Molecular Mechanism of Biased Ligand Conformational Changes in

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

Molecular Mechanism of Biased Ligand Conformational Changes in CC Chemokine Receptor 7 Zied Gaieb, David D. Lo, and Dimitrios Morikis J. Chem. Inf. Model., Just Accepted Manuscript • DOI: 10.1021/acs.jcim.6b00367 • Publication Date (Web): 16 Aug 2016 Downloaded from http://pubs.acs.org on August 18, 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.

Journal of Chemical Information and Modeling 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 55

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

Journal of Chemical Information and Modeling

1

Molecular Mechanism of Biased Ligand

2

Conformational Changes in CC Chemokine

3

Receptor 7

4

Zied Gaieb1, David D. Lo2, Dimitrios Morikis1, *

5

1

Department of Bioengineering, 2Division of Biomedical Sciences, School of Medicine,

6

University of California, Riverside, CA, 92521, USA

7

KEYWORDS: G protein-coupled receptors, chemokines, chemokine receptors, CCR7, CCL19,

8

CCL21, computational modeling, molecular dynamics, biased ligands, molecular mechanism

9

ACS Paragon Plus Environment

1


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 55

1

ABSTRACT

2

Biased ligand binding to G protein-coupled receptors enables functional selectivity of

3

intracellular effectors to mediate cellular function. Despite the significant advances made in

4

characterizing the conformational states (transmembrane helical arrangements) capable of

5

discriminating between G protein and arrestin binding, the role of the ligand in stabilizing such

6

conformations remains unclear. To address this issue, we simulate microsecond dynamics of CC

7

chemokine receptor 7 (CCR7) bound to its native biased ligands, CCL19 and CCL21, and detect

8

a series of molecular switches that are mediated by various ligand-induced allosteric events.

9

These molecular switches involve three tyrosine residues (Y1123.32, Y2556.51, and Y2887.39), three

10

phenylalanine residues (F1163.36, F2085.47, and F2486.44), and a polar interaction between Q2526.48

11

and R2947.45 in the transmembrane domain of CCR7. Conformational changes within these

12

switches, particularly hydrogen bond formation between Y1123.32 and Y2556.51, lead to global

13

helical movements in the receptor’s transmembrane helices and contribute to the transitioning of

14

the receptor to distinct states. Ligand-induced helical movements in the receptor highlight the

15

ability of biased ligands to stabilize the receptor in different states through a dynamic network of

16

allosteric events.

17


2

Page 3 of 55

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

1


INTRODUCTION

2

Conformational diversity in G protein-coupled receptors (GPCRs) is a fundamental component

3

of the receptor’s selective properties in binding intracellular effectors, such as G protein and

4

arrestin.1 The conformational ensemble of the receptor comprises of distinguishable

5

conformational states characterized by its seven-transmembrane (TM) helical arrangements,

6

critical to coordinating the receptor’s selective binding to intracellular (INC) effectors.1–3 To

7

enable the complex function of a GPCR, bound ligands are capable of stabilizing the receptor in

8

many distinct conformational states ranging from inactive to fully active, spanning G protein,

9

arrestin, or GPCR kinase (GRK) biased conformations.1,3–6 Additionally, receptor conformations

10

are capable of discriminating between the different effector subtypes and arrestin binding

11

modes.1,7 These attributes emphasize the importance of the receptor’s conformational diversity

12

and the role of the ligands in shifting the receptor’s equilibrium in sampling its different

13

conformational states.

14

Here, we examine a key regulator of the adaptive immune response in the CC chemokine

15

receptor family, CC chemokine receptor 7 (CCR7).8 Its ligands, CCL19 and CCL21, have

16

distinct roles in the homing and functional compartmentalization of T cells and antigen-

17

presenting dendritic cells to and within the secondary lymph nodes as a result of their differential

18

chemotactic behaviors.8–10 To initiate CCR7’s cellular function, CCL19 and CCL21 have been

19

shown to selectively induce distinct signaling pathways in the cell.8 While both ligands mediate

20

their signaling through binding of intracellular Gi protein and GRK6 to CCR7, only CCL19

21

induces CCR7 internalization and desensitization through receptor phosphorylation by GRK3

22

and recruitment of β-arrestin 2.4,11 This differential binding of CCR7 to β-arrestin 2, GRK3, and


3


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 55

1

GRK6 suggests the presence of selective conformational states in CCR7 induced by its biased

2

ligands, CCL19 and CCL21.

3

Ligand binding to a GPCR drives the receptor to its corresponding conformational states

4

through different molecular switches and helical rearrangements.12 Crystallographic and

5

NMR spectroscopy studies have identified specific TM helices involved in the INC

6

rearrangements of the receptor. Specifically, upon ligand binding, TM5, TM6, and TM7 undergo

7

large helical displacements to accommodate effector binding.1,3,13,14 A recent study by Liu et al.

8

shows that β2-adrenergic receptor explores two independent equilibrium conformations in TM6

9

and TM7;1 where, ligands are capable of fine-tuning the relative population of equilibrium

10

conformational states adopted by TM6 and TM7 to accommodate biased or unbiased binding of

11

G protein and β-arrestin.1

19

F-

12

Despite these fundamental advances in characterizing the role of TM helices in selectively

13

binding INC effectors, the role of the ligands and their associated molecular mechanisms and

14

pathways in stabilizing such conformations, in particular the relative positions of TM5, TM6,

15

and TM7, remains unclear. Previous studies using molecular dynamics (MD) simulations have

16

focused on the activation mechanism in response to unbiased agonist,15–18 to the exception of a

17

few studies that discuss “known microscopic characteristics” of the biased conformational states

18

in β2-adrenergic receptor.19 However, to the best of our knowledge, detailed ligand-specific

19

structural allosteric pathways have not been fully characterized in any GPCR, let alone

20

chemokine receptors. Such structural detail is essential for not only gaining a deeper

21

understanding of how ligands operate in a selective manner, but also aids in the rational design

22

of drugs targeting desired signaling pathways.


4

Page 5 of 55

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


1

In the present study, we apply conventional MD (cMD) using Anton20 and accelerated MD

2

(aMD)21,22 to delineate the conformational changes in CCR7 in response to its biased ligands,

3

CCL19 and CCL21. From all simulations, we identify various molecular switches that undergo

4

various ligand-specific conformational changes. Additionally, during the cMD simulation, we

5

capture ligand-associated allosteric pathways, starting at the ligand-binding site, and propagating

6

to these molecular switches. Highly correlated helical movements are coupled to the

7

conformational changes occurring within the molecular switches and illustrate transitioning of

8

the receptor between two distinct states.

9


5


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

1

METHODS

2

Nomenclature

Page 6 of 55

3

Residues are represented as a one-letter amino acid code and a number corresponding to their

4

order in the sequence. Additional sequence information is represented in the superscripts.

5

Superscripts of ligand residues denote the ligand it belongs to; and receptor residues are

6

numbered according Ballesteros–Weinstein and convey the helix number and position of each

7

residue relative to the most conserved residue in the helix.23

8

System Setup for Molecular Dynamics Simulations

9

The structure of CCR7 and CCL19/CCL21-bound CCR7 were modeled after the newly

10

determined crystal structure of the chemokine receptor CXCR4 bound to a viral chemokine

11

antagonist vMIP-II (PDB code 4RWS).24 The structures of the CCL19 and CCL21 ligands and

12

CCR7 receptor were generated separately and were subsequently assembled into complexes

13

using the vMIP-II-CXCR4 crystal structure as a template. First, the mouse sequences of CCR7,

14

CCL19, and CCL21 were extracted from UniProt (http://www.uniprot.org/).25 Alignments of

15

receptor and ligand sequences with CXCR4 (PDB 4RWS) and vMIP-II (PDB 4RWS)

16

respectively

17

(http://www.ebi.ac.uk/Tools/msa/clustalw2/).26,27 Sequence alignments revealed a percent

18

identity of 27% between vMIP-II and CCL19 and 23% between vMIP-II and CCL21. Despite

19

the low sequence identity, vMIP-II was used as a template since all chemokine ligands have a

20

common structural motif that is structurally maintained by two disulfide bridges and a flexible

21

N-terminal domain28 (ligand-NTD) (residues 1-7). The chemokine ligand has a canonical fold: a

22

N-loop, followed by three antiparallel β-strands, an α-helix, and a flexible C-terminal domain

23

(CTD). The three β-strands are connected by two loops: the 30s-loop, and 40s-loop. Unlike

were

then

performed

using

the

ClustalW2

server

6


Page 7 of 55

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


1

CCL19, the full sequence of CCL21 comprises of a long and positively charged CTD that allows

2

binding to glycosaminoglycans. Given that CTD does not interact with CCR7 and not involved

3

in receptor activation,8 a truncated version of CCL21 (residues 1-83) was modeled along with the

4

full sequence of CCL19 (residues 1-83) using Modeller9.11.29 Both models assume a robust core

5

forming the canonical chemokine motif and show low rmsd to their respective experimental

6

structures: rmsd between CCL19 and CCL21 models and their respective NMR structures is

7

1.143 Å and 1.184 Å, respectively.28,30 Modeling of the ligands using vMIP-II was done to obtain

8

three-dimensional structures of each ligand in their bound conformation to CCR7.

9

Sequence alignment of CCR7 with CXCR4 (sequence extracted from PDB file 4RWS) showed

10

a 36% identity and 47% similarity (with 39% identity and 63% similarity for the transmembrane

11

domain

12

[http://imed.med.ucm.es/Tools/sias.html],

13

physicochemical properties of aligned amino acids: aromatic (F, Y, W), hydrophobic (V, I, L, M,

14

A, F, W), aliphatic (V, I, L), positively charged (R, K, H), negatively charged (D, E), polar (Not

15

charged) (N, Q), or small (A, T, S)). The receptor is composed of three domains: the

16

transmembrane (TMD, residues 26-305), N-terminal (CCR7-NTD, residues 1-25), and C-

17

terminal (CTD, residues 306-354) domains that have different structural motifs and were

18

therefore modeled separately. The CCR7-NTD is an intrinsically disordered segment of the

19

receptor, that could not be detected in several crystallographic and NMR studies of many

20

GPCRs.24,31–34 Therefore, the CCR7-NTD was modeled using Modeller9.11 in a random coil

21

conformation and was positioned away from the extracellular loops to avoid clashes between the

22

CCR7-NTD and the ligand when positioning CCL19 and CCL21. In contrast, the CTD contains

23

an α-helical sequence motif F(RK)xx(FL)xxx(LF) and an intrinsically disordered segment.32

(TMD)).

(Sequence

similarities by

taking

were into

calculated consideration

using the

SIAS following

7



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 55

1

Given the success of i-TASSER in predicting protein structure, the full CTD sequence was

2

modeled using the i-TASSER server (http://zhanglab.ccmb.med.umich.edu/I-TASSER/).35,36 The

3

top generated model had a relatively low confidence score of -2 (C-score range [-5, 2]), which is

4

due to the intrinsically disordered segment of the CTD. However, more importantly, the

5

produced i-TASSER model still generated the canonical helical domain, known as helix 8

6

(residues 309-319), present in many other GPCRs such as CCR5 (PDB 4MBS),32 CXCR1 (PDB

7

2LNL),34 and β2-adrenergic receptors (PDB 3P0G, 3SN6, 3NY8, 2RH1).14,33,37,38 Lastly, The

8

TMD of CCR7 is composed of 7 TM helices (TM1: residues 27-61; TM2: residues 68-96; TM3:

9

residues 102-134; TM4: residues 144-172; TM5: residues 194-228; TM6: residues 239-267;

10

TM7: residues 274-303) connected by three extracellular loops (ECL) and three intracellular

11

loops (ICL). Templates exhibiting more than 35% sequence homology can be used to generate

12

highly reliable GPCR models.39 Therefore, with a 39% sequence identity, CXCR4 was used as a

13

template to generate a three-dimensional structure of CCR7’s TMD (residues 19-303) using

14

Modeller9.11.29

15

The modeled ligands and receptor were then used to construct the complex structures

16

following the published binding mode of the CXCR4-vMIP-II crystal structure (PDB code

17

4RWS).24 The ligands bind the receptor via a two-site mechanism where the CCR7-NTD binds

18

the chemokine ligand, site I, followed by ligand-NTD binding inside the receptor’s extracellular

19

pocket, site II.24 For site I, various NMR studies of chemokine ligands with their receptor NTDs

20

show similar interactions where the receptor-NTD interacts with the N-loop, 40s loop, and third

21

β-strand of the ligand.28,30,40,41 Therefore, chemokine-binding to its receptor involves two

22

experimentally determined structural constraints: the receptor-NTD and extracellular binding

23

pocket in sites I and II respectively. Published crystallographic structures of chemokine-GPCR 8


Page 9 of 55

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


1

complexes (PDB 4RWS and 4XT1/3) show very similar binding modes (root-mean-square

2

deviation (rmsd) of 1.3 Å) and comply with the aforementioned structural constraints.24,42 In light

3

of the new chemokine-GPCR structures, CCL19 and CCL21 were positioned using CXCR4-

4

vMIP-II crystal structure (PDB code 4RWS)24 as a template using Chimera.43 The binding mode

5

complies with previously published NMR chemical shift perturbations between CCR7-NTD and

6

its ligands CCL19 and CCL21.28,30 These chemical shift perturbation studies, carried by Love et

7

al, indicate the presence of a binding interface in each ligand (N-loop, 40s loop, and third β-

8

strand) that interacts with the CCR7-NTD.

9

To assess the stability of the receptor terminal domains, we calculate rmsd time series of both

10

domains (CCR7-NTD and CCR7-CTD) in all of our simulations (cMD and aMD for each of the

11

two complexes). All rmsd plots show stabilization of both domains in all simulations (Fig. S1 in

12

Supporting Material). Furthermore, contact maps between the CCR7-NTD and the ligand

13

calculated from both aMD and Anton simulations reveal very close agreement with NMR data of

14

CCL19 and CCL21 with CCR7-NTD (Fig. S2).28,30

15

In site II, both ligand-NTDs adopt different poses in the receptor-binding pocket during our

16

simulations (Fig. S3). Fig. S3 displays contacts between CCR7 and its ligand-NTDs at less than

17

5 Å and a residence time of 50% or more of the equilibrated cMD recorded frames (equilibrated

18

time domain are specified in Fig. 1 caption). CCL21 N-terminus forms a salt bridge with

19

E1694.60, which facilitates the ligand’s interaction with TM4, TM5, and TM6. In addition, D2CCL21

20

forms a salt bridge with the second extracellular loop (ECL2), R185ECL2; and D6CCL21 interacts

21

with K261.24 and R301.28. Unlike CCL21, CCL19 forms no interactions with TM4, TM5, and TM6

22

and its charged residues are stabilized with the charge-complementary residues in TM1: D4CCL19

23

interacts with K261.24, R301.28, K331.31, and E941.64; E6CCL19 interacts with R185ECL2 and E181ECL2; 9



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 10 of 55

1

D7CCL19 interacts with E181ECL2, and K261.24; CCL19’s N-terminus interacts with E941.64, D2857.36,

2

R301.28, and K331.31.

3

Our generated complex models were equilibrated in a model of a lipidic membrane patch

4

through our multi-step molecular dynamics simulations protocol described below and

5

subsequently used to initiate long timescale MD simulations.

6

Conventional Molecular Dynamics Simulations

7

Equilibration MD simulations of CCR7-CCL19 and CCR7-CCl21, were performed using

8

NAMD, version 2.9.44 Initial protein structure files were prepared using the PSFGEN utility in

9

VMD45 and the CHARMM36 forcefield.46–49 All disulfide bonds were maintained during the

10

simulations: two disulfide bonds in CCR7 (C24-C274 and C105-C186), two disulfide bonds in

11

CCL19 (C8-C34 and C9-C50), and two in CCL21 (C8-C34 and C9-C52). The receptor region

12

embedded in the membrane was determined using the Positioning of Proteins in Membrane

13

(PPM) server (http://opm.phar.umich.edu/server.php) and used to position the lipid bilayer

14

around the receptor.50 The palmitoyl-oleoyl-phosphatidyl-choline (POPC) bilayer was generated

15

using the Membrane plugin in VMD45 and all overlapping lipid molecules within 1 Å were

16

removed. The lipid-protein system was embedded into a water box using the VMD utility

17

SOLVATE and the TIP3P model for the water molecules. The water box dimensions of the

18

ligand bound receptor was 105 Å × 105 Å × 120 Å respectively. The system was neutralized

19

using sodium and chloride counterions at an ionic strength of 150 mM. The final ligand bound

20

CCR7 systems contained ~270 lipid molecules, ~200 Na+, ~214 Cl-, and ~35,000 water

21

molecules, for a total of ~149,000 atoms.

22

NAMD44 was used to equilibrate the system in a series of minimization, melting, and

23

equilibration steps at 1 atm pressure and 310 K. The system was freely minimized for 1000 steps 10


Page 11 of 55

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


1

of conjugate gradient energy minimization. The protein, solvent, ions, and lipid heads were then

2

fixed and the system was minimized for 1000 steps and simulated for 0.5 ns to allow the lipids

3

tails to equilibrate at 1 fs/step. The following simulation series were run at 2 fs/step. The system

4

was simulated with protein restrained at 5 kcal/mol to allow the environment to relax in 1000

5

steps minimization and 0.5 ns simulation. The system was then subjected to a minimization step

6

for 1000 steps and five equilibration stages (1 ns each) with all protein atoms harmonically

7

constrained (using force constants of 5, 4, 3, 2, and 1 kcal/mol/Å2, respectively) to their post-

8

minimization positions, and a final unconstrained equilibration stage of 15 ns.

9

All simulations were performed using periodic boundary conditions and particle-mesh Ewald

10

electrostatics for long-range electrostatic interactions with a grid point density of 1/Å.

11

Nonbonded van der Waals interactions and short-range electrostatic interactions were calculated

12

with an interaction cutoff of 12 Å and switching distance of 10 Å. The SHAKE algorithm was

13

employed to fix the length of all hydrogen-containing bonds, enabling the use of 2 fs integration

14

time steps. Coordinates were sampled every 2 ps.

15

Final output velocities, dimensions, and coordinates from equilibration NAMD simulations

16

were used as input to simulate our systems on Anton,20 a special purpose supercomputer for

17

biomolecular simulation designed and constructed by D. E. Shaw Research (DESRES). All

18

Anton simulations were performed under the NPT ensemble using a multigrator (310K using a

19

Nosé-Hoover thermostat and an isotropic pressure of 1 atm using the Martyna-Tobias-Klein

20

barostat).51 Multigrator with thermostat interval of 24 ps and barostat interval of 240 ps were

21

used. All bond lengths to hydrogen atoms were constrained using the M-SHAKE. A RESPA

22

integrator was used with a time step of 2 fs for bonded, VDW, and short-range electrostatic

23

interactions, and 6 fs for long-range electrostatic interactions. Long-range electrostatic 11



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 55

1

interactions were handled with the k-space Gaussian Split Ewald (GSE) method and a 64 × 64 ×

2

64 grid. Interaction parameters such as GSE parameters, and nonbonded cutoffs were determined

3

systematically using Anton scripts, designed to optimize accuracy and performance. For CCL19-

4

CCR7, σ = 3.19 Å, σs = 1.92 Å, Rcut = 13.76 Å, Rspread = 8.30 Å; and for CCL21-CCR7, σ =

5

3.17 Å, σs = 1.84 Å, Rcut = 13.67 Å, Rspread = 7.97 Å. Initial configuration of our three

6

systems produced root mean-squared force errors of no more than 0.0023 kcal/mol/Å.

7

Accelerated Molecular Dynamics Simulations

8

Accelerated molecular dynamics simulations were performed using NAMD2.9 at the dual-

9

boost acceleration level. The dual-boost applies a boost potential to all dihedral angles and all

10

atoms in the system using the input parameters: Edihed = Vdihed_avg + 0.3*Vdihed_avg, αdihed =

11

0.3*Vdihed_avg/5, Etotal = Vtotal_avg + 0.2*Natoms, and αtotal = 0.2*Natoms. Natoms is the number of

12

atoms in each system; and Vdihed_avg and Vtotal_avg are the average dihedral and total energies

13

respectively, extracted from 45 ns equilibration cMD simulations performed as described above.

14

Accelerated MD simulations were performed for 150 ns each by restarting from the 45 ns

15

equilibration conventional MD simulations.21,22

16

Analysis Protocols

17

System snapshots were extracted at a rate of 180 ps and 2 ps during Anton and aMD

18

production simulations, respectively. All frames have been analyzed to extract different

19

measures and create time series of data as shown in figures. These measures include hydrogen

20

bond, torsion angle, and atomic distance calculations. Analysis of the MD trajectories was

21

performed with in-house scripts using R programming language,52 Python,53 Chimera,43 the

22

Bio3D library,54,55 and TimeScapes.56

12


Page 13 of 55

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


1

Hydrogen bonds were calculated with Chimera, using hydrogen bond criteria as described in

2

ref.43,57 Backbone and side-chain torsion angles were calculated using Bio3D. Atomic distances

3

were calculated between non-hydrogen side-chain atoms using TimeScapes56 with a cutoff of 5

4

Å. Interacting residue distance time series are determined by calculating the minimum distances

5

of the set of atomic distance time series of two interacting residues. Calculating side-chain

6

contacts using the minimum-distance interacting atoms allows for an accurate separation

7

distance between both residues, as compared to using the centroid or a representative atom of the

8

side-chain. Distance time series between Cα atoms were also calculated using TimeScapes.56

9

Distance probability distributions of time series were calculated by computing Gaussian kernel

10

density estimates using R. Percent occupancy of an interaction was calculated as the percentage

11

of frames in which an interaction is within 5 Å in the MD trajectory. Pairwise cross-correlation

12

was calculated between all Cα distance time series using Bio3D.54,55 Cross-correlation coefficient

13

cutoff of 0.95 was used to cluster correlated Cα distance time series; and correlated time series

14

that show a coupled behavior to the different measures in Fig. 2 are extracted as ligand-induced

15

global helical motions. Rmsds are calculated using Bio3D based on residue side-chains or Cα

16

carbons distance time series rather than atomic coordinates.

17

Π-Stacking Interactions

18

Aromatic residue trimers within proteins in the protein data bank (PDB) have been studied

19

extensively and found to form two distinct geometrical clusters: symmetrical and ladder

20

conformations.58 To monitor the π-stacking arrangement within the tyrosine triad, we measured

21

the angle formed between the Cζ carbons in each of the tyrosines, and was found to adopt an

22

obtuse (between 90° and 120°) and acute (around 60°) θ-angle for ladder and symmetric

23

conformations, respectively.58 Molecular mechanics studies and PDB surveys indicate a π 13



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 55

1

stacking interaction upper-limit distance of 7.5 Å (distance between benzene ring centroids),

2

where the energy drops below the Boltzmann temperature factor.58,59

3

Sequence Alignment

4

The mouse sequences of all CC chemokine receptors were extracted from UniProt

5

(http://www.uniprot.org/).25 Alignment of all receptors with CCR7 was then performed using the

6

Clustal Omega server (http://www.ebi.ac.uk/Tools/msa/clustalo/).60 Sequence logos are

7

generated using the WebLogo3 (http://weblogo.berkeley.edu/logo.cgi).61,62 Mouse proteins were

8

selected to enable the use of our study to generate testable hypotheses for experimental mouse

9

models.

10

14


Page 15 of 55

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

1


RESULTS

2

With the recent available chemokine-bound crystal structure of CXCR4 (vMIP-II-bound

3

CXCR4, PDB code 4RWS),24 we set out to study the structural mechanism of CCR7 dynamics in

4

response to its native biased chemokine ligands, CCL21 and CCL19. Using the vMIP-II-CXCR4

5

complex, we modeled the CCL21 and CCL19-bound structures of CCR7 (see Methods), and

6

using Anton, we then performed cMD simulations on the CCL19-CCR7 and CCL21-CCR7

7

structures for 7 μs each, and we simulated both systems using dual-boost aMD for 150 ns each.

8

Accelerated MD was designed as an enhanced sampling MD method and shown to be able to

9

capture microsecond timescale events in various molecular systems.21,63–65 Using both MD

10

methods, we aim to identify microsecond timescale conformational changes within the TMD of

11

the receptor in response to each of its native biased ligands.

12

Chemokine ligands carry their function by binding to the extracellular region of the receptor

13

following a two-site model.24 In addition to binding the receptor’s NTD, the ligand interacts with

14

the extracellular loops (ECL) of the receptor, ensuring specificity of the ligand to its receptor,

15

and the N-terminal domain of the ligand (ligand-NTD: residues 1-7 preceding the first two

16

conserved cysteines) interacts inside the receptor’s extracellular binding pocket, and carries

17

receptor activation.28,66,67 This is emphasized further in our system by experimental truncation of

18

the CCL19-NTD which has converted the ligand to an antagonist; confirming that the ligand-

19

NTD is responsible for receptor activation.67 Therefore, we delineate the different binding poses

20

adopted by both ligand-NTDs in the receptor’s binding pocket (Fig. S3) and their effect on

21

receptor dynamics.

22

Given the physicochemical properties of the receptor binding pocket and ligand-NTD, both

23

CCL21-NTD and CCL19-NTD stabilize into different binding poses (Fig. S3). The electrostatic 15



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 55

1

profile of both ligand-NTDs displays different positions of its charged residues. CCL21

2

(S1DGGGQD7CC) has three charges distributed on either sides of the CCL21-NTD heptapeptide,

3

with one negatively charged residue at position 7 (D) and two charged residues, one positive at

4

the backbone N-terminus, and one negative at position 2 (D). However, CCL19

5

(G1ANDAED7CC) has four charged residues distributed uniformly along the CCL19-NTD

6

heptapeptide, with one N-terminal positive charge, and three negatively charged residues at

7

positions 4 (D), 6 (E), and 7 (D). Correspondingly on the receptor side, CCR7 displays two

8

distinct electrostatic regions in its binding pocket, one formed by a salt bridge between TM3 and

9

TM4 (K1133.33 and E1694.60), and another formed within TM1, TM2, and TM7 (K261.24, K271.25,

10

R301.28, E942.64, D2857.36) (Fig. S3C). Consequently, both ligand-NTDs stabilize in different

11

binding pocket of the receptor by forming complementary electrostatic interactions (Fig. S3).

12

CCL21’s missing middle negative charge, together with the flexibility of the three consecutive

13

glycines, allows its CCL21-NTD to stretch across and interact with both electrostatic patches in

14

the receptor, while the presence of the middle charge, D4, anchors the CCL19-NTD to interact

15

only with TM1, TM2, TM3, and TM7. The binding of each ligand-NTD in different pockets in

16

CCR7 agrees with experimental mutagenesis data obtained by Ott et al. showing the differential

17

effect of K1133.33 mutation to alanine in CCR7, where the mutation affects binding and activation

18

by 3.5- and 22-fold, respectively, in response to CCL21 but shows no effect to CCL19.66

19

CCL21 and CCL19 Induce Different Conformational Changes Within Ligand-Specific

20

Molecular Switches in CCR7. From both conventional and accelerated molecular dynamics of

21

CCL21-bound and CCL19-bound CCR7 complexes, we depict multiple regions in the

22

transmembrane domain (TMD) of CCR7 that show distinct conformational behavior (Fig. 1).

23

These functionally relevant regions act as molecular switches in the receptor and include: the tri 16


Page 17 of 55

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


1

tyrosine switch (Y1123.32, Y2556.51, and Y2887.39), the tri-phenylalanine switch (F1163.36, F2085.47,

2

and F2486.44), and the polar bridge (Q2526.48 and R2947.45). Each can take on multiple

3

conformations, and each can adopt ligand-specific conformations, which demonstrates the

Figure 1. Molecular switches in the CCR7 transmembrane domain adopt ligand-specific conformations. Molecular switches with multiple states in our simulations are represented in dark and light colors in their corresponding molecular graphics. The color code in the molecular graphics and population density plots is green for CCL21-bound and purple for the CCL19-bound simulations. (A) Molecular graphics of CCR7 bound to CCL21 and CCL19 with the three molecular switches outlined in boxes. Residues involved in the switches are shown as stick models. Each highlighted region is labeled with a letter corresponding to panels (B)-(E). (B) Side view of representative structures of the tri-tyrosine switch. Sidechain distances between Y1123.32 and Y2556.51 are plotted as population densities. (C) Top view of representative structures of the tri-tyrosine switch in the CCL21- and CCL19-bound CCR7 simulations illustrating π-stacking interactions. θ is the angle formed by the Cζ atoms of the three tyrosines. Population densities of the θ angle are plotted. (D) Representative structures of the tri-phenylalanine switch in the CCL21-bound and CCL19-bound CCR7 simulations. Side-chain distances between F1163.36 and F2486.44 are plotted as population densities. (E) Molecular graphics of the polar bridge. Side-chain distances between Q2526.48 and R2947.45 are plotted as population densities for the CCL21- and CCL19-bound CCR7 simulations. Criteria for inter-residue distances shown as population densities in all figures are outlined in Methods. 17



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

1

Page 18 of 55

ligand’s ability to regulate the arrangements within its molecular switches.

2

The multimodal distributions in Fig. 1 illustrate the effect that each ligand has on the various

3

conformational states sampled by each of the molecular switches. Within the tri-tyrosine switch,

4

all simulations show the formation of π-stacking interactions (Fig. 1C) following the loss of a

5

hydrogen bond between Y1123.32 and Q2526.48. (π-stacking interactions are monitored by

6

measuring the angle θ formed between the Cζ carbons in each of the tyrosines.) The bimodal

7

distributions of the population density plots of angle θ indicate the presence of conformational

8

changes that result in the formation and loss of π-stacking interactions (Fig. 1C). These

9

conformational changes result in equally distributed sampling of both states in the CCL21-bound

10

simulations, while the equilibrium is shifted towards the loss of π-stacking interactions in the

11

CCL19-bound simulations. Similarly, only the CCL21-bound simulations display hydrogen bond

12

formation between Y1123.32 and Y2556.51 (Fig. 1B). Probability density plots in Fig. 1B show a 3-

13

Å peak associated with hydrogen bond formation in the CCL21 and not CCL19 simulations.

14

Moreover, only the CCL21-bound simulations involve conformational changes within the tri-

15

phenylalanine switch, where the interaction between F1163.36 and F2486.44 is lost, moving from 4

16

Å to 7 Å in the CCL21-bound but not the CCL19-bound simulations (Fig. 1D). Lastly, Fig. 1E

17

displays the polar interaction between Q2526.48 and R2947.45 as another characteristic molecular

18

switch. This polar bridge resides deeper in the TMD of the receptor and its multi-peak

19

distributions show higher variability than the previous switches and a clear difference in

20

arrangements between the CCL21-bound and CCL19-bound simulations (Fig. 1E). Both cMD

21

and aMD simulations illustrate the effect of the ligand on each of the molecular switches. In

22

particular, the aMD simulations provide independent reproducibility of conformational changes

23

is each of the ligand-bound systems. Here, probability density plots of the aMD simulations in 18


Page 19 of 55

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


1

Fig. S4 show similar ligand-dependent sampling of the conformational states of each of the

2

molecular switches including: hydrogen bond formation between Y1123.32 and Y2556.51 in CCL21

3

and not CCL19 (Fig. S4A), equally distributed sampling of both gain and loss of π-stacking

4

interactions in the CCL21, while the equilibrium is shifted towards the latter in the CCL19 (Fig.

5

S4B), and clear difference in arrangements within the polar bridge (Fig. S4D).

6

The different positions of the ligand-NTDs play a critical role in initiating different allosteric

7

mechanisms in the receptor by disrupting and establishing specific contacts within the binding

8

pocket. In that aspect, the Anton simulations capture various allosteric events that are clearly

9

coupled to the transitions between the different conformational states in the molecular switches.

10

In contrast, even though the aMD simulations were able to reproduce the different

11

conformational changes within the molecular switches (Fig. S4), they display uncoupled and

12

stochastic variability within these switches that lacks the correlation documented in the Anton

13

simulations (Figs. 2 and 3). This is owed to the fact that the aMD consists of an enhanced

14

sampling method that applies a dual-boost to the all atoms and dihedral angles in the system.21,65

15

The energy boost might disrupt the fine-tuned transitions within the receptor and blur the energy

16

gaps governing the conformational transitions observed in the Anton cMD.

17

Both ligand-bound structures are each simulated for 7 μs using cMD on the Anton

18

supercomputer.20 Rmsd calculations of CCR7 display an initial increase related to the

19

stabilization of the initial receptor conformations used in both simulations (Fig. S5B).

20

Subsequently, CCL19-bound CCR7 shows a stable structure in the remainder of the simulation

21

(Fig. S5B, right panel), while CCL21-bound CCR7 exhibits a sharp increase at 4 μs related to a

22

large conformational transition in the receptor (Fig. S5B, left panel).

19



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 55

1

Conformational changes within the characterized molecular switches are coupled to

2

different allosteric events initiated by CCL21-NTD. During the 7-μs CCL21-bound cMD

3

simulation, we capture multiple conformational changes in suboptimal hydrogen bonds and van

4

der Waals tertiary interactions. These types of interactions represent two of the main physical

5

forces stabilizing membrane protein structures, and changes within these interactions constitute

6

allosteric events that are critical for signal propagation in the receptor.31 The time series of all

7

detected allosteric events are listed in Fig. 2 and clearly depict different conformational states

8

separated at 1.9 μs, 4 μs, and 5 μs. These states represent the initial (0-1.9 μs), intermediate (1.9-

9

5 μs), and final (5-7 μs) states of all depicted events. The intermediate state is a transitional

10

phase where different arrangements occur at different times and is further divided to two

11

intermediate substates I (1.9-4 μs) and II (4-5 μs).

12

At the ligand-binding pocket, CCL21 induces the formation of a hydrogen-bonding network

13

within Q6CCL21, N2666.62, Q2626.58, and N2817.32 (Fig. 2B, and Fig. S6). The formation of the

14

critical hydrogen bond between Q2626.58 and N2817.32 is synchronized to various other events in

15

the receptor at 4 μs (Fig. 2).

16

The stabilization of the hydrogen bond is associated with a decrease in distance between TM6

17

and TM7 at the ligand-binding pocket, which in turn prompts an equilibrium shift in the

18

backbone φ-torsion angle of P2546.50. CCR7 has multiple prolines in its TM helices forming

19

helical kinks. The highly conserved P2546.50 in TM6 is part of the WxPF/Y motif and is

20

positioned in the middle of the helix and considered an important dynamic component in the

21

rearrangements of helices in receptor activation.6,68 In Fig. 2C, a gradual transition of P2546.50 φ-

22

torsion angle from a bimodal to a unimodal distribution is illustrated in the population density of

23

torsion angles at different states. The initial state has a clear bimodal distribution of φ-torsion 20


Page 21 of 55

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

1


angles around -70° and -45°. Then, at intermediate state I, there is an equilibrium shift towards

21



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 55

Figure 2. During the cMD simulation, CCL21 induces a series of allosteric events that prompt equilibrium shifts in the discrete conformational states of CCR7’s molecular switches. (A) Molecular graphics of CCL21-bound CCR7 structure. Residues involved in the various allosteric events are shown as stick models and are outlined. Each highlighted region is labeled with a letter corresponding to panels (B)-(E). (B) Q2626.58, N2666.62, N2817.32, and Q6CCL21, shown as stick models, are involved in a hydrogen-bonding network in the ligandbinding site. The molecular graphics is a representative structure of the dominant conformation showing a hydrogen bond between Q2626.58 and N2817.32 (4-7 μs). The bar plot displays the hydrogen bond distance time series between Q2626.58 and N2817.32 side-chains. The bar plots of remaining hydrogen bonds are displayed in Fig. S6. (C) φ-torsion angle of P2546.50 population density (left) and time series (right) plots. The time series is broken down to four time segments labeled accordingly as initial, intermediate (I and II) and final states. Distributions of the φ-torsion angle are plotted for the following time segments: solid line (01.9 μs), dashed line (1.9-4 μs), and a dash-dot line (4-7 μs). (D) Molecular graphics showing representative structures of the dominant conformation in intermediate state I (1.9-4 μs) in dark green and the final state (5-7 μs) in light green. Residues involved in the tri-tyrosine switches and neighboring allosteric events are shown as stick models and labeled accordingly. Bar plots display the hydrogen bond distance time series of Y1123.32 with Q2526.48 and Y2556.51. Time series ranges are divided as shown in panel C and indicated by the vertical red lines. (E) Molecular graphics showing the tri-phenylalanine switch region using the same conformations as in panel D. Side-chain distance population densities are plotted for different time segments of the simulation as shown in panel C. The atoms used to calculate the hydrogen bond distance are marked in panels B and D. Hydrogen bond distances of 0 Å represent the absence of a hydrogen bond. Data for CCL19-bound cMD simulation, corresponding to panels B, C, D, and E are displayed in Fig. S8. 1

an angle distribution around a torsion angle of -70°, which in turn fully stabilizes to a clear

2

unimodal distribution around -75° at 4 μs.

3

On the same TM6 helix, P2546.50 is flanked by two critical residues: Q2526.48 and Y2556.51.

4

Both residues are capable of forming hydrogen bonds with Y1123.32 on TM3. Throughout the

5

simulation, the Y1123.32 side-chain hydrogen bond transitions from Q2526.48 to Y2556.51 (Fig. 2D).

6

This transition occurs twice during the simulation (at 3.1 μs and 4 μs) and is coupled with the

7

bimodal to unimodal transition observed in P2546.50 φ-torsion angle (Fig. 2C, and D). The

8

disruption of a hydrogen bond between Y1123.32 and Q2526.48 is associated with a φ-torsion angle

9

of -75° in P2546.50. At 4 μs, we observe a disruption of the hydrogen bond between Y1123.32 and

10

Q2526.48 occurring simultaneously with the proline φ-torsion angle adjustment followed by the 22


Page 23 of 55

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


1

second intermediate state experiencing a microsecond delay before the formation of a hydrogen

2

bond between Y1123.32 and Y2556.51.

3

During that delay, we detect the formation of π-stacking interactions within the tri-tyrosine

4

switch, Y1123.32, Y2556.51, and Y2887.39, between 4 and 5 μs (Fig. S7A). This interaction

5

stabilizes the transitional state resulting from both, the loss of the hydrogen bond between

6

Y1123.32 and Q2526.48, and the adjustment of P2546.50 φ angle. In the CCL21-bound cMD

7

simulation, a wide distribution of angles in intermediate state I indicates a disordered

8

arrangement between the three tyrosines and a lack of π-stacking interactions (Fig. S7A). Then,

9

these tyrosines are stabilized in intermediate state II by forming π-stacking interactions, where

10

Y2556.51 interacts with Y2887.39 and Y1123.32 at distances of 5.5 Å and 6 Å, respectively (less than

11

the upper-limit of 7.5 Å).58,59 These π-stacking interactions adopt a ladder conformation where θ

12

has an angle distribution around 110°. At 5 μs, π-π interactions between Y2556.51 and Y2887.39

13

are maintained. However, centroid distance between Y1123.32 and Y2556.51 increase to 7.5 Å,

14

indicating a loss in π-stacking (Fig. 2D and Fig. S7A). In the final state, the tri-tyrosine π-

15

stacking interactions are lost, resulting in hydrogen bond formation between Y1123.32 and

16

Y2556.51.

17

Another functionally relevant feature in CCR7 is the tri-phenylalanine switch, which resides

18

within the vicinity of the tri-tyrosine switch and is composed of three phenylalanines (F1163.36,

19

F2085.47, and F2486.44) in TM3, TM5, and TM6 (Fig. 2E). At 4 μs, F1163.36 reorients its side-chain

20

away from F2486.44 (increasing the distance from 4 Å to 7.5 Å) as well as F2085.47 (increasing the

21

distance from 3.5 Å to 4.5 Å). Meanwhile, we also observe changes within TM4 and TM5 where

22

L1654.56 and G2075.45 show a clear change in both, distance distribution and variance after 4 μs.

23



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 55

1

In contrast to the allosteric events observed in the CCL21-bound simulation, the CCL19-bound

2

cMD simulation do not exhibit any of the events characterized in CCL21 and all torsion angles

3

and interactions appear stable throughout the entirety of the simulation with the exception of the

4

hydrogen bond between Y1123.32 and Q2526.48 explored below (Fig. S8). This difference in

5

receptor behavior is a result of the differential binding and biased nature of both ligands.

6

Conformational changes within the characterized molecular switches are coupled to

7

different allosteric events initiated by CCL19-NTD. Despite the stability of CCR7’s rmsd in

8

the CCL19-bound cMD simulation (Fig. S5B), at 5.1 μs, we observe a brief series of side-chain

9

rearrangements (Fig. 3). Starting at the ligand-binding pocket, we observe the formation of a

10

hydrogen bond between CCL19’s N-terminus and Y411.39 side-chain at 5.1 μs and again 6.9 μs.

11

The hydrogen bond formation induces a series of synchronized allosteric events in the TMD of

12

CCR7. The time series listed in Fig. 3 clearly depict different conformational states of CCR7’s

13

tri-tyrosine switches (states I and II).

14

The hydrogen bond formation between G1CCL19 and Y411.39 is synchronized with a change in

15

the χ1-torsion angle of Y2887.39 from -70° to -170° (Figs. 3B and C). The side-chain orientation

16

of Y2887.39 allows for a reorganization of the tri-tyrosine switch (Y1123.32, Y2556.51, and Y2887.39)

17

to form π-stacking interactions, where Y1123.32 interacts with Y2887.39 and Y2556.51 at distances

18

of 6 Å or less (less than the upper-limit distance of 7.5 Å).58,59 We further measure the θ-angle

19

formed by the Cζ carbons of each of the three tyrosines to identify the symmetrical/ladder

20

arrangement of π-stacking interactions (Lanzarotti et al. 2011) (Fig. S7B). The difference in

21

angle between states I and II illustrates the transition of π-stacking interactions from symmetrical

22

in state I (angles sampled around 60°) to ladder in state II (angles sampled around 140°). This

23

transition prompts a distance increase between Y2887.39 and Y2556.51 to 9 Å.58,59 The π-stacking 24


Page 25 of 55

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


1

interactions reached in response to CCL19 show similar arrangement to the tri-tyrosine switch in

2

CCL21-bound cMD simulation (Fig. S7A).

3

25



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 55

Figure 3. During the cMD simulation, CCL19 induces a series of allosteric events that prompt equilibrium shifts in the discrete conformational states of CCR7’s tri-tyrosine switch. (A) Molecular graphics of CCL19-bound CCR7 structure. Residues involved in the various allosteric events are shown as stick models and outlined. Each highlighted region is labeled with a letter corresponding to panels (B)-(D). (B) G1CCL19 and Y411.39 are shown as stick models as a representative structure of the dominant conformation displaying the hydrogen bond between G1CCL19 and Y411.39 (5.1-6.4 μs). The bar plot displays the hydrogen bond distance time series. The atoms used to calculate the hydrogen bond distance are marked top of the panel. Hydrogen bond distances of 0 Å represent the absence of a hydrogen bond. (C) Molecular graphics showing representative structures of the dominant CCL19-induced conformation in state I (dark purple) and state II (light purple). The χ1-torsion angle of Y2887.39 time series is plotted. The time series is broken down to two time segments labeled accordingly. (D) Molecular graphics shows a side view of the tri-tyrosine switch region using the same conformations as panel C. Y1123.32, Y2556.51, Y2887.39, and Q2526.48 are shown as stick models and labeled accordingly. Q2526.48 and Y1123.32 distance population densities are plotted for both CCL19 states. 1

In CCL21-bound cMD simulation, intermediate state II is associated with the loss of a

2

hydrogen bond between Y1123.32 and Q2526.48 and result in the formation of the tri-tyrosine

3

ladder π-stacking interactions. Similarly, in our CCL19-bound cMD simulation state II, even

4

though the hydrogen bond between Y1123.32 and Q2526.48 persists (Fig. S8C), we see an

5

equilibrium shift in the side-chain distance between both residues. Population density of the side-

6

chain distance between Y1123.32 and Q2526.48 shows an increase in the distance sampled

7

(population increase for a distance of more than 4 Å in Fig. 3D), illustrating a weakening and

8

loss of hydrogen bonding in a large portion of state II. The characterized transitions in the

9

allosteric sites occur twice during in the simulation (at 5.1 μs and 6.9 μs), highlighting the

10

coupled nature of these allosteric events (Fig. 3). However, hydrogen bond loss between Y1123.32

11

and Q2526.48 is loosely coupled to the remaining allosteric events and only shows a slight

12

equilibrium shift. This results in an even weaker coupling to the hydrogen bond formation

13

between Y1123.32 and Y2556.51 characterized in the CCL21-bound simulation. Indeed, the

14

hydrogen bond between Y1123.32 and Y2556.51 is not formed in the CCL19-bound cMD and is

15

briefly present during the aMD. 26


Page 27 of 55

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


1

Both ligands induce large conformational changes in CCR7’s helical domains. In CCL21-

2

bound cMD simulation, the involvement of TM2, TM3, TM4, TM5, TM6, and TM7 in global

3

helical movements is synchronized with the aforementioned localized events (Fig. 4). This

4

coupling of local side-chain fluctuations and global helical motion is direct evidence of the

5

ligand-induced mechanism involved in the transformation of the receptor global structure.

6

Large helical motions are characterized as correlated Cα-Cα distance time series. We identify

7

multiple sets of distance time series that are not only correlated at a coefficient of 0.95 (using

8

Pearson's inner-product correlation calculation), but also show coupled changes with the

9

characterized molecular switches at ~3.1, 4, and 5 μs (Fig. 4). In the extracellular (EXC) region,

10

we detect a set of correlated distance time series involving TM3 interacting with TM2, TM6, and

11

TM7 (Fig. 4A). For this set of distance time series, an increase in distance between TM3 and its

12

neighboring helices (TM2, TM6, and TM7) is coupled to the formation and loss of hydrogen

13

bond between Y1123.32 and Y2556.51 (Figs. 2D and 4).

14

We also observe changes within TM4, TM5, and TM6 where distances of TM5 interacting

15

with TM4 and TM6 increase and decrease, respectively (Fig. 4A). Both changes are coupled

16

with the rearrangement observed in the tri-phenylalanine switch (Fig. 2E). TM4 and TM5 move

17

apart due to the detachment between L1654.56 and G2075.45, changing in distance distribution and

18

variance at 4 μs (Fig. 2E). In contrast, TM5 and TM6 move closer together, and the distance

19

between TM6 and TM7 decreases as a result of the CCL21-induced formation of the hydrogen

20

bond between Q2626.58 and N2817.32 (Figs. 2 and 4A).

27



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 55

1

28


Page 29 of 55

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


Figure 4. CCL21 induces global helical motions in CCR7 cMD simulation. (A) Global motions in the EXC region (upper half) of CCR7. Sets of Cα-Cα distance time series correlated at a coefficient of 0.95 are illustrated in different colors on the molecular structure. Each set is represented by a group of edges and each edge connects two Cα atoms whose distance time series is within a correlated set. A top view and side views of the receptor illustrate the different sets of correlated distance time series and display the TM helices containing the residues involved in the illustrated sets. A representative distance time series from each of the sets is plotted with color code matching that in the molecular graphics. Time series are divided to the same states defined in Fig. 2. Regions of characterized molecular switches (tri-tyrosine and tri-phenylalanine switches) are highlighted with dashed box. (B) Global motions in the INC region (lower half) of CCR7. Global motions are represented similarly to panel A. 1

Concurrently, we observe larger helical motions within the INC region of TM2, TM4, TM5,

2

TM6, and TM7. We show an interchange of inter-helical behavior between TM2, TM4, and

3

TM6, where, at 4 μs, the distance between TM2 and TM4 is stabilized, while the distance

4

between TM2 and TM6 undergoes periodic fluctuations (Fig. 4B). More importantly, critical

5

helices, TM5, TM6, and TM7, are involved in helical movements in the INC region of the

6

receptor, where TM6 experiences a shift in its equilibrium position towards TM5 and away from

7

TM7 (Fig. 4B). As a result of the characterized rearrangements, the INC end of TM6 shows an

8

increase in its distance with other neighboring helices (TM2 and TM7). We observe periodic

9

fluctuations between TM6 and TM2 and an increase in the distance between TM6 and TM7 (Fig.

10

4B).

11

TM4 is involved in movements with different TM helices in the EXC and INC region of the

12

receptor. As a result, we highlight the complementary role of TM4 in propagating the helical

13

movements from the EXC and INC regions of CCR7. The TM4 topology in GPCRs allows it to

14

interact with TM5 and TM2 in the EXC and INC regions respectively. In our CCL21-bound

15

simulation, the separation of TM4 and TM5 in the EXC region (Fig. 4A) allows for tighter

16

interactions between TM2 and TM4 in the INC region, where the interaction is stabilized (Fig. 29



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 55

1

4B). This stabilization complements the increased fluctuations between TM2 and TM6

2

associated with TM6 conformational changes.

3

Overall, a series of CCL21-induced molecular switches (Figs. 2B, C, and D) prompt subtle

4

rearrangements in the movement of TM6 towards TM5 and away from TM2 and TM7, when

5

compared to the larger helical motions in the EXC half of the receptor (Fig. 4). Similarly, in the

6

aMD simulation, hydrogen bond formation between Y1123.32 and Y2556.51 induces large helical

7

motions in the TMDs as characterized in the CCL21-bound cMD simulation (Fig. S9).

8

On the other hand, in the CCL19-bound simulation, the lack of hydrogen bond formation

9

between Y1123.32 and Y2556.51 result in minor global helical motions in CCR7 in both the cMD

10

and aMD simulations. CCL19-induced molecular rearrangements show inconsequential

11

movements of the receptor’s TM5, TM6, and TM7 EXC regions in cMD simulation, where these

12

movements are synchronized with the localized events in state II (Fig. S10). As CCL19

13

transitions from states I to II, TM5 exhibits an outward movement away from TM6 and TM7.

14

DISCUSSION

15

Molecular dynamic simulations have made substantial advances in the past years in describing

16

the structural mechanism of GPCR activation.15–19 However, the majority of these studies focus

17

on the unbiased activation of GPCRs; and the molecular basis of ligand-biased signaling remains

18

to be elucidated. In this study, using Anton microsecond dynamics and accelerated MD, we aim

19

to delineate the structural elements that are responsible for mediating the ligand-specific function

20

of chemokine receptors. This entails the characterization of ligand-specific structural events in a

21

biased chemokine receptor system: CCR7 and its endogenous ligands CCL19 and CCL21.

22

During the MD simulations of both ligand-bound systems, we depict functionally relevant

23

regions (molecular switches) in CCR7 that adopt ligand-specific arrangements, and, using the 30


Page 31 of 55

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


1

cMD simulations, we detail ligand-induced allosteric pathways capable of mediating the ligand-

2

specific conformational changes within these molecular switches in CCR7.

3

Both conventional and accelerated MD simulations capture conformational changes within the

4

molecular switches in the TMD of CCR7 involving: the tri-tyrosine switch (Y1123.32, Y2556.51,

5

and Y2887.39), the tri-phenylalanine switch (F1163.36, F2085.47, and F2486.44), and the polar bridge

6

(Q2526.48 and R2947.45). Additionally, cMD simulations capture a clear connection between

7

ligand-binding and conformational changes within the molecular switches in CCR7 through a

8

series of allosteric events, following a relay of interactions. In CCL21-bound CCR7, changes in

9

the tri-tyrosine and tri-phenylalanine switches are induced through a hydrogen bond between

10

Q2626.58 and N2817.32 and an adjustment of the φ-torsion angle of P2546.50. On the other hand,

11

CCL19 is capable of only disrupting the tri-tyrosine switch, and instead it acts through hydrogen

12

bond formation between its N-terminus and Y411.39 and the reorientation of the χ1-torsion angle

13

in Y2887.39. Allosteric events induced by the ligands form allosteric paths between the ligand-

14

NTD and the molecular switches with varying degrees of coupling. Indeed, the tri-tyrosine

15

switch appears to be strongly coupled to the CCL21-induced allosteric events and result in the

16

hydrogen bond formation between Y1123.32 and Y2556.51, while it is loosely coupled to the

17

CCL19-induced allosteric events and do not result in the hydrogen bond formation between

18

Y1123.32 and Y2556.51.

19

Within the CCL21 allosteric path, the adjustment of the P2546.50 φ-torsion angle and the loss of

20

the hydrogen bond between Y1123.32 and Q2526.48 belong to the same group of strongly coupled

21

events, and show synchronized conformational changes at ~3 and 4 μs (Fig. 2). Conformational

22

changes within this group are a result of an equilibrium shift at 4 μs induced by hydrogen bond

23

formation between Q2626.58 and N2817.32. As Y1123.32 is involved in a relay of hydrogen bonds 31



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 55

1

between Q2526.48 and Y2556.51, the hydrogen bond formation between Y1123.32 and Y2556.51

2

occurs stochastically after a 1 μs delay. This hydrogen bond is favored through the formation of

3

π-stacking interactions within the tri-tyrosine switch and hydrogen bond loss between Y1123.32

4

and Q2526.48. In contrast, the same hydrogen bond (Y1123.32 and Y2556.51) does not occur in the

5

CCL19-bound simulations, which is due to the loose coupling of the crucial event of hydrogen

6

bond loss between Y1123.32 and Q2526.48 to its CCL19-induced allosteric events. Within the

7

CCL19 allosteric pathway, the detected allosteric events (hydrogen bond formation between

8

CCL19’s N-terminus and Y411.39 and the reorientation of the χ1-torsion angle in Y2887.39) belong

9

to the same group as highly coupled events occurring at 5.1 μs and 6.9 μs. These events are

10

loosely coupled to the hydrogen bond loss between Y1123.32 and Q2526.48, whose distance only

11

shows a slight shift in equilibrium upon side-chain rotation in Y2887.39.

12

Another characteristic molecular switch in CCR7 is the tri-phenylalanine switch.

13

Rearrangements in F1163.36, F2085.47, and F2486.44 occur in the CCL21-bound and not CCL19-

14

bound simulations.

15

The third featured molecular switch, the polar bridge, involves a polar interaction between

16

Q2526.48 and R2947.45 and belongs to a cluster of conserved polar residues that reside within the

17

TMD of the receptor.69 This polar-residue network accommodates a nearly continuous water-

18

filled passage connecting the EXC and INC side of the receptor, where key rearrangements

19

affect the flow of the water and are critical for receptor activation.69–73 Our simulations (both

20

cMD and aMD) show a highly variable polar bridge involving Q2526.48 and R2947.45 (Fig. 1E).

21

These residues connect TM6 and TM7 and contribute to INC rearrangements of the TM helices

22

and to the connectivity of the hydrogen-bonding network regulating the presence of water

23

molecules within the TMD of the receptor. Our cMD simulations exhibit different equilibrium 32


Page 33 of 55

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


1

distances between Q2526.48 and R2947.45 side-chains in the CCL19- and CCL21-bound CCR7

2

(Fig. S11). The CCL21-bound simulation shows an increase in distance of 3 Å between the

3

initial and final state of the receptor through a highly variable intermediate state, whereas the

4

CCL19-bound simulation lacks any conformational changes within the polar bridge and shows a

5

stable distance of 4 Å (Fig. S11). These conformational changes indicate a differing degree of

6

coupling between the polar bridge and the characterized molecular switches and allosteric events

7

when comparing the ligand-bound cMD simulations.

8

Overall, different parts of the receptor appear to operate through loosely coupled clusters of

9

interlocked allosteric events. In other words, each event (or group of events) is capable of

10

shifting the equilibrium of the various functional regions of the receptor to induce various

11

conformational changes. It is important to note that the degree of coupling between the different

12

events changes in a ligand-specific manner; and it is this concerted coupling that allows the

13

receptor to induce a variety of ligand-specific conformational states that may contribute to its

14

biased signaling. Our simulations may not be sufficiently long to account for equilibrium

15

sampling or to capture all the different allosteric pathways within the receptor. However, we

16

highlight the dependence of each allosteric pathway to its ligand-NTD’s interactions inside the

17

receptor-binding pocket, which provides a structural dependence to the differently coupled

18

conformational changes within the switches. It may be that further sampling might result in

19

conformational changes within the tri-tyrosine and tri-phenylalanine switches in the CCL19-

20

bound simulation similar to those in the CCL21-bound simulation. However, these

21

conformational changes may still show different degrees of coupling due to the ligand-NTD’s

22

conformation and interactions within the binding pocket of the receptor.

33



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 34 of 55

1

Molecular switches characterized in CCR7 belong to critical TM helices involved in the

2

relative helical conformations responsible for selective-binding of intracellular effectors, namely

3

TM5, TM6, and TM7. Using Cα-Cα distance time series cross-correlation analysis, we detect

4

correlated global helical motions in the ligand and effector binding sites that are synchronized

5

with the aforementioned series of molecular switch rearrangements at the side-chain level (Fig.

6

4). In both conventional and accelerated MD simulations, one notable allosteric event is the

7

hydrogen bond formation between Y1123.32 and Y2556.51 that is highly coupled to large helical

8

motions in the EXC region of the TMD (Figs. 2 and 4, and Fig. S9). This interaction occurs in

9

the CCL21 simulations only, and explains the lack of correlated large helical motions in CCL19-

10

bound simulations. Both local side-chain rearrangements and EXC global helical motions result

11

in minor conformational changes in the INC region. These changes involve a slight adjustment

12

of the position of TM6 INC region relative to the rest of its neighboring helices, where TM6 is

13

adjusted for preferential interaction with TM5, and separation from TM2 and TM7. These

14

changes are loosely coupled to the events in the EXC region and do not show any outward

15

movement of TM6 that is indicative of receptor activation.14,74

16

The identified molecular switches are consistent with experimentally reported key

17

conformational changes during the activation of several receptors. Residues at positions 5.50,

18

3.40, and 6.44 form the “P-I-F” motif and experience discrete conformational states, indicative

19

of active and inactive conformations.17,75–77 However, in the M2 receptor, activation involves the

20

breaking of the interaction between F1955.47 and V1995.51 with an unchanged P-I-F motif.15,77

21

These studies illustrate that residues involved in the various side-chain rearrangements are not

22

consistent across receptors. However, these alterations occur within the same vicinity. Similarly,

23

in CCR7, even though the P-I-F motif remains unchanged, phenylalanine residues, F2486.44 and 34


Page 35 of 55

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


1

F2085.47, are part of CCR7’s tri-phenylalanine switch, where F1163.36 flips its side-chain away

2

from F2486.44 and F2085.47 (Fig. 2E). Another key molecular change involves W6.48 (part of

3

WxPF/Y motif) in several GPCRs.15,78 This residue is mutated to Q2526.48 in CCR7 and was

4

shown to participate in key hydrogen bonding and polar interactions with Y1123.32 and R2947.45 in

5

TM3 and TM7, respectively. Additionally, P2546.50 and Y2556.51, which are also part of the

6

WxPF/Y motif, participate in signal propagation in CCR7.68 Our simulations illustrate the

7

presence of several allosteric events that are in line with previously determined molecular

8

switches in the literature.

9

Additionally, sequence alignment of this family of receptors (Fig. 5) shows high conservation

10

of the residues involved in the observed molecular switches, which emphasizes their importance

11

in the ligand mediated signal propagation in the receptor. In the tri-tyrosine switch, both Y2556.51

12

and Y1123.32 (involved in large helical motions) are conserved in the CC chemokine family.

13

However, the third tyrosine, Y2887.39, is mutated to E7.39 in all other receptors. The presence of

14

a negatively charged residue at that position would disrupt the differential binding poses between

15

CCL19 and CCL21. Therefore, Y2887.39 is a critical residue unique to CCR7 in inducing its

16

ligands’ differential binding. Markedly, this residue is involved in CCL19’s allosteric events and

17

not CCL21’s. Polar residues at positions Q2526.48 and R2947.45 in CCR7 are mutated to

18

tryptophan and histidine, respectively, in other CC chemokine receptors. Despite these

19

mutations, important hydrogen bonding capability (tryptophan indole group) and the presence of

20

a positive charge (histidine), that are critical to the characterized allosteric events, are conserved.

21

Lastly, the tri-phenylalanine switch is conserved throughout all CC chemokine receptors with

22

few conserved mutations of leucine to alanine or phenylalanine to leucine (with the exception of

23

CCR2 and CCR5). Additional across species alignment (mouse, human, and bovine) illustrates 35



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 36 of 55

1

the conservation of all molecular switch residues in CCR7 and ligand-NTD domains of CCL19

2

and CCL21 with the exception of less conserved mutations from G6 to A6 in CCL21 and A2 to

3

T2 in CCL19 (data not shown). ! !

CCR7 CCR8 CCR4 CCR2 CCR5 CCR1 CCR3 CCR10 CCX-CKR CCR6 CCR9 CXCR4 CCR7 CCR8 CCR4 CCR2 CCR5 CCR1 CCR3 CCR10 CCX-CKR CCR6 CCR9 CXCR4

(30%) (36%) (34%) (34%) (34%) (34%) (36%) (37%) (40%) (42%) (34%)

80 90 95 110 120 160 170 200 210! ...DILFLLILPFWAYSE-A...GIYKLSFFSGM...MLALFLSIPEL...QVAQMVFGFLV...! ...DLLFVLSIPFQTHNL-L...GLYYIGFFSSM...LAAVTATIPLM...HFEINALGLLL...! ...DLLFVLSLPFWGYYA-A...WMYLVGFYSGI...SVAVFASLPGL...SLEINVLGLLI...! ...DLLFLLTLPFWAHYA-A...GLYHIGYFGGI...VVAVFASLPGI...TIMRNILSLIL...! ...DLLFLLTLPFWAHYA-A...GLYHIGYFGGI...AVAVFASLPEI...TLKMVILSLIL...! ...DLVFLFTLPFWIDYKLK...GFYYLGLYSEI...ALAILASMPAL...ALKLNLLGLIL...! ...DLLFLFTVPFWIHYVLW...GFYYLALYSEI...GLAGLAALPEF...ALRMNIFGLAL...! ...DLLLALTLPFAAAGA-L...GLYSASFHAGF...LLSLFLALPAL...AVAQVVLGFAL...! ...DLLLLITLPFWAVNA-V...ALYTVNFVSGM...MAAILLSIPQL...QMLEIGIGFVV...! ...DILFVLTLPFWAVTHAT...GTYAVNFNCGM...FISIIISSPTF...MGLELFFGFFT...! ...DLLFLATLPFWAIAA-A...SMYKMNFYSCV...VMAAVLCTPEI...LILKVTLGFFL...! ...DLLFVITLPFWAVDA-V...VIYTVNLYSSV...IPALLLTIPDF...QFQHIMVGLIL...! 245 250 260 285 290 300 ! ...VVFIVFQLPYNGVVLA...VTYSLASVRCCVNPFLYAF! ...IVSLLFWVPFNVALFL...VTEVISFTHCCVNPVIYAF! ...VLFLGFWTPYNVVLFL...ATETLAFIHCCLNPVIYFF! ...IVYFLFWTPYNIVLFL...VTETLGMTHCCINPVIYAF! ...IVYFLFWTPYNIVLLL...ATETLGMTHCCLNPVIYAF! ...LLFFLLWTPYNLSVFV...VTEVIAYTHCCVNPIIYVF! ...IVFFIFWTPYNLVLLF...VTEVIAYTHCCINPVIYAF! ...VAFVVLQLPYSLALLL...VTGGLTLVRCSLNPVLYAF! ...VVFIVTQLPYNVVKFC...VTESIALFHSCLNPILYVF! ...LVFLACQIPHNMVLLV...VAEVLAFLHCCLNPVLYAF! ...TVFIMSQFPYNSILVV...VTQTIAFFHSCLNPVLYVF! ...LAFFACWLPYYIGISI...ITEALAFFHCCLNPILYAF!

Figure 5. Multiple sequence alignment of residues in CC chemokine receptor family and CXCR4 displays conserved motifs within the characterized switches. Percent identities of each chemokine receptor to CCR7 are reported between parentheses. The transmembrane helical domains are colored in blue in CCR7’s sequence and CCR7’s residue numbers are displayed above the alignment. Residues involved in the characterized switches are highlighted in different colors: red (W902.60 and tri-tyrosine switch: Y1123.32, Y2556.51, and Y2887.39), orange (polar bridge: Q2526.48 and R2947.45), purple (P2546.50), and green (L1654.56, G2075.45 and tri-phenylalanine switch: F1163.36, F2085.47, and F2486.44). The sequences are truncated for clarity and the full sequence alignment is displayed in Fig. S12. 4

The absence of the outward motion of the INC region of TM6 indicates that the receptor

5

remains inactive during our simulations despite the manifestation of conformational changes

6

within the molecular switches and large helical motions in the EXC region of the receptor. We

7

postulate that the characterized networks of allosteric events fine-tune CCR7’s molecular

8

switches to drive the receptor to different bound states that show ligand-specific behavior. 36


Page 37 of 55

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


1

Similar networks of coupled allosteric events have been characterized previously as a sequential

2

model of activation in rhodopsin, where initial small conformational changes induced by the

3

ligands are converted to larger changes in the EXC regions of the receptor and induce the

4

canonical TM6 outward motion.74,79,80 However, despite the presence of a sequential model of

5

conformational changes in CCR7, our simulations show loose allosteric coupling between the

6

EXC and the INC regions. This loose coupling between the agonist-binding site and the

7

intracellular interface has been shown to be responsible for the complex signaling behavior

8

observed for β2-adrenergic receptor.18 These differences in activation mechanism between

9

rhodopsin and β2-adrenergic receptors are a result of their inherent biological function, where

10

rhodopsin ensures the rapid and efficient detection of a photon,80 while the β2-adrenergic receptor

11

evolved for a more complex signaling function.18 In that aspect, unlike CCR7, β2-adrenergic

12

receptor evolved to function through unbiased small molecules hormones,81 whereas CCR7

13

functions through biased peptidic chemokine ligands. This finding might justify the existence of

14

a hybrid model in CCR7, consisting of a “rhodopsin-like” sequential network of allosteric events

15

and the “β2-adrenergic-like” loose coupling between the EXC and INC regions of the receptor.

16

We speculate that the ability of the biased chemokine ligands to regulate the molecular switches

17

ensures efficient transitioning of the receptor to ligand-associated states to carry a ligand-specific

18

function, while the loose coupling between the ligand-binding and INC sites maintains the

19

receptor’s ability to carry out its complex signaling. This presents a hypothesis on how biased

20

ligands may modulate the conformation of the receptor. Given that the simulations are based on a

21

model of the ligand-bound receptor, further experimental validation is required. Ultimately, high-

22

resolution structures of the ligand-bound receptors would depict ligand-stabilized conformations

23

of the receptor’s molecular switches. 37



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 38 of 55

1

Knowledge of the structural and dynamic features of CCR7 provides a framework to aid in the

2

rational design of therapeutics to modulate cell migration or receptor silencing. Here, we identify

3

ligand-dependent conformations in CCR7 that provide the structural detail necessary to

4

rationally design functionally selective drugs. Targeting CCR7 may provide the needed

5

regulation of acute and chronic inflammatory responses involved in many autoimmune diseases

6

and assist in the development of cancer therapies, as CCR7 has been shown to be involved in

7

solid tumors metastasis to the lymph nodes.8,82–85

8

38


Page 39 of 55

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


1

ASSOCIATED CONTENT

2

Supporting Information (SI), containing a document and several supporting figures is attached to

3

this manuscript. The SI document discusses previously published mutagenesis data. SI figures 1-

4

12 are references and discussed throughout the text. This material is available free of charge at

5

http://pubs.acs.org.

6 7

AUTHOR INFORMATION

8

*Corresponding Author

9

Tel.: 951-827-2696

10

Email: [email protected]

11 12

AUTHOR CONTRIBUTIONS

13

ZG, DDL and DM designed research; ZG performed research; ZG and DM analyzed data and

14

wrote the manuscript.

15 16

CONFLICT OF INTEREST STATEMENT

17

The authors declare no conflict of interest associated with this publication.

18

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

1

Page 40 of 55

ACKNOWLEDGEMENTS

2

Anton computer time was provided by the National Center for Multiscale Modeling of

3

Biological Systems (MMBioS) through Grant P41GM103712-S1 from the National Institutes of

4

Health and the Pittsburgh Supercomputing Center (PSC). The Anton machine at PSC was

5

generously made available by D.E. Shaw Research.

6 7

ABBREVIATIONS

8

GPCR, G protein-coupled receptor; CCR7, CC chemokine receptor 7; CCL19, CC chemokine

9

ligand 19; CCL21, CC chemokine ligand 21; GRK, GPCR kinase; cMD, conventional molecular

10

dynamics; aMD; accelerated molecular dynamics; PDB, protein data bank; TMD,

11

transmembrane domain; NTD, N-terminal domain; CTD, C-terminal domain; INC, intracellular;

12

EXC, extracellular.

13 14

40


Page 41 of 55

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


1

REFERENCES

2

(1)

3 4

Liu, J. J.; Horst, R.; Katritch, V.; Stevens, R. C.; Wüthrich, K. Biased Signaling Pathways

in β2-Adrenergic Receptor Characterized by 19F-NMR. Science. 2012, 335, 1106–1110. (2)

Reiter, E.; Ahn, S.; Shukla, A. K.; Lefkowitz, R. J. Molecular Mechanism of β-Arrestin-

5

Biased Agonism at Seven-Transmembrane Receptors. Annu. Rev. Pharmacol. Toxicol. 2012, 52,

6

179–197.

7

(3)

Rahmeh, R.; Damian, M.; Cottet, M.; Orcel, H.; Mendre, C.; Durroux, T.; Sharma, K. S.;

8

Durand, G.; Pucci, B.; Trinquet, E.; Zwier, J. M.; Deupi, X.; Bron, P.; Baneres, J.-L.; Mouillac,

9

B.; Granier, S. Structural Insights into Biased G Protein-Coupled Receptor Signaling Revealed

10 11

by Fluorescence Spectroscopy. Proc. Natl. Acad. Sci. 2012, 109, 6733–6738. (4)

Zidar, D. A.; Violin, J. D.; Whalen, E. J.; Lefkowitz, R. J. Selective Engagement of G

12

Protein Coupled Receptor Kinases (GRKs) Encodes Distinct Functions of Biased Ligands. Proc.

13

Natl. Acad. Sci. 2009, 106, 9649–9654.

14 15 16 17 18

(5)

Boguth, C. A.; Singh, P.; Huang, C.; Tesmer, J. J. G. Molecular Basis for Activation of G

Protein-coupled Receptor Kinases. EMBO J. 2010, 29, 3249–3259. (6)

Katritch, V.; Cherezov, V.; Stevens, R. C. Structure-Function of the G-Protein-Coupled

Receptor Superfamily. Annu. Rev. Pharmacol. Toxicol. 2013, 53, 531–556. (7)

Shukla, A. K.; Westfield, G. H.; Xiao, K.; Reis, R. I.; Huang, L.-Y.; Tripathi-Shukla, P.;

19

Qian, J.; Li, S.; Blanc, A.; Oleskie, A. N.; Dosey, A. M.; Su, M.; Liang, C.-R.; Gu, L.-L.; Shan,

20

J.-M.; Chen, X.; Hanna, R.; Choi, M.; Yao, X. J.; Klink, B. U.; Kahsai, A. W.; Sidhu, S. S.;

21

Koide, S.; Penczek, P. A.; Kossiakoff, A. A.; Woods Jr, V. L.; Kobilka, B. K.; Skiniotis, G.; 41 ACS Paragon Plus Environment


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 42 of 55

1

Lefkowitz, R. J. Visualization of Arrestin Recruitment by a G-Protein-Coupled Receptor.

2

Nature. 2014, 512, 218–222.

3 4

(8)

Förster, R.; Davalos-Misslitz, A. C.; Rot, A. CCR7 and Its Ligands: Balancing Immunity

and Tolerance. Nat. Rev. Immunol. 2008, 8, 362–371.

5

(9)

Haessler, U.; Pisano, M.; Wu, M.; Swartz, M. A. Dendritic Cell Chemotaxis in 3D under

6

Defined Chemokine Gradients Reveals Differential Response to Ligands CCL21 and CCL19.

7

Proc. Natl. Acad. Sci. 2011, 108, 5614–5619.

8

(10) Schumann, K.; Lämmermann, T.; Bruckner, M.; Legler, D. F.; Polleux, J.; Spatz, J. P.;

9

Schuler, G.; Förster, R.; Lutz, M. B.; Sorokin, L.; Sixt, M. Immobilized Chemokine Fields and

10

Soluble Chemokine Gradients Cooperatively Shape Migration Patterns of Dendritic Cells.

11

Immunity. 2010, 32, 703–713.

12

(11) Kohout, T. A.; Nicholas, S. L.; Perry, S. J.; Reinhart, G.; Junger, S.; Struthers, R. S.

13

Differential Desensitization, Receptor Phosphorylation, -Arrestin Recruitment, and ERK1/2

14

Activation by the Two Endogenous Ligands for the CC Chemokine Receptor 7. J. Biol. Chem.

15

2004, 279, 23214–23222.

16

(12) Kahsai, A. W.; Xiao, K.; Rajagopal, S.; Ahn, S.; Shukla, A. K.; Sun, J.; Oas, T. G.;

17

Lefkowitz, R. J. Multiple Ligand-Specific Conformations of the β2-Adrenergic Receptor. Nat.

18

Chem. Biol. 2011, 7, 692–700.

19

(13) Kang, Y.; Zhou, X. E.; Gao, X.; He, Y.; Liu, W.; Ishchenko, A.; Barty, A.; White, T. A.;

20

Yefanov, O.; Han, G. W.; Xu, Q.; de Waal, P. W.; Ke, J.; Tan, M. H. E.; Zhang, C.; Moeller, A.;

21

West, G. M.; Pascal, B. D.; Van Eps, N.; Caro, L. N.; Vishnivetskiy, S. A.; Lee, R. J.; Suino 42


Page 43 of 55

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


1

Powell, K. M.; Gu, X.; Pal, K.; Ma, J.; Zhi, X.; Boutet, S.; Williams, G. J.; Messerschmidt, M.;

2

Gati, C.; Zatsepin, N. A.; Wang, D.; James, D.; Basu, S.; Roy-Chowdhury, S.; Conrad, C. E.;

3

Coe, J.; Liu, H.; Lisova, S.; Kupitz, C.; Grotjohann, I.; Fromme, R.; Jiang, Y.; Tan, M.; Yang,

4

H.; Li, J.; Wang, M.; Zheng, Z.; Li, D.; Howe, N.; Zhao, Y.; Standfuss, J.; Diederichs, K.; Dong,

5

Y.; Potter, C. S.; Carragher, B.; Caffrey, M.; Jiang, H.; Chapman, H. N.; Spence, J. C. H.;

6

Fromme, P.; Weierstall, U.; Ernst, O. P.; Katritch, V.; Gurevich, V. V.; Griffin, P. R.; Hubbell,

7

W. L.; Stevens, R. C.; Cherezov, V.; Melcher, K.; Xu, H. E. Crystal Structure of Rhodopsin

8

Bound to Arrestin by Femtosecond X-Ray Laser. Nature. 2015, 523, 561–567.

9

(14) Rasmussen, S. G. F.; DeVree, B. T.; Zou, Y.; Kruse, A. C.; Chung, K. Y.; Kobilka, T. S.;

10

Thian, F. S.; Chae, P. S.; Pardon, E.; Calinski, D.; Mathiesen, J. M.; Shah, S. T. A.; Lyons, J. A.;

11

Caffrey, M.; Gellman, S. H.; Steyaert, J.; Skiniotis, G.; Weis, W. I.; Sunahara, R. K.; Kobilka, B.

12

K. Crystal Structure of the β2 Adrenergic receptor–Gs Protein Complex. Nature. 2011, 477,

13

549–555.

14

(15) Miao, Y.; Nichols, S. E.; Gasper, P. M.; Metzger, V. T.; McCammon, J. A. Activation

15

and Dynamic Network of the M2 Muscarinic Receptor. Proc. Natl. Acad. Sci. 2013, 110, 10982–

16

10987.

17

(16) Dror, R. O.; Arlow, D. H.; Maragakis, P.; Mildorf, T. J.; Pan, A. C.; Xu, H.; Borhani, D.

18

W.; Shaw, D. E. Activation Mechanism of the β2-Adrenergic Receptor. Proc. Natl. Acad. Sci.

19

2011, 108, 18684–18689.

20

(17) Nygaard, R.; Zou, Y.; Dror, R. O.; Mildorf, T. J.; Arlow, D. H.; Manglik, A.; Pan, A. C.;

21

Liu, C. W.; Fung, J. J.; Bokoch, M. P.; Thian, F. S.; Kobilka, T. S.; Shaw, D. E.; Mueller, L.;

43



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 44 of 55

1

Prosser, R. S.; Kobilka, B. K. The Dynamic Process of β2-Adrenergic Receptor Activation. Cell.

2

2013, 152, 532–542.

3

(18) Manglik, A.; Kim, T. H.; Masureel, M.; Altenbach, C.; Yang, Z.; Hilger, D.; Lerch, M.

4

T.; Kobilka, T. S.; Thian, F. S.; Hubbell, W. L.; Prosser, R. S.; Kobilka, B. K. Structural Insights

5

into the Dynamic Process of β2-Adrenergic Receptor Signaling. Cell. 2015, 161, 1101–1111.

6

(19) Tikhonova, I. G.; Selvam, B.; Ivetac, A.; Wereszczynski, J.; McCammon, J. A.

7

Simulations of Biased Agonists in the β2 Adrenergic Receptor with Accelerated Molecular

8

Dynamics. Biochemistry (Mosc.). 2013, 52, 5593–5603.

9

(20) Shaw, D. E.; Dror, R. O.; Salmon, J. K.; Grossman, J. P.; Mackenzie, K. M.; Bank, J. A.;

10

Young, C.; Deneroff, M. M.; Batson, B.; Bowers, K. J.; Chow, E.; Eastwood, M. P.; Ierardi, D.

11

J.; Klepeis, J. L.; Kuskin, J. S.; Larson, R. H.; Lindorff-Larsen, K.; Maragakis, P.; Moraes, M.

12

A.; Piana, S.; Shan, Y.; Towles, B. Millisecond-Scale Molecular Dynamics Simulations on

13

Anton. In Proceedings of the Conference on High Performance Computing Networking, Storage

14

and Analysis; IEEE: Portland; OR; 2009; pp 1–11.

15

(21) Hamelberg, D.; Mongan, J.; McCammon, J. A. Accelerated Molecular Dynamics: A

16

Promising and Efficient Simulation Method for Biomolecules. J. Chem. Phys. 2004, 120, 11919–

17

11929.

18 19

(22) Wang, Y.; Harrison, C. B.; Schulten, K.; McCammon, J. A. Implementation of Accelerated Molecular Dynamics in NAMD. Comput. Sci. Discov. 2011, 4, 015002.

20

(23) Ballesteros, J. A.; Weinstein, H. Integrated Methods for the Construction of Three-

21

Dimensional Models and Computational Probing of Structure-Function Relations in G Protein 44


Page 45 of 55

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


1

Coupled Receptors. In Methods in Neurosciences; Sealfon, S. C., Ed.; Receptor Molecular

2

Biology; Academic Press, 1995; Vol. 25, pp 366–428.

3

(24) Qin, L.; Kufareva, I.; Holden, L. G.; Wang, C.; Zheng, Y.; Zhao, C.; Fenalti, G.; Wu, H.;

4

Han, G. W.; Cherezov, V.; Abagyan, R.; Stevens, R. C.; Handel, T. M. Crystal Structure of the

5

Chemokine Receptor CXCR4 in Complex with a Viral Chemokine. Science. 2015, 347, 1117–

6

1122.

7 8

(25) Apweiler, R.; Bairoch, A.; Wu, C. H. Protein Sequence Databases. Curr. Opin. Chem. Biol. 2004, 8, 76–80.

9

(26) Larkin, M. A.; Blackshields, G.; Brown, N. P.; Chenna, R.; McGettigan, P. A.;

10

McWilliam, H.; Valentin, F.; Wallace, I. M.; Wilm, A.; Lopez, R.; Thompson, J. D.; Gibson, T.

11

J.; Higgins, D. G. Clustal W and Clustal X Version 2.0. Bioinformatics. 2007, 23, 2947–2948.

12

(27) Goujon, M.; McWilliam, H.; Li, W.; Valentin, F.; Squizzato, S.; Paern, J.; Lopez, R. A

13

New Bioinformatics Analysis Tools Framework at EMBL–EBI. Nucleic Acids Res. 2010, 38,

14

W695–W699.

15

(28) Love, M.; Sandberg, J. L.; Ziarek, J. J.; Gerarden, K. P.; Rode, R. R.; Jensen, D. R.;

16

McCaslin, D. R.; Peterson, F. C.; Veldkamp, C. T. Solution Structure of CCL21 and

17

Identification of a Putative CCR7 Binding Site. Biochemistry (Mosc.). 2012, 51, 733–735.

18 19

(29) Šali, A.; Blundell, T. L. Comparative Protein Modelling by Satisfaction of Spatial Restraints. J. Mol. Biol. 1993, 234, 779–815.

20

(30) Veldkamp, C. T.; Kiermaier, E.; Gabel-Eissens, S. J.; Gillitzer, M. L.; Lippner, D. R.;

21

DiSilvio, F. A.; Mueller, C. J.; Wantuch, P. L.; Chaffee, G. R.; Famiglietti, M. W.; Zgoba, D. 45



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 46 of 55

1

M.; Bailey, A. A.; Bah, Y.; Engebretson, S. J.; Graupner, D. R.; Lackner, E. R.; LaRosa, V. D.;

2

Medeiros, T.; Olson, M. L.; Phillips, A. J.; Pyles, H.; Richard, A. M.; Schoeller, S. J.; Touzeau,

3

B.; Williams, L. G.; Sixt, M.; Peterson, F. C. Solution Structure of CCL19 and Identification of

4

Overlapping CCR7 and PSGL-1 Binding Sites. Biochemistry (Mosc.). 2015, 54, 4163–4166.

5 6

(31) Venkatakrishnan, A. J.; Deupi, X.; Lebon, G.; Tate, C. G.; Schertler, G. F.; Babu, M. M. Molecular Signatures of G-Protein-Coupled Receptors. Nature. 2013, 494, 185–194.

7

(32) Tan, Q.; Zhu, Y.; Li, J.; Chen, Z.; Han, G. W.; Kufareva, I.; Li, T.; Ma, L.; Fenalti, G.;

8

Li, J.; Zhang, W.; Xie, X.; Yang, H.; Jiang, H.; Cherezov, V.; Liu, H.; Stevens, R. C.; Zhao, Q.;

9

Wu, B. Structure of the CCR5 Chemokine Receptor–HIV Entry Inhibitor Maraviroc Complex.

10

Science. 2013, 341, 1387–1390.

11

(33) Cherezov, V.; Rosenbaum, D. M.; Hanson, M. A.; Rasmussen, S. G. F.; Thian, F. S.;

12

Kobilka, T. S.; Choi, H.-J.; Kuhn, P.; Weis, W. I.; Kobilka, B. K.; Stevens, R. C. High-

13

Resolution Crystal Structure of an Engineered Human β2-Adrenergic G Protein–Coupled

14

Receptor. Science. 2007, 318, 1258–1265.

15

(34) Park, S. H.; Das, B. B.; Casagrande, F.; Tian, Y.; Nothnagel, H. J.; Chu, M.; Kiefer, H.;

16

Maier, K.; De Angelis, A. A.; Marassi, F. M.; Opella, S. J. Structure of the Chemokine Receptor

17

CXCR1 in Phospholipid Bilayers. Nature. 2012, 491, 779–783.

18 19

(35) Roy, A.; Kucukural, A.; Zhang, Y. I-TASSER: A Unified Platform for Automated Protein Structure and Function Prediction. Nat. Protoc. 2010, 5, 725–738.

20 21

(36) Zhang, Y. I-TASSER Server for Protein 3D Structure Prediction. BMC Bioinformatics. 2008, 9, 40. 46


Page 47 of 55

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


1

(37) Wacker, D.; Fenalti, G.; Brown, M. A.; Katritch, V.; Abagyan, R.; Cherezov, V.;

2

Stevens, R. C. Conserved Binding Mode of Human β2 Adrenergic Receptor Inverse Agonists

3

and Antagonist Revealed by X-Ray Crystallography. J. Am. Chem. Soc. 2010, 132, 11443–

4

11445.

5

(38) Rasmussen, S. G. F.; Choi, H.-J.; Fung, J. J.; Pardon, E.; Casarosa, P.; Chae, P. S.;

6

DeVree, B. T.; Rosenbaum, D. M.; Thian, F. S.; Kobilka, T. S.; Schnapp, A.; Konetzki, I.;

7

Sunahara, R. K.; Gellman, S. H.; Pautsch, A.; Steyaert, J.; Weis, W. I.; Kobilka, B. K. Structure

8

of a Nanobody-Stabilized Active State of the β2 Adrenoceptor. Nature. 2011, 469, 175–180.

9 10

(39) Schmidt, T.; Bergner, A.; Schwede, T. Modelling Three-Dimensional Protein Structures for Applications in Drug Design. Drug Discov. Today. 2014, 19, 890–897.

11 12

(40) Skelton, N. J.; Quan, C.; Reilly, D.; Lowman, H. Structure of a CXC ChemokineReceptor Fragment in Complex with Interleukin-8. Structure. 1999, 7, 157–168.

13

(41) Millard, C. J.; Ludeman, J. P.; Canals, M.; Bridgford, J. L.; Hinds, M. G.; Clayton, D. J.;

14

Christopoulos, A.; Payne, R. J.; Stone, M. J. Structural Basis of Receptor Sulfotyrosine

15

Recognition by a CC Chemokine: The N-Terminal Region of CCR3 Bound to CCL11/Eotaxin-1.

16

Structure. 2014, 22, 1571–1581.

17

(42) Burg, J. S.; Ingram, J. R.; Venkatakrishnan, A. J.; Jude, K. M.; Dukkipati, A.; Feinberg,

18

E. N.; Angelini, A.; Waghray, D.; Dror, R. O.; Ploegh, H. L.; Garcia, K. C. Structural Basis for

19

Chemokine Recognition and Activation of a Viral G Protein–coupled Receptor. Science. 2015,

20

347, 1113–1117.

47



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 48 of 55

1

(43) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E.

2

C.; Ferrin, T. E. UCSF Chimera-A Visualization System for Exploratory Research and Analysis.

3

J. Comput. Chem. 2004, 25, 1605–1612.

4

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

5

Skeel, R. D.; Kalé, L.; Schulten, K. Scalable Molecular Dynamics with NAMD. J. Comput.

6

Chem. 2005, 26, 1781–1802.

7 8

(45) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. J. Mol. Graph. 1996, 14, 33–38.

9

(46) MacKerell, A. D.; Bashford, D.; Bellott, M.; Dunbrack, R. L.; Evanseck, J. D.; Field, M.

10

J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F.

11

T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher, W. E.; Roux, B.;

12

Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.; Watanabe, M.; Wiórkiewicz-Kuczera, J.; Yin,

13

D.; Karplus, M. All-Atom Empirical Potential for Molecular Modeling and Dynamics Studies of

14

Proteins †. J. Phys. Chem. B 1998, 102, 3586–3616.

15

(47) Mackerell, A. D.; Feig, M.; Brooks, C. L. Extending the Treatment of Backbone

16

Energetics in Protein Force Fields: Limitations of Gas-Phase Quantum Mechanics in

17

Reproducing Protein Conformational Distributions in Molecular Dynamics Simulations. J.

18

Comput. Chem. 2004, 25, 1400–1415.

19

(48) Best, R. B.; Zhu, X.; Shim, J.; Lopes, P. E. M.; Mittal, J.; Feig, M.; MacKerell, A. D.

20

Optimization of the Additive CHARMM All-Atom Protein Force Field Targeting Improved

48


Page 49 of 55

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


1

Sampling of the Backbone ϕ, ψ and Side-Chain χ1 and χ2 Dihedral Angles. J. Chem. Theory

2

Comput. 2012, 8, 3257–3273.

3

(49) Klauda, J. B.; Venable, R. M.; Freites, J. A.; O’Connor, J. W.; Tobias, D. J.; Mondragon-

4

Ramirez, C.; Vorobyov, I.; MacKerell, A. D.; Pastor, R. W. Update of the CHARMM All-Atom

5

Additive Force Field for Lipids: Validation on Six Lipid Types. J. Phys. Chem. B 2010, 114,

6

7830–7843.

7

(50) Lomize, M. A.; Pogozheva, I. D.; Joo, H.; Mosberg, H. I.; Lomize, A. L. OPM Database

8

and PPM Web Server: Resources for Positioning of Proteins in Membranes. Nucleic Acids Res.

9

2012, 40, D370–D376.

10

(51) Lippert, R. A.; Predescu, C.; Ierardi, D. J.; Mackenzie, K. M.; Eastwood, M. P.; Dror, R.

11

O.; Shaw, D. E. Accurate and Efficient Integration for Molecular Dynamics Simulations at

12

Constant Temperature and Pressure. J. Chem. Phys. 2013, 139, 164106.

13 14

(52) R. Core Team. A Language and Environment for Statistical Computing.; R Foundation for Statistical Computing, Vienna, Austria, 2014.

15 16

(53) Van Rossum, G.; Drake, F. L.; The Python Language Reference Manual. Network Theory Ltd: United Kingdom. 2011.

17

(54) Grant, B. J.; Rodrigues, A. P. C.; ElSawy, K. M.; McCammon, J. A.; Caves, L. S. D.

18

Bio3d: An R Package for the Comparative Analysis of Protein Structures. Bioinformatics. 2006,

19

22, 2695–2696.

20 21

(55) Skjærven, L.; Yao, X.-Q.; Scarabelli, G.; Grant, B. J. Integrating Protein Structural Dynamics and Evolutionary Analysis with Bio3D. BMC Bioinformatics. 2014, 15, 399. 49



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 50 of 55

1

(56) Wriggers, W.; Stafford, K. A.; Shan, Y.; Piana, S.; Maragakis, P.; Lindorff-Larsen, K.;

2

Miller, P. J.; Gullingsrud, J.; Rendleman, C. A.; Eastwood, M. P.; Dror, R. O.; Shaw, D. E.

3

Automated Event Detection and Activity Monitoring in Long Molecular Dynamics Simulations.

4

J. Chem. Theory Comput. 2009, 5, 2595–2605.

5 6

(57) Mills, J. E. J.; Dean, P. M. Three-Dimensional Hydrogen-Bond Geometry and Probability Information from a Crystal Survey. J. Comput. Aided Mol. Des. 1996, 10, 607–622.

7

(58) Lanzarotti, E.; Biekofsky, R. R.; Estrin, D. A.; Marti, M. A.; Turjanski, A. G. Aromatic–

8

Aromatic Interactions in Proteins: Beyond the Dimer. J. Chem. Inf. Model. 2011, 51, 1623–1633.

9

(59) McGaughey, G. B.; Gagné, M.; Rappé, A. K. π-Stacking Interactions ALIVE AND

10

WELL IN PROTEINS. J. Biol. Chem. 1998, 273, 15458–15463.

11

(60) Sievers, F.; Wilm, A.; Dineen, D.; Gibson, T. J.; Karplus, K.; Li, W.; Lopez, R.;

12

McWilliam, H.; Remmert, M.; Söding, J.; Thompson, J. D.; Higgins, D. G. Fast, Scalable

13

Generation of High-quality Protein Multiple Sequence Alignments Using Clustal Omega. Mol.

14

Syst. Biol. 2011, 7, 539.

15 16

(61) Crooks, G. E.; Hon, G.; Chandonia, J.-M.; Brenner, S. E. WebLogo: A Sequence Logo Generator. Genome Res. 2004, 14, 1188–1190.

17 18

(62) Schneider, T. D.; Stephens, R. M. Sequence Logos: A New Way to Display Consensus Sequences. Nucleic Acids Res. 1990, 18, 6097–6100.

19 20

(63) Wereszczynski, J.; McCammon, J. A. Nucleotide-Dependent Mechanism of Get3 as Elucidated from Free Energy Calculations. Proc. Natl. Acad. Sci. 2012, 109, 7759–7764. 50


Page 51 of 55

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


1

(64) Gasper, P. M.; Fuglestad, B.; Komives, E. A.; Markwick, P. R. L.; McCammon, J. A.

2

Allosteric Networks in Thrombin Distinguish Procoagulant vs. Anticoagulant Activities. Proc.

3

Natl. Acad. Sci. 2012, 109, 21216–21222.

4

(65) Pierce, L. C. T.; Salomon-Ferrer, R.; Augusto F. de Oliveira, C.; McCammon, J. A.;

5

Walker, R. C. Routine Access to Millisecond Time Scale Events with Accelerated Molecular

6

Dynamics. J. Chem. Theory Comput. 2012, 8, 2997–3002.

7

(66) Ott, T. R.; Pahuja, A.; Nickolls, S. A.; Alleva, D. G.; Struthers, R. S. Identification of CC

8

Chemokine Receptor 7 Residues Important for Receptor Activation. J. Biol. Chem. 2004, 279,

9

42383–42392.

10

(67) Ott, T. R.; Lio, F. M.; Olshefski, D.; Liu, X.-J.; Struthers, R. S.; Ling, N. Determinants of

11

High-Affinity Binding and Receptor Activation in the N-Terminus of CCL-19 (MIP-3β).

12

Biochemistry (Mosc.). 2004, 43, 3670–3678.

13

(68) Schwartz, T. W.; Frimurer, T. M.; Holst, B.; Rosenkilde, M. M.; Elling, C. E. Molecular

14

Mechanism of 7tm Receptor Activation—a Global Toggle Switch Model. Annu. Rev.

15

Pharmacol. Toxicol. 2006, 46, 481–519.

16 17

(69) Katritch, V.; Fenalti, G.; Abola, E. E.; Roth, B. L.; Cherezov, V.; Stevens, R. C. Allosteric Sodium in Class A GPCR Signaling. Trends Biochem. Sci. 2014, 39, 233–244.

18

(70) Angel, T. E.; Gupta, S.; Jastrzebska, B.; Palczewski, K.; Chance, M. R. Structural Waters

19

Define a Functional Channel Mediating Activation of the GPCR, Rhodopsin. Proc. Natl. Acad.

20

Sci. 2009, 106, 14367–14372.

51



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 52 of 55

1

(71) Angel, T. E.; Chance, M. R.; Palczewski, K. Conserved Waters Mediate Structural and

2

Functional Activation of Family A (rhodopsin-Like) G Protein-Coupled Receptors. Proc. Natl.

3

Acad. Sci. 2009, 106, 8555–8560.

4

(72) Yuan, S.; Filipek, S.; Palczewski, K.; Vogel, H. Activation of G-Protein-Coupled

5

Receptors Correlates with the Formation of a Continuous Internal Water Pathway. Nat. Commun.

6

2014, 5, 4733.

7

(73) Valentin-Hansen, L.; Frimurer, T. M.; Mokrosinski, J.; Holliday, N. D.; Schwartz, T. W.

8

Biased Gs versus Gq and β-Arrestin Signaling in the NK1 Receptor Determined by Interactions

9

in the Water Hydrogen-Bond Network. J. Biol. Chem. 2015, 290, 24495-24508.

10

(74) Altenbach, C.; Kusnetzow, A. K.; Ernst, O. P.; Hofmann, K. P.; Hubbell, W. L. High-

11

Resolution Distance Mapping in Rhodopsin Reveals the Pattern of Helix Movement due to

12

Activation. Proc. Natl. Acad. Sci. 2008, 105, 7439–7444.

13

(75) Huang, W.; Manglik, A.; Venkatakrishnan, A. J.; Laeremans, T.; Feinberg, E. N.;

14

Sanborn, A. L.; Kato, H. E.; Livingston, K. E.; Thorsen, T. S.; Kling, R. C.; Granier, S.;

15

Gmeiner, P.; Husbands, S. M.; Traynor, J. R.; Weis, W. I.; Steyaert, J.; Dror, R. O.; Kobilka, B.

16

K. Structural Insights into µ-Opioid Receptor Activation. Nature. 2015, 524, 315–321.

17

(76) Wacker, D.; Wang, C.; Katritch, V.; Han, G. W.; Huang, X.-P.; Vardy, E.; McCorvy, J.

18

D.; Jiang, Y.; Chu, M.; Siu, F. Y.; Liu, W.; Xu, H. E.; Cherezov, V.; Roth, B. L.; Stevens, R. C.

19

Structural Features for Functional Selectivity at Serotonin Receptors. Science. 2013, 340, 615–

20

619.

52


Page 53 of 55

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


1

(77) Kruse, A. C.; Hu, J.; Pan, A. C.; Arlow, D. H.; Rosenbaum, D. M.; Rosemond, E.; Green,

2

H. F.; Liu, T.; Chae, P. S.; Dror, R. O.; Shaw, D. E.; Weis, W. I.; Wess, J.; Kobilka, B. K.

3

Structure and Dynamics of the M3 Muscarinic Acetylcholine Receptor. Nature. 2012, 482, 552–

4

556.

5 6

(78) Deupi, X.; Standfuss, J. Structural Insights into Agonist-Induced Activation of G-ProteinCoupled Receptors. Curr. Opin. Struct. Biol. 2011, 21, 541–551.

7

(79) Park, P. S.-H.; Lodowski, D. T.; Palczewski, K. Activation of G Protein–Coupled

8

Receptors: Beyond Two-State Models and Tertiary Conformational Changes. Annu. Rev.

9

Pharmacol. Toxicol. 2008, 48, 107–141.

10 11

(80) Smith, S. O. Structure and Activation of the Visual Pigment Rhodopsin. Annu. Rev. Biophys. 2010, 39, 309–328.

12

(81) Rajagopal, S.; Ahn, S.; Rominger, D. H.; Gowen-MacDonald, W.; Lam, C. M.; DeWire,

13

S. M.; Violin, J. D.; Lefkowitz, R. J. Quantifying Ligand Bias at Seven-Transmembrane

14

Receptors. Mol. Pharmacol. 2011, 80, 367–377.

15

(82) Legler, D. F.; Uetz-von Allmen, E.; Hauser, M. A. CCR7: Roles in Cancer Cell

16

Dissemination, Migration and Metastasis Formation. Int. J. Biochem. Cell Biol. 2014, 54, 78–82.

17

(83) Shields, J. D.; Kourtis, I. C.; Tomei, A. A.; Roberts, J. M.; Swartz, M. A. Induction of

18

Lymphoidlike Stroma and Immune Escape by Tumors That Express the Chemokine CCL21.

19

Science. 2010, 328, 749–752.

53



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 54 of 55

1

(84) Stacker, S. A.; Williams, S. P.; Karnezis, T.; Shayan, R.; Fox, S. B.; Achen, M. G.

2

Lymphangiogenesis and Lymphatic Vessel Remodelling in Cancer. Nat. Rev. Cancer 2014, 14,

3

159–172.

4 5

(85) Tutunea-Fatan, E.; Majumder, M.; Xin, X.; Lala, P. K. The Role of CCL21/CCR7 Chemokine Axis in Breast Cancer-Induced Lymphangiogenesis. Mol. Cancer. 2015, 14, 35.

6 7 8

54


Page 55 of 55

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


1

Table of Content Graphic

2

For Table of Content Use Only

3

Molecular Mechanism of Biased Ligand Conformational Changes in CC Chemokine Receptor 7 Zied Gaieb1, David D. Lo2, Dimitrios Morikis1, *

4

55


Molecular Mechanism of Biased Ligand Conformational Changes in

Recommend Documents