. (2001)
2+
Synaptotagmin VII as a Plasma Membrane Ca Sensor in Exocytosis. Neuron 30, 459-473.
56. Sugita, S.; Shin, O. H.; Han, W.; Lao, Y.; Siidhof, T. C. (2002) Synaptotagmins Form Hierarchy of Exocytotic Ca2+-sensors with Distinct Ca2+-affinities. EMBOJ. 21, 270-280.
57. Nalefski, E. A.; Falke, J. J. (1996) The C2 Domain Calcium-binding Motif: Structural and Functional Diversity. Protein Sci. 5, 2375-2390.
58. Murray, D.; Honig, B. (2002) Electrostatic Control of the Membrane Targeting of C2 Domains. Mol. Cell 9, 145-154.
59. Bhalla, A.; Tucker, W. C.; Chapman, E. R. (2005) Synaptotagmin Isoforms Couple Distinct Ranges of Ca2+, Ba2+, mid Sr2+ Concentration to SNARE-mediated Membrane Fusion. Mol. Biol. Cell 16, 4755-4764.
63


60. Corbin, J. A.; Evans, J. H.; Landgraf, K. E.; Falke, J. J. (2007) Mechanism of Specific
2b
Membrane Targeting by C2 Domains: Localized Pools of Target Lipids Enhance Ca Affinity. Biochemistry 46, 4322-4336.
61. Tongraar, A.; Liedl, K. R.; Rode, B. M. (1997) Solvation of Ca2+ in Water Studied by Bom-oppenheimer ab initio QM/MM Dynamics. J. Phys. Chem. A 101, 6299-6309.
62. Jalilehvand, F.; Spangberg, D.; Lindqvist-Reis, P.; Hermansson, K.; Persson, L; Sandstrom, M. (2001) Hydration of the Calcium Ion. An EXAFS, Large-Angle X-ray Scattering, and Molecular Dynamics Simulation Study. J. Am. Chem. Soc. 123 (3), 431-441.
63. Pirc, E. T.; Zidar, J.; Bukovec, P. (2012) A Computational Study of Calcium(II) and Copper(II) Ion Binding to the Hyaluronate Molecule. Int. J. Mol. Sci. 13 (9), 12036-12045.
64


Full Text

PAGE 1

C ALCIUM AND MEMBRANE ASSOCIATION FOR THE C2 DOMAINS OF SYNAPTOTAGMIN 1 AND SYNAPTOTAGMIN 7 BY MOLECULAR DYNAMICS by NARA LEE CHON B.S., University of Colorado 201 4 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 Program 201 8

PAGE 2

ii This thesis for the Master of Science degree by Nara Lee Chon has been approved for the Chemistry Program by Hai Lin, Chair Jefferson Knight Michael Crowley Date: May 12, 2018

PAGE 3

iii Chon Nara Lee (M.S., Chemistry) Calcium and Membrane A ssociation for the C2 Domains of Synaptotagmin 1 and Synaptotagmin 7 by Molecular Dynamics Thesis directed by Professor Hai Lin ABSTRACT The C2 domain i s an important membrane binding motif in cell signaling pathway s found ubiquitously in mammals T wo synaptotagmin (Syt) proteins Syt1 and Syt7 each employ two C2 domains (C2A and C2B) to interact with membrane s d uring exocytosis acting as Ca 2+ sensors To explore Ca 2+ ion binding and membrane association of the Syt1 and Syt7 C2 domains we performed molecular dynamics simulations on a series of model systems First, i n the solvated protein s simulations we found that Ca 2+ ions were chelated by the protein oxygen atoms in the Ca 2+ binding loops (CBLs): the side chain oxygen atoms in the aspartate and the serine residue s, and backbone oxygen atoms However the number of bound Ca 2+ ions and their coordination shell compositions varie d Syt1 C2A and Syt7 C2A and C2B each bound 3 Ca 2+ ions tightly, but Syt1 C2B only bound 2 Ca 2+ ions tightly and the outermost Ca 2+ ion weakly. W ater molecules and Cl ions were also recruited from the bulk solution to complete the Ca 2+ solvation shells Two chimeric C2 domains (Syt7:1C2A CH and Syt1:7C2B CH ) were also studied as compar isons with the wild type proteins Syt7:1C2A CH is a hybrid of the Syt7 C2A body and the Syt1 C2A CBLs whereas Syt1:7C2B CH the Syt 1 C2B body and the Syt 7 C2B CBLs Our data suggest s that Syt7:1C2A CH shows similar Ca 2+ binding to the wild type C2A of Syt1 and Sy t7 In contrast, the Ca 2+ ion binding of Syt1:7C2 B CH resembles Syt1 C2B in agreement with

PAGE 4

iv experiment. Second, we generated four of Syt7 C2A to 3:1 POPC/POPS membrane associated models F229 in CBL3 inserted deep in to the membrane shortly after the beginning of the simulations on the other hand, F167 in CBL1 enter ed and left the membrane constantly We suspect that this oscillating motion is caused by the competition between the electrostatic attractions ( between the polybasic region near the 4 strand and lipid head groups) and hydrophobic interactions ( between F167 and the lipid acyl chains). W e conclude that the affinity of the two Syt isoforms for Ca 2+ ions is important predictor of the Syt mediated membrane fusion The form and content of this abstract are approved. I recommend its publication Approved: Hai Lin

PAGE 5

v DEDICATION T o my husband Song Mo and lovely two year old daughter Dana for their unconditional love and solicitude that allow me to write this thesis weekdays and weekends with full of joy To my famil y in Seoul and here in Denver for their spiritual support. To Adam Duster for taking his time to review this thesis and giving me honest and friendly feedback. To Hai Lin and all members of the Lin Lab for making me fall in love with computational chemistry and enjoy my life as a computational chemist.

PAGE 6

vi A CKNOWLEDGMENTS I would like to thank my advisor, Dr. Hai Lin, for his guidance and immense support in preparing this thesis. I am thankful to Dr. Jefferson Knight Dr. Nathalie Reuter and Dr. A run A nantharam and their group members who have widened my perspectives by sharing their knowledge and enriching ideas with me. This work is supported by National Science Foundation (CHE 0952337) and Camille & Henry Dreyfus Foundation (TH 14 028). This work used the Extreme Science and Engineering Discovery Environment (XSEDE) under grant CHE 140070 and MCB160138 which are supported by National Science Foundation grant number ACI 1053575 and the National Energy Research Scientific Computing Centre (NERSC) u nder grant m2495, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE AC02 05CH11231. This research is also supported by the University of Colorado Denver Chemistry Departmental St udent Research Fellowship.

PAGE 7

vii TABLE OF CONTENTS CHAPTER I. I NTRODUCTION ................................ ................................ ................................ .................... 1 1.1. Membrane Association of C2 domain containing Synaptotagmins ................................ 1 1.2. Sequence and Structural C omparisons between Syt1 and Syt7 ................................ ...... 3 1.3. Unanswered Questions and Aims of the Present Study ................................ .................. 5 II. C OMPUTATIONAL METHODS ................................ ................................ ........................... 7 2.1. Model Preparation ................................ ................................ ................................ ........... 7 2.1.1. Standalone Membrane or Protein ................................ ................................ ............. 7 2.1.1.1. Standalone Membrane ................................ ................................ ....................... 7 2.1.1.2. Standa lone C2 Domains ................................ ................................ .................... 8 2.1.1.2.1. Wild type Syt1 Protein: Syt1C2A WT 3Ca 2+ Syt1C2B WT 2Ca 2+ and Syt1C2B WT 3Ca 2+ ................................ ................................ .................. 13 2.1.1.2.2. Wild type Syt7 Protein: Syt7C2A WT 2Ca 2+ Syt7C2A WT 3Ca 2+ and Syt7C2B WT 3Ca 2+ ................................ ................................ .................. 1 6 2.1.1.2.3. Chimeric Syt Proteins: Syt7:1C2A CH 3Ca 2+ Syt1:7C2B CH 2Ca 2+ and Syt1:7C2B CH 3Ca 2+ ................................ ................................ ............... 19 2.1.2. Syt7 C2A Protein Membrane Complex Models ................................ .................... 20 2.1.2.1. Construction and Solvation of the Pre insertion Models ................................ 2 1 2.1.2.2. Construction and Solvation of the Emb edded Models ................................ ... 2 3

PAGE 8

viii 2.2. Data Analysis ................................ ................................ ................................ ................ 2 4 2.2.1. Area per Lipid (APL) and Lipid Order Parameter ( S CH ) ................................ ........ 2 4 2.2.2. Poisson Boltzmann Calculation ................................ ................................ ............. 2 5 III. R ESULTS AND DISCUSSION ................................ ................................ ............................. 26 3.1. Simulations of the Standalone 3:1 PC/PS Membrane ................................ ................... 26 3.2. Simulations of the Standalone C2 Domains ................................ ................................ .. 27 3.2.1. Comparisons of 2 Ca 2+ with 3 Ca 2+ Models in Each C2 WT Domain ..................... 28 3.2.2. Comparisons of WT and CH Domains ................................ ................................ .. 39 3.3. Simulations of Syt7 C2A Protein membrane Association ................................ ............ 42 3.3.1. Comparisons of the Pre insertion with the Embedded Syt7C2A WT membrane ..... 43 3.3.2. Comparisons of Simulation with Experim ent ................................ ........................ 52 3.4. Coordination of the Ca 2+ Ions in the Standalone Syt Proteins and Syt7C2A WT membrane Complexes ................................ ................................ ................................ ... 53 IV. C ONCLUSION ................................ ................................ ................................ ...................... 58 R EFERENCES ................................ ................................ ................................ ................................ ......... 59

PAGE 9

ix L IST OF ABBREVIATIONS alpha APL a rea per lipid beta CBL Ca 2+ binding loop CHARMM c hemistry at H ar vard m acromolecular m echanics CHARMM GUI CHARMM, a web based graphical user interface (www.charmm gui.org) CH c himeric or chimera CMAP cross term for the based energy correction maps EPR electron paramagnetic resonance HMMM highly mobile membrane mimetic S CH l ipid order parameter MD m olecular dynamics NAMD n anoscale m olecular d ynamics POPC or PC 1 palmitoyl 2 oleoyl sn glycero 3 phosphocholine POPS or PS 1 palmitoyl 2 oleoyl sn glycero 3 phospho L serine PO 4 p hosphate group PDB protein data bank (www.rcsb.org) PQR PDB data with atomic charge and radius parameters RMSD root mean square deviation Syt synaptotagmin Syt1 synaptotagmin 1 Syt7 synaptotagmin 7 VMD v isual m olecular d ynamics WT wild type

PAGE 10

1 CHAPTER I I NTRODUCTION 1.1. Membrane Association of C2 domain containing Synaptotagmins The C2 domain first identified in protein kinase C as the second conserved domain 1 is found to be a key component of many cell signaling proteins such as synaptotagmin (Syt) protein s 2 The Syt proteins contain two C2 domains C2A and C2B Each domain compr is e s approximately 130 residues All C2 domain s have eight beta strands ( 1 8) connected by flexible loops including three calcium(Ca 2+ ) binding loops (CBLs, CBL1 3) ( Figure 1 ) 3 With two to four Ca 2+ ions bound in the CBLs synaptotagmin 1 (Syt1) and synaptotagmin 7 (Syt7) bind to membranes containing negatively charged phospholipid s such as phosphatidylserine ( Figure 2 ) 4 5 6 7 Syt1 and Syt7 both of which serv e as Ca 2+ ion sensor s, are the two well studied protein s in the Syt family They are responsible for exocytosis in synaptic vesicles and in secretory granules of endocrine cells Despite the ir structural similarit ies th e se two proteins facilitate vesicle release in marked ly different ways In neurons, Syt 1 primary responds to the rapid synchronous neurotransmitter release while Syt7 facilitates the slow asynchronous release by jointly regulat ing the exocytosis with other Syt isoforms. 8 9 10 11 Both C2A domains in Syt1 and Syt7 bind three Ca 2+ ions to the CBLs. However, t he C2A domain of Syt7 has stronger electrostatic attractions between its poly basic lysine cluster region and the phospholipid head groups than Syt1. 12 13 14 The coordination of the Ca 2+ ions by the protein and lipid and insertion of hydrophobic residues into the membrane allow Syt7 C2A to dock deeply to the membrane. 15

PAGE 11

2 A relative role for the C2 domains of Syt1 and Syt7 in Ca 2+ triggered neuro transmi tter release is reversed. For Syt1, Ca 2+ binding to the C2B domain is more important for synaptic transmission than Ca 2+ binding to the C2A domain 16 T he isolated Syt1 C2B domain can simultaneously bind to two membrane s: one membrane through the CBL region at the bottom of the C2B domain and ano ther membrane through the top face loops in the presence of Ca 2+ ions supporting that the Syt 1 protein may trigger neurotransmitter release by bringing the synaptic vesicle and plasma membranes together mainly via C2B domains 17 18 19 For Syt7, however, the C2A domain actively militate s in favor of fusion opening but the C2B domain is selectively essential for neurotransmitter release. 20 8 Figure 1 The structures of the C2 domains in Syt7 in gray strand in orange) ( A ) C2A and ( B ) C2B domains are with each of the highlighted Ca 2+ binding loops (CBL1 3; red, yellow, and blue spheres). The Ca 2+ ions are represented by small green spheres with labels from 1 to 3. The side chains of the critical residues in Syt7 C2A (F167 and F229) are only shown as sticks (white, H; cyan, C). Figure modified from Fig. 2 in Ref. 21

PAGE 12

3 Figure 2 The chemical structures of ( A ) 1 palmitoyl 2 oleoyl sn glycero 3 phosphocholine and ( B ) 1 palmitoyl 2 oleoyl sn glycero 3 phospho L serine Figure modified from Fig. 1 in Ref. 21 1.2. S equence and Structural C omparisons between Syt1 and Syt7 The two Syt isoforms, Syt1 and Syt7, each contain a set of two Ca 2+ activated C2 domains. Figure 3 A displays the sequence alignment for the aforementioned proteins. T he Syt1 and Syt7 C2A domains can each accommodate three Ca 2+ ions in the binding sites. 22 23 T he residues of CBL2 and CBL 3 are well conserved across Syt isoform ( Figure 3B ) T wo residues (arginine and phenylalanine) located in the middle of CBL3 are strictly conserved across these proteins ; i n particular, F234 in Syt1 and F229 in Syt7 play a key role in the protein binding to the membrane 3 24 25 The CBL1 residues in Syt1 and Syt7 are less strictly conserved, with M173 in Syt1 correspond ing to F167 in Syt7 The CBL region is known to be essential for Ca 2+ dependent phospholipid binding This binding is mediated by CBL1 and CBL3 primarily through electrostatic attractions and secondarily through hydrophobic interactions. When the protein approaches the membrane surface, th e two hydrophobic residues in CBL1 and CBL3 of Syt7 (F167 and F229) act as anchors to the hydrophobic lipid tails However, M173 in Syt1 C2A does

PAGE 13

4 not exhibit the hydrophobic interactions with the lipid tails as strong as the phenylalanine does H ence the binding affinity of Syt1 is less than of Syt7. 15 21 14 Just like C2A, b oth C2B domains in Syt1 and Syt7 have a cluster of the acidic residue s such as the aspartate residue in the CBLs H owever the number of bound Ca 2+ ions are different : two and three Ca 2+ ions bind to Syt1 and Syt7, respectively 26 27 A serine residue in CBL3 of Syt7 C2B (S362) is not present in Syt1 C2B domain Crystallography revealed the third Ca 2+ ion binds to the CBLs in Syt1 C2B but at very high Ca 2+ ion concentration with coordination of the side chain oxygen atom in the asparagines residue 28 Moreover, Syt1 C2B contains two helices : one of these helices is in Syt7 C2B while the other helix near the C terminal loop is predicted to be present in only Syt 1 26 27 29 The functional differentiation in Syt1 and Syt7 arises in part f rom subtle sequence variations

PAGE 14

5 Figure 3 Sequence alignment and comparison. ( A ) C 2 domains of all proteins employed in this work, showing each three of Ca 2+ binding loops (CBLs) with red shaded boxes. The critical residues in the CBL1 and CBL3 of Syt C2A domains are also highlighted with the red font. For the chimeric ( CH ) domains, Syt7:1C2A CH contain s the Syt7 C2A sequence but with the CBLs replaced by the counterpart in Syt1 C2A domain Syt1:7C2B CH contain s the Syt1 C2B sequence but with the CBLs from Syt7 C2B. ( B ) The C2A domain s equence s of CBL 1 to CBL3 regions among mammalian Syt isoforms using Clustal Omega. Boxed regions indicate positions at the tip of CBL1 and CBL3 Figure taken from Fig. S1 in Ref 21 1.3. Unanswered Q uestions and A ims of the Present S tudy D etailed mechanism s of the membrane association s by Syt1 and Syt7 are not well understood at the molecular level A tomistic molecular dynamics (MD) simulations could provide some much needed missing information about th e functions of the proteins There are several questions that we seek to answer:

PAGE 15

6 (1) How many Ca 2+ ions are most likely bound to Syt7 C2A? It has been experimentally shown that the C2A domain in Syt7 can bind three Ca 2+ ions but the affinity is much lower for the third Ca 2+ than the first two Ca 2+ ions 30 The available NMR structure of Syt7 C2A is a Ca 2+ free form 31 Because it is essential for Syt7 C2A to bind Ca 2+ ions before insert ing in to the membrane it is important to determine the most optimal stoichiometry o f Ca 2+ ion binding (2) What are the structural features of Syt1 and Syt7 at membrane binding sites ? Recently Rao et al. found that S yt1 and Syt7 are sorted to different pop ulations of chromaffin granules and fusion pores of granules harboring Syt 1 expand more rapidly than pores of granules expressing Syt7 32 33 Here we performed wild type (WT) versus chimera (CH) s imulations by generating the Syt7:1C2A CH and Syt1:7C2B CH domains as a preliminary study The Syt7:1 C2A CH denotes that the C2A domain contains the Syt7 C2A body with Syt1 C2A CBLs while Syt1:7 C2B CH contains the Syt1 C2B body with Syt7 C2B CBLs (3) What is the orientation of Syt7 C2A when it docks to the membrane? We will characterize the geometry of the protein membrane complex. Our results will be compared with the experiment s from e lectron paramagnetic resonance (EPR) depth parameter measurements 34 T h e results presented in this thesis were from two independent research works: (i) Ca 2+ binding for C2A WT of Syt1 and Syt7 and membrane association for Syt7 C2A, 21 34 and (ii) Ca 2+ binding for C2A CH C2B CH and C2B WT of Syt1 and Syt7. 35 Therefore, the simulation settings in corresponding to the experiments are different from each work.

PAGE 16

7 CHAPTER II C OMPUTATIONAL METHODS 2.1. Model Preparation 2.1.1. Standalone Membrane or Protein 2.1.1.1. Standalone Membrane A p hospholipid bilayer model was generated by mixing 1 palmitoyl 2 oleoyl sn glycero 3 phosphocholine (POPC PC ) and 1 palmitoyl 2 oleoyl sn glycero 3 phospho L serine (POPS PS ) at a molar ratio of 3:1 employing Membrane Builder in CHARMM GUI 36 T he 3:1 PC/PS membrane model is composed of 192 POPC lipids, 64 POPS lipids, 64 Na + ions as counter ion s and 9219 water molecules were prepared by the C HARMM 36 37 force fields and TIP3P 38 water mo del Using NAMD 39 the model was equilibrated at 310 K and 1 bar for 160 ns under the NpT ensemble with a periodic boundary. 40

PAGE 17

8 Table 1. Numbers of atoms and dimensions for protein, membrane, an d protein membrane complex models employed in this work. Protein Water Lipid Ca 2+ Na + K + Cl Initial cube d imension [ 3 ] Standalone membrane 3:1 PC/PS 0 27 657 33 856 a 0 64 0 0 92 92 79 Standalone protein WT Syt1C2A 3Ca 2+ 2 083 19 839 0 3 0 0 4 80 80 50 Syt1C2B 2Ca 2+ 2 410 75 657 0 2 0 71 81 101 126 103 Syt1C2B 3Ca 2+ 2 410 75 651 0 3 0 71 83 101 126 103 Syt7C2A 3Ca 2+ 2 181 12 150 0 3 0 0 9 50 65 65 Syt7C2B 3Ca 2+ 2 300 75 762 0 3 0 71 89 101 126 103 CH Syt1 :7 C2B 2Ca 2+ 2 413 75 246 0 2 0 71 82 101 126 96 Syt1 :7 C2B 3Ca 2+ 2 413 75 648 0 3 0 71 84 101 126 103 Syt7 :1 C2A 3Ca 2+ 2 160 75 864 0 3 0 71 78 101 126 103 Protein membrane Pre insertion Lying down 2 181 75 951 33 856 a 3 64 0 9 92 94 120 Standing up 2 181 75 747 33 856 a 3 64 0 9 91 95 118 Embedded 1 st 2 181 65 805 33 454 b 3 64 0 9 92 94 106 2 nd 2 181 74 898 33 454 b 3 64 0 9 91 94 118 a 256 lipids, including 192 POPC (134 atoms each) and 64 POPS (127 atoms each). b 253 lipids, including 189 POPC and 64 POPS (deleted 3 of PC lipids close to protein). 2.1.1.2. Standalone C2 Domains All s tructural manipulations were performed using VMD 41 M inimizations and MD s imulati ons were performed using NAMD as detailed below. The Syt proteins were solvated in either water ( for WT C2A domain models ) or 0.15 M of KCl solution ( for WT C2B domain and all chimera, CH C2 domain models ), described by C HARMM 22 force field s with the CMAP correction 42 43 44 and T IP3P 38 water model The number of water molecules and counter ions (K + or Cl ) in the systems varied across models but all systems were charge neutral ( Table 1 2 ). For

PAGE 18

9 the CH Syt C2 domain, homology models were created from amino acid sequence s ( translated from the DNA sequence s through Translate Tool in ExPASy 45 ) using the Swiss Model 46 program The procedure of constructing the Syt protein models is illustrated in Figure 4 and the detailed information will be discussed in the following sections. The minimization and MD simulations were performed at 298 K and 1 bar under the NpT ensemble with a periodic boundary. H armonic restraints were imposed upon the distances between Ca 2+ ions and their coordinating oxygen atoms during the early stage of equilibration ; the force constants were 50 kcal mol 1 2 for Syt7:1C2A CH 3Ca 2+ and 200 kcal mol 1 2 for both Syt1:7C2B CH 2 and 3Ca 2+ models The simulation times are listed in Table 3

PAGE 19

10 Table 2. The net charges (in e ) of secondary structure s and the whole proteins. Structure Protein C2A C2B True ID Relative Charge True ID Relative Charge WT CH WT CH Overall: Start End Syt1 Syt7 E140 K267 E134 G265 1 128 1 132 2 +3 +1 K272 V419 S266 A403 1 148 1 138 +6 +12 +7 Syt1 Syt7 K144 D152 R138 Y145 5 13 5 12 1 +1 +1 D275 L280 E269 277 4 9 4 12 1 1 1 Syt1 Syt7 N157 A166 L152 Q160 18 27 19 27 0 +2 +2 L289 K297 S282 R291 18 26 17 26 0 +2 0 CBL1 Syt1 Syt7 L168 T176 L162 T170 29 37 29 37 1 0 1 K301 D309 A295 D303 30 38 30 38 1 2 Syt1 Syt7 P179 L186 P173 L180 40 47 40 47 +1 +1 +1 P310 Q318 P304 Y312 39 47 39 47 +1 +1 +1 Syt1 Syt7 K192 E194 K186 E188 53 55 53 55 0 0 0 R322 K327 K315 K321 51 56 50 56 +5 +4 +5 CBL2 Syt1 Syt7 H198 N203 K192 N197 59 64 59 64 +2 +3 +2 K332 P337 K326 P331 61 66 61 66 +1 +2 Syt1 Syt7 E208 F212 E202 F206 69 73 69 73 1 1 1 Y338 F345 I332 D340 67 74 67 75 1 2 1 Syt1 Syt7 T223 D230 I218 D225 84 91 85 92 1 1 1 V355 D363 T349 D357 84 92 84 92 1 1 1 CBL3 Syt1 Syt7 F231 H237 Y226 N232 92 98 93 99 +1 +1 +1 Y364 D371 K358 D365 93 100 93 100 0 +1 Syt1 Syt7 I239 P246 D233 P241 100 107 100 108 0 2 2 A372 G379 V366 L372 101 108 101 107 +1 +1 +1 Syt1 Syt7 V255 D261 Q251 K255 116 122 118 122 2 +1 +1 I401 T406 V395 Q400 130 135 130 135 0 0 0

PAGE 20

11 Figure 4 The construction procedure of C2 domain model systems. When building a 3 Ca 2+ bound model, the addition of all th r e e or the outermost (3 rd ) Ca 2+ ion(s) from one structure to another is carried out after the superposition ( on backbone heavy atoms: N, C, O, and C ) of the two structures The MD simulations are performed as listed in Table 3

PAGE 21

12 Table 3 The settings of NAMD in MD simulation for all models. Equilibration times a (R estraints b ) C2A w/o membrane Syt1C2A WT 3Ca 2+ 20ps 20ps 200ps 200ns Syt7C2A WT 2Ca 2+ 20ps 20ps 200ps 200ns Syt7C2A WT 3Ca 2+ 20ps 20ps 200ps 200ns Syt7:1C2A CH 3Ca 2+ 500ps (50) 500ps (50) 1ns 100ns C2B w/o membrane Syt1C2B WT 2Ca 2+ 500ps 500ps 1ns 100ns Syt1C2B WT 3Ca 2+ 500ps 500ps 1ns 100ns Syt7C2B WT 3Ca 2+ 500ps 500ps 1ns 100ns Syt1:7C2B CH 2Ca 2+ 500ps (200) 500ps (200) 1ns 100ns Syt1:7C2B CH 3Ca 2+ 500ps (200) 500ps (200) 1ns 100ns C2A w/ membrane Syt7C2A WT 3Ca 2+ 400ps (100, 2000) 400ps (100, 2000) 2ns 500ns a The equilibration was performed in four stages In the first two stages the temperature and pressure controls were turned on sequentially while the protein backbone atoms were frozen and the Ca 2+ solvation shell s were restrained Then the backbone was allowed to move freely in the third stage I n the final s tage t he MD simulations were performed without any constraint or restraint. b The harmonic f or ce constants (in kcal mol 1 2 ) of the restraint potentia ls for the Ca 2+ O distan ce (see Table 4 ). For the C2A with membrane simulations, 100 kcal mol 1 2 w ere put upon the Ca 2+ O distances in pre insertion models and 2000 kcal mol 1 2 were in the embedded models

PAGE 22

13 2.1.1.2.1. Wild type Syt1 Protein s : Syt1C2A WT 3Ca 2+ Syt1C2B WT 2Ca 2+ and Syt1C2B WT 3Ca 2+ The experimental ly determined structures for the Ca 2 + bound Syt1 C2A and C2B domains [Protein Data Bank (PDB) entr ies 1BYN 22 and 1K5 W 26 respectively ] were utilized to build the solvated models. In this Syt1 C2A structure, the coordination of three Ca 2+ ion s is listed in Table 4 The C2B domain of Syt1, however, only have 2 Ca 2+ ions in the binding si te 26 24 23 Thu s, the first Ca 2+ ion is coordinated by the oxygen atoms of the D303, D309, D363, and D365 side chains, and the Y364 backbone; and the second Ca 2+ ion by the D303, D363, D365, and D371 side chains, and the M302 backbone

PAGE 23

14 Table 4 The distances between Ca 2+ and their coordinating oxygen atoms in the binding site of 3 Ca 2+ bound Syt C2 domain models Experimental distances given in parenthesis a WT C2A CH C2A Syt1 Syt7 Syt7:1 Residue Atom Distance [ ] Residue Atom Distance [ ] Residue Atom Distance [ ] 1 st Ca 2+ D172 OD1 2.6 0.2 (2.81) D166 OD1 2.2 0.1 D166 OD1 2.1 0.1 D172 OD2 2.1 0.1 ( 2.82 ) D166 OD2 2.5 0.2 D166 OD2 2.8 0.3 D178 OD2 2.1 0.1 ( 2.81 ) D172 OD2 2.1 0.1 D172 OD1 2.1 0.1 D230 OD1 2.1 0.1 ( 2.81 ) D225 OD1 2.1 0.1 D225 OD1 2.1 0.1 F231 O 2.3 0.1 ( 2.81 ) Y226 O 2.3 0.1 F226 O 2.3 0.1 D232 OD1 2.2 0.1 ( 2.82 ) D227 OD1 4.2 0.1 D227 OD1 2.2 0.1 2 nd Ca 2+ L171 O 2.3 0.1 (2.79) K165 O 2.3 0.1 L165 O 4.8 0.2 D172 OD1 2.2 0.1 (2.83) D166 OD1 2.2 0.1 D166 OD1 4.1 0.1 D230 OD1 3.9 0.1 (3.31) D225 OD1 3.9 0.1 D225 OD1 4.0 0.1 D230 OD2 2.1 0.1 (2.84) D225 OD2 2.1 0.1 D225 OD2 2.2 0.1 D232 OD1 2.5 0.2 (2.80) D227 OD1 2.2 0.1 D227 OD1 2.4 0.1 D232 OD2 2.2 0.1 (2.83) D227 OD2 2.5 0.1 D227 OD2 2.2 0.1 D238 OD2 2.1 0.1 (2.82) D233 OD2 2.1 0.1 3 rd Ca 2+ D232 OD2 6.0 0.2 (2.84) D227 OD2 6.6 0.1 D227 OD2 4.8 0.6 S235 OG 6.9 0.8 (2.84) S230 OG 5.7 0.1 S230 OG 5.3 0.8 K236 O 4.8 0.7 (2.81) R231 O 4.9 0.1 K231 O 4.3 0.3 D238 OD1 2.1 0.1 (2.84) D233 OD1 2.1 0.1 D233 OD1 3.6 0.4 D238 OD2 4.2 0.2 (2.84) D233 OD2 3.7 0.1 D233 OD2 2.1 0.1 WT C2B CH C2B Syt1 Syt7 Syt1:7 1 st Ca 2+ D303 OD1 2.2 0.1 (2.59) D297 OD1 2.4 0.1 (2.67) D303 OD1 2.7 0.3 D303 OD2 2.6 0.2 (2.82) D297 OD2 2.2 0.1 (2.43) D303 OD2 2.2 0.1 D309 OD2 2.1 0.1 (2.23) D303 OD2 2.1 0.1 (2.37) D309 OD2 2.1 0.1 D363 OD1 2.1 0.1 (2.81) D357 OD1 2.1 0.1 (2.40) D363 OD1 2.1 0.1 D363 OD2 3.1 0.2 (3.31) K358 O 2.3 0.1 (2.35) K364 O 2.3 0.1 Y364 O 2.3 0.1 (2.82) D359 OD1 2.3 0.1 (2.32) D365 OD1 2.2 0.1 D365 OD1 2.2 0.1 (2.80) 2 nd Ca 2+ M302 O 2.3 0.1 (2.82) M296 O 2.3 0.1 (2.32) M302 O 2.3 0.1 D303 OD1 4.1 0.1 (2.81) D297 OD1 2.2 0.1 (2.42) D303 OD1 2.2 0.1 D363 OD2 2.1 0.0 (2.80) D357 OD2 2.1 0.0 (2.30) D363 OD2 2.1 0.1 D365 OD1 2.4 0.1 (2.81) D359 OD1 2.3 0.1 (2.50) D365 OD1 2.5 0.1 D365 OD2 2.2 0.1 (2.81) D359 OD2 2.3 0.1 (2.49) D365 OD2 2.2 0.1 D371 OD2 2.1 0.1 (2.81) D365 OD2 2.1 0.1 (2.32) D371 OD2 2.1 0.1 3 rd Ca 2+ D365 OD2 4.3 0.2 D359 OD2 2.2 0.1 (2.38) D365 OD2 4.6 0.2 K369 O 4.8 0.5 S362 OG 2.3 0.1 (2.44) S368 OG 5.1 0.6 D371 OD1 2.1 0.0 R363 O 2.2 0.1 (2.32) R369 O 4.1 0.2 D371 OD2 3.5 0.2 D365 OD1 2.1 0.1 (2.45) D371 OD1 2.1 0.1 a For Syt7 C2A WT and the CH domains and the third Ca 2+ ion in Syt1C2B WT the experimental values are not available.

PAGE 24

15 Two 3 Ca 2+ model s (Syt1C2A WT 3Ca 2+ and Syt1C2B WT 3Ca 2+ ) and one 2 Ca 2+ model (Syt1C2B WT 2Ca 2+ as mentioned earlier ) was generated for Syt1. Among the 20 NMR structures available in the PDB (1BYN), the first was selected for building the Syt1C2A WT 3Ca 2+ model. Just like Syt1 C2A domain, t he first model of 20 NMR structures in the PDB ( 1K5W ) for Syt1 C2B WT 2Ca 2+ model w as used followed by solvation and adding ions for charge neutrality To build the three Ca 2+ bound C2B model Syt1C2B WT 2Ca 2+ was superimpos ed with Syt7C2B WT 3Ca 2+ ( PDB entry 3 N5A 27 ) and the third Ca 2+ in the Syt7C2B WT 3Ca 2+ binding pocket was manually copied to the Syt 1 C2B WT 3Ca 2+ model. Note that the C2B domains of Syt1 and Syt7 are very similar with a RMSD of only 0. 9 for the backbone heavy ( N C O and C ) atoms ( Figure. 5 ) By construction the third Ca 2+ ion in Syt1C2B WT 3Ca 2+ interact s with the side chain oxygen atoms of the D365 and D371 and the backbone oxygen atom of the K369 ( Table 4 ), which are corresponding oxygen atoms in Syt7C2B WT 3Ca 2+ T he serine residue that is expressed in CBL3 of Syt7 C2B is not present in Syt1 C2B domain Therefore, in all 3 Ca 2+ models except for Syt1C2B WT 3Ca 2+ model, the third Ca 2+ ion was chelated by the hydroxyl oxygen atom in the serine along with the carboxyl oxygen atoms in the aspartate and the carbonyl oxygen atoms in the protein backbone.

PAGE 25

16 Figure 5 ( A ) Superposition of the Syt1 and Syt7 C2B domains based upon the root mean square deviation (RMSD) per residue. As the RMSD values become small to big, the color range goes from blue to red. The last 10 residues (E410 EEVDAMLA V419; a red helix in the figure; green arrow) in Syt1 C2B were ignored during the comparison, because Syt7 C2B does no t have the corresponding residues. ( B ) RMSD per residue with respect to the start of the Syt7 C2B domain The residue numbers in the context of the full length protein can be obtained by adding 271 and 265 for the C2B domains in Syt1 and Syt7, respectively. 2.1.1.2.2. Wild type Syt7 Protein s : Syt7C2A WT 2Ca 2+ Syt7C2A WT 3Ca 2+ and Syt7C2B WT 3Ca 2+ The solvate d Syt7 C2A with Ca 2+ ions were acquired from the experimental NMR structures (PDB entr y 2D8K 31 ) While the presence of up to three Ca 2+ ions is known to stabilize the Syt7 C2A domain, no Ca 2+ ions were present in the experimental structure. 27 30 31 In order to find the most feasible Ca 2+ bound C2A structures, two models were rendered by augment ing two and three Ca 2+ ion s, respectively (Syt7C2A WT 2Ca 2+ and Syt7C2A WT 3Ca 2+ ) using the first model of 20 NMR structures First, Syt1C2A WT 3Ca 2+ was aligned with Ca 2+ free Syt7 C2A. Then the coordinates of all 3 Ca 2+ ions were extracted from Syt1 and added to Syt7, resulting in Syt7C2A WT 3Ca 2+ All three Ca 2+ ions were coordinate d by the oxygen atoms as listed in Table 4 F or Syt7C2A WT 2 Ca 2+ the third Ca 2+ ion was simply deleted from Syt7C2A WT 3Ca 2+ F igure

PAGE 26

17 1A offers an overview of the Syt7 C2A with two critical residues in CBL1 and CBL3. The conformations of the CBLs with Ca 2+ ions after minimization were presented in Figure 6 showing that some Ca 2+ ions in the binding pocket required not only the protein oxygen atoms but also water molecules to form a complete solvation shell (see Section 3.4 Coordination of the Ca 2+ Ions in the Standalone Syt Proteins and Syt7C2A WT membrane Complexes ). A m odel of Syt7 C2B was prepared in a similar way Th e th ree Ca 2+ bound c rystal structure (PDB entry 3N5A, Syt7C2B WT 3Ca 2+ ) was used as an initial structure The protein was solvated and ions were added for neutraliz ing the excess charge The coordination of all three Ca 2+ ions was again described in Table 4

PAGE 27

18 Figure 6 Geometries of Ca 2+ ions in the binding sites after minimization ( A ) C2A domains in gray ( Syt1C2A WT 3 Ca 2+ top; Syt7 C2A WT 2 Ca 2+ middle; Syt7C2A WT 3 Ca 2+ bottom) and ( B ) C2B domains in purple ( Syt1C2 B WT 2 Ca 2+ top; Syt 1 C2 B WT 3 Ca 2+ middle; Syt7C2 B WT 3 Ca 2+ bottom). A ll Ca 2+ ions were in green spheres with surrounding residues shown as sticks ( N, blue; O, red; C, cyan ), and the coordinating oxygen atoms for Ca 2+ ions shown as red spheres. Protein H omitted for clarity Figure modified from Fig 2 in Re f 21

PAGE 28

19 2.1.1.2.3. Chimeric Syt Proteins: Syt7:1C2A CH 3Ca 2+ Syt1:7C2B CH 2Ca 2+ and Syt1:7C2B CH 3Ca 2+ A chimeric ( CH ) C2A domain with 3 Ca 2+ ions in the Ca 2+ binding pocket (Syt7 : 1C2A CH 3Ca 2+ ) was constructed, where Syt7 : 1 denotes the combin ation of Syt7 C2A bod y and Syt1 C2A Ca 2+ binding loops The homologous model for Syt7 : 1C2A CH 3Ca 2+ was built using the Ca 2+ free NMR experimental structure of Syt7 C2A as a template T he model consists of 126 residues E134 P259 as S260 GPSS G265 in the template sequence was excluded T he se 6 residues were added at the C terminus of Syt7 : 1C2A CH using the molefracture plugin in VMD which assists the compar isons between Syt7 : 1C2A CH 3Ca 2+ and Syt7C2A WT 3Ca 2+ in the presence of the p hospholipid bilayer F inally, the equilibrated Syt7C2A WT 3 Ca 2+ structure supplied the coordinates of three Ca 2+ ions in the binding site to the CH C2A model. T wo models of CH C2B ( Syt1 : 7C2B CH 2Ca 2+ and Syt1 : 7C2B CH 3Ca 2+ ) were created by mixing the bodies of Syt1 C2B and the CBL region of Syt7 C2B The CH C2B domains were constructed using the Syt1 C2B NMR structure as a template The three Ca 2+ ions were copied from the crystal structure of Syt7 C2B and added to the homologous model to generat e the Syt1 : 7C2B CH 3Ca 2+ model By construction t he overall Ca 2+ binding geometry in Syt1 : 7C2B CH 3Ca 2+ is similar to that in S yt 7 C2B WT 3Ca 2+ with a one to one correspondence between the Ca 2+ coordinating residues in the two models The coordination of three Ca 2+ ions was given in Table 4 The Syt1 : 7C2B CH 2 Ca 2+ model was then c onstructed from the equilibrated Syt1 : 7C2B CH 3Ca 2+ trajectory after deleting the third Ca 2+ ion

PAGE 29

20 2.1.2. Syt7 C2A Protein Membrane Complex Models All protein membrane complex models were built using CHARMM 42 and VMD program s and minimization and MD simulations were performed using NAMD Two types of protein membrane relative geometries were studied for Syt7 C2A WT : the protein above the membrane ( pre insertion ) and ins erted into the membrane ( embed ded ) in Figure 7 The equilibrated 3:1 PC/PS membrane was used for constructing all Syt7C2A WT membrane complexes. After the equilibrated standalone protein and membrane were merged, the energy minimization was carried out to relax the structures with harmonic restraints on selected atoms ( Table 5 ). After adding water molecules and counter ions into the systems, the initial protein membrane structures were minimized again using NAMD for 18 000 steps MD simulations were carried out under the NpT ensemble at 298 K and 1 bar with four st age s of sequential equilibration: first two stages for 800 ps with a frozen protein backbone and restrained Ca 2+ solvation shells then for 2 ns after releasing the restraints on the protein backbone The length s of the final stage of equilibration varied from system to system ( Table 3 ) Table 5 The harmonic restraints on the selected atoms of the protein and membrane in the energy minimizations through CHARMM program for building the initial protein membrane complex structures Force constants (kcal mol 1 2 ) Selected atoms 150 protein backbone atoms ( N, C, O, and C ) water molecules in the standalone membrane model counter ions (Na + ions) in the standalone membrane model 100 membrane atoms located >5 away from the protein protein side chain atoms located >10 away from the membrane 50 membrane atoms located < 5 from the protein 1 protein side chain atoms located <10 from the membran e

PAGE 30

21 2.1.2.1. Construction and Solvation of the Pre insertion Models The equilibrated structure of Syt7C2A WT 3Ca 2+ model was extracted from the trajectory (at t = 6 ns) generated from the 200 ns standalone solvated protein simulation The water molecules and ions were then deleted The Ca 2+ bound protein was placed above the equilibrated 3:1 PC/PS membrane in two different orientations name ly the and orientations in which the tilt angle between the long axis of the C2A domain and the membrane surface was set to 23 and 36 respectively ( Figure 7 A ). The tilt angle for the protein with respect to the membrane plane is defined in Figure 8 A In the lying down model, the Syt7 C2A tilt angle is similar to a previously reported Syt1 C2A experimental docking geometry 47 The cri tical residues F167 in CBL1 and F229 in CBL3 of the lying down model were in the aqueous phase above the membrane, but CBL3 was more closely located to the membrane surface In contrast to the lying down model, CBL1 and CBL3 in t he standing up model were similar ly positioned f rom the membrane in order to place the residue s F167 and F229 in the aqueous phase right above the membrane ; t his required n o lipid molecule dele tion ( Table 1 ) T hen the models were solvated in water and ion s were added to neutral ize the excess charge

PAGE 31

22 Figure 7 Initial models used for simulation s of Syt7C2A WT membrane c omplexes Two pre insertion models were made : ( A left ) lying down and ( A, right ) standing up models have slightly different tilt angles with respect to the membrane plane and initial penetration depths of F229 of CBL3 ( 23 and 6.9 for the lying down model and 36 and 4.2 for the standing up model respectively). ( B ) In the emb edded models, both CBL1 and CBL3 were inserted into the membrane by 1.8 and the protein remained approximately perpendicular to the membrane plane throughout the simulation The protein structures are shown as cartoon in gray strands in orange, and the critical residues F167 and F229 for Syt7C2A WT are represented by van der Waals spheres ) Lipids are shown as lines and water molecules and counter ions are hidden for clarity Color code: N, blue; O, red; C, cyan; H, white; and P, brown. Figure modified from Fig 3 in Ref 21 Figure 8 ( A ) The tilt angle ( ) of the protein (gray, cartoon) with respect to the membrane surface (green shaded area, M). The plane of the protein (blue shaded area, P) was defined by the three center of mass vectors of the outer mos t sh eets, 5 (red), 6 (blue), and 8 (purple). Each normal vector of the membrane surface and the protein was denoted as M and P respectively. ( B ) The penetration depth d between the C atom (yellow sphere) in the residue s F167 and F229 for Syt7C2A WT (the conserved residue F452 in Slp4C2A WT given in parentheses ) and the average phosphate plane of the membrane (black horizontal line). Lipids in this figure are shown as lines except phosphate atoms as licorice. Water molecules and counter ions are hidden for clarity Color code: N, blue; O, red; C, cyan; H, white; and P, brown. Figure modified from Fig 3 4 in Ref. 21

PAGE 32

23 2.1.2.2. Construction and Solvation of the Embedded Models simulations were made with different initial structures of the equilibrated Syt7 C2A solvated model (at t = 6 and 8 ns). E ach protein was shallowly inserted into the membrane and hence three overlapping POPC lipids were deleted I n addition water molecules and ions (not Ca 2+ ion ) which were close to (< 3 ) or contacted directly with the proteins were eliminated. I n both embedded models, the C for F167 of CBL1 was placed above the average phosphate plane by 1.4 while the C for F229 of CBL3 was below by 1.8 on average. Importantly, h owever, both side chain phenyl rings were embedded in the membrane ( Figure 7 B ). The initial tilt angles were 48 and 35 for the first and second embedded models respectively. The starting protein conformations in the two embedded models were similar to each other except the amine group in the side chain of K184 which was near the membrane surface in the first model but oriented away from the membrane in the second model ( Figure 9 ). Figure 9 Two orientations of the residue K184 that observed during the embedded models simulations; when the side chain of K184 was ( A ) located in the bulk water and ( B ) interacting with the head groups of lipids. The protein is shown as cartoo n in gray, with strand in orange K184 as ball and (located in the loop between and strands ), and lipid s (without H ) as licorice. Color code: N, blue; O, red; C, cyan; H, white; and P, brown .) Water and ions are o mitted for clarity. During the simulations of Syt7C2A WT sometimes strand was represented by a loop like in (A) due to its short sequence containing K186 L187 E188 residues Figure taken from Fig. S2 in Ref. 21

PAGE 33

24 2.2. Data Analysis 2.2.1. Area per Lipid (APL) and Lipid Order Parameter ( S CH ) The area per lipid (APL) describes the average size taken up by a lipid leaflet area in the membrane For the 3:1 PC/PS membrane the APL value was obtained by dividing the average area of one leaflet in the membrane by the total number of lipids in the leaflet ( N lip = 128) as seen in Equation 1 : Equation 1 where x t and y t are the dimensions of the membrane in the x y plane at the instantaneous time t O nly the reference APL values for the pure PC ( APL POPC ) and PS ( APL POP S ) are available from previous experi mental 48 and computational 49 50 studies ( Table 6 ) which were used to estimate the reference APL for the 3:1 PC/PS mixture membrane with the following expression : Equation 2 Table 6 Comparisons of the a rea p er l ipid (AP L) in 2 between the literature and s imulated m embranes Pure 3:1 PC/PS POPC POP S E xperiment 64.3 1.3 a N/A N/A Computation 65.5 1.1 b 55 1.0 c 62.9 0.9 d / 62.6 0.9 e a From R ef. 48 b From R ef. 49 c From R ef. 50 d The estimated APL using the values from b and c e The calculated APL of the 3:1 PC/PS membrane using the last 120 ns simulation.

PAGE 34

25 The order parameter ( S CH ) is a measure of structural orientation or flexibility of lipids in a membrane. The lipid head groups, which are charged, tend to interact with other head groups, ions or water molecules restrict ing their motions. On the other hand the motions of the tails occur in a more fluid environment T hus the order parameters of different carbons in the membrane give the information about the orientation anisotropy The value of S CH wa s evaluated by measur ing the spatial orientation of a C H bond, described by the angl e be tween the C H vector and the membrane normal ( Eq 3 ) Equation 3 where denotes the average cosine square angle s between each C H bond in a m ethylene group (or a methane group near cis double bond) and the membrane norm over all lipids. 2.2.2. Poisson Boltzmann Calculation The Poisson Boltzmann (PB ) calculations for the protein structures were performed using the Ad aptive Poisson Boltzmann Solver 51 Two sets (Ca 2+ bound vs. Ca 2+ free) of PB calculations were performed for the 3 Ca 2+ bound WT and CH C2 domains of S yt proteins. The Ca 2+ bound Syt structures for C2 domains were provided by utilizing the last snapshot of the protein standalone simulations, followed by the generation of the PQR (PDB data with the charge and radius parameters ) files at a p H of 7.4 with C HARMM 22 force field Finally the electrostatic potential s were described in 0.15 M of KCl.

PAGE 35

26 CHAPTER II I R ESULTS AND DISCUSSION 3.1. Simulations of the S tandalone 3:1 PC/PS Membrane The standalone simulation for 3:1 PC/PS membrane w as performed in water with the counter ions of 64 Na + ions for 160 ns in order to calculate t he area per lipid (APL) and order parameter ( S CH ) for the membrane ( Figure 10 ) Because the instantaneous APL dropped quickly in the first 10 ns and became stable afterwards only the data from the last 120 ns of the trajectory was used to calculate the average APL and S CH T he average APL was 62.6 2 and was similar to the APL of 62 .9 2 estimated by utilizing the literature data taken from the NMR measurements ( Table 6 ) 48 49 50 The S CH for each C in the palmitoyl and oleo yl chains ( Figure 2 ) of the membrane model were computed using the same data set as APL calculations. In Figure 10 B S CH g enerally f e ll between those for the pure POPC and pure POPS. 52 53 Note that for the oleoyl chain the re a re only a few reference S CH Nevertheless the calculated values still agreed with the literature values Overall our results of APL and S CH agreed well with the literature values suggesting that the simulated membrane structure was well equilibrated af ter 40 ns.

PAGE 36

27 Figure 10 ( A ) Area per lipid (APL) as a function of simulation time for the standalone 3:1 PC/PS membrane for 160 ns. ( B ) Order parameter ( S CH ) as a function of carbon atom index of the palmitoyl (top) and oleoyl (bottom) chains. The values from the NMR measurements (pure POPC 52 in green and POPS 53 in blue curves) and this study (black curves) were compared. Only the values of S CH for the palmitoyl chain in a pure POPS membrane were available. Figure modified from Fig. S3 in Ref. 21 3.2. Simulations of the S tandalone C2 Domains Both C2 domains of Syt1 and Syt7 have the ability to bind to a phospholipid membrane in the presence of Ca 2+ ions 54 55 56 The exact number of bound Ca 2+ ions to the Ca 2+ binding loops (CBLs) are unknown, but up to four Ca 2+ ions can bind to the CBLs. 23, 30, 57 The NMR and crystal structures were used for modeling WT C2A and C2B domains for S yt1 and Syt7 to study those different Ca 2+ sensitivities 22 26 2 7 31 Experimentally, Ca 2+ bound struc ture s are available for each domain of Syt1 (Syt1C2A WT 3Ca 2+ and Syt1C2B WT 2Ca 2+ ) and C2B domain of Syt7

PAGE 37

28 (Syt7C2B WT 3Ca 2+ ), but not for Syt7 C2A Therefore, the two models for Ca 2+ bound Syt7 C2A were created by adding two and three Ca 2+ ions to the Ca 2+ free Syt7 C2A experimental structure (Syt7C2A WT 2Ca 2+ and Syt7C2A WT 3Ca 2+ ) respectively Furthermore, one Ca 2+ ion was added in Syt1C2B WT 2Ca 2+ model to build the 3 Ca 2+ bound Syt1 C2B model (Syt1C2B WT 3Ca 2+ ), as described in Computational Methods 2.1.1.2 A series of homologous CH C2 domains (Syt7:1C2A CH 3Ca 2+ Syt1:7C2B CH 2Ca 2+ and Syt1:7C2B CH 3Ca 2+ ) w as also created to provide additional information on the role of three loops in Ca 2+ bind ing 31 26 3.2.1. Comparisons of 2 Ca 2+ with 3 Ca 2+ M odels in E ach C2 WT D omain For Syt7C2A WT 2Ca 2+ and 3Ca 2+ and Syt1C2B WT 2Ca 2+ and 3Ca 2+ models, t he initial positions of the first two Ca 2+ ions in the CBLs were the same However their positions were rearranged during the minimization s leading to the different ways of coordination with the protein oxygen atoms in the CBLs ( Figure 6 ) I n the Syt7C2A WT 3Ca 2+ model, t he first Ca 2+ ion was coordinated by the side chains of D166, D172, D225, and D227, the backbone oxygen atom of Y226, and one water molecule ; the second Ca 2+ ion by the side chains of D166, D225, D227, and D233, the backbone oxygen atom of K165, and one water molecule ; and the third Ca 2+ ion by the side chains of D227, S230, and D233 and the backbone oxygen atom of R231. In the Syt7C2A WT 2 Ca 2+ model, however, the second Ca 2+ ion moved toward s CBL3, so one extra water molecule entered the empty space. The overall conformation of the side chain s in CBL3 was very different from that in the 3 Ca 2+ model. T he first Ca 2+ ion was coordinated by the side chains of D166 and D172, the backbo ne oxygen atoms of K165 and Y226, and two water

PAGE 38

29 molecules The second Ca 2+ ion was coordinated solely by the side chains of D225, D227, S230, and D233. The C2B domains of Syt1 with 2 and 3 Ca 2+ are illustrated in Figure 6 B I n the 3 Ca 2+ model of Syt1 C2B domain, t he first Ca 2+ ion was coordinated by the side chains of D 303 D 309 D 363 and D 365 the backbone oxygen atom of Y 364 and two water molecules ; the second Ca 2+ ion by the side chains of D 303 D 363 D 365 and D 371 the backbone oxygen atom of M302 and two water molecules (n ote that one of the water molecules here was also shared by the first Ca 2+ ion ) ; and the third Ca 2+ ion by the side chains of D 365 and D371 the backbone oxygen atom of K369, and three water molecules In this model, the position of the third Ca 2+ ion is located somew here between the binding pocket and the bulk water, as the ion interacts with less protein oxygen atoms (three) than in Syt1 C2A (four) or Syt7 C2B (four). The loose binding of the third Ca 2+ ion is primarily due to t he absence of a serine residue in CBL3 of Syt1 C2B (S235 in Syt1 C2A, S230 in Syt7 C2A, and S362 in Syt7 C2B), which plays a key role in trapping the third Ca 2+ ion inside the CBLs in Syt1 C2A and Syt7 C2AB The importance of this serine residue will be covered in Section 3.2.2. Comparisons of WT and CH D omains In the 2 Ca 2+ model, just like in the Syt7C2A WT 2Ca 2+ model the second Ca 2+ ion moved toward s CBL3 and allowed one water molecule to come in by creating more space. T herefore, t he first Ca 2+ ion was surrounded by the side chains of D 303, D 309 D363, and D365, the backbo ne oxygen atom of Y364 and two water molecules and the second Ca 2+ ion by the side chains of D 303 D 363 D365, and D 371, the backbone oxygen atom of M302, and two water molecules Un like the Syt7C2A WT 2Ca 2+ model, here the solvation shells of the first and second Ca 2+ ion s do not share the same water molecules

PAGE 39

30 After minimization, a 200 ns trajectory for e ach of the three C2A domains (Syt1 C2A WT 3Ca 2+ Syt 7 C2A WT 2Ca 2+ and Syt 7 C2A WT 3Ca 2+ ) and a 100 ns trajectory for each of the three C2B domains ( Syt1 C2 B WT 2Ca 2+ Syt 1 C2 B WT 3Ca 2+ and Syt 7 C2 B WT 3Ca 2+ ) were generated The overall conformation of the protein was quite stable, as can be seen from the consistently small root mean square deviation ( RMSD ) with respect to the starting geometry of the protein backbone N, C, O, and C as a function of the simulation time ( Figure 11 A ) As listed in Table 7 the average RMSD of the backbone atoms are 1.1 for Syt1 C2A WT 3Ca 2+ 2.6 for Syt7C2A WT 2Ca 2+ and 1.8 for Syt7C2A WT 3Ca 2+ The smallest average RMSD value and the fluctuations (0.1) were found in Syt1C2A WT 3Ca 2+ model implying that the Syt1 C2A model which started from the Ca 2+ bound NMR structure was the most stable during the simulation time The largest RMSD fluctuations occurred in the 2 Ca 2+ model of Syt7 C2A in which two Ca 2+ ions in t he CBL s moved substantially compared to the equivalent Ca 2+ ions in the 3 Ca 2+ model of Syt7 C2A ( the respective RMSD values are 3.2 and 1.8 for the two Ca 2+ ions in the 2 Ca 2+ model and 1.0 and 0.5 in the 3 Ca 2+ model ). Recall that the NMR structure for Syt7 C2A was Ca 2+ free. The addition of multiple Ca 2+ ions into the Ca 2+ free Syt7 C2A model s required the adaptations of the protein, leading to the greater protein movements than observed in other models In the case of Syt1C2B WT 3Ca 2+ model, only the third Ca 2+ ion from Syt7C2B WT 3Ca 2+ was added to Syt1C2B WT 2Ca 2+ likely requiring less protein conformational changes For the C2B domains, t he average RMSD of the backbone atoms and the fluctuations over simulation time are substantially small er ( 1.4 for Syt1C2B WT 2Ca 2+ 1.1 for Syt1C2B WT 3Ca 2+ and 0.9 for Syt7C2B WT 3Ca 2+ ) than for the C2A domains The RMSD value for the 2 Ca 2+ model of Syt1C2B WT was larger than for the 3 Ca 2+ model but the difference is minor. Interestingly, the first and second Ca 2+ ions in the CBLs of the 2 Ca 2+ model

PAGE 40

31 exhibited smaller movement s compared to those of the 3 Ca 2+ model ( RMSD values for the two Ca 2+ ions in the 2 Ca 2+ model are 0.4 and 0.5 and RMSD values in the 3 Ca 2+ model are 0.9 and 0.6 ), in contrast to the 2 and 3 Ca 2+ models of Syt7C2A WT T he Ca 2+ ion different stabilit ies in the Syt7C2A WT and Syt1C2B WT were found in agreement with experiments 26 30 Figure 11 Root mean square with respect to the first frame of the simulation, as a function of simulation time for all ( A ) standalone protein s (top C2A ; bottom C2B) and ( B ) Syt7 C2A membrane complexes models ( A ) The traces are shown for both C2 domains of Syt1 WT 3Ca 2+ i n red and Syt7 WT 3Ca 2+ in navy. For 2 Ca 2+ models for Syt7C2A WT and Syt1C2B WT the traces are represented by light blue a nd pink respectively. For the chimeric (CH) domains, Syt7:1 C2A CH is shown as green, Syt1:7 C2B CH 2Ca 2+ as gray and Syt1:7 C2B CH 3Ca 2+ as black. The simulation times are different per domain; C2A WT for 200 ns and C2B WT and all CH for 100 ns. ( B ) The t races are shown for two Syt7C2A pre in sertion models; lying down in red and standing up in yellow, and for two Syt7C2A embedded models; 1 st in green and 2 nd in navy for 500 ns each simulation.

PAGE 41

32 Table 7. Root mean square deviations (RMSD) in averaged over the trajectories for selected atoms in the simulations of standalone Syt1 and Syt7 C2A domains. The maximum RMSD values are given in parentheses. a Backbone b First Ca 2+ Second Ca 2+ Third Ca 2+ Standalone C2A WT Syt1 3Ca 2+ 10 ns c 1.0 0.1 (1.4) 0.6 0.2 (1.1) 0.6 0.2 (1.1) 3.8 1.2 (6.1) 200 ns d 1.1 0.1 (1. 5 ) 0.5 0.2 (1.1) 0.5 0.2 (1.3) 4.1 0.3 (6.0) Syt7 2Ca 2+ 10 ns c 1.9 0.5 (2.8) 1.7 0.5 (3.0) 1.2 0.4 (2.4) N/A 200 ns d 2.6 0.2 (3.5) 3.2 0.6 (5.1) 1.8 0.5 (3.6) N/A Syt7 3Ca 2+ 10 ns c 1.6 0.4 (2.3) 0.8 0.2 (1.4) 0.7 0.3 (1.6) 1.6 0.8 (3.6) 200 ns d 1.8 0.1 (2.4) 1.0 0.2 (1.8) 0.5 0.2 (1.2) 2.4 0.4 (4.0) Standalone C2B WT Syt1 2Ca 2+ 100 ns e 1.4 0.1 (1.8) 0.4 0.2 (1.2) 0.5 0.2 (1.3) N/A Syt1 3Ca 2+ 100 ns e 1.1 0.1 (1.4) 0.9 0.2 (1.7) 0.6 0.2 (1.4) 1.6 0.5 (3.2) Syt7 3Ca 2+ 100 ns e 0.9 0.1 (1.2) 0.5 0.2 (1.0) 0.5 0.2 (1.1) 0.5 0.2 (1.4) Standalone CH Syt7:1 C2A 3Ca 2+ 100 ns e 1.9 0.1 (2.4) 1.1 0.2 (1.7) 1.4 0.2 (2.1) 3.6 0.6 (5.4) Syt1:7 C2B 2Ca 2+ 100 ns e 0.9 0.1 (1.2) 0.6 0.2 (1.3) 0.6 0.2 (1.5) N/A Syt1:7 C2B 3Ca 2+ 100 ns e 1.2 0.1 (1.5) 0.5 0.2 (1.3) 0.6 0.2 (1.3) 2.8 0.3 (3.9) Syt7C2A WT membrane Pre insertion Lying down 500 ns f 1.5 0.1 (2.2) 0.4 0.2 (1.2) 0.4 0.2 (1.4) 2.0 0.4 (4.7) Stand up 500 ns f 1.4 0.1 (2.0) 0.4 0.2 (1.0) 0.4 0.2 (1.1) 2.4 0.3 (3.8) Embed First 500 ns f 1.1 0.1 (1.6) 0.5 0.2 (1.4) 0.5 0.2 (1.3) 1.1 0.5 (2.8) Second 500 ns f 1.2 0.2 (1.9) 0.5 0.2 (1.3) 0.7 0.2 (1.6) 1.1 0.3 (2.3) a Mean S.D. calculated with the initial structure as reference and averaged over the trajectory o f 10 ns and 200 ns simulations for the protein standalone models, and 500 ns simulations for the protein membrane complex models b Protein backbone atoms were defined as N, O, C, and C c Calculated using the last 6 ns of trajectories of 10 ns standalone SytC2A WT simulations. d Calculated using the last 8 0 ns of trajectories of 200 ns standalone SytC2A WT simulations. e Calculated using the last 2 0 ns of tr ajectories of 1 00 ns standalone SytC2B WT and chimeric C2 domain simulations. f Calculated using the last 400 ns of trajectories of 5 00 ns standalone Syt7C2A membrane simulations.

PAGE 42

33 The coordination of three Ca 2+ ions in the CBLs per C2 WT domain closely resembles each other as depicted by the distances between the Ca 2+ ions and their coordinating oxygen atom s from the domains as a function of simulation time ( Figure 12 ) The first two Ca 2+ ions in both C2A WT models were relatively stable in the binding site, maintaining approximately constant distances with the chelating protein oxygen atoms most of the time Substantial conformational change only occurred early in the simulations. For example, at t = 3 ns, the distance of the first and second Ca 2+ ions with the carboxyl oxygen atom of the D227 in Syt7C2A WT (the red curves with markers in Figure 12 A ) quickly drop ped from 6 to 4 and from 4 to 2 respectively As the oxygen atoms in D227 side chain which was initially located in the middle of the second and third Ca 2+ ions ( Figure 6 A bottom panel ), went closer to the first and second Ca 2+ ions, one water molecule diffused in and took the position The addition of this water molecule initiated the conformational changes of S230 and R231, resulting in the distance s of the third Ca 2+ ion with the hydroxyl oxygen atom in S230 and with carbonyl oxy gen atom in R231 increase from 2 to 6 and from 2 to 5 respectively, at t = 6 ns. S ubsequently up to 5 water molecules moved near the third Ca 2+ ion during the simulation. The average distances between three Ca 2+ ions and their coordinating protein oxygen atoms are provided in Table 4 T he se results indicate that the first two Ca 2+ ions in the Syt1 an d Syt7 C2A WT models are strongly enfolded in the binding site while the third Ca 2+ ion binds less tightly This interaction occurs primarily through interactions with the side chain oxygen atoms of the D238 in Syt1 and the D233 in Syt7 and secondarily with t he backbone oxygen atoms of the K236 in Syt1 and the R231 in Syt7. Th e structural and functional similarit ies be tween two C2A WT domains of Syt1 and Syt7 propose that Syt7C2A WT binds three Ca 2+ ions in a similar way to that of Syt1 C2A WT as suggested by a previous NMR measurement ( Figure 13 top panel ) 30 T he RMSD plot for the

PAGE 43

34 third Ca 2 + in Syt7C2A WT as a function of the 200 ns simulation time shows the stable binding in the CBLs. T herefore, the 3 Ca 2+ model for Syt7C2A WT w as used for the protein membrane association simulation in this study Next, f or the C2 B WT models, the first two Ca 2+ ions were substantial ly held by the coordinating oxygen atoms, as evidenced by the relatively consistent distances between the atoms ( Figure 12 B ) The third Ca 2+ ion in Syt1C2B WT model, however, behaved very differently to that in Syt7C2B WT model I n Syt7C2B WT the side chain oxygen atom of the S362 clung tightly to the third Ca 2+ ion in collaboration with the side chain oxygen atoms in the D359 and D365, and the backbone oxygen atom in the R363. A s mentioned earlier t his serine is not found in Syt1C2B WT T he third Ca 2+ ion in Syt1C2B WT was surrounded by five water molecules from the beginning of the equilibration ( three water molecules after minimization, Figure 6 B middle panel ) For comparison, the third Ca 2+ ion in Syt7C2B WT was surrounded by two water molecules ( one water molecule after the minimization, Figure 6 B bottom panel ). Both the previous experiment 26 and thi s computation ( Figure 13 bottom panel ) suggest that Syt1C2B WT exhibits a low affinity for the third Ca 2 + ion

PAGE 44

35 Figure 12 Distances between each Ca 2+ ion and its coordinating protein oxygen atoms, as a function of time of standalone ( A ) C2A and ( B ) C2B domains with three Ca 2+ ions. In each column (left, Syt1 WT ; center, Syt7 WT ; right, chimeric domains) the top, middle, and bottom panels are for the first, second, and third Ca 2+ ions, respectively The consistent c olor code s are used for the equivalent residue atoms in the order of Syt1C2A Syt7C2A Syt7:1 CH C2A Syt1C2B Syt7C2B Syt1:7 CH C2B domains : L171 K165 L165 M302 M296 M302 O, black; D172 D166 D166 D303 D297 D303 OD1, red; D172 D166 D166 D303 D297 D303 OD2, green; D178 D172 D172 D309 D303 D309 OD2, gray; D230 D225 D225 D363 D357 D363 OD1, blue; D230 D225 D225 D363 D357 D363 OD2, purple; F231 Y226 F226 Y364 K358 K364 O, black with markers; D232 D227 D227 D365 D359 D365 OD1, red with markers; D232 D227 D227 D365 D359 D365 OD2, green with markers; S235 S230 S230 N/A S362 S368 OG, gray with markers; K236 R231 K231 K369 R363 R369 O, brown with markers; D238 D233 D233 D371 D365 D371 OD1, blue with markers; D238 D233 D233 D371 D365 D371 OD2, purple with markers.

PAGE 45

36 Figure 13 Root mean square deviation (RMSD) of the third Ca 2+ ion with respect to the first frame of the simulation, as a function of simulation time for all 3 Ca 2+ models of WT and CH C2A domains for 200 ns ( top ) and C2B domain for 100 ns simulations ( bottom ). The traces are shown for both C2 domains of Syt1 WT i n red and Syt7 WT in green, and chimeric domains in yellow. Each C2A WT and C2B WT domain in Syt1 and Syt7 displays generally similar electrostatic potentials from Poisson Bolt zmann cal culations ( Figure 14 ) The Ca 2+ binding sites in all domains have negative potentials in the Ca 2+ free structures due to exposed aspartate residues and backbone oxygen atoms in the CBLs ( Figure 6 ) However this region has a positive potential after three Ca 2+ ions are bound o n the me mbrane binding Moreover, t he region from 3 loop to CBL2 has a positive potential in both of the Ca 2+ free and Ca 2+ bound forms due to a cluster of lysine and arginine residues T he net charg e of this region is rather positive ( Table 2 ) The positive potentials of the CBLs in the presence of Ca 2+ ions and the 3 4 strands with the

PAGE 46

37 adjacent loops promote the anionic membrane protein association through electrostatic interactions. 15, 23 58 S ubtle differences still existed between do mains. For the two C2A WT domains, the overall electrostatic potential of Syt1 C2A is more negative than that of Syt7 C2A This is consistent with the protein net charges reported in Table 2 : 2 for Syt1 and +3 for Syt7 C2A WT Particularly, the CBL1 region in Syt1 has a net charge of 1 while i n Syt7 it is charge neutral. This lead s to the more negative potentials in CBLs of Syt1C2A WT (black arrows in Figure 14 ). Similarly Syt1 C2 B is less positive than Syt7 C2 B : t he total charges are +6 for Syt1C2B WT and +12 for Syt7C2B WT However a n heli x in the C terminus of Syt1C2B WT ( but not in Syt7C2B WT ) has a net charge of 4, making its entire C2B domain less positive (green arrows in Figure 14 ) If only from 1 to 8 ( with interconnecting loops ) in those two C2B WT domains are compared the net charges will be +11 for Syt1C2B WT and +12 for of Syt7C2B WT Overall, Syt7 C2AB WT domains have more positive potentials than Syt1 C2AB WT domains, implying that Syt7 possibl y possesses greater electrostatic interactions with anionic lipid membranes in the presence of Ca 2+ ions 56, 59

PAGE 47

38 Figure 14 Electrostat ic potential isosurfaces for Syt C2AB and Slp4 C2A domains T he protein s are shown strands colored orange, C a 2+ ions as yellow spheres T he potential maps computed with and without the three Ca 2+ ions for Syt proteins B lue for +50 mV and red for 50 mV equipotential contours ( assuming 0.15 M KCl and pH 7.4 ).

PAGE 48

39 3.2.2. Comparisons of WT and CH D omains One C2A and two C2B chimeric ( CH ) mo dels were created in 0.15 M KCl solution The C2A CH model retained the Syt7 body with Syt1 CBLs while the C2B CH models retained the Syt1 body with Syt7 CBLs specifically for the purpose of corresponding with the collaborative experiments 35 For C2A CH model, both Syt1 and Syt7 C2A domains require to have three Ca 2+ in CBLs, so only a 3 Ca 2+ model was built. However, Syt1 and Syt7 have two and three Ca 2+ ions in the binding sites of C2B domains T hus 2 and 3 Ca 2+ models were generate d for C2B CH models. Figure 11 A illustrates the overall protein stabilities of the CH domains The average RMSD values are all smaller than 2 : 1.9 for Syt7:1C2A CH 3Ca 2+ 0.9 for Syt1:7C2B CH 2Ca 2+ and 1.2 for Syt1:7C2B CH 3Ca 2+ ( Table 7 ) The ~0.6 of the average RMSD for both the first and second Ca 2+ ions in the two Syt1:7C2B CH models indicates that t he major difference was observed for the third Ca 2+ ion s in Syt1:7C2B CH 3Ca 2+ Furthermore, t he third Ca 2+ ions in Syt7:1C2A CH 3Ca 2+ and Syt1:7C2B CH 3Ca 2+ models have even greater RMSD values than the ir first two Ca 2+ ions (respective RMSD values are 3.6 and 2.8 for the third Ca 2+ ions) The fluctuation s of the RMSD for the third Ca 2+ ion in C2A CH however, is doubled in magnitude of 0.6 compared to the fluctuation s for the third Ca 2+ ion in Syt1:7C2B CH 3Ca 2+ (0.3) implying that the third Ca 2+ ion binding in C2B CH is more stable than C2A CH I n Figure 13 top panel the first big jump in the C2A CH RMSD plot appear s at t = 6 ns for the oxygen atoms of S230 side chain and K231 backbone Those oxygen atoms moved away from the third Ca 2+ ion as two water molecules diffused in to the site rapidly (the distance s between the third Ca 2+ ion and the oxygen atoms doubled from 2 to 4 ). At t = 78 ns, the second jump occurred when one water molecule

PAGE 49

40 approached the third Ca 2+ ion from bulk causing the side chain of D227 to slightly move away from the third Ca 2+ ion to wards the first and second Ca 2+ ions. This interaction has been observed in C2A WT as mentioned earlier. The two carboxyl oxygen atoms in D233 closely interacted with the third Ca 2+ ion but o ne of the oxygen atom s began to interact with the second Ca 2+ ion after the last 5 ns of the simulation. The overall conformation of CBL3 is somewhat similar to the two 3 Ca 2+ C2A WT models B ut longer simulation s are required to compare the C2A domains between the WT and CH proteins to reduce the statistical uncertainti es O verall the RMSD for the third Ca 2+ ion in the C2B CH model resembles that for Syt1 C2B WT but the ion s solvation shell is different in the CH and WT model s In the first snapshot of the C2B CH equilibration, the third Ca 2+ ion was surrounded by the side chain oxygen atoms in D3 65 S3 68 and D3 71 the backbone oxygen atom in R36 9 and three water molecules. An early RMSD change occurred a t t = 9 ns when the aspartate residue (D365 in Syt1:7C2B CH ) mov ed toward s the first and second Ca 2+ ions and two water molecule s diffus ed in to fill the evacuated space Subsequently t he side chain oxygen atom in S36 8 moved away from the third Ca 2+ ion while the backbone oxygen atom in R36 9 came nearer to the third Ca 2+ ion. As the distances between the third Ca 2+ ion and D3 65 and S36 8 became longer more water molecules entered the binding site, and the third Ca 2+ ion was coordinated by a total of 6 water molecules in 20 ns. After that, two of these water molecules w ere replaced by Cl ion s at t = 20 and 40 ns. This Cl recruitment was only detected in C2B CH model, not in C2A CH and we found that the third Ca 2+ ion in Syt1C2B WT model also recruit ed one Cl ion during the simulation The details will be discussed more in Section 3.4. Coordination of the Ca 2+ Ions in the Standalone Syt Proteins and Syt7C2A WT membrane Complexes

PAGE 50

41 The side chain of S368 in Syt1:7C2B CH ( G368 was in Syt1C2B WT instead ) showed t wo orientations in the simulations ( Figure 15 ) R otations of the side chain in S368 caused the distances between the third Ca 2+ ion and its hydroxyl oxygen atom (OG) to vary The distances tended to be longer when the hydroxyl group pointed in ward from the CBL region, characterized by a gauche (+) conformation with 60 of the 1 angle in S368 ( Figure 16 ). As the side chain exhibits the gauche (+) conformation, one more water molecule diffused in between the hydroxyl oxygen atom and the third Ca 2+ ion But when the serine residue had the trans conformation (180 o f the 1 angle), the oxygen atom helped to hold the third Ca 2+ ion, preventing it s escaping from the binding site into the bulk water. Th ese changes impact the RMSD of the third Ca 2+ ion ( Figure 13 bottom panel ). Figure 15 Two orientations of the residue S368 that observed during the standalone Syt1C2B CH 3Ca 2+ model when the side chain of S368 faced ( A ) inward and ( B ) outward the CBLs (green circles) The orientations of S368 represent the gauch e (+) conformation in (A) see Figure 16 T he t hree Ca 2+ ions and were in green spheres with surrounding residues in CBL3 ( sticks : N, blue; O, red; C, cyan ), and the coordinating oxygen atoms (red sphere) and Cl ions (blue sphere) for the third Ca 2+ ions are shown. Protein H is omitted for clarity The number under the dotted line in (A) and (B) indicates that the distance in between the third Ca 2+ ion and the hydroxyl oxygen atom in the serine residue.

PAGE 51

42 Figure 16 ( A ) D istances between the side chain oxygen atom (OG) of S368 and the third Ca 2+ ion (gray curves with markers) and the backbone nitrogen atom (N) of S368 (blue curves) in Syt1:7C2B CH model. ( B ) T he dihedral 1 angles (N C C OG) for the sam e S368 The serine residue had mostly trans conformations but had g auche (+) at ~50 ns when the angles of ~60 3.3. S imulations of Syt7 C2A Protein membrane Association Th e membrane targeting mechanisms of C2A domains in Syt7 require detection of multiple signals such as Ca 2+ ions and target lipids ( usually anionic phospholipids ) 47 25 60 T he MD simulations of 500 ns were performed for each of the four Syt7C2A WT membrane complexes ( two pre insertion models and two embedded models ) T he initial structure of Syt7C2A WT in the pre insertion stand ing up model was used to build the embedded models.

PAGE 52

43 3.3.1. Comparisons of the Pre insertion with the Embedded Syt7C2A WT m embrane In order to characterize the orientation of the membrane associated Syt7 C2A WT the protein tilt angle s and the penetration depth s of F167 and F229 were calculated as a function of simulation time ( Table 8 ) The tilt angle is defined to be the angle between the membrane plane and the norm of a protein plane containing the three centers of mass (COM ) of the three outermost ( Figure 8 A ) T he average til t angles were 13.5 and 37 .1 fo r th e pre insertion l ying down and stand ing up models and 32.8 and 32.2 for two embedded models respectively There were no significant angle differences between each model except the lying down model These angles indicated that the ly ing down and standing up protein s were relatively more parallel and perpendicular to the membrane surface respectively S trikingly as the Syt7C2A WT domain start ed to lay down, it tilt ed to wards the side that contains the 3 4 strands wh ich are associated with positive electrostatic potentials (see Section 3.2.1 Comparisons of 2 Ca 2+ with 3 Ca 2+ M odels in E ach C2 WT D omain ) We found s trong interactions of this polybasic region ( formed by the cluster of the lysine residues ) with the lipid head group in all simulations Interestingly, this polybasic region is identified to be essential for the strong Ca 2+ dependent binding of Syt1 to PS lipid containing membrane 12 The protein orientation was general ly inferred from the tilt angle calculation while the penetration depth calculation was used to describe which residue s have significant interacti ons with membrane. The depths of the critical residues are defin ed by the distances between the ir C atom and the average phosphate (PO 4 ) plane of the membrane as exemplified in Figure 8 B for F167 in CBL1 and F229 in CBL3 A p ositive penetration depth indicates the C atom of the residue i s present on the lipid face of the average PO 4 plane It is evident that F167 often pass ed in and out of the membrane interface This type of motion of F167 was more common in the pre insertion

PAGE 53

44 stand ing up and two embedded models ( Figure 17 A ). In the ly ing down model, F167 remain ed in the aqueous phase all the time following the initial position ( initial depth : 15.9 in the bulk water ). For F229 the average penetrated depths are similar in all models: 0.2 for the l ying dow n and 2.2 for the stand ing up pre insertion models and 2.8 for the first and 2.4 for the second embedded models, respectively, suggesting that F229 in CBL3 drives the penetration process prior to F167 in CBL1. In Figure 18 the penetration depths for the pre insertion models of individual residues at various simulation times were plotted for the two pre insertion models The overall trend of the protein as a whole approaching the average PO 4 was evident when compar ing the trace s of the first 50 ns (red curves) with those of the last 50 ns (purple curves). However, the actual penetration process is far more complicated. For example, in the lying down model, the protein moved back to the aqueou s phase after 300 ns while CBL1 and the polybasic region of 3 4 strands moved near the membrane D uring the last 1 00 ns, CBL1 constantly returned to the bulk water while the polybasic region was stuck by the interaction with the lipid head groups For the standing up model, the depths for CBL1 (particularly for F167) and the polybasic region were similar at ~250 ns. After that, CBL1 started to anchor deeper to the membrane surface while the polybasic region moved back to wards the bulk water In contrast, t he region s of CBL3 in both pre insertion models exhibited steep increase in depth ( approach ing and bind ing to membrane ) over time ( Figure 18 ).

PAGE 54

45 Table 8 The averaged (final 400 ns from 500 ns simulations ) tilt angles and the penetration depths with respect to the membrane surface. The initial values are given in parentheses The selected atoms are positioned below the average PO 4 of the membrane when the depth values are positive (depth of avg. PO 4 = 0). T ilt angle ( ) D epth ( ) F167 F229 Pre insertion L ying down 13.510.0 (23.1) C a 9.82.5 ( 15.9) 0.22.2 ( 6.9) phenyl b 7.8 2.7 ( 14.1) 3.12.3 ( 4.1) S tanding up 37.1 9.5 (36.1) C a 3.13.3 ( 7.1) 2.22.0 ( 4.2) phenyl b 1.03.9 ( 4.9) 4.92.0 ( 3.9) E mbedded 1 st model 32.8 14.4 (48.2) C a 2.03.0 ( 1.3) 2.81.6 (2.3) phenyl b 0.93.1 (1.8) 4.81.7 (5.4) 2 nd model 32.2 9.1 (34.7) C a 2.62.7 ( 1.4) 2.41.8 (1.3) phenyl b 0.23.2 (1.4) 4.31.9 (4.9) a The penetration depths were calculated using the C position. b The penetration depths were calculated using the center of mass of the side chain phenyl ring.

PAGE 55

46 Figure 17 The t ilt angles (black) and penetration depths of the protein center of mass (gray) and the critical residues ( F167 in red and F229 in blue ) for different Syt7C2A WT membrane models over the 500 ns simulation time Note that the penetration depth of 0 corresponds to the position of the average PO 4 plane. Figure taken from Fig. 7 in Ref. 21

PAGE 56

47 Figure 18 P enetration depths for the lying down ( top ) and standing up ( bottom ) Syt7C2A WT pre insertion models, as a function of the protein sequence. Each of 10 traces per plot represent s on the average position of the residue for 50 ns increments of the 500 ns trajectory. The rainbow colored order indicates the simulation ti me (red for the first 50 ns to purple for the last 50 ns). Three pink regions on the graph correspond to the regions of the CBLs (left to right, CBL1 to 3) and the blue region for the polybasic region of the 4 strand.

PAGE 57

48 Interestingly, all simulations depicted large swinging motions of the protein between a more parallel (smaller tilt angle) and a more perpendicular (greater tilt angle) orientation ( Figure 17 ) F or two pre insertion models, even with large swinging motion occurred ( angle > + 40 at t = ~170 ns for the lying down and t = ~70 ns for the standing up models ), the stable bin ding of F229 maintained. For the two embedded models, the large swinging motion s ( angle > +50 ) appeared at t = ~70 ns for the 1 st embedded and t = ~140 ns for the 2 nd embedded models but a gain F229 stayed inside the membrane. T he overall stability of F167 binding to the membrane was different from the 1 st to 2 nd embedded models It is likely th at the insertion is only partial ly completed over the 500 ns simulation time Tilt angles were correlat ed with F167 penetration depths ( Figure 19 ) I t is less predominant for the tilt angle and the depth of F229 because of the stability of F229 in membrane Visual inspection of the trajectories revealed that a s the depth of F167 bec a me more positive after 200 ns the tilt angle increased T he protein orientation in the pre insertion standing up model began to resemble the two embedded models. When the depths of F167 approached 0 with approximately 45 of the tilt angles th e correlation manifested in all but the lying down model Th e polybasic region of the protein s experienced strong electrostatic attractions to the negatively charged head groups in PS lipids. T hus CBL1 ( represented by F167 ) moved away from the membrane leading to the more parallel protein orientation On the other hand, when the protein tilt angle increased the interactions between F167 in CBL1 and the hydrophobic environment of the membrane also increased The competition between the electrostatic attr actions (the polybasic region near 4 strand with the lipid head groups) and the hydrophobic interactions (F167 in CBL1 and F229 in CBL3 with the lipid acyl tails) led to seesaw like movements of Syt7C2A WT domain with F229 as the fulcrum

PAGE 58

49 Th is type of seesaw like movement of Syt7 C2A WT domain was also observed in the two embedded model simulations. We counted the interacting pair of the heavy atoms in protein and lipids as a simple way to measure the electrostatic attraction s (by positively charged resid ues ) and the hydrophobic interaction s (by F167 and F229 ) W hen the distance was shorter than 5.0 th e pair was counted ( Table 9 ) The residues in the CBLs ( mostly in CBL1 and CBL 3) produced most of the lipid contacts in line with the depth calculations. Two residues at the flexible C terminal loop S264 and G265, are treated as exception s because they belong to the connecting loop between C2A and C2B domains Both F167 and F229 were surrounded by the hydrophilic residues (D166 and S168 for F167 ; R228, S230, and R231 for F229). The sums of the lipid head group and acyl chain contacts were 66.8 and 64.0 for F167 and F229 respectively Among these contacts, about 30 % and 70 % are from the acyl chain carbon atoms for F167 and F229, re spectively More specifically, F229 had the lipid acyl chain contacts of 45.5 which is the greate st among all residues, because the aforementioned residue had the deepest penetration depth (2.6 ). The second greatest 19.9 is found for for F167 Un like F229, F167 stayed above the membrane on average, but the big standard deviation (2. 0 ) of its penetration depth This is because F167 frequent ly travel ed into and out of the acyl chain region due to the seesaw like movement s described above.

PAGE 59

50 Table 9 The protein residues with the largest numbers of lipid contact, averaged from the two embedded model simulations. a Residue Lipid contacts Depth ( ) Head group Acyl chain F167 46.9 19.9 2.3 2.0 R231 39.3 7.4 1.8 1.3 G265 32.3 0.1 4.9 1.5 R228 28.1 3.8 0.3 1.2 S230 20.5 8.6 1.0 1.2 S264 18.8 0.0 6.5 1.8 F229 18.5 45.5 2.6 1.2 S168 17.8 1.4 3.5 2.1 K194 14.8 0.1 8.4 1.6 D166 12.7 0.0 4.5 1.6 a Numbers of contacts are averaged over the last 400 ns of the simulations of the first and second embedded models. A cutoff of 5.0 was used in counting the heavy atom contacting pairs; i.e., if the distance between a heavy atom of the protein and a heavy atom of the lipids was less than 5.0 this pair was counted. Entries for residues that average more than three interactions with acyl chains are shown in bold. The penetration depths are measured from the average phosphate atom is located below (deeper than) the phosphate plane.

PAGE 60

51 Figure 19 The correlation between the p enetration depths of F167 (red) and F229 (blue) and tilt angles of Syt7C2A WT associated with the membrane, for each simulation of the indicated p rotein membrane complex models Figure taken from Fig 8 in Ref. 21

PAGE 61

52 3.3.2. Comparison s of Simulation with Experiment P enetration depths for critical r esidues of Syt7 C2A from the embedded protein membrane complex simulations were linearly correlated ( R 2 = 0.8) with e lectron paramagnetic resonan ce (EPR) depth parameter measurements 34 ( Figure 20 ) This suggests that our docking results are consistent with the experiments although the experimental depths are on average deeper by roughly 4 Note that our calculations were taken from the protein membrane complex models, while t he experimental data were derived from EPR spin labeling depth measurements employing the NMR structure 31 (which is the Ca 2+ free form) and a solvated protein structure from our standalone protein simulations Our calculated i nsertion depths of CBL 1 and CBL 3 are also consistent with available computational observation t hrough MD simulations using the highly mobile membrane mimetic (HMMM) model although the HMMM allow s slightly deeper penetration into the membrane relative to all atomistic membrane model. 14 Figure 20 Comparisons between experimental and computational membrane penetration depths of selected residues In the left panel, the R 2 value of the regression line is 0.8. The horizontal error bars indicate experimental statistical errors, and vertical error bars indicate computational statistical standard deviations. The right panel shows the side by side comparisons of penetration dept hs for each residue (red, experimental; blue, computational). Figure modified from Fig. 10 in Ref. 21

PAGE 62

53 3.4. Coordination of the Ca 2+ I on s in the S tandalone Syt P roteins and Syt7C2A WT M embrane C omplexes The presence of the Ca 2+ ions is essential for the Syt protein membrane association process Th e results in this work suggest th ree Ca 2+ ions bind to both C2A WT domains of Syt1 and Syt7, two to Syt1C2B WT and three to Syt7C2B WT in good agreement with experiments. 26 30 To further probe the roles of the Ca 2+ ions in the protein binding to the anionic lipid containing membrane the integrated coordination number (ICN) was computed with the radial pair distribution function tool in VMD 41 ( Table 10 ) T he ICN count s the average number of electronegative atoms (protein oxygen, lipid oxygen, or solvent atoms ; water oxygen and/ or Cl ion) within the first solvation shell of Ca 2+ ion from the ion ) providing important information about the Ca 2+ ion coordination The literature value for the Ca 2+ ion ICN ranges from 6 8. 61 62 63 This agrees well with the ICN across all standalone 3 Ca 2+ protein models, which the ICN of 7.9 (6.9 from protein oxygen and 1.0 from water oxygen atoms) and 7.0 (6.0 from protein oxygen and 1.0 from water oxygen atoms) for the first and second Ca 2+ ions, respectively A minor variation was found for the second Ca 2+ ion in Syt7:1C2A CH model which has 5.0 of protein oxygen and 2.0 of water oxygen atoms. The third Ca 2+ ions had distinct ICNs from domain to domain and from WT to CH. For the solvated C2A domain models, t he three domains (Syt1C2A WT Syt7C2A WT and Syt7:1C2A CH ) showed differen t ICN : 7.0, 7.1, and 7.3, respectively, where the water oxygen atoms contribute d more to the coordination shell than the protein oxygen atoms. For the solvated C2B domain models, the I CN for Syt7C2B W T was similar to those for the C2A domains (7.0)

PAGE 63

54 and the contributions are mostly from the protein oxygen atoms as expected from the tight binding ( Figure 12 B ) The Syt1C2B WT and Syt1:7C2B CH each carry one protein oxygen atom three or four water oxygen atoms, and one or two Cl ions Therefore, the Ca 2+ ion coordination in the CH C2B domain seem ed to resemble Syt1C2B WT to a great extent even though the CBLs in C2B CH came from Syt7C2B WT This striking findings were in agreement with measurements by Bendahmane et al. 35 Both Syt1C2B WT and Syt1:7C2B CH recruited Cl ions to complete the coordinate shell of the outermost Ca 2+ ion ( Figure 21 ). If anion (here Cl ) recruitment is analogous to the anionic lipid membrane binding 3 these resu lts suggest that the existence of the third Ca 2+ ion in the CH C2B model will lead to a larger affinity for the membrane than that in Syt1C2B WT In the presence of the anionic lipid membrane (3:1 PC/PS) the third Ca 2+ ions in the C2A WT domains had different interactions with the coordinating oxygen atoms based upon the initial position s of the protein s The ICNs were 7.1, 7.0, and 7.3 for the third Ca 2+ ions in the pre insertion lying down model, and pre insertion standing up model and the ( average ) embedded model, respectively T he embedded model had a similar ICN for the third Ca 2+ ion to the standalone models However, the solvation shell composition changed. In the standalone models, 4.0 water molecules participated in the Ca 2+ coordination but as the proteins approached the membrane surface, one lipid oxygen atom replaced the water I n the pre insertion standing up model, 3.4 water molecules and 0.5 lipid oxygen atom were present in the third Ca 2+ ion s solvation shell, while in the average embedded model, 2.5 water molecules and 1.3 lipid oxygen atoms in the solvation shell Noticeably t he contribution of the lipid oxygen atoms was larger when the protein associated with membrane. The change in the coordination shell composit ion of

PAGE 64

55 the third Ca 2+ ions suggest s that the third Ca 2+ ion may not only prepare the protein for membrane association by modifying the CBLs electrostatic potential but also directly participate in C2A lipid binding. Based on the similarities of the se results for the solvated C2A and C2B domains such interaction s between the outermost Ca 2+ ion and membrane lipid s may also exist in Syt1 C2A and C2B domains. 57 60

PAGE 65

56 Table 10 Integrated coordination number s for the first solvation shell for each Ca 2+ ion in the standalone 3 Ca 2+ bound protein models and protein membrane complex models. Ligand First Ca 2+ Second Ca 2+ Third Ca 2+ Syt1C2A WT Protein O 6.9 6.0 3.0 Water O 1.0 1.0 4.0 Cl N/A N/A N/A Sum 7.9 7.0 7.0 Syt7C2A WT Protein O 6.9 6.0 3.3 Water O 1.0 1.0 3.8 Cl N/A N/A N/A Sum 7.9 7.0 7.1 Syt7:1C2A CH Protein O 6.9 5.0 3.3 Water O 1.0 2.0 4.0 Cl 0.0 0.0 0.0 Sum 7.9 7.0 7.3 Syt1C2B WT Protein O 6.9 6.0 1.2 Water O 1.0 1.0 4.0 Cl 0.0 0.0 1.0 Sum 7.9 7.0 6.2 Syt7C2B WT Protein O 6.9 6.0 5.0 Water O 1.0 1.0 2.0 Cl 0.0 0.0 0.0 Sum 7.9 7.0 7.0 Syt1:7C2B CH Protein O 6.9 6.0 1.1 Water O 1.0 1.0 3.0 Cl 0.0 0.0 2.0 Sum 7.9 7.0 6.1 Pre insertion: Lying down Protein O 6.9 6.0 3.1 Water O 1.0 1.0 4.0 Lipid O 0.0 0 .0 0.0 Sum 7.9 7.0 7.1 Pre insertion: Standing up Protein O 6.9 6.0 3.1 Water O 1.0 1.0 3.4 Lipid O 0.0 0 .0 0.5 Sum 7.9 7.0 7.0 1 st embedded Protein O 6.9 6.0 3.4 Water O 1.0 1.0 3.0 Lipid O 0 .0 0 .0 1.3 Sum 7.9 7.0 7.6 2 nd embedded Protein O 6.9 6.0 3.6 Water O 1.0 1.0 2.0 Lipid O 0.0 0 .0 1.3 Sum 7.9 7.0 6.9

PAGE 66

57 Figure 21 The distances between the third Ca 2+ and Cl ion s as a function of the 100 ns simulation time indicating the Cl recruitment by the third Ca 2+ ions. The third Ca 2+ ions are completely stabilized by ( A ) one Cl ion in Syt1C2B WT and ( B ) two in Syt1:7C2B CH after ~40 ns.

PAGE 67

58 CHAPTER I V C ONCLUSION In this work, we investigated the C2 domains in Syt1 and Syt7 protein s through MD simulations. The main conclusions are as follows: i. Syt7C2A WT preferentially binds three Ca 2+ ions in agreement with the experiment 15 and its third Ca 2+ ion is coordinated by a combinati on of protein residues in the CBLs, water and lipid if present ed Syt7:1C2A CH also binds three Ca 2+ ions in the CBLs and its Ca 2+ ion affinity somewhat fell somewhat between the Syt1C2A WT and Syt7C2A WT models ii. The Syt1C2B WT domain, which is known to bind two Ca 2+ ion s can hold one additional Ca 2+ ion in the CBLs. The binding of the third Ca 2+ ion to the protein, however, is less firm than that in Syt1C2A WT or Syt7C2B WT as the third Ca 2+ ion is coordinat ed by fewer protein residues and more water molecules. T he WT and CH C2B domains of Syt1 recruit Cl ions from the bulk solution if the third Ca 2+ ion is bound. In particular, Syt1:7C2B CH tend ed to attract more Cl ions. The stronger anion recruitment by Syt1:7C2B CH implies a greater anionic lipid affinity than Syt1C2B WT iii. W hen the CBLs with three bound Ca 2+ ions in Syt7C2A WT are sufficiently close to an anionic phospholipid membrane, the protein spontaneously approaches and inserts into the membrane. The orientation of the protein is determined by the competition between the electrostatic attractions and hydrophobic interactions. Three of the four simulated geometries of the protein membrane complex ( not the lying down model) are in good agreements with the experimental EPR measurements 3 4

PAGE 68

59 R EFERENCES 1. Coussens, L.; Parker, P. J.; Rhee, L.; Yang Feng, T. L.; Chen, E.; Waterfield, M. D.; Francke, U.; Ullrich, A. ( 1986 ) Multiple, Distinct Forms of Bovine and Human Protein Kinase C Suggest Diversity in Cellular Signaling Pathways. Science 233 859 866. 2. Nalefski, E. A.; Wisner, M. A.; Chen, J. Z.; Sprang, S. R.; Fukuda, M.; Mikoshiba, K.; Falke, J. J. ( 2001 ) C2 Domains fro m Different Ca 2+ Signaling Pathways Display Functional and Mechanistic Diversity. Biochemistry 40 3089 3100. 3. Sutton, R. B.; Davletov, B. A.; Berghuis, A. M.; S dhof, T. C.; Sprang, S. R. ( 1995 ) Structure of the First C2 Domain of Synaptotagmin I: A Novel Ca 2+ /Phospholipid binding Fold. Cell 80 929 938. 4. Brose, N.; Petrenko, A. G.; S dhof, T. C.; Jahn, R. ( 1992 ) Synaptotagmin: A Calcium Sensor on the Synaptic Vesicle Surface. Science 256 1021 10 25. 5. Perin, M. S.; Fried, V. A.; Mignery, G. A.; Jahn, R.; S dhof, T. C. ( 1990 ) Phospholipid Binding by a Synaptic Vesicle Protein Homologous to the Regulatory Region of Protein Kinase C. Nature 345 260 263. 6. Chapman, E. R. ( 2008 ) How Does Synaptotagmin Tri gger Neurotransmitter Release? Annu. Rev. Biochem. 77 615 641. 7. Rickman, C.; Craxton, M.; Osborne, S.; Davletov, B. ( 2004 ) Comparative Analysis of Tandem C2 Domains from the Mammalian Synaptotagmin Family. Biochemistry 378 681 686. 8. Bacaj, T.; Wu, D.; Yang X.; Morishita, W.; Zhou, P.; Xu, W.; Malenka, R. C.; S dhof, T. C. ( 2013 ) Synaptotagmin 1 and Synaptotagmin 7 Trigger Synchronous and Asynchronous Phases of Neurotransmitter Releas. Neuron 80 947 959. 9. S dhof, T. C. ( 2013 ) Neurotransmitter Release: The Last Millisecond in the Life of a Synaptic Vesicle. Neuron 80 (3), 675 690. 10. Luo, F.; Bacaj, T.; S dhof, T. C. ( 2015 ) Synaptotagmin 7 is Essential for Ca 2+ triggered Delayed Asynchronous Release But Not for Ca 2+ Dependent Vesi cle Priming in Retinal Ribbon Synapses. J. Neurosci. 35 11024 11033. 11. Dolai, S.; Xie, L.; Zhu, D.; Liang, T.; Qin, T.; Xie, H.; Kang, Y.; Chapman, E. R.; Gaisano, H. Y. ( 2016 ) Synaptotagmin 7 Functions to Replenish Insulin Granules for Exocytosis in Human Cells. Diabetes 65 1962 1976. 12. Takahashi, H.; Shahin, V.; Henderson, R. M.; Takeyasu, K.; Edwardson, J. M. ( 2010 ) Interaction of Synaptotagmin with Lipid Bilayers, Analyzed by Single Molecule Force Spectroscopy. Biophys J. 99 (8), 2550 2558.

PAGE 69

60 13. Hui, E .; Bai, J.; Wang, P.; Sugimori, M.; Llinas, R. R.; Chapman, E. R. ( 2005 ) Three Distinct Kinetic Groupings of the Synaptotagmin Family: Candidate Sensors for Rapid and Delayed Exocytosis. Proc. Natl. Acad. Sci. U S A 102 5210 5214. 14. Vermaas, J. V.; Tajkhorshid, E. ( 2017 ) Differential Membrane Binding Mechanics of Synaptotagmin Isoforms Observed in Atomic Detail. Biochemistry 56 281 293. 15. Brandt, D. S.; Coffman, M. D.; Falke, J. J.; Knight, J. D. ( 2012 ) Hydrophobic Contributions to the Membrane Dockin g of Synaptotagmin 7 C2A Domain: Mechanistic Contrast between Isoforms 1 and 7. Biochemistry 51 (39), 7654 76 64. 16. Mackler, J. M.; Drummond, J. A.; Loewen, C. A.; Robinson, I. M.; Reist, N. E. ( 2002 ) The C 2 B Ca 2+ binding Motif of Synaptotagmin Is Required fo r Synaptic Transmission in vivo Nature 418 340 344. 17. Rizo, J.; Chen, X.; Arac, D. ( 2006 ) Unraveling the Mechanisms of Synaptotagmin and SNARE Function in Neurotransmitter Release. Trends Cell Biol. 16 339 350. 18. Fernandez Chacon, R.; Konigstorfer, A.; Gerber, S. H.; Garcia, J.; Matos, M. F. ( 2001 ) Synaptotagmin I Functions as a Calcium Regulator of Release Probability. Nature 410 41 49. 19. Voleti, R.; Tomchick, D. R.; Sdhof, T. C.; Rizo, J. ( 2017 ) Exceptionally Tight Membrane binding May Explain the Key Role of the Synaptotagmin 7 C2A Domain in Asynchronous Neurotransmitter Release. PNAS 114 (40), E8518 E8527. 20. Segovia, M.; Als, E.; Montes, M. A.; Bonifas, I.; Jemal, I.; Lindau, M.; Maximov, A.; Sdhof, T. C.; Alvarez de Toledo, G. ( 2010 ) Push and pull Re gulation of the Fusion Pore by Synaptotagmin 7. PNAS 107 (44), 19032 19037. 21. Chon, N. L.; Osterberg, J. R.; Henderson, J.; M., K. H.; Reute, N.; Knight, J. D.; Lin, H. ( 2015 ) Membrane Docking of the Synaptotagmin 7 C2A Domain: Computation Reveals Interplay between Electrostatic and Hydrophobic Contributions. Biochemistry 54 5696 5711. 22. Shao, X.; Fernandez, I.; S dhof, T. C.; Rizo, J. ( 1998 ) Solution Structures of the Ca 2+ Free and Ca 2+ Bound C2A Domain of Synaptotagmin I: Does Ca 2+ Induce a Conformational Ch ange? Biochemistry 37 (46), 16106 16115. 23. Ubach, J.; Zhang, X.; Shao, X.; Sdhof, T. C.; Rizo, J. ( 1998 ) Ca 2+ Binding to Synaptotagmin: How Many Ca 2+ Ions Bind to the Tip of a C2 domain? EMBO J. 17 3921 3930. 24. Shao, X.; Davletov, B. A.; Sutton, R. B.; Sdhof, T. C.; Rizo, J. ( 1996 ) Bipartite Ca 2+ Binding Motif in C2 Domains of Synaptotagmin and Protein Kinase C. Science 273 248 251.

PAGE 70

61 25. Vasquez, J. K.; Chantranuvatana, K.; Giardina, D. T.; Coffman, M. D.; Knight, J. D., Lateral Diffusion of Proteins on Sup ported Lipid Bilayers: Additive Friction of Synaptotagmin 7 C2AC2B Tandem Domains. Biochemistry 2014, 53 7904 7913. 26. Fernandez, I.; Arac, D.; Ubach, J.; Gerber, S. H.; Shin, O.; Gao, Y.; Anderson, R. G. W.; S dhof, T. C.; Rizo, J. ( 2001 ) Three dimensional Structure of the Synaptotagmin 1 C2B domain: Synaptotagmin 1 as a Phospholipid Binding Machine. Neuron 32 1057 1069. 27. Xue, M.; Craig, T. K.; Shin, O. H.; Li, L.; Brautigam, C. A.; Tomchick, D. R.; S dhof, T. C.; Rosenmund, C.; Rizo, J. ( 2010 ) Structural an d Mutational Analysis of Functional Differentiation between Synaptotagmins 1 and 7. Plos One 5 12544. 28. Cheng, Y.; Sequeira, S. M.; Malinina, L.; Tereshko, V.; Sllnerr, T. H.; Patel, D. J. ( 2004 ) Crystallographic Identification of Ca 2+ and Sr 2+ Coordination Sites in Synaptotagmin I C2B Domain. Protein Science 13 2665 2672. 29. Ubach, J.; Garcia, J.; Nittler, M. P.; S dhof, T. C.; Rizo, J. ( 1999 ) Structure of the Janus faced C2B Domain of Rabphilin. Nat. Cell Biol. 1 106 112. 30. Maximov, A.; Lao, Y.; Li, H.; Chen, X.; Rizo, J.; Srensen, J. B.; Sdhof, T. C. ( 2008 ) Genetic Analysis of Synaptotagmin 7 Function in Synaptic Vesicle Exocytosis. Proc Natl Acad Sci U S A. 105 3986 3991. 31. Nagashima, T.; Hayashi, F.; Yokoyama, S. (2015) Solution Structure of the First C2 Domain of Synaptotagmin VII unpublished results 32. Rao, T. C.; Passmore, D. R.; Peleman, A. R.; Das, M.; Chapman, E. R.; Anantharam, A. ( 2014 ) Distinct Fusion Properties of Synaptotagmin 1 and Synaptotagmin 7 Bearing Dense Core Granules. Mol. Biol. Cell 25 2416 2427. 33. Rao, T. C.; Santana Rodriguez, Z.; Bradberry, M. M.; Ranski, A. H.; Dahl, P. J.; Schmidtke, M. W.; Jenkins, P. M.; Axelrod, D.; Chapman, E. R.; Giovannucci, D. R.; Anantharam, A. ( 2017 ) Synaptotagmin Isoforms Confer Distinct Activ ation Kinetics and Dynamics to Chromaffin Cell Granules. J. Gen. Physiol. 149 763 780. 34. Osterberg, J. R.; Chon, N. L.; Boo, A.; Maynard, F. A.; Lin, H.; Knight, J. D. ( 2015 ) Membrane Docking of the Synaptotagmin 7 C2A Domain: Electron Paramagnetic Resonance Measurements Show Contributions from Two Membrane Binding Loops. Biochemistry 54 5684 5695. 35. Bendahmane, M.; Bohannon, K. P.; Rao, T. C.; Schmidtke, M. W.; Bradberry, M. M .; Abbineni, P.; Chon, N. L.; Tran, S.; Lin, H.; Chapman, E. R.; Knight, J. D.; Anantharam, A. ( 2018 ) The Synaptotagmin 1 C2B Domain Ca 2+ binding Loops Modulate the Rate of Fusion Pore Expansion. Mol. Cell 29 1 44. 36. Jo, S. K., T.; Iyer, V. G.; Im, W. ( 2008 ) CHARMM GUI: A Web based Graphical User Interface for CHARMM. J. Comput. Chem. 29 1859 1865.

PAGE 71

62 37. Klauda, J. B.; Venable, R. M.; Freites, J. A.; O'Connor, J. W.; Tobias, D. J.; Mondragon Ramirez, C.; Vorobyov, I.; MacKerell, A. D.; Pastor, R. W. ( 2010 ) Update of the CHARMM All atom Additive Force Field for Lipids: Validation on Six Lipid Types. J. Phys. Chem. B 114 (23), 7830 7843. 38. Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. ( 1983 ) Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 79 (2), 926 935. 39. Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. ( 2005 ) Scalable Molecular Dynamics with NAMD. J. Comput. Chem. 26 (16), 1781 1 802. 40. Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. ( 1995 ) A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 103 (19), 8577 85 93. 41. Humphrey, W.; Dalke, A.; Schulten, K. ( 1996 ) VMD: Visual Molecular Dynamics. J. Molec. Graphics 14 (1), 33 38. 42. Brooks, B. R.; Brooks, C. L., 3rd; Mackerell, A. D., Jr.; Nilsson, L.; Petrella, R. J.; Roux, B.; Won, Y.; Archontis, G.; Bartels, C.; Boresch, S.; Caflisch, A.; Caves, L.; Cui, Q.; Dinner, A. R.; Feig, M.; Fischer, S.; Ga o, J.; Hodoscek, M.; Im, W.; Kuczera, K.; Lazaridis, T.; Ma, J.; Ovchinnikov, V.; Paci, E.; Pastor, R. W.; Post, C. B.; Pu, J. Z.; Schaefer, M.; Tidor, B.; Venable, R. M.; Woodcock, H. L.; Wu, X.; Yang, W.; York, D. M.; Karplus, M. ( 2009 ) CHARMM: The Biomo lecular Simulation Program. J. Comput. Chem. 30 (10), 1545 1 614. 43. Mackerell, A. D., Jr.; Bashford, D.; Bellott, M.; Dunbrack, R. L.; Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher, W. E., III; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.; Watanabe, M.; Wiorkiewicz Kuczera, J.; Yin, D.; Karplus, M. ( 1998 ) All Atom Empirical Potential for M olecular Modeling and Dynamics Studies of Proteins. J. Phys. Chem. B 102 (18), 3586 3 616. 44. Mackerell, A. D., Jr.; Feig, M.; Brooks, C. L. ( 2004 ) 3rd, Extending the Treatment of Backbone Energetics in Protein Force Fields: Limitations of Gas Phase Quantum Me chanics in Reproducing Protein Conformational Distributions in Molecular Dynamics Simulations. J. Comput. Chem. 25 (11), 1400 14 15. 45. Gasteiger, E.; Gattiker, A.; Hoogland, C.; Ivanyi, I.; Appel, R. D.; Bairoch, A. ( 2003 ) ExPASy: The Proteomics Server for In depth Protein Knowledge and Analysis. Nucleic Acids Res. 31 3784 3788. 46. Biasini, M.; Bienert, S.; Waterhouse, A.; Arnold, K.; Studer, G.; Schmidt, T.; Kiefer, F.; Cassarino, T. G.; Bertoni, M.; Bordoli, L.; Schwede, T. ( 2014 ) SWISS MODEL: Modelling Protein Tertiary and Quaternary Structure Using Evolutionary Information. Nucleic Acids Res. 42 W252 W258.

PAGE 72

63 47. Frazier, A. A.; Roller, C. R.; Havelka, J. J.; Hinderliter, A.; Cafiso, D. S. ( 2003 ) Membrane bound Orientation and Position of the Synaptotagmin I C2A Domain by Site Directed Spin Labeling. Biochemistry 42 (1), 96 105. 48. N.; Nieh, M. P.; Katsaras, J. ( 2011 ) Fluid Phase Lipid Areas and Bilayer Thicknesses of Commonly Used Phosphatidylcholines as a Function of Temperatur e. Biochim. Biophys. Acta, Biomembr. 1808 2761 2771. 49. H. ( 2003 ) Effect of Sodium Chloride on a Lipid Bilayer. Biophys. J. 85 1647 1655. 50. Mukhopadhyay, P.; Monticelli, L.; Tieleman, D. P. ( 2004 ) Molecula r Dynamics Simulation of a Palmitoyl Oleoyl Phosphatidylserine Bilayer with Na + Counterions and NaCl. Biophysical J. 86 (3), 1601 1609. 51. Baker, N. A.; Sept, D.; Joseph, S.; Holst, M. J.; McCammon, J. A. ( 2001 ) Electrostatics of Nanosystems: Application to M icrotubules and the Ribosome. Proc. Natl. Acad. Sci. U. S. A. 98 (18), 10037 100 41. 52. Seelig, A.; Seelig, J. ( 1974 ) Dynamic Structure of Fatty Acyl Chains in a Phospholipid Bilayer Measured by Deuterium Magnetic Resonance. Biochemistry 13 (23), 4839 48 45. 53. Venable, R. M.; Luo, Y.; Gawrisch, K.; Roux, B.; Pastor, R. W. ( 2013 ) Simulations of Anionic Lipid Membranes: Development of Interaction specific Ion Parameters and Validation using NMR Data. J. Phys. Chem. B 117 10183 10192. 54. Li, C.; Ullrich, B.; Zhang, J Z.; Anderson, R. G.; Brose, N.; Sdhof, T. C. ( 1995 ) Ca 2+ dependent and independent Activities of Neural and Non neural Synaptotagmins. Nature 375 594 599. 55. Sugita, S.; Han, W.; Butz, S.; Liu, X.; Fernndez Chacn, R.; Lao, Y.; Sdhof, T. C. ( 2001 ) Syna ptotagmin VII as a Plasma Membrane Ca 2+ Sensor in Exocytosis. Neuron 30 459 473. 56. Sugita, S.; Shin, O. H.; Han, W.; Lao, Y.; Sdhof, T. C. ( 2002 ) Synaptotagmins Form Hierarchy of Exocytotic Ca 2+ sensors with Distinct Ca 2+ affinities. EMBO J. 21 270 280. 57. N alefski, E. A.; Falke, J. J. ( 1996 ) The C2 Domain Calcium binding Motif: Structural and Functional Diversity. Protein Sci. 5 2375 2390. 58. Murray, D.; Honig, B. ( 2002 ) Electrostatic Control of the Membrane Targeting of C2 Domains. Mol. Cell 9 145 154. 59. Bhalla, A. ; Tucker, W. C.; Chapman, E. R. ( 2005 ) Synaptotagmin Isoforms Couple Distinct Ranges of Ca 2+ Ba 2+ and Sr 2+ Concentration to SNARE mediated Membrane Fusion. Mol. Biol. Cell 16 4755 4764.

PAGE 73

64 60. Corbin, J. A.; Evans, J. H.; Landgraf, K. E.; Falke, J. J ( 2007 ) Mechanism of Specific Membrane Targeting by C2 Domains: Localized Pools of Target Lipids Enhance Ca 2+ Affinity. Biochemistry 46 4322 4336. 61. Tongraar, A.; Liedl, K. R.; Rode, B. M. ( 1997 ) Solvation of Ca 2+ in Water Studied by Born oppenheimer ab in itio QM/MM Dynamics. J. Phys. Chem. A 101 6299 6309. 62. D.; Lindqvist Reis, P.; Hermansson, K.; Persson, I.; Sandstrom, M. ( 2001 ) Hydration of the Calcium Ion. An EXAFS, Large Angle X ray Scattering, and Molecular Dynamics Simula tion Study. J. Am. Chem. Soc. 123 (3), 431 4 41. 63. Pirc, E. T.; Zidar, J.; Bukovec, P. ( 2012 ) A Computational Study of Calcium(II) and Copper(II) Ion Binding to the Hyaluronate Molecule. Int. J. Mol. Sci. 13 (9), 12036 120 45.