Inner and Outer Coordination Shells of Mg2+ in ... - ACS Publications

Jan 4, 2016 - Graduate School of Nanoscience and Technology, Chulalongkorn University, ... Department of Chemistry, Faculty of Science, Chulalongkorn ...
0 downloads 0 Views 2MB Size
Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article 2+

Inner and Outer Coordination Shells of Mg in CorA Selectivity Filter from Molecular Dynamics Simulations Sunan Kitjaruwankul, Pattama Wapeesittipan, Panisak Boonamnaj, and Pornthep Sompornpisut J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.5b10925 • Publication Date (Web): 04 Jan 2016 Downloaded from http://pubs.acs.org on January 13, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Inner and Outer Coordination Shells of Mg2+ in CorA Selectivity Filter from Molecular Dynamics Simulations

Sunan Kitjaruwankul,†,‡ Pattama Wapeesittipan,‡ Panisak Boonamnaj,‡ and Pornthep Sompornpisut*,‡ †

Graduate School of Nanoscience and Technology, Chulalongkorn University, Bangkok 10330,

Thailand ‡

Department of Chemistry, Faculty of Science, Chulalongkorn University, Bangkok 10330,

Thailand *

Corresponding author:

Pornthep Sompornpisut Department of Chemistry, Faculty of Science, Chulalongkorn University, Bangkok 10330, Thailand. Phone: 662-2187604; Fax: 662-2187598

AUTHOR EMAIL ADDRESS: [email protected] TITLE RUNNING HEAD: Molecular dynamics simulation of CorA

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 39

Abstract Structural data of CorA Mg2+ channels show that the five Gly-Met-Asn (GMN) motifs at the periplasmic loop of the pentamer structure form a molecular scaffold serving as a selectivity filter. Unfortunately, knowledge about the cation selectivity of Mg2+ channels remains limited. Since Mg2+ in aqueous solution has a strong first hydration shell and apparent second hydration sphere, the coordination structure of Mg2+ in a CorA selectivity filter is expected to be different from that in bulk water. Hence, this study investigated the hydration structure and ligand coordination of Mg2+ in a selectivity filter of CorA using molecular dynamics (MD) simulations. The simulations reveal the inner-shell structure of Mg2+ in the filter is not significantly different from that in aqueous solution. The major difference is the characteristic structural features of the outer shell. The GMN residues engage indirectly in the interactions with the metal ion as ligands in the second shell of Mg2+. Loss of hydrogen bonds between inner- and outer-shell waters observed from Mg2+ in bulk water is mostly compensated by interactions between waters in the first solvation shell and the GMN motif. Some water molecules in the second shell remain in the selectivity filter and become less mobile to support the metal binding. Removal of Mg2+ from the divalent cation sensor sites of the protein had an impact on the structure and metal binding of the filter. From the results, it can be concluded that the GMN motif enhances the affinity of the metal binding site in the CorA selectivity filter by acting as an outer coordination ligand.

2 ACS Paragon Plus Environment

Page 3 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Introduction Magnesium ion (Mg2+) is present abundantly in living cells in the form of divalent cation. It plays a significant role in many physiological and biological activities including DNA/RNA transcription and replication, energy metabolism, protein translation, cell regulation and nerve and muscle function.1-3 A low Mg2+ intake has been shown to relate with several health problems, for instance, diabetes, cardiovascular disease and osteoporosis.4-6 Because of its significance in nutritional development and medicine, there is an increased interest in understanding the transport mechanism of this divalent cation. The transport of Mg2+ across lipid bilayer cell membranes is regulated and facilitated by a class of specialized membrane proteins, which include MgtA, MgtB, MgtE and CorA (the cobalt-resistance mutant) in the prokaryotic system3,

7-8

. CorA is the divalent cation channel responsible for Mg2+ homeostasis. It is

considered to be the major Mg2+ transport protein found in most eubacteria and archaea and represents a useful molecular model to study insight into the structural basis of the regulatory mechanisms for humans.9-11 In addition, the CorA family facilitates the influx of Co2+ and Ni2+ with lower efficiency. The CorA Mg2+ channel has functional homologs with eukaryotic Mrs2 and Alr1 magnesium transport proteins, but they have a low sequence conservation. Nevertheless, Mg2+ channels in the CorA/Mrs2/Alr1 superfamily are characterized by the highly conserved signature GMN sequence present in the extracellular loop (Figure 1A).12-13 Despite its biological significance, little information is available regarding the molecular mechanism underlying Mg2+ homeostasis. The crystal structures of CorA from Thermotoga maritima (TmCorA) provide an excellent structural framework for understanding Mg2+ channel properties.14-16 The TmCorA structure is a funnel-shaped homopentamer with a five-fold symmetry axis perpendicular to the membrane (Figure 1B). Each monomer contains a large N-

3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 39

terminal cytoplasmic domain, a long stalk helix and two transmembrane segments: the inner pore-forming TM1 and the outer TM2. The pore-forming TM1 helices are surrounded by TM2 helices facing the hydrophobic lipid bilayer. TM1 and TM2 are joined by the periplasmic loop, which contains the conserved signature GMN sequence responsible for the selectivity of divalent cations. The cytoplasmic domain forms an αβα sandwich-like structure that consists of seven parallel and antiparallel β-strands laid between two sets of three α-helices. The stalk and TM1 helices of each monomer assemble to form a long hydrophobic constriction pore (~40 Å). The available crystal structures of CorA show that the ion-conduction pore has a diameter ranging from 2 to 7 Å.

Figure 1 (A) Sequence alignment of signature GMN sequence region of TmCorA with other CorA, Mrs2 and Alr1 family. The numbering is based on the TmCorA sequence. (B) Crystal structure of TmCorA Mg2+ channel (4EED) highlighting a molecular scaffold 4 ACS Paragon Plus Environment

Page 5 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

formed by the five GMN motifs at the extracellular loop and one of the five M1/M2 binding sites at the divalent cation sensor sites

A balance of Mg2+ inside and outside of the cell is very important for the viability of the cell and also for its proper functioning. Despite its biological importance, knowledge of the ion regulation and selectivity mechanisms on a structural level is incomplete. It has been suggested that the functional activity of TmCorA is regulated by intracellular Mg2+ concentration17. All crystallographic models that have been obtained at high concentrations of divalent cations revealed a narrow, long and hydrophobic pore, supporting the idea of a closed or non-conducting conformation. This observation is also supported by the crystal structures showing the Mg2+ ions bound to the two cytoplasmic domains of the adjacent protomers acting as a clamp to stabilize the channel in a non-conductive conformation (Figure 1). Additionally around the pentameric ring, a total of five metal-binding sites, each of which can host up to two metal ions (denoted as M1 and M2 sties) has been proposed to act as metal ion-sensing regulatory by acting as part of a divalent cation sensor (DCS, Figure 1B).16 The M1 binding site is in close proximity to two coordinating acidic residues, D89 and D253' (“ ' ” refers to the adjacent subunit). It has been shown to play a primary role in conformational control of protein regulation. The second M2 binding site located at a distance of 7Å from M1 is surrounded by the backbone carbonyl group of L12 and the side chain carboxylate oxygen of E88, D175 and D253. Experimental and theoretical studies suggest that the Mg2+-protein interactions at the DCS sites secures the structure of TmCorA in a closed state conformation. When the intracellular Mg2+ concentration is deficient, the bound Mg2+ ions are released from the DCS sites, resulting in an open channel

5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 39

conformation. According to molecular dynamics studies, removal of Mg2+ from the DCS site causes structural rearrangement of the protein, leading to opening the pore of the channels.17-19 The structural basis of the metal coordination complex that makes ion channels selective is of prime interest for understanding their biological function. Apart from the presence of Mg2+ in the DCS sites, the crystal structures of TmCorA show the existence of another important binding site which is related to the channel regulatory function: the site at the signature GMN sequence at the extracellular loop. It has been proposed that this conserved GMN motif is responsible for the ion recognition function.20 Crystallographic data of TmCorA show that the five GMN motifs at the periplasmic loop of the pentamer structure form a molecular scaffold serving as the divalent cation binding site (Figure 1B). It has been proposed to form the entry point for the channel by acting as the selectivity filter for Mg2+ recognition.18, 21 Furthermore, the functional study shows that the mutation within this signature sequence abolishes Mg2+ uptake while this might result in changing the specificity to other cations.20 It has generally been inferred that Mg2+ in aqueous solution exhibits a well-defined first hydration shell and apparent second hydration sphere. Therefore, the hydration structure of the Mg2+ bound to the GMN motif could be different from that of the free Mg2+ in bulk. Furthermore, because of a larger volume in the CorA selectivity filter with respect to the potassium or sodium channels, the structure basis of the ion selectivity of the TmCorA Mg2+ channel is expected to be different from those monovalent cation channels. In the closed KcsA channel, the backbone of tetrameric carbonyl groups of the selectivity filter is about 3.0-3.4Å from K+ and thus bind to dehydrated K+ ions. In contrast, the crystallographic data of TmCorA reveals that the GMN residues appear to bind to the Mg2+ ion in a hydrated form10 since the distance from the metal center to the five carbonyl oxygen atoms of G312 is in a range of 3.8 to 5.0 Å and 4.2 to 5.2 for the five asparagine side

6 ACS Paragon Plus Environment

Page 7 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

chains of N314 (Figure 1B). Unfortunately, the detailed hydration structure of Mg2+ within the selectivity filter of TmCorA remains elusive. It is anticipated that the interactions of these residues affects to some extent the hydration structure of Mg2+, especially the coordination structure of the second shell. Based on the current medium-resolution crystal structures (2.93.9Å), it is still unclear how the water molecules are arranged in the Mg2+-GMN filter complex of TmCorA. In this study, we investigated the possible structure of coordination shells of Mg2+ in the selectivity filter of TmCorA and the effects of the metal-unbound state of the DCS sites on the coordination shell structure of Mg2+ within the filter. It is hypothesized that the coordination structure of the Mg2+ ion in the GMN selectivity filter is expected to be different from that in aqueous solution. MD simulations of the Mg2+-GMN complex in the form of the truncated model as well as the full-length TmCorA Mg2+ channel have been conducted to address questions about i) what a major difference should be between the coordination structure of Mg2+ in the selectivity filter and that in bulk water, and ii) how the selectivity filter GMN residues interacts with the Mg2+ ion. The simulations reveal that while interactions of the inner coordination shell of Mg2+ in the filter are similar to those in aqueous solution, interactions of the outer coordination shell engage the GMN motif and a number of water molecules. The results have clearly demonstrated the role of the GMN residues as outer coordination ligands for the Mg2+ complex in the CorA selectivity filter. The study should provide the opportunity to better understand the molecular principle underlying the selectivity of Mg2+ channels.

Methods Structure models

7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 39

The TmCorA crystal structure with the PDB ID code 4EED was used as a full-length model. The missing periplasmic loop (residue number from 316 to 325) was modeled by using the loop prediction program LOOPY22. It should be noted that structure coordinates of the GMN signature residues (G312-M313-N314) were obtained directly from the x-ray crystal structure. For the missing hydrogens atoms in the model, the PSFGEN plugin of VMD was used23. The assignment of ionizable side chains was based on the pKa values at a pH of 7 using PROPKA24. In addition, to reduce the size of the simulated systems and improve the sampling by focusing on the important parts of the selectivity filter, we constructed a truncated selectivity filter model, which contains only residues 310 to 327 (the selectivity filter and the periplasmic loop) and a Mg2+ ion. To explore the effect of Mg2+ ions in the DCS sites on the structure of the selectivity filter, MD simulations of the full-length model were conducted using three different model systems: i) the crystal structure (denoted as FLCorA), ii) the crystal structure with a removal of Mg2+ ions in the DCS sites and remaining only Mg2+ in the selectivity filter (denoted as FLCorAnoMgDCS) and iii) the crystal structure with a removal of Mg2+ ions in every site (FLCorAnoMgall). For the truncated model, the simulations were carried out with and without constraining backbone atoms of the residues of the loop (denoted as CtrnLoop-Mg and FlexLoop-Mg) to compare degree of model uncertainty regarding the loop structure prediction. Furthermore, the simulation of Mg2+ in aqueous solution (denoted as Free-Mg) was also performed to serve for comparison. A total of six different systems (Table 1) were constructed for the study.

Simulation setup

8 ACS Paragon Plus Environment

Page 9 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

All simulated systems were built using the VMD v1.9 package. MD simulations of the full-length model were conducted in a phospholipid bilayer. The preparation of the system followed the membrane protein tutorial.25 Each model was inserted into a pre-equilibrated lipid bilayer, which comprised 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and TIP3P water model. The simulated model was solvated by placing the lipid bilayer in the xy plane, and the central axis of the pore-forming TM domains of TmCorA was oriented to align with the zaxis of the bilayer. The protein charge was neutralized by adding counterions (Na+ and Cl-) to the system at 0.1 M concentration using VMD’s autoionize plugin. For the full-length model, the simulation box consists of ∼300,000 atoms. The simulations of CtrnLoop-Mg and FlexLoop-Mg were conducted in aqueous solution since the periplasmic loop of TmCorA was mostly located on the extracellular side of the channel. To incorporate the solvent, the truncated model was solvated in a box of TIP3P water and its charge was neutralized by the counterions using the VMD plugin scripts. For the simulation of the truncated selectivity filter model, the backbone atoms of the two terminal residues were constrained with respect to their initial positions to mimic the immobility of the missing transmembrane and intracellular parts of the protein. The CtrnLoop-Mg simulation was carried out by constraining the position of backbone atoms with respect to their initial coordinates, while no constrained atoms were applied in the FlexLoop-Mg simulation. To validate the consistency of the simulations and improve sampling, both simulations were performed with three replications using independent starting configuration.

Table 1. Summaries of detailed molecular dynamics simulations system

#Mg2+

#water

#POPC

Box dimension(Å3)

Total atoms

Time (ns) × 9

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

FLCorA FLCorA-no Mgall FLCorA-no MgDCS FlexLoop-Mg CtrnLoop-Mg Free-Mg

14 0 4 1 1 1

66763 62059 66743 9298 9298 826

493 493 493 0 0 0

131 × 124 × 170 128 × 128 × 169 139 × 140 × 144 78 × 73 × 52 78 × 73 × 52 30 × 30 × 30

Page 10 of 39

293648 286402 293598 29517 29517 2481

repetition 100×1 100×1 100×1 20×3 20×3 10×1

All MD simulations were carried out using the NAMD 2.8 software26. The CHARMM22 and CHARMM27 parameter sets were, respectively, used for protein and for lipid and the metal ions (including Mg2+, Ca2+, Na+ and K+)27, and the TIP3P model was used for water.28 The dimension of each periodic box in all simulated systems are summarized in Table 1. Simulations were carried out employing periodic boundary conditions. The particle mesh Ewald summation was employed to treat long-range electrostatic interactions with a distance cut-off of 12 Å and pair list distance of 13.5 Å.29 The bonds between the hydrogen and heavy atoms were kept as rigid using the SHAKE and SETTLE algorithms.30-31 The simulations were conducted at a constant pressure of 1 atm using a Nośe-Hooven Langevin Piston and a constant temperature of 300K.32-33 During the initial run, temperature of the system was slowly increased from 100 K to 300 K. Then, the constant temperature was maintained at 300 K using Langevin dynamics with the damping coefficient of 1 ps-1. The details of all MD systems are shown in Table 1.

Molecular dynamics simulations The structure of the initial configuration was relaxed using energy minimizations and restrained MD simulations. In this step, the systems were minimized and equilibrated with every atom, except for those of lipid tails that were kept in the fix position. Subsequently, protein atoms and Mg2+ ions were restrained for about 6ns, allowing the remaining part of the system such as waters, lipids and counterions to be relaxed. During this step, the hydration of the system 10 ACS Paragon Plus Environment

Page 11 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

was monitored to ensure that water molecules filled the space in the protein, especially the cavity in the selectivity filter. Finally, free molecular dynamics were carried out to further equilibrate the whole system. MD simulations were conducted with the time step of 2fs. MD snapshot of the trajectory was recorded every 2 ps. The simulations were conducted for 100 ns for the full-length model and about 20 ns for the truncated model. The simulation of free Mg2+ in aqueous solution was performed for 10ns.

Analysis of the MD trajectory The resulting trajectories of MD production run were mainly analyzed through the execution of VMD scripts. This includes the root-mean-square-deviation (RMSD) of protein atoms, the root-mean-square-fluctuation (RMSF), radial distribution function (RDF), probability distribution of coordination number, hydrogen-bonding, inter-atomic distances and its distribution, principal component analysis (PCA) and the mean residence time (MRT). The pore radius of the protein models was estimated using the HOLE program.34 Principal component analysis (PCA) A principal component analysis was performed on MD trajectories of the FLCorA, FLCorA-no MgDCS and FLCorA-no Mgall systems to identify essential dynamics of motions to see how presence or absence of Mg2+ induces changes in CorA. PCA was performed for Cα atoms, whose structure from the trajectory was aligned to its initial structure to filter out all the trivial translations and rotations. About 1500-2000 structural snapshots at 20 ps intervals were chosen from the last 40ns (60 – 100 ns) MD trajectories. The first two principal components of motion, PC1 and PC2, which correspond to the first two Eigen vectors of the covariance matrix,

11 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 39

were calculated using the WORDOM software35, The interpolated structures along the first two principal components and movies were created using VMD with in-house scripts.

Mean residence time (MRT) The exchange rate between water molecules in the solvation shell and those in the bulk solution was used to determine the mean residence time. The time correlation function (C(t)) is defined by Eq. (1):

C (t ) =

T (0)T (t ) T

(1)

where a binary function, T(t), takes the value of 1 if the considered water is still in the site for a time t, and 0 otherwise, T(0) is equal to 1 at time zero when T is a period of time to be determined. For the long-time decay, C(t) approaches ~0. Here, the mean residence time of water, τ, in the hydration shell was estimated from integration of the residence time correlation function. The time period was divided into every 30ps duration. A fit of the residence-time decay curve of C(t) to the exponential function e-t/τ yields the value of τ.

Results First and second hydration shell of Mg2+ The RDF or g(r) of the Mg-O(water) from the simulation of Mg2+ in aqueous solution (Free-Mg) displays a well-separated peak of the first hydration shells with a sharp Mg-O peak at 2.0 Å (Figure 2A). The second RDF peak is relatively broad covering the region from about 3.5

12 ACS Paragon Plus Environment

Page 13 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

up to about 4.8 Å with the maximum of the RDF at 4.2 Å. As expected for strong ion-solvent interactions, a well separation of RDF between the first and the second hydration shells was observed. This suggested no exchange of water molecule between the inner and the outer shells of Mg2+ during the 10ns simulation time. The experimentally-determined residence times of the inner-shell water were found to be on the order of microsecond time scale at 25oC.36 The RDF between the second and bulk was considerably separated, but still shows a non-zero value at minimum. This indicates that water molecules in the second hydration shell and the bulk undergo solvent exchange. The integration of RDF or n(r) of the first shell indicates six water molecules with their oxygen atoms point toward the Mg2+center. Visual inspection reveals the inner coordination shell of Mg2+ is hexacoordinated with an octahedral arrangement. The rigidity of the inner coordination structure of Mg2+ can be illustrated by 100 percentage frequency of occurrence (Figure 2B). Furthermore, the integration number, n(r) of the water molecules in the second shell, is up to ~17. The hydration number of the outer coordination sphere varies by

several waters because of the water exchange during the simulation. This suggests the interactions of the second-shell waters to the metal ion are relatively weak. Figure 2B illustrates the probability distributions of coordination numbers of the first and second solvation shells calculated from the RDF graph. The results clearly show that the first solvation shell comprises a coordination number (CN) of 6. For the outer coordination sphere of Mg2+, the distribution ranges from 8 to 17 with the most preference of 13 (followed by 12, 14 and 11 with slightly lower probability). The corresponding MRT of the outer-shell waters exhibits a value of 10.2ps (Table 2). From this simulation, the results are in overall good agreement with x-ray

diffraction and QM/MM MD studies.37-39 The computational studies of hydrated gas-phase magnesium ion suggested that the most probable hydration structure of Mg2+ in aqueous

13 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 39

solution is in the form of Mg[H2O]62+ [H2O]12 arranged approximately as a dodecahedral cluster.40-41 As expected, orientations of several water molecules of the inner and outer coordination spheres satisfy hydrogen bond criteria (Figure 3B). It should also be noted that the

observed hydrogen bonding of water molecules in the inner hydration shell also contributes to the stability of the hydration structure of Mg2+.

Figure 2. (A) g(r) of the Mg-O(water) distance of Free-Mg2+ (black), FlexLoop-Mg (red) and CtrnLoop-Mg(blue) and (B) the corresponding probability distribution of coordination number for the first and second shells (CN1 and CN2).

Table 2. Hydration number with the most occurrence and mean residence time (MRT) of water in the inner and outer shells of Mg2+ in aqueous solution, truncated and full-length CorA model at 300K. System Hydration numbera MRT (ps) Free-Mg 6 (13) neb (10.2) FlexLoop-Mg 6 (3) ne (29.4) CtrnLoop-Mg 6 (5) ne (29.4) FLCorA 6 (8) ne (47.1) FLCorA_noMgDCS 6 (10) ne (35.2) a Numbers in parentheses are the values for the outer hydration shell. bne=No exchange of water during the simulation time.

14 ACS Paragon Plus Environment

Page 15 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

For the CtrnLoop-Mg and FlexLoop-Mg simulations (Figure 2), the coordination structure of the Mg2+-GMN complex can be illustrated by the corresponding g(r) of Mg-O(water) and their CN distribution (averaged over triplicated-simulations). It is clear that the first-shell peak and the integration coordination number of both simulations are significantly identical to that observed from the simulation of Mg2+ in aqueous solution. However, an important difference between the ion-water RDFs from the simulations of Mg2+ in the periplasmic loop containing the GMN sequence and that in aqueous solution is explicitly pronounced by the second-shell peak with a lower intensity of RDF. An analysis of the CN distribution shows that the most preferable number of water molecules in the outer coordination sphere is 5 and 3 for FlexLoop-Mg and CtrnLoop-Mg, respectively. Apparently, a considerable loss of the secondshell water molecules was found for the hydrated Mg2+ in the selectivity filter of TmCorA in comparison with those found from the simulation of Mg2+ in aqueous solution. Evaluation of the MRT of the remaining waters at the second coordination sphere yields a value of 29.4ps for both CtrnLoop-Mg and FlexLoop-Mg. By comparing with Free-Mg2+, this longer residence time of water exchange rate indicates that the second-shell waters located in the selective filter of TmCorA become more localized and harder to exchange. The water elimination for the outer layer of the hydrated Mg2+ may be a result of repulsive interactions of the GMN residues as its metal-binding pocket is too narrow to accommodate the whole hydration structure of Mg2+ present in the bulk water. Because hydrogen bonding of the second-shell waters to the hexahydrated Mg2+ is relatively weak, the interactions can be substituted via hydrogen bond acceptor atoms (oxygen or nitrogen) from amino acid residues in the selectivity filter.

15 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 39

Figure 3 (A) g(r) of the Mg-O of G312 and of N314 and (B) inner and outer shells of the coordination structure taken from MD snapshots shown together with hydrogen bonds (dashed line). Mg2+ is colored as the green sphere. The first-shell waters are drawn as a ball and stick, and the second-shell waters are shown as a stick in a white sphere. G312 and N314 side chains are drawn as a stick presentation. Residues located within the second shell are indicated by a pink sphere.

To demonstrate amino acid substitution at the second coordination sphere, an analysis of Mg-O(G312) and Mg-Oδ(N314) RDFs (Figure 3A) obtained from CtrnLoop-Mg and FlexLoopMg simulations clearly shows a single peak with maximum about 4.2-4.3Å from the metal center. It is recognizable that by the distance of about 4.2Å, the GMN residues do not engage the interactions as ligand in the first coordination shell of Mg2+. The location of the RDF peak is essentially the same in comparison with the second RDF peak of Mg-O(water) observed from the Free-Mg2+ simulation. This implies that G312 and N314 could serve as the second-shell ligand 16 ACS Paragon Plus Environment

Page 17 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

when the metal ion is bound to the TmCorA selectivity filter. The integration of the peaks gives a value of coordination number between 4 and 5 for each RDFs, which is consistent with the composition of five G312 and N314 residues from the pentamer structure. With a visual inspection, the orientation of O(G312) and Oδ(N314) atoms appears to interact with hydrogen atoms of the first-shell water through hydrogen bonding (Figure 3B). To illustrate such hydrogen-bonding interactions, hydrogen bonds between ligands in the first and second coordination shell of Mg2+ were analyzed by employing a geometric definition of hydrogen bonds with the criteria of the distance between H/donor (N or O) and H/acceptor (usually O) not exceeding 3.0 Å and the hydrogen bond angle (H/donor/acceptor) larger than 65°. Using the criteria above, the MD trajectories were used to count the number of hydrogen bonding contacts from six waters of the inner shell to waters of the outer shell as well as the occupancy of hydrogen bond for the pairs of the first-shell water and the five G312 (denoted as w1-G312) and N314 (denoted as w1-N314). The results are shown in Figure 4.

Figure 4. (A) Probability distribution of the number for hydrogen bonding between the first-shell water (w1) and the second-shell waters (w2) and (B) occupancy of hydrogen bond for the pairs of w1-G312 and w1-N314.

17 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 39

Figure 4A shows the number of hydrogen bonds between the inner-- and outer-shell waters (denoted as w1-w2) obtained from Free-Mg, CtrnLoop-Mg and FlexLoop-Mg. It can be clearly seen that the number of water-water hydrogen bonds observed from the Free-Mg simulation is greater than that found for CtrnLoop-Mg and FlexLoop-Mg. Figure 4B illustrates interactions between the inner-shell waters and the outer-shell residues (G312 and N314) of each chain. This confirms that the loss of waters in the outer shell of Mg2+ in aqueous solution is compensated by hydrogen bond formation between the first-shell water and a ring of five carbonyl oxygen atoms of G312 as well as of five amide oxygen atoms of N314. Both CtrnLoopMg and FlexLoop-Mg results show that the signature GMN sequence hydrogen bonded to water molecules in the inner coordination sphere substitutes the interactions of the inner and outer-shell waters (w1-w2). It should also be noted that due to the rigidity of the backbone employed during the simulation, the hydrogen bonds observed from CtrnLoop-Mg exhibit a higher occupancy with respect to FlexLoop-Mg.

Simulations of the full-length protein The 100ns MD trajectories of three TmCorA systems consisting of FLCorA, FLCorAnoMgall and FLCorA-noMgDCS were generated. The backbone RMSD with respect to the starting structure of MD simulation was inspected to monitor behavior of the simulation systems (Figure 5). The low RMSD value during the first ~6ns was due to the restraints imposed on the protein backbone. RMSD started to increase after removing the restraints. During the course of MD simulations, all three systems showed a gradual increase in RMSD, except for FLCorA-noMgall, which exhibits a rapid change in the first 40ns. However, they reach a steady state after 50ns and

18 ACS Paragon Plus Environment

Page 19 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

remain stable until the end of the simulation (100ns) for all simulated systems. A small RMSD fluctuation between 3 and 4 Å suggests the overall stability of the protein structure and a wellbehaved system of the simulations.

Figure 5 Backbone RMSD relative to the starting structure versus MD time for FLCorA, FLCorA-noMgDCS and FLCorA-noMgall.

A further analysis of structural and dynamical fluctuation of the protein is focused on five different regions: cytoplasmic domains (residues 30-239), stalk (residues 240-290), inner transmembrane helices (or TM1 residues 291-310), the GMN sequence (residue 312-314) and outer transmembrane helices (or TM2 residues 327-349). Compared with RMSDs, the FLCorAnoMgall simulation shows, during the first 40ns, a rapid increase in RMSD of cytoplasmic 19 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 39

domain greater than FLCorA and FLCorA-noMgDCS simulations, which exhibit a gradual increase in RMSD. However, after 40ns of simulations, FLCorA-noMgall simulation shows a drop in RMSD to a value comparable to the other two simulated systems. When the simulation time reaches 100 ns, this domain displays the RMSD profile approaching a similar degree of structure deviation for all simulated systems. For the structure of transmembrane pore, it appears that TM1 has the smallest deviation for all simulated systems. A noticeable effect due to the absence of Mg2+ in DCS can be clearly seen by the higher RMSD values of the stalk obtained from FLCorA-noMgall and FLCorA-noMgDCS simulations compared to that of FLCorA. During the first 40ns of simulations, the RMSD of the stalk helix domain of FLCorA-noMgall rises to 4Å, showing the highest difference in RMSD fluctuation compared to the other two systems. However, the stalk-RMSD of FLCorA-noMgall drops for the next 40ns and rises back to about 4Å again during the last 20ns. On the other hand, RMSD of FLCorA-noMgDCS rises gradually to the average value of ~4Å during the last 20ns. It should be noted that both simulations (FLCorAnoMgall and FLCorA-noMgDCS) revealed the structure of stalk with RMSD greater than that of FLCorA over the 100ns of simulation. This implies that the absence of Mg2+ in DCS of TmCorA produces an apparent impact on the motion of the stalk domain. It should be noted that the RMSD of TM2 in the FLCorA_noMgall simulation exhibits large fluctuations. This is probably due to the fact that TM2 is a short segment and being connected to a periplasmic loop, which appears to be highly mobile. Nevertheless, our simulations did not show an apparent relationship between the fluctuation of TM2 and its impact on coordination structure of Mg2+ in the selectivity filter. Coordination structure of Mg2+ in the selectivity filter in full-length TmCorA

20 ACS Paragon Plus Environment

Page 21 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The two RDFs of Mg-O(water) from FLCorA and FLCorA-noMgDCS simulations (Figure 6A) show a sharp well-defined first peak and a broader second peak, presenting a similar pattern in comparison with the g(r) of Mg-O(water) of Free-Mg, CtrnLoop-Mg and FlexLoop-Mg systems (Figure 2). The n(r) of the first peak corresponds to six water molecules that strongly bound to Mg2+ in an octahedral arrangement. During the simulation, no exchange phenomenon was observed for waters located within the first coordination sphere. For the second-shell RDF peak, there is no clear separation at the minimum of the peak, implying a swap of water molecules between the outer coordination sphere and the bulk. These results suggest that water molecules in the outer shell are relatively more mobile. The coordination number of the outershell waters with the highest occurrence is 8 for FLCorA and 10 for FLCorA-noMgDCS (Figure 6B). It appears that the greater number of the second-shell water found in FLCorA-noMgDCS simulation is associated with the side chain rearrangement of the GMN residues, giving rise to an increase in the cavity size of the selectivity filter (Figure 7B). The MRT of the outer-shell water is of 47.1ps for FLCorA, which appears to be considerably longer than that of 35.2ps for FLCorA-noMgDCS. A shorter water residence time indicates more mobile water in the selectivity filter of the FLCorA-noMgDCS system.

Figure 6 (A) g(r) and n(r) of the Mg-O(water) distance of FLCorA and FLCorA-noMgDCS 21 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 39

and (B) the corresponding probability distribution of coordination number of the inner and outer shells (CN1 and CN2).

The impact of missing Mg2+ in the DCS site on the metal binding at the GMN pocket can be further demonstrated by the Mg-O(G312) and Mg-Oδ(N314) RDFs with their corresponding integration number and distance distribution plots (Figure 7A). In the case of the Mg-O(G312) RDFs, only one broad peak is visible with a maximum at the distance of ~4.2Å from the metal center. The MD of full-length TmCorA revealed again that all the five G312 residues do not bind to the Mg2+ ion directly. Instead, they interact with the hexahydrated Mg2+. For the FLCorA simulation, the Mg-O(G312) RDF gives the corresponding integration number of about four out of five G312 residues. An evaluation of hydrogen bond contacts from the MD trajectory indicates that interactions between the backbone carbonyl oxygen of G312 and the first-shell water molecules are maintained in majority. For the FLCorA_noMgDCS system, it was, however, found that only two out five G312 residues of the GMN motif participate in hydrogen bond interactions with the first-shell waters. A considerable decrease of the interactions to the innershell waters suggests the selectivity filter loosely binds to the metal ion. Comparison of the distance distributions of Mg-O(G312) between FLCorA and FLCorA_noMgDCS shows the second-shell ligands of the FLCorA_noMgDCS system tend to drift away from the metal center (inset figure 7A). This indicates that the absence of Mg2+ in the DCS site can provoke disruption of hydrogen bonding between the GMN motif to the first-shell waters of Mg2+.

22 ACS Paragon Plus Environment

Page 23 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 7 (A) g(r) of the Mg-O of G312 and of N314 for FLCorA and FLCorA-noMgDCS systems. The inset figures show the Mg-O distance distribution of the G312 and N314 residues. Each chain (A, B, C, D and E) is colored individually. (B) Inner and outer shells of the coordination structure taken from MD snapshots. The model representation and color scheme used in this figure are the same as that used in Fig 4.

The impact of a metal vacancy in the DCS site to the binding of Mg2+ at the selectivity filter is clearly seen in the Mg-Oδ(N314) RDFs. For FLCorA_noMgDCS, hydrogen bonds made up by five N314 residues with the inner-shell waters are all disrupted as the Mg-Oδ(N314) peak is shifted to a greater radius and becomes featureless with a maximum roughly at 5.1Å (Figure 7A). One can see a broader distribution of the Mg-Oδ(N314) distance compared to the narrower

23 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 39

distance distribution obtained from FLCorA. By visual inspection, hydrogen bond interactions between the N314 side chains of FLCorA and the inner-shell water are still detectable, whereas all five amide side chains of N314 of the FLCorA_noMgDCS system point away from the center of the filter (Figure 7B). From MD results, it can be concluded that missing Mg2+ in the DCS site of the cytoplasmic domain affects the structure of the GMN scaffold, giving rise to a decrease in the interactions between the selectivity filter residues and the hexahydrated Mg2+. Another interesting observation is that the interactions of the G312 appear to be more stable over simulation time than N314. This is most clearly seen by a comparison of the RDFs and distance distribution plots of the Mg-O(G312) and those of the Mg-Oδ(N314) as shown in Figure 7A. The Mg-Oδ(N314) plots have less well-defined shape in comparison to the MgO(G312). This implies that the G312 residues contribute to greater stability of the hydrated Mg2+-GMN complex than the N314 residues. From the simulations, the absence of Mg2+ in DCS produces a significant impact on the structural scaffold at the GMN binding site, giving rise to the disruption of hydrogen bonds between water molecules in the inner shell and G312 and N314 of each protomer. Interactions between the hexahydrated Mg2+ and the selectivity filter become weaker. This is the result of a change in the backbone conformations of G312 as well as the orientation of side chain of N314 in the periplasmic loop, leading to the opening of the extracellular mouth of the channel. It has been observed that at the simulation time of ~50ns, the hexahydrated Mg2+ ion tends to move towards the permeation pore (Figure 8A). The Z-position distribution plot shows a majority of events that for FLCorA_noMgDCS, the Mg2+ positions are closer to the pore entry of the channel than that of the FLCorA simulation. The results again confirm that the G312 and N314 are responsible for the stability of the hexahydrated Mg2+ in the selectivity filter.

24 ACS Paragon Plus Environment

Page 25 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 8 (A) fluctuation and distribution of the corresponding z-position of Mg2+ in the selectivity filter (SF) and (B) a pore diameter plot of the MD snapshot

One can see that the diameter of the GMN scaffold structure of FLCorA_noMgDCS increases substantially (Figure 8B). This suggests that the pocket at the selectivity filter becomes wider. Nevertheless, the Mg2+ across the membrane was not observed because the pore of the channel remains unchanged. As expected, the study’s simulations also show a significant structure perturbation of the cytoplasmic domain and the stalk helices upon removal of Mg2+ from the DCS sites. Since the metal ions at the DCS sites are surrounded by several negatively charged residues of the two adjacent chains, the absence of Mg2+ from the DCS sites possibly leads to electrostatic repulsions among these residues, unlocking the channel to a conductive state. However, during the course of the simulations, the channel pore did not open wide enough to permit translocation of a hydrated Mg2+ ion. Principal component analysis To identify the specific conformational changes and motional processes associated with the presence or absence of Mg2+ in DCS sites, post-processing MD including per-residue RMSF 25 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 39

and essential dynamics analysis were performed between 60 and 100ns of MD trajectories of the full-length protein systems. Figure 9 shows the residue-based Cα RMSFs of the FLCorA, FLCorA_noMgDCS and FLCorA_noMgall over the last 40ns of the simulation time. It is evident that the FLCorA_noMgDCS trajectory reveals significantly more motion in the cytoplasmic (residues 150-195) and stalk helices (residues 245-290) as compared to other simulated systems. In addition, the inset plots show the largest fluctuation of metal-coordinating residues including E88, D89, D175 and D253 in the FLCorA_noMgDCS trajectory. The interactions between the metal ions and the coordinated residues are expected to lock the helix bundle together to stabilize the channel in a non-conductive conformation. In the absence of Mg2+ at the DCS site, charge repulsion of the negatively charged carboxyl groups of these residues could play a dominant role in facilitating the stalk helix to move toward the open conformation.

Figure 9 Per-residue Cα RMSFs from the FLCorA (black), FLCorA_noMgDCS(red) and FLCorA_noMgall (blue) trajectories analyzed between 60 and 100 ns. The insets highlight an increase of RMSF for the metal-binding residues (E88, D89, D175 and D253) in the DCS site.

26 ACS Paragon Plus Environment

Page 27 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

PCA approach was applied to find the essential protein motions. In most cases, the first few eigenvectors or principal components are sufficient to capture the dominant modes of largescale motions. As shown in Figure 10A and 10B, the interpolated structures along the first principal component (PC1) were generated to visualize the comparative motion difference between the FLCorA and FLCorA_noMgDCS trajectories. PC1 result of both trajectories clearly show that the cytoplasmic domain and extracellular loop appear to be the first dominant motion, which corresponds to rocking and twisting movement. As a comparison, PC1 can capture a different magnitude of the motion in the stalk region, where the range of motion observed in the FLCorA_noMgDCS trajectory is considerably greater than that in the FLCorA trajectory. Projections of the trajectories onto the conformational subspace described by PC1 and PC2 are illustrated in Figure 10C. It is obvious that the protein is sampling different conformational space during the simulation. To better understand these functional motion, structures and animations representing the major collective motions along PC1 and PC2 are provided as Supporting Information.

27 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 39

Figure 10 Collective motions of CorA from the FLCorA (A) and FLCorA_noMgDCS (B) trajectories, analyzed between 60 and 100 ns, along the PC1. Motions that highlight the significant regions; stalk (side view) and periplasmic loop (extracellular view), are illustrated as linear interpolations between the extreme projections of the structures onto the PC1. Color scale from blue-white-red depicts low to high atomic displacement. Black arrows indicate the position of Mg2+ in DCS site. Dotted areas highlights a different magnitude in the stalk motion compared between the two trajectories. Cytoplasmic domain of three protomers has been omitted for clarity. (C) PCA scatter plot of structural snapshots along the first two principal components with a color scale running from blue (at 60ns) to red (the end at 100 ns).

Discussion 28 ACS Paragon Plus Environment

Page 29 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

It has generally been accepted that Mg2+ forms an octahedral complex in the first hydration sphere. This is supported by several sources of information, especially from X-ray diffraction methods37, Cambridge Structural Database42 and computational studies at various levels of accuracy39, 41, 43. Within the context of this study’s model approach, it was found that the Mg2+ ion in the selectivity filter of TmCorA binds strongly to six water molecules as an octahedral first coordination sphere. In comparison with Mg2+ in aqueous solution, no significant change was observed in the first-shell structure. As expected, the results are generally in good agreement with the conception that the most common first-shell ligand of Mg2+ is water. An apparent difference is the structure of the second coordination layer that consists of water molecules, G312 and N314. This study’s simulations revealed that the GMN motif of TmCorA utilizes the backbone oxygen of G312 and the side-chain amide oxygen of N314 as the secondshell ligand for stabilizing the hexahydrated Mg2+ through hydrogen bonds with the inner-shell waters. In order for the Mg2+ recognition by TmCorA, G312 and N314 residues are placed at a proper distance and orientation to enhance the metal binding affinity in the selectivity filter. This conclusion is further supported by evidence from additional simulations of Free-M systems where M = other metal ion (Ca2+, K+ and Na+). The structure features of the first and second peaks of g(r) of the M-O(water) show a substantial difference in the effective hydrated sphere of the metal ions compared to those of Mg2+ (supplementary Fig. S1), suggesting a different coordination structure and geometry of the metals. Moreover, MD simulations of FlexLoop-M model systems show that only Mg2+ remains in the selectivity filter as it possesses the minimal displacement along the symmetry axis of the channel (Figure 11). On the other hand, other three cations exhibit a large fluctuation. The selectivity filter cannot trap these metal ions within the GMN scaffold. Instead, they immediately leave the filter after a few ten or hundred picoseconds

29 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 39

of the simulations. These results indicate that the metal binding pocket in the GMN selectivity filter does not match the size and coordination geometry of these metal ions in solution.

Figure 11 Fluctuation of z-position of the metal ions with respect to its initial z-coordinates. The coordinates were taken from MD simulations of FlexLoop-M model systems (where M = K+ , Na+, Ca2+ and Mg2+).

Amino acid residues as the second-shell ligand of metals have been commonly found in several metalloproteins. 44-50 A statistical analysis in PDB shows that peptide backbone and polar or charged residues of the proteins are located at the second coordination layer, serving as the second-shell ligand.

51

It has been proposed that the important role of residues in the second

coordination layer in metal binding selectivity can be described the combined effects of charge, size, coordination number and geometry of the metal ions, although not as direct as for the first shell. Among these, Asn side chains are commonly found in the metal second coordination layer that can engage in hydrogen-bond formation. It should be noted that not all water molecules in the second shell are, however, removed. The full-length simulations as well as truncated models confirm some water molecules remain 30 ACS Paragon Plus Environment

Page 31 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

within the second coordination sphere to support the binding of G312 and N314 residues to the metal. In addition, it was found that the amide nitrogen of N314 does not interact directly with the first shell water since it prefers to act as a hydrogen bond donor atom. This notification is confirmed by this study’s simulations showing that the N314 amide group interacts only with hydrogen bond acceptor atoms of water molecules in the second shell. From these results, both G312 and N314 contribute to the stability of the inner coordination structure of Mg2+ greater than Ca2+, K+ and Na+ ions. From this study, it can be concluded that the G312 and N314 residues act as the second-shell ligand, responsible for stabilizing the inner-shell ligands and fine-tuning of the orientation of metal-ligand conplex. The coordination size and coordination geometry of the metal ions should be taken into account for the selectivity property of CorA Mg2+ channels.

Conclusions This study’s simulations have demonstrated the role of the GMN residues as the secondshell ligand in TmCorA Mg2+ channel by exploring the coordination structure of Mg2+ within the selectivity filter. The crystal structure of the full-length and of periplasmic loop of TmCorA were applied in the study. The results show that the structure of the inner coordination shell of Mg2+ in the selectivity filter is not significantly different from Mg2+ in bulk water. However, a noticeable difference is found at the outer coordination shell. The hydration number of the second shell of Mg2+ in the selectivity filter decreases with respect to that in bulk water. The results clearly show that G312 and N312 residues act as the second-shell ligand by interacting with the inner-shell water although not all five GMN are involved in the interactions. An impact on the structure of the selectivity filter due to removal of Mg2+ ions from the DCS sites of the protein can be seen

31 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 39

by the loss of Mg2+ binding to the GMN residues. This study thus confirms that the GMN signature sequence of TmCorA is essential for divalent cation selectivity.

Acknowledgments This research has been supported by the Ratchadaphiseksomphot Endowment Fund, Chulalongkorn University (CU-56-408-HR) to PS, the 90th Anniversary of Chulalongkorn University to SK, the 100th Anniversary Chulalongkorn University Fund for Doctoral Scholarship to PB and the Development and Promotion of Science and Technology Talented Project (DPST) for Doctoral Scholarship to SK and PW.

Supporting Information Supplementary materials include Figures and Movies.

32 ACS Paragon Plus Environment

Page 33 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

References (1)

Maguire, M. Magnesium: A Regulated and Regulatory Cation. Met Ions Biol Syst. 1990,

26, 135–153. (2)

Maguire, M.; Cowan, J. Magnesium Chemistry and Biochemistry. Biometals 2002, 15,

203-210. (3)

Maguire, M. E. Magnesium Transporters: Properties, Regulation and Structure. Front.

Biosci. 2006, 11, 3149-3163. (4)

Rude, R. K. Magnesium Deficiency: A Cause of Heterogeneous Disease in Humans. J.

Bone Miner. Res. 1998, 13, 749-758. (5)

Saris, N. E.; Mervaala, E.; Karppanen, H.; Khawaja, J. A.; Lewenstam, A. Magnesium.

An Update on Physiological, Clinical and Analytical Aspects. Clin. Chim. Acta 2000, 294, 1-26. (6)

Hartwig, A. Role of Magnesium in Genomic Stability. Mutat. Res. 2001, 475, 113-121.

(7)

Moncrief, M. B. C.; Maguire, M. E. Magnesium Transport in Prokaryotes. J. Biol. Inorg.

Chem. 1999, 4, 523-527. (8)

Smith, R. L.; Maguire, M. E. Microbial Magnesium Transport: Unusual Transporters

Searching for Identity. Mol. Microbiol. 1998, 28, 217-226. (9)

Bui, D. M.; Gregan, J.; Jarosch, E.; Ragnini, A.; Schweyen, R. J. The Bacterial

Magnesium Transporter Cora Can Functionally Substitute for Its Putative Homologue Mrs2p in the Yeast Inner Mitochondrial Membrane. J. Biol. Chem. 1999, 274, 20438-20443. (10)

Moomaw, A. S.; Maguire, M. E. Cation Selectivity by the Cora Mg2+ Channel Requires

a Fully Hydrated Cation. Biochemistry 2010, 49, 5998-6008. (11)

Kehres, D. G.; Maguire, M. E. Structure, Properties and Regulation of Magnesium

Transport Proteins. Biometals 2002, 15, 261-270.

33 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(12)

Page 34 of 39

Xia, Y.; Lundback, A. K.; Sahaf, N.; Nordlund, G.; Brzezinski, P.; Eshaghi, S. Co2+

Selectivity of Thermotoga Maritima Cora and Its Inability to Regulate Mg2+ Homeostasis Present a New Class of Cora Proteins. J. Biol. Chem. 2011, 286, 16525-16532. (13)

Niegowski, D.; Eshaghi, S. The Cora Family: Structure and Function Revisited. Cell.

Mol. Life Sci. 2007, 64, 2564-2574. (14)

Eshaghi, S.; Niegowski, D.; Kohl, A.; Martinez Molina, D.; Lesley, S. A.; Nordlund, P.

Crystal Structure of a Divalent Metal Ion Transporter Cora at 2.9 Angstrom Resolution. Science 2006, 313, 354-357. (15)

Lunin, V. V.; Dobrovetsky, E.; Khutoreskaya, G.; Zhang, R.; Joachimiak, A.; Doyle, D.

A.; Bochkarev, A.; Maguire, M. E.; Edwards, A. M.; Koth, C. M. Crystal Structure of the Cora Mg2+ Transporter. Nature 2006, 440, 833-837. (16)

Payandeh, J.; Pai, E. F. A Structural Basis for Mg2+ Homeostasis and the Cora

Translocation Cycle. EMBO J. 2006, 25, 3762-3773. (17)

Dalmas, O.; Sompornpisut, P.; Bezanilla, F.; Perozo, E. Molecular Mechanism of Mg2+-

Dependent Gating in Cora. Nat Commun 2014, 5, 3590. (18)

Pfoh, R.; Li, A.; Chakrabarti, N.; Payandeh, J.; Pomes, R.; Pai, E. F. Structural

Asymmetry in the Magnesium Channel Cora Points to Sequential Allosteric Regulation. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 18809-18814. (19)

Chakrabarti, N.; Neale, C.; Payandeh, J.; Pai, E. F.; Pomes, R. An Iris-Like Mechanism

of Pore Dilation in the Cora Magnesium Transport System. Biophys. J. 2010, 98, 784-792. (20)

Palombo, I.; Daley, D. O.; Rapp, M. Why Is the Gmn Motif Conserved in the

Cora/Mrs2/Alr1 Superfamily of Magnesium Transport Proteins? Biochemistry 2013, 52, 48424847.

34 ACS Paragon Plus Environment

Page 35 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(21)

Guskov, A.; Nordin, N.; Reynaud, A.; Engman, H.; Lundback, A. K.; Jong, A. J.;

Cornvik, T.; Phua, T.; Eshaghi, S. Structural Insights into the Mechanisms of Mg2+ Uptake, Transport, and Gating by Cora. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 18459-18464. (22)

Xiang, Z.; Soto, C. S.; Honig, B. Evaluating Conformational Free Energies: The Colony

Energy and Its Application to the Problem of Loop Prediction. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 7432-7437. (23)

Humphrey, W.; Dalke, A.; Schulten, K. Vmd: Visual Molecular Dynamics. J. Mol.

Graph. Model. 1996, 14, 33-38. (24)

Perozo, E.; Cortes, D. M.; Sompornpisut, P.; Kloda, A.; Martinac, B. Open Channel

Structure of Mscl and the Gating Mechanism of Mechanosensitive Channels. Nature 2002, 418, 942-948. (25)

Aksimentiev, A.; Sotomayor, M.; Wells, D. Membrane Proteins Tutorial. University of

Illinois at Urbana-Champaign Beckman Institute for Advanced Science and Technology Theoretical and Computational Biophysics Group Computational Biophysics Workshop 2012, 143. (26)

Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.;

Skeel, R. D.; Kalé, L.; Schulten, K. Scalable Molecular Dynamics with Namd. J. Comput. Chem. 2005, 26, 1781-1802. (27)

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., et al. Charmm: The Biomolecular Simulation Program. J. Comput. Chem. 2009, 30, 1545-614.

35 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(28)

Page 36 of 39

Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L.

Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926-935. (29)

Darden, T.; York, D.; Pedersen, L. Particle Mesh Ewald - an N.Log(N) Method for

Ewald Sums in Large Systems. J. Chem. Phys. 1993, 98, 10089-10092. (30)

Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. Numerical Integration of the Cartesian

Equations of Motion of a System with Constraints: Molecular Dynamics of N-Alkanes. J Comput Phys 1977, 23, 327-341. (31)

Miyamoto, S.; Kollman, P. A. Settle: An Analytical Version of the Shake and Rattle

Algorithm for Rigid Water Models. J. Comput. Chem. 1992, 13, 952-962. (32)

Feller, S. E.; Zhang, Y. H.; Pastor, R. W.; Brooks, B. R. Constant-Pressure Molecular-

Dynamics Simulation - the Langevin Piston Method. J. Chem. Phys. 1995, 103, 4613-4621. (33)

Martyna, G. J.; Tobias, D. J.; Klein, M. L. Constant-Pressure Molecular-Dynamics

Algorithms. J. Chem. Phys. 1994, 101, 4177-4189. (34)

Smart, O. S.; Neduvelil, J. G.; Wang, X.; Wallace, B. A.; Sansom, M. S. Hole: A

Program for the Analysis of the Pore Dimensions of Ion Channel Structural Models. J. Mol. Graph. 1996, 14, 354-360. (35)

Seeber, M.; Cecchini, M.; Rao, F.; Settanni, G.; Caflisch, A. Wordom: A Program for

Efficient Analysis of Molecular Dynamics Simulations. Bioinformatics 2007, 23, 2625-7. (36)

Helm, L.; Merbach, A. E. Inorganic and Bioinorganic Solvent Exchange Mechanisms.

Chem. Rev. 2005, 105, 1923-1959. (37)

Caminiti, R.; Licheri, G.; Piccaluga, G.; Pinna, G. X-Ray Diffraction Study of

Magnesium Chloride Aqueous Solutions. J. Appl. Crystallogr. 1979, 12, 34-38.

36 ACS Paragon Plus Environment

Page 37 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(38)

Tongraar, A.; Rode, B. M. The Role of Non-Additive Contributions on the Hydration

Shell

Structure

of

Mg2+

Studied

by

Born-Oppenheimer

Ab

Initio

Quantum

Mechanical/Molecular Mechanical Molecular Dynamics Simulation. Chem. Phys. Lett. 2001, 346, 485-491. (39)

Tongraar, A.; Rode, B. M. Structural Arrangement and Dynamics of the Hydrated Mg2+:

An Ab Initio Qm/Mm Molecular Dynamics Simulation. Chem. Phys. Lett. 2005, 409, 304-309. (40)

Bock, C. W.; Markham, G. D.; Katz, A. K.; Glusker, J. P. The Arrangement of First- and

Second-Shell Water Molecules around Metal Ions: Effects of Charge and Size. Theor. Chem. Acc. 2006, 115, 100-112. (41)

Markham, G. D.; Glusker, J. P.; Bock, C. W. The Arrangement of First- and Second-

Sphere Water Molecules in Divalent Magnesium Complexes: Results from Molecular Orbital and Density Functional Theory and from Structural Crystallography. J. Phys. Chem. B 2002, 106, 5118-5134. (42)

Allen, F. H. The Cambridge Structural Database: A Quarter of a Million Crystal

Structures and Rising. Acta Crystallographica Section B-Structural Science 2002, 58, 380-388. (43)

Jiao, D.; King, C.; Grossfield, A.; Darden, T. A.; Ren, P. Y. Simulation of Ca2+ and

Mg2+ Solvation Using Polarizable Atomic Multipole Potential. J. Phys. Chem. B 2006, 110, 18553-18559. (44)

Maynard, A. T.; Covell, D. G. Reactivity of Zinc Finger Cores: Analysis of Protein

Packing and Electrostatic Screening. J. Am. Chem. Soc. 2001, 123, 1047-1058. (45)

Dudev, T.; Lim, C. Factors Governing the Protonation State of Cysteines in Proteins: An

Ab Initio/Cdm Study. J. Am. Chem. Soc. 2002, 124, 6759-6766.

37 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(46)

Page 38 of 39

Lesburg, C. A.; Lloyd, M. D.; Cane, D. E.; Christianson, D. W. Crystallization and

Preliminary X-Ray Diffraction Analysis of Recombinant Pentalenene Synthase. Protein Sci. 1995, 4, 2436-2438. (47)

Mertz, P.; Yu, L.; Sikkink, R.; Rusnak, F. Kinetic and Spectroscopic Analyses of Mutants

of a Conserved Histidine in the Metallophosphatases Calcineurin and Lambda Protein Phosphatase. J. Biol. Chem. 1997, 272, 21296-21302. (48)

He, Q. Y.; Mason, A. B.; Woodworth, R. C.; Tam, B. M.; MacGillivray, R. T.; Grady, J.

K.; Chasteen, N. D. Mutations at Nonliganding Residues Tyr-85 and Glu-83 in the N-Lobe of Human Serum Transferrin. Functional Second Shell Effects. J. Biol. Chem. 1998, 273, 1701817024. (49)

Vipond, I. B.; Moon, B. J.; Halford, S. E. An Isoleucine to Leucine Mutation That

Switches the Cofactor Requirement of the Ecorv Restriction Endonuclease from Magnesium to Manganese. Biochemistry 1996, 35, 1712-1721. (50)

Marino, S. F.; Regan, L. Secondary Ligands Enhance Affinity at a Designed Metal-

Binding Site. Chem. Biol. 1999, 6, 649-655. (51)

Dudev, T.; Lin, Y. L.; Dudev, M.; Lim, C. First-Second Shell Interactions in Metal

Binding Sites in Proteins: A Pdb Survey and Dft/Cdm Calculations. J. Am. Chem. Soc. 2003, 125, 3168-3180.

38 ACS Paragon Plus Environment

Page 39 of 39

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

TOC graphic

39 ACS Paragon Plus Environment