Unraveling Allosteric Mechanisms of Enzymatic Catalysis with an

6 hours ago - ... Mechanisms of Enzymatic Catalysis with an Evolutionary Analysis of Residue-residue Contact Dynamical Changes. Phuoc J. Vu, Xin-Qiu Y...
0 downloads 12 Views 2MB Size
Subscriber access provided by READING UNIV

Article

Unraveling Allosteric Mechanisms of Enzymatic Catalysis with an Evolutionary Analysis of Residue-residue Contact Dynamical Changes Phuoc J. Vu, Xin-Qiu Yao, Mohamed Faizan Momin, and Donald Hamelberg ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b04263 • Publication Date (Web): 01 Feb 2018 Downloaded from http://pubs.acs.org on February 2, 2018

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.

ACS Catalysis 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 36 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

ACS Catalysis

1 2

Unraveling Allosteric Mechanisms of Enzymatic Catalysis with an Evolutionary Analysis of Residue-residue Contact Dynamical Changes

3 4

Phuoc Jake Vu1, Xin-Qiu Yao1, Mohamed Momin, Donald Hamelberg*

5

Department of Chemistry, Georgia State University, Atlanta, Georgia 30303-2515, USA.

6 7 8 9 10

1

Equal contributions

*Corresponding to: Prof. Donald Hamelberg; Department of Chemistry, Georgia State University, 29 Peachtree Center Ave NE, Atlanta, Georgia 30303-2515, USA. Telephone: (404) 413-5564; E-mail: [email protected].

11 12

1 ACS Paragon Plus Environment

ACS Catalysis 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 36

13

Abstract

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

The evolution of protein conformational dynamics contains important information about protein function and regulation. Here, we describe an approach to dynamical-evolution analysis based on multiple microsecond molecular dynamics simulations and residueresidue contact analysis. We illustrate our approach by comparing three human cyclophilin isoforms, cyclophilin A, D, and E, which belong to a family of enzymes catalyzing peptidyl-prolyl cis-trans isomerization. Our results reveal that despite distinct overall equilibrium conformations between cyclophilins under substrate-free conditions, functional dynamical changes resembling substrate-binding and catalytic processes tend to be conserved. Key residues displaying either concerted or specific dynamical changes among isoforms during the reactions are identified, which delineate two distinct allosteric pathways for cyclophilin function consistent with recent nuclear magnetic resonance experiments. A sequence-based coevolution analysis is also employed for further understanding dynamical consequences. Our results collectively provide a framework where both common and specific functional mechanisms of a protein family can be elucidated.

29 30 31

Keywords: evolution, allosteric regulation, enzyme dynamics, residue-residue contact, cyclophilin, molecular dynamics

2 ACS Paragon Plus Environment

Page 3 of 36 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

ACS Catalysis

32

Introduction

33

Protein internal motions or dynamics have been increasingly recognized to play a crucial

34

role in protein function and regulation.1-3 Typical protein dynamics span a broad range of

35

spatiotemporal scales, from subtle backbone and sidechain fluctuations (~10-9s) and domain

36

motions (~10-6s) to large-amplitude conformational changes normally observed in

37

biological molecular machines (10-3-102s). Notable examples in which dynamics determine

38

function include the dynamical rearrangements in enzymes to facilitate catalytic turnover,4

39

the conformational changes in transporters to pump small molecules in and out of cell

40

membrane,5 the force-producing structural changes in molecular motors,6 and the prevailing

41

roles of dynamics in the allosteric regulation during signal transduction.7-9 Despite

42

numerous efforts in inspecting protein dynamics with both experimental and computational

43

methods, such as nuclear magnetic resonance (NMR)10 and molecular dynamics (MD)

44

simulation,11 how protein molecules harness thermal fluctuations for function remains

45

elusive.

46

47

Protein dynamics reflect the underlying energy landscape, which is predominantly

48

determined by protein sequence. In this perspective, evolution modified protein dynamics

49

and function by altering the energy landscape. During evolution, certain patterns of protein

50

dynamics must be conserved to retain the core protein function that arose early. Similar to

51

how multiple sequence alignment screens out variable residues to detect structurally or

52

functionally critical sites, comparing protein dynamics at aligned residues across a protein

3 ACS Paragon Plus Environment

ACS Catalysis 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 36

53

family identifies dynamically conserved sites that are intimately related to the core function.

54

Meanwhile, variable dynamics in synergy with variable sequences across family members

55

underlie the subtle functional diversity, and hence, comparative analysis of dynamics also

56

helps to foster our understanding of protein functional specificity (e.g., selective substrate

57

binding or distinct kinetics). Indeed, this new paradigm of evolution-based approach to

58

delineating the complex sequence-structure-dynamics-function relationship has gained

59

increasing interests. By comparing the crystallographic structures of several homologous

60

proteins under distinct functional states, Babu and colleagues recently derived the universal

61

mechanisms governing the activation of heterotrimeric guanine nucleotide-binding proteins

62

(G proteins) and G protein-coupled receptors.12-13 With X-ray crystallography and NMR

63

spectroscopy, Wright and colleagues compared the dynamics of human and Escherichia

64

coli dihydrofolate reductase (DHFR), an enzyme that catalyzes the NADPH-dependent

65

reduction of dihydrofolate to tetrahydrofolate, and identified the residue determinants

66

underlying the distinct catalytic efficiency and robustness between the human and the

67

bacterial enzymes.14 In addition to the comparisons between extant proteins, the structures

68

and dynamics of ancestral proteins have been modeled with a computational ancestor

69

reconstruction method, which empowers a direct tracking of the evolution of protein

70

dynamics.15-17 These recent advances demonstrate how a study of protein dynamics in the

71

context of evolution provides unprecedented insights into functional mechanisms. The

72

knowledge obtained complements that derived from a large-scale evolutionary analysis

73

aimed at inferring sequence and structural traits underlying conserved and divergent

74

functions of enzyme superfamilies18-19 and can be further leveraged to develop new

4 ACS Paragon Plus Environment

Page 5 of 36 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

ACS Catalysis

75

algorithms for rational protein design and drug discovery. However, a general approach to

76

the evolutionary analysis of conformational dynamics for most biological systems is still

77

lacking.

78

79

Cyclophilin A (CypA) is a ubiquitous protein where function is strongly coupled with

80

dynamics.4, 20 As a peptidyl-prolyl cis-trans isomerase (PPIase), cyclophilin A catalyzes the

81

interconversion between the cis (with the peptidyl-prolyl torsion angle ω=0º) and the trans

82

(ω=±180º) states of the prolyl peptide bond in Xaa-Pro motifs, where Xaa represents any

83

amino acid (Figure 1A & B). Humans have 17 cyclophilin isoforms (Figure S1), among

84

which CypA is the best characterized, both experimentally and computationally.

85

Cyclophilins are known targets for immunosuppressant, and their function is critical to

86

many important cellular processes, including protein folding and signal transduction.21

87

Cyclophilin D (CypD) and cyclophilin E (CypE) are the family members closest to CypA,

88

with high sequence identity (68-75%) and almost identical structures to CypA (the

89

backbone root mean square deviation, or RMSD, is within 0.50-0.65 Å; see Figure 1A).

90

The three isoforms of cyclophilin have the same PPIase activity; however, they function in

91

different subcellular locations: CypA is generally found in the cytosol, CypD in the

92

mitochondria matrix, and CypE in the nucleus.22 We recently examined the dynamical

93

properties, represented by the breaking and formation of residue-residue contacts, of CypA

94

under various substrate-binding and mutational conditions using a combined approach of

95

MD simulations and NMR experiments.23 In particular, we found an interesting ‘dynamic

5 ACS Paragon Plus Environment

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

Page 6 of 36

96

cluster’ containing residues showing substantial dynamical changes upon substrate binding

97

located ~15 Å away from the active site. This finding reveals a new allosteric mechanism in

98

CypA, but its generality across the cyclophilin family is not known.

99

100

In this work, we develop a new approach for general evolutionary analysis of protein

101

conformational dynamics based on multiple microsecond-long MD simulations and the

102

similar contact analysis method previously described.23 We illustrate our approach by

103

comparing the dynamics between human CypA, CypD, and CypE derived from MD

104

simulations (totaling 26 µs). For each cyclophilin isoform, three functional states

105

collectively representing substrate-binding and catalytic isomerization processes are

106

examined. This allows us to evaluate the conservation of dynamics across isoforms with

107

respect to distinct enzymatic processes. Dissecting the contact dynamics further enables us

108

to identify the key residues determining the common and isoform-specific dynamical

109

changes. We also perform a sequence coevolution analysis to identify correlated amino acid

110

substitutions between residues in the cyclophilin family and find that dynamically

111

conserved contacts do not always represent the most highly coevolving residue pairs,

112

suggesting that dynamics is the missing link and sequence and structure alone cannot fully

113

describe function. The consensus groups of residues derived from the dynamical

114

conservation analysis are consistent with recent NMR experiments.24

115

116

6 ACS Paragon Plus Environment

Page 7 of 36 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

ACS Catalysis

117 118 119 120 121 122 123 124 125 126

Figure 1. Backbone and active site structures are highly similar across CypA, CypD, and CypE. (A) The substrate-free crystallographic structures of CypA (white; PDB: 3K0M), CypD (red; PDB: 4O8H), and CypE (green; PDB: 3UCH) are superimposed and are represented as cartoon. The modeled substrate (See Methods) is displayed as cartoon (yellow) with ‘Gly-Pro’ motif shown as licorice and colored by atom types. The enlarged view shows the substrate-binding pocket represented as white transparent surface. Side chains of active site residues are displayed as sticks and are color-coded the same as backbone. (B) Schematic cis-trans isomerization of the ‘Gly-Pro’ motif. Ts, transition state.

127

Results and Discussion

128

Protein backbone and residue side chain dynamics under the substrate-free state are

129

distinct among CypA, CypD, and CypE.

130

Fluctuation analysis and principal component analysis (PCA) of MD simulation trajectories

131

reveal that backbone dynamics are different between CypA, CypD, and CypE. Multiple

132

long-time (2.2-2.7 µs) MD simulations were performed under substrate-free conditions for

7 ACS Paragon Plus Environment

ACS Catalysis 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 36

133

each cyclophilin isoform. Snapshots of the latter 2-µs of each simulation trajectory were

134

analyzed. Residue-wise averaged root mean square fluctuation (RMSF) of backbone atoms

135

derived from the simulations shows that, whilst the atomic fluctuations at most residues

136

look similar across isoforms, the fluctuations located at the β4-β5 loop (the loop between

137

the fourth and the fifth β-strands), the β5-β6 loop, the β6-β7 loop, and the α2-β8 loop (the

138

loop between the second α-helix and the eighth β-strand) are apparently different (Figure

139

2A). In general, CypA and CypE are more flexible than CypD. PCA was then performed on

140

the Cartesian coordinates of backbone atoms from the simulations to examine backbone

141

conformational distributions at equilibrium (Figure 2B). It reveals that CypE has the

142

broadest distribution and samples multiple conformational states in the subspace spanned

143

by the top two principal components (i.e., PC 1 and PC 2, which collectively capture nearly

144

50% of total atomic mean displacements or variance), indicating the overall highest

145

flexibility of CypE among the isoforms. CypA has a distribution with modest broadness but

146

it still samples at least two conformational states, and it samples the space largely

147

overlapped with that populated by CypE. In contrast, CypD has the narrowest distribution,

148

which is clearly separated from those of CypA and CypE, and samples only one

149

conformational state, indicating that the backbone of CypD is rigid and has distinct

150

equilibrium properties from CypA and CypE.

151

152

Inspection of residue-residue contacts reveals distinct side chain dynamics among CypA,

153

CypD, and CypE. A pair of residues is in contact whenever their minimal non-hydrogen

8 ACS Paragon Plus Environment

Page 9 of 36 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

ACS Catalysis

154

atomic distance is at most 4.5 Å and they are separated by at least three residues in

155

sequence (i to i+n, n≥3). The 4.5 Å-threshold employed here represents a well defined

156

interaction range for most amino acids and has been applied to contact analysis in previous

157

simulation studies.23, 25-26 Other parameters were also used to test the robustness of our

158

results (See discussion all over Results and Discussion). For each residue pair, the

159

probability of contact formation during the simulation was calculated and compared

160

between cyclophilin isoforms. It shows that the residue pairs with substantial difference in

161

contact formation probability (i.e., |df=fCypY-fCypX|≥0.1, where fCypX and fCypY are contact

162

probabilities calculated for the two compared isoforms and the threshold 0.1 represents an

163

error estimate of f23) distribute all over the cyclophilin molecule for all the comparisons

164

(Figure 2C-E), indicating that side chain dynamics are distinct among the isoforms. In

165

summary, although in both sequence and structure cyclophilin isoforms are very similar,

166

they possess different conformational ensembles at equilibrium at least under the substrate-

167

free state.

168

9 ACS Paragon Plus Environment

ACS Catalysis 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 36

169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187

Figure 2. Backbone and sidechain dynamics under the substrate-free state are different between CypA, CypD, and CypE. (A) Residue-wise averaged root mean square fluctuation (RMSF) of backbone atoms derived from the MD simulations under substratefree conditions. Residue numbers are based on CypA. Secondary structure elements are annotated as black (α helices) and grey (β strands) rectangles at the top and bottom of the plot. (B) PCA performed on the Cartesian coordinates of backbone atoms from the simulations for CypA (grey), CypD (red), and CypE (green). Simulation-generated conformational snapshots are projected in the subspace spanned by the two principal components capturing the largest structural variance (PC1 and PC2; the number in the axis label indicates the percentage of variance captured by the corresponding PC). Probability density distributions of the conformational samples are represented as contour lines. The sampled space of the CypA and CypD simulations are also outlined. (C-E) Contact probability difference under the substrate-free state between isoforms (df=fCypY-fCypX, where f is the probability of contact formation and CypX and CypY are the corresponding cyclophilins under the comparison CypX/CypY) mapped to the crystal structure of CypA (PDB: 1M9F). Blue and red cylinders represent contacts with df≥0.1 and df≤-0.1, respectively, where cylinder radius is proportional to |df|. The yellow star indicates the active site.

10 ACS Paragon Plus Environment

Page 11 of 36 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

ACS Catalysis

188

189

Dynamical changes of residue-residue contacts during substrate binding are conserved

190

between CypA and CypE but are not conserved between CypA/E and CypD.

191

Although the conformational dynamics under an individual state (i.e., the substrate-free

192

state) are very different between the cyclophilin isoforms, dynamical changes from the

193

substrate-free to the substrate-bound state display a certain extent of similarity. Multiple

194

2.4-µs additional simulations were performed under substrate-bound conditions where the

195

peptidyl-prolyl torsion angle (ω) was in the cis-conformation (termed ‘cis-bound state’).

196

The probability difference of contact formation during simulations between the substrate-

197

free and the cis-bound states was used to characterize the dynamical changes during

198

substrate binding. Analysis of the CypA simulations reveals a site ~15 Å away from the

199

active site showing substantial dynamical changes (|df=fY-fX|≥0.1, where fX and fY are

200

contact probabilities calculated for the two compared states) of residue contacts (Figure

201

3A). This observation resembles the dynamic cluster identified in the previous simulation

202

study of CypA,23 although the substrate employed in the present work is five-residue longer

203

than that used in the previous study.23 This consistency indicates that the revealed

204

dynamical changes are an intrinsic dynamical characteristic of CypA independent from the

205

identity of the bound substrate. Intriguingly, overall similar patterns of dynamical changes

206

are observed between CypA and CypE (Figure 3A & C). A large portion of contacts that

207

are either more often formed (with an increase of contact probability) or more often broken

208

(decrease of contact probability) upon substrate binding in CypA are shown to have the

11 ACS Paragon Plus Environment

ACS Catalysis 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 36

209

same direction of changes in CypE (Figure 3E). These contacts with the same trends of

210

changes from one state to the other between isoforms are defined as ‘dynamically

211

conserved contacts.’

212

213

To quantitatively measure the overall similarity of dynamical changes between isoforms,

214

we developed a dynamical conservation index (DCI), which is defined by the percentage of

215

dynamically conserved contacts. In the calculation of DCI, contacts showing small

216

dynamical changes (absolute contact probability difference |df|