Characterization of the Polyspecific Transferase of Murine Type I Fatty

Subscriber access provided by READING UNIV

Article

Characterization of the polyspecific transferase of murine type I fatty acid synthase (FAS) and implications for polyketide synthase (PKS) engineering Alexander Rittner, Karthik S. Paithankar, Khanh Vu Huu, and Martin Grininger ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.7b00718 • Publication Date (Web): 12 Jan 2018 Downloaded from http://pubs.acs.org on January 13, 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 Chemical Biology 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 34 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 Chemical Biology

This study characterizes the transferase domain (MAT) of the murine fatty acid synthase (FAS) and explains its substrate promiscuity on a molecular basis. By engineering FAS based constructs, among other generating bimodular constructs as shown in the Figure, it is demonstrated that properties of MAT can be harnessed for polyketide synthase (PKS) engineering. 39x28mm (300 x 300 DPI)

ACS Paragon Plus Environment

ACS Chemical Biology 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 34

2

Characterization of the polyspecific transferase of murine type I fatty acid synthase (FAS) and implications for

3

polyketide synthase (PKS) engineering

1

4 5

Alexander Rittner, Karthik S. Paithankar, Khanh Vu Huu and Martin

6

Grininger*

7 8

Author affiliation

9 10 11 12 13

Institute of Organic Chemistry and Chemical Biology, Buchmann Institute for Molecular Life Sciences, Cluster of Excellence for Macromolecular Complexes, Goethe University Frankfurt, Max-von-Laue-Str. 15, 60438 Frankfurt am Main, Germany.

14

Correspondence

15

*[email protected]

16 17

Abstract

18

Fatty acid synthases (FASs) and polyketide synthases (PKSs) condense acyl

19

compounds to fatty acids and polyketides, respectively. Both, FASs and PKSs,

20

harbor acyltransferases (ATs), which select substrates for condensation by β-

21

ketoacyl synthases (KSs). Here, we present the structural and functional

22

characterization of the polyspecific malonyl/acetyltransferase (MAT) of murine FAS.

23

We assign kinetic constants for the transacylation of the native substrates, acetyl-

24

and malonyl-CoA, and demonstrate the promiscuity of FAS to accept structurally and

25

chemically diverse CoA-esters. X-ray structural data of the KS-MAT didomain in a

26

malonyl-loaded state suggests a MAT-specific role of an active site arginine in

27

transacylation. Owing to its enzymatic properties and its accessibility as a separate

28

domain, MAT of murine FAS may serve as versatile tool for engineering PKSs to

29

provide custom-tailored access to new polyketides that can be applied in antibiotic

30

and antineoplastic therapy.

31


1

Page 3 of 34 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


32

Introduction

33

Fatty acids are key compounds of cell membranes, metabolism and signaling,

34

and are obtained either directly from the diet, or synthesized de novo in a repeating

35

cyclic reaction catalyzed by fatty acid synthases (FASs). The chemistry of fatty acid

36

synthesis is largely conserved from bacteria to eukaryotes. Acetyl- and malonyl-CoA

37

compounds are first loaded by acyltransferases (ATs) and then condensed by a

38

β-ketoacyl synthase (KS) in a decarboxylative Claisen condensation. Subsequently,

39

the β-ketoester is modified by three active sites in the ketoreductase (KR),

40

dehydratase (DH) and enoylreductase (ER) domains. A first cycle delivers a fully

41

reduced ACP-bound butyryl group, which is elongated with a malonyl moiety in

42

further cycles to a select chain length (typically C16). During this multi-step catalytic

43

process, the acyl carrier protein (ACP) domain shuttles the fatty acid cargo between

44

the different enzymatic domains (Figure 1A).

45

The enzymatic functions of fatty acid synthesis are contained in remarkably

46

different structural frames. In general, type II FAS systems comprised of separate

47

enzymes are found mostly in plants, bacteria and in mitochondria. In case of

48

eukaryotes, cytosolic de novo fatty acid synthesis is performed by large multidomain

49

type I FAS.1, 2 For example, fungal FASs exhibit large barrel-shaped complexes of

50

about 2.6 MDa, in which the synthesis is enclosed in reaction chambers. In contrast,

51

animal FASs show an open, X-shaped architecture of about 540 kDa.3-7

52

The synthetic concept of fatty acid synthesis is exploited in polyketide

53

synthases (PKSs) for the production of a large set of bioactive compounds.8 Similarly

54

to FASs, PKSs do occur in different structural arrangements with type II and III

55

consisting of individual proteins. However, type I PKSs are contained within

56

multidomain polypeptides creating complex assembly lines of up to several MDa in

57

size.9 Structural information, recently collected on PKS modules, illustrates structural

58

similarity of PKSs and animal FASs.10-13 Contrary to FASs, PKSs vary the degree of


2


Page 4 of 34

59

β-carbon processing from module to module and utilize a broad set of acyl-CoA

60

priming and elongation compounds, which leads to a gigantic structural and chemical

61

versatility of polyketide natural compounds.

62

Animal FASs employ an AT domain with the dual function of priming fatty acid

63

synthesis with acetyl-CoA and elongating fatty acid chains with malonyl-CoA. The

64

domain is termed malonyl/acetyltransferase (MAT) (see Figure 1A).14 Earlier studies

65

revealed that MAT also transfers unusual CoA-esters, as for example priming

66

substrates like phenylacetyl-CoA,15 and extender substrates like methylmalonyl-

67

CoA.16, 17 The substrate promiscuity of MAT is not apparent in the narrow product

68

spectrum of mainly C16 fatty acids of animal FAS. This may not surprise, as previous

69

reports have shown that product specificity in fatty acid synthesis is determined by

70

multiple factors, including substrate availability and the substrate specificity of the KS

71

domain.18, 19

72

ATs of modular PKSs have a more diversified function as MAT of animal

73

FAS. As part of each elongation module, ATs are responsible for the selection of the

74

appropriate acyl-substrate in each elongation cycle, which is typically a malonyl- or a

75

methylmalonyl-CoA. As part of loading modules, ATs prime the synthesis of

76

polyketides. Both of these transferases are primary targets for engineering, leading

77

to the regiospecific modification of polyketides. The impact of the transferase

78

specificity on the PKS product spectra has been demonstrates on the loading-AT of

79

the avermectin PKS and the elongation-AT of the monensin PKS (module 5).

80

Insertion of these domains in heterologous PKS systems as well as the modulation of

81

these domains in their substrate specificity can induce the production of novel

82

compounds.20-25 Despite its occurrence in animal FAS only, MAT can be an attractive

83

target for custom polyketide synthesis, when its polyspecificity is harnessed for PKS

84

engineering. To this end, we recently started to characterize structure and function of

85

MAT. We aimed at revealing the molecular basis of MAT substrate polyspecificity to

86

understand MAT substrate promiscuity, as well as at making MAT available for PKS


3

Page 5 of 34 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


87

engineering.

88

Here, we report the in vitro characterization of the MAT domain of murine

89

FAS. To the best of our knowledge, this is the first time an extensive enzyme kinetic

90

analysis has been performed on MAT with both native substrates and multiple non-

91

canonical CoA-esters. Utilizing a continuous enzyme-coupled assay, this study

92

provides a profound quantitative description of the polyspecificity in MAT-mediated

93

transacylation.15, 17, 26, 27 Kinetic analysis further describes the distinctive features of

94

remarkably low hydrolysis rates and fast transfer rates. We further report the 2.9 Å

95

resolution structure of the KS-MAT didomain trapped in its malonyl-loaded state,

96

which helps to reveal the molecular foundation for the polyspecific character of

97

transacylation. Moreover, in a rational engineering approach, we express and purify

98

MAT as stand-alone domain in Escherichia coli and demonstrate its functionality. We

99

express and purify FAS-based bimodular constructs, which mimic loading-module–

100

module 1 PKS assemblies, to underline the direct usability of MAT in PKS

101

engineering. Murine MAT is a valuable new tool in the toolkit for PKS engineering

102

that may enlarge the spectrum of accessibly molecules in tailored polyketide

103

synthetic pathways.

104


4


105

Page 6 of 34

RESULTS AND DISCUSSION

106

Expression and purification of the KS-MAT didomain and ACP. Murine

107

FAS was chosen after extensive screening of multiple human and animal FAS

108

constructs in recombinant expressions, as it gave soluble and relatively large

109

quantities of protein (Figure S1). E. coli expression yielded about 2 mg of purified

110

protein after tandem affinity purification from 1 L of culture. Comprehensive enzyme

111

kinetic analysis of MAT and structural data were collected on the didomain

112

subconstruct KS-MAT (2–853), bearing a KS knockout (C161G) when used for

113

functional studies. Both subconstructs as well as constructs extending the murine

114

FAS fold have finally been characterized. Murine FAS (UniProt accession code

115

P19096) is highly homologous to human, porcine and rat FAS (with a sequence

116

identity of 81 %, 78 % and 94 %, respectively to murine FAS when aligned using the

117

program BLAST).28

118

Using recombinant expression in E. coli, ACP was not post-translationally

119

modified and remained in the apo state. For full activation, different co-expression

120

strategies with the 4'-phosphopantetheinyl transferase from Bacillus subtilis (Sfp)

121

were tested. As observed in this study, only bicistronic co-expression reproducibly

122

delivered protein at high yield and quality. For construct design and expression yields

123

in E. coli, see Supporting Information Note S1 and Figure S2. For more information

124

on protein preparation, see Experimental Procedures.

125

Characterization of transacylation polyspecificity. MAT catalyzes the acyl

126

transfer via a ping-pong bi-bi mechanism with an acylated serine being the central

127

intermediate of the reaction (Figure 1B).29 All kinetic investigations of MAT were

128

performed with an enzyme-coupled, continuous assay using α-ketoglutarate

129

dehydrogenase (αKGDH) to couple the release of free coenzyme A (CoA) to the

130

reduction of nicotinamide adenine dinucleotide (NAD+), which was monitored

131

fluorometrically (Figure S3).26, 27 In the established assay we ensured that initial rates


5

Page 7 of 34 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


132

of transferase reactions indeed depend on MAT activity and not on the coupled

133

downstream reaction. The mutant protein S581A was inactive, confirming the

134

specificity of MAT-mediated transacylation. The assay was further able to reproduce

135

the phenomenon of a compromised transfer of malonyl-CoA when the active site

136

arginine R606 is mutated to alanine (Figure S3B and Table 1, for more information

137

see also Experimental Procedures).30

138

In the next step, we determined MAT-mediated substrate hydrolysis rates to

139

enable appropriate background subtraction for transacylation (see Figure 1B).

140

Hydrolysis was slow with turnover rates of 9.8 ± 1.7 × 10–3 s–1 and 9.3 ± 0.8 × 10–3 s–1

141

for malonyl- and acetyl-CoA, respectively. These values agree well with previously

142

reported data for rat FAS (4.1 ± 0.8 nmol min–1 mg–1 protein (3.2 × 10–3 s–1)).30

143

Transacylation was about 3-4 magnitudes faster than hydrolysis, and allowed

144

ignoring MAT-mediated hydrolysis for background correction. Depending on the acyl-

145

CoA, significant non-enzymatically catalyzed hydrolysis or self-acylation of murine

146

ACP was observed,31 which required background corrections (see Experimental

147

procedure).

148

We selected 12 acyl-CoA substrates for analyzing the specificity of

149

transacylation, by determining the apparent kcat/KM values (Table 1). Substrates were

150

tested in a concentration range of 0.1 to 100 µM at fixed ACP concentrations. While

151

being most specific for malonyl-CoA, MAT was capable of transferring all tested

152

CoA-esters, except palmitoyl-CoA, with specificity constants in the same order of

153

magnitude. The high efficiency of MAT in the transacylation of a set of chemically

154

and structurally diverse compounds illustrates that MAT is not just tolerant in its

155

transacylation reaction, but indeed polyspecific (Table 1 and Figure S4). To ensure

156

that rates for non-native CoA-esters originate from transacylation and not from a

157

possible increased hydrolysis rate, the MAT-mediated hydrolysis rate was also

158

determined for every substrate. Although many non-native CoA-esters, for example,


6


Page 8 of 34

159

crotonyl- and succinyl-CoA, were hydrolyzed up to three times faster than acetyl- or

160

malonyl-CoA, the apparent kcat of the transacylation reaction remained 102-103 orders

161

of magnitude faster. Hence, we can rule out a “proof reading” function of MAT by

162

hydrolysis. The differences in hydrolysis rates rather imply slightly different transition

163

states for non-native substrates, which are less protected from hydrolysis. In this

164

aspect, MAT may differ from special ATs from PKSs, where hydrolysis was

165

suggested to determine specificity 32, 33

166

This set of acyl-CoA substrates was carefully chosen both for experimentally

167

exploring MAT transacylation promiscuity and for their commercial availability.

168

However, several of the tested compounds are also physiologically relevant, and

169

such data can provide interesting insight into animal fatty acid synthesis. (This is

170

outlined in Supporting Information Note S2).

171

Crystal structure of KS-MAT with bound malonyl moiety. The purified

172

murine KS-MAT didomain was crystallized, and crystals were soaked with malonyl-

173

CoA. X-ray diffraction data were collected to a resolution of 2.9 Å, and the resulting

174

structural model refined to R/Rfree of 0.18/0.22 (Table S1). The asymmetric unit

175

contains four molecules (referred to chains A, B, C and D), arranged as two

176

biological dimers wedged in the cleft between the KS and the linker domain. The

177

individual KS-MAT dimers are observed in a conformation that superimposes well

178

with the KS-MAT architectures of the natively purified porcine FAS,34 and the

179

corresponding didomain structure of human FAS (average r.m.s.d (Cα) of biological

180

dimers to porcine FAS of 1.6 and 1.7 Å; human KS-MAT superimposes to porcine

181

FAS with an average r.m.s.d 1.3 and 1.4 Å) (Figure 2A).35, 36 The four molecules in

182

the asymmetric unit show overall similar structure and superimpose with an r.m.s.d

183

(Cα) of 1, 0.8, and 0.9 Å (for chains A, B and C vs. chain D, respectively).

184

As observed in KS-MAT structures before,34, 35 the subdomains undergo rigid-

185

body movement (Figure 2B). A KS domain-based superposition (average r.m.s.d


7

Page 9 of 34 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


186

(Cα) 0.1 Å over residue range 1–407) indicates the variability in the relative

187

positioning of KS vs. MAT. A superposition of Cα atoms of the MAT α/β-hydrolase

188

subdomain (MAT without the ferredoxin-like fold, average r.m.s.d (Cα) is less than

189

0.4 Å over residue ranges 488–615 and 685–806) shows pronounced conformational

190

flexibility for the ferredoxin-like fold, represented by a rotational movement of the

191

subdomain of 10.6, 6.3 and 10.9 degrees (values given for chains A, B and C vs.

192

chain D). The MAT entry tunnel and the active site are comprised by residues of the

193

α/β-hydrolase and the ferredoxin-like subdomain (see Figure 2B).34, 37

194

Although the four molecules in the asymmetric unit superimpose well onto

195

one another, we found striking differences in substrate binding. Upon soaking KS-

196

MAT crystals with malonyl-CoA, the malonyl group was transferred onto the MAT

197

domain resulting in a covalent malonyl-enzyme complex in two of the four molecules

198

(chain A and D). Malonyl loading confirms MAT activity in the crystallized

199

conformation. In chain C, one molecule malonyl-CoA is non-covalently bound at the

200

entrance of the active site tunnel, whereas no substrate density could be detected in

201

chain B (Figure 2C).

202

Mechanistic insight into MAT-mediated substrate transfer. The active site

203

is characterized by a narrow shape and a hydrophobic environment (Figure 3A). The

204

active site serine (S581) is positioned in a strand-turn-helix motif (identified by the

205

consensus sequence GHSXG), termed the ‘nucleophile elbow’,38 where its

206

nucleophilicity is enhanced by a helix dipole moment.39 The tightness of the strand-

207

turn-helix motif induces the nucleophilic residue to adopt energetically unfavorable

208

main chain torsion angles, and imposes steric restrictions on residues located in its

209

proximity. This is very important for transition state stabilization by the oxyanion

210

hole.38 The catalytic dyad is accomplished by the presence of H683 at short distance,

211

as the general acid/base catalyst of the transferase reaction. For catalysis, Nπ of the

212

imidazole ring accepts the S581 hydroxyl group proton, whereas Nτ is in H-bond


8


Page 10 of 34

213

distance to the backbone carbonyl groups of residue N738, L739 and S741. After

214

nucleophilic attack of S581 at the thioester carbonyl group a tetrahedral intermediate

215

state is formed. The X-ray structural data on MAT in its malonyl-loaded state allows

216

localizing the oxyanion hole that stabilizes the transition state by charge-dipol

217

interactions. Here, the carbonyl oxygen is in H-bond distance to the backbone amide

218

of M499 in chain A and D (2.9 Å).37 The distance to the backbone amide of L582 is

219

3.6 Å in both malonyl-loaded active sites (chains A and D), which exceeds a typical

220

H-bond length. This may indicate that in MAT the trapped acyl-enzyme

221

conformational state has already moved away from the transition state conformation

222

and the oxyanion hole is disassembling. In the E. coli malonyl transferase FabD,

223

structural data revealed a slightly different embedment of malonyl in the oxyanion

224

hole (carbonyl oxygen at distances of 3.0 and 3.1 Å; distance of nitrogen atoms

225

5.2 Å) (Figure S5A).40

226

The active site arginine R606 is located at the bottom of the active site tunnel.

227

Earlier data have identified R606 to be important in malonyl transfer,41 as interacting

228

with free carboxyl group of the malonyl moiety.33,

229

densities of the non-substrate-bound active sites (chain B and C) are less well

230

defined in this region, R606 is observed pointing towards S581. Such an orientation

231

has also been observed in the unloaded active site of human KS-MAT, which was

232

solved at a higher resolution of 2.15 Å (Figure 3B and S5B).35 In the malonyl-loaded

233

active sites (chains A and D), the guanidinium group of R606 is in bidendate

234

interaction with the carboxyl of the malonyl moiety. Contrary to structural data on the

235

malonyl transferases FabD,40 and DynE8,33 where an end-on coordination is

236

observed, the active site arginine R606 in murine KS-MAT forms a side-on bidentate

237

interaction (Figure 3C). In addition, R606 of MAT is embedded in an H-bond network

238

that is less extended than observed in FabD and DynE. A strongly positionally

239

conserved phenylalanine (F553, murine FAS numbering), which replaces glutamine

240

of FabD and DynE, is largely responsible for the reduced H-bonding of R606

40

Although the overall electron


9

Page 11 of 34 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


241

(Figure 3C).

242

We assume that a positionally rather variable guanidinium group of R606 is

243

able to coordinate polar groups and may even swing out for the accommodation of

244

bulky residues (see Figure S5C). Herein, the side-on coordination of the malonyl

245

moiety by the guanidinium group may well reflect a unique active site geometry of

246

MAT that provides the necessary conformational variability of R606.

247

Transacylation of malonyl vs. acetyl moieties. To aid in interpreting

248

structural data, we expanded the enzyme kinetic analysis of both substrates.

249

Particularly, we were interested in refining the understanding of R606 in

250

transacylation of malonyl vs. acetyl groups. The assay in hand allowed addressing

251

absolute steady-state kinetic parameters by double reciprocal fits. We determined

252

apparent Km and kcat constants for malonyl- and acetyl-CoA at five different ACP

253

concentrations by fitting the raw data with the Hanes-Woolf equation (see Figure S4).

254

Individual reciprocal plotting of these constants against the reciprocal ACP

255

concentration resulted in the absolute Km of 8.8 µM and the absolute kcat of 119 s–1

256

for malonyl-CoA, and 12 µM (Km) and 99.2 s–1 (kcat) for acetyl-CoA (Figure 4). The

257

difference in transition state energies between acetyl and malonyl transfer ∆∆GES‡

258

can be calculated to 1.1 kJ mol–1. Molecular dynamics (MD) simulation recently

259

yielded

260

∆∆GES‡ = +14.2 ± 1.1 kJ mol–1 for the mutation of arginine to lysine in the

261

malonyl/palmitoyltransferase (MPT) of the yeast FAS I.42 To gain further insight into

262

the role of R606, we determined apparent kinetic constants for the R606A mutant at

263

a fixed ACP concentration (Table 1). Whereas hydrolysis rates remained nearly

264

unaltered, the specificity constant for malonyl-CoA decreased significantly being forty

265

fold smaller than the wildtype constant (kcat/Km of 0.3 ± 0.1 M–1 s–1). This corresponds

266

to

267

(2.3 kcal mol-1). Intriguingly, enzyme kinetic data did disclose a lower impact of

268

carboxyl coordination in malonyl binding and transacylation of murine FAS MAT, than

a

an

difference

estimated

in

transition

weakened

state

free

malonyl-CoA

energy


of

binding

of

∆∆GES‡ = 9.2 kJ mol–1

10


Page 12 of 34

269

estimated from MD simulations on the yeast FAS MPT domain and visual inspection

270

of structural data on malonyl-loaded type II AT (see Figure 3C).33, 40 Data indicate

271

that transacylation of the malonyl vs. the acetyl moiety does not differ in the plain

272

contribution of binding energy by bidentate coordination of the free carboxyl group of

273

the malonyl moiety, but includes a more global reshaping of the active site. A

274

possible explanation for the role of R606 may lie in the compensation for the

275

desolvation of substrates in the hydrophobic environment of the active site (see

276

Figure 3A).43 Such a role may be embedded in a molecular mode of substrate

277

accommodation that is based on the clamping of acyl-CoA compounds between the

278

positionally variable α/β-hydrolase and ferredoxin-like fold rather than their diffusion

279

via a substrate binding channel. The different occupancy of the active sites, observed

280

in the four molecules of the asymmetric unit (see Figure 2C), could then be

281

interpreted by the sensitivity of substrate accommodation towards the relative

282

positioning of the α/β-hydrolase and ferredoxin-like fold (see Figure 2B).

283

Engineering of MAT and murine FAS. To make MAT transacylation

284

properties available for PKS engineering, we assessed whether murine FAS MAT

285

can be used as a discrete, KS-independent (stand-alone) domain. As such, it may be

286

employed in trans-AT systems,44 or embedded in AT-ACP didomain structures of

287

loading-modules. Guided by our structural data and by comparison with homologous

288

proteins, two constructs were engineered, in which the MAT domain is liberated from

289

its structural embedment in the rigid KS-MAT fold. A LD-MAT construct was

290

designed that contained the MAT-domain and the linker domain (LD). It resembles a

291

construct of the human FAS reported in a previous study (Figure 5A).37 Another MAT

292

construct (486–808) was designed by mainly comparing structural data on the

293

avermectin synthase loading-AT (see Figure 5A; for construct design see

294

Figure S6).21 While MAT expressed at low concentrations, a fusion construct with a

295

maltose binding protein (MBP) attached as domain to the C-terminal helix


11

Page 13 of 34 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


296

significantly increased yields.

297

As a next step, we constructed two bimodular systems that resemble loading-

298

module–module1 assemblies as found in modular PKS. With these constructs, we

299

sought to collect initial information whether such bimodular fusions can, in principle,

300

deliver intact proteins. As a “module1”, we worked with full-length murine FAS. Both

301

the LD-MAT and the MAT domain, the latter without the MBP fusion, were N-

302

terminally fused to full-length FAS. Additionally, the ACP domain was integrated in

303

both constructs, as found in loading-modules (see Figure 5A and Figure S7). We

304

would like to note that such bi- and multimodular FAS constructs may be interesting

305

for FAS-mediated compound synthesis. They outline the perspective of FAS-

306

mediated custom synthesis of small molecules.42, 45, 46

307

The structural integrity of truncated and bimodular constructs was checked by

308

SDS-PAGE, phosphopantetheinylation with fluorescent CoA (FAS and bimodular

309

constructs),

310

(Figure 5B-G; see also Supporting Information Note S1 and Figure S8). Although

311

fusion of loading didomains did not reduce the general protein quality of both

312

bimodular constructs, the overall activity of the inherent FAS fold was decreased

313

compared to the wildtype FAS as judged by NADPH consumption. To evaluate

314

whether the additional loading module interferes structurally or catalytically with fatty

315

acid synthesis, we tested three controls of the LD-MAT-ACP-FAS construct providing

316

either loading module domain MAT or ACP or both as functional knockouts

317

(Figure S8). All three controls regained activity of wildtype FAS, as read out by

318

NADPH consumption rates. These data are interesting, since they imply the

319

structural integrity of module 1 (the integral FAS fold) in the bimodular constructs.

320

The comprised activity of the non-mutated wildtype bimodular construct (no knock-

321

out in MAT and ACP domains) further demonstrates a functional communication of

322

the loading module with the integral FAS fold.

323

size

exclusion

chromatography

(SEC)

and

enzymatic

activity

Implications for PKS design. Murine FAS MAT is relevant for PKS


12


Page 14 of 34

324

engineering as it has unique catalytic properties. (i) MAT fulfills its physiological

325

function in loading murine FAS with acetyl- and malonyl moieties by lose substrate

326

specificity. As shown in this study, murine MAT is truly polyspecific in its

327

transacylation properties holding out to prospect to load PKS with a broad set of acyl

328

compounds in chimeric systems. (ii) Murine MAT is fast in transacylation, as

329

demonstrated by absolute turnover rates (kcat) of 119 s–1 for malonyl and 99.2 s–1 for

330

acetyl groups. Absolute kinetic values for transacylation rates have not been

331

determined for other ATs yet. (iii) Murine MAT shows remarkable low hydrolytic

332

activity. Rates are more than an order of magnitude lower as compared to ATs of

333

PKS ((9.8 ± 1.7 × 10–3 s–1 and 9.3 ± 0.8 × 10–3 s–1 for malonyl- and acetyl-CoA,

334

respectively; DEBS AT3 of 3.3 × 10–2 s–1 for methylmalonyl-CoA and AveAT0 of

335

2.5 × 10–1 s–1 for 2-methylbutyryl-SNAC).21,

336

relevance of murine MAT for engineering, paving the way to an energetically less

337

costly incorporation of designer acyl-CoA substrates. (iv) Murine MAT is available in

338

the heterologous host E. coli and, as such, easily amenable to in vitro engineering

339

and mutational studies.

27

Low hydrolysis rates increase the

340

Murine MAT may be harnessed for PKS engineering in two aspects. The

341

provided substrate-bound structure may serve as a template to broaden the

342

substrate spectrum of a very specific PKS AT by introducing prominent active site

343

residues, like F553. Or, on the other hand, this particular polyspecific domain can be

344

introduced in a chimeric FAS/PKS system. The latter has been shown in this study

345

by the construction of bimodular FAS constructs. If the structural independence of

346

MAT from the KS-fold also allows an exchange of PKS extender-ATs remains

347

unknown.

348

Broadly substrate-accepting ATs are valuable in PKSs as they can pave the

349

way to regiospecific modification of polyketide lead compounds with non-canonical

350

acyl substrates. To incorporate designer acyl substrates, as e.g. provided by the Erb

351

lab,47 a promiscuous enzyme as MAT may be better suited for adaptive protein


13

Page 15 of 34 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


352

engineering than highly evolved, substrate specific ATs from existing modular

353

PKSs.48

354 355


14


Page 16 of 34

356

Methods

357

Cloning and construct expression. Fasn gene from Mus musculus was

358

purchased from Source BioScience (cDNA clone: IRAVp968A0187D). We essentially

359

performed ligation independent cloning with the In-Fusion HD Cloning Kit (Clonetech)

360

(Table S2). The sequence of all plasmids was confirmed with the “dye terminator”

361

method. All constructs were expressed in E. coli BL21gold(DE3) cells at 20 °C and

362

180 rpm, after induction with 0.25 mM IPTG. Proteins were essentially purified by

363

tandem affinity (Ni-chelating and Strep-tactin) chromatography. Active FAS and

364

active bimodular constructs were obtained by co-expressing with Bacillus subtils Sfp.

365

α-Ketoglutarate dehydrogenase coupled activity assay and fatty acid

366

synthase activity assay. The enzyme-coupled assay performed in this study was

367

adapted from references.26, 27 NADH fluorescence was monitored in a 96-well set-up.

368

Eight data points were collected that cover substrate concentrations (Sol 3) of

369

0.2 × Km; 0.3 × Km; 0.5 × Km; 0.75 × Km; 1.25 × Km; 2 × Km; 3 × Km; 5 × Km. Overall

370

fatty acid synthase activity was measured fluorometrically or by the decrease of

371

absorbance at 340 nm following the oxidation of NADPH at 25 °C.

372

Crystallization. Crystals were obtained at 0.2 M potassium-sodium tartrate,

373

25 % (w/v) PEG 3350 at 20 °C growing to sizes of about 75 × 75 × 75 µm3. Drops

374

with the crystals were supplemented with 0.5 µl of 10 mM malonyl-CoA (Sigma-

375

Aldrich) for up to 2 minutes and subsequently treated with a cryosolution containing

376

20 % (v/v) glycerol in the mother liquor.

377

Data collection and processing. Crystals were exposed to single

378

wavelength X-radiation at the Swiss Light Source (X06SA), and maintained at 100 K,

379

while data were recorded onto a detector (DECTRIS EIGER 16M). All diffraction data

380

are publicly available at https://zenodo.org/record/. The phase problem was solved

381

using molecular replacement method with the program Molrep,49 using the structural

382

model of a monomer from the human FAS KS-MAT didomain (pdb accession code


15

Page 17 of 34 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


383

3hhd).35 The model was built using Coot.50 Data collection and refinement statistics

384

are given in Table S1.

385

Accession codes. Atomic coordinates and structure factors have been deposited

386

with the Protein Data Bank under the accession code 5my0.

387 388 389

Author contribution A.R.

performed

molecular

cloning,

protein

expression,

purification

390

experiments, enzymatic assays and analyzed corresponding data. A.R. conceived

391

the project. M.G. designed the research. K.V. helped establishing enzyme kinetics

392

and crystallization. Crystallization was performed by A.R., K.V. and K.S.P. Crystal

393

structure was solved by K.S.P. The authors A.R., K.S.P. and M.G. analyzed data and

394

wrote the manuscript.

395 396

Acknowledgements

397

We thank F. Bourdeaux for measurements with HPLC-MS and the Swiss

398

Light Source for the beamtime. We are also grateful to S. Müller-Knapp for carefully

399

reading the manuscript.

400 401

Supporting Information

402

Supporting Information supplies detailed description of Methods, 8 Figures

403

and 2 Tables, and Supplementary Notes on the construct design and expression

404

yields as well as on the implication of MAT polyspecificity for mammalian fatty acid

405

synthesis.

406

This material is available free of charge via the internet.

407


16


408

Page 18 of 34

Funding sources

409

This work was supported by a Lichtenberg grant of the Volkswagen

410

Foundation to M.G. (grant number 85701). This project was further supported by the

411

LOEWE program (Landes-Offensive zur Entwicklung wissenschaftlich-ökonomischer

412

Exzellenz) of the state of Hesse and was conducted within the framework of the

413

MegaSyn Research Cluster.

414 415

References

416 417

[1] White, S. W., Zheng, J., Zhang, Y.-M., and Rock, C. O. (2005) The Structural Biology of Type II Fatty Acid Biosynthesis, Annu Rev Biochem. 74, 791-831.

418 419

[2] Maier, T., Leibundgut, M., Boehringer, D., and Ban, N. (2010) Structure and function of eukaryotic fatty acid synthases, Q. Rev. Biophys. 43, 373-422.

420 421 422

[3] Jenni, S., Leibundgut, M., Boehringer, D., Frick, C., Mikolasek, B., and Ban, N. (2007) Structure of fungal fatty acid synthase and implications for iterative substrate shuttling, Science 316, 254-261.

423 424 425

[4] Leibundgut, M., Jenni, S., Frick, C., and Ban, N. (2007) Structural Basis for Substrate Delivery by Acyl Carrier Protein in the Yeast Fatty Acid Synthase, Science 316, 288-290.

426 427 428

[5] Lomakin, I. B., Xiong, Y., and Steitz, T. A. (2007) The Crystal Structure of Yeast Fatty Acid Synthase, a Cellular Machine with Eight Active Sites Working Together, Cell 129, 319-332.

429 430 431

[6] Johansson, P., Wiltschi, B., Kumari, P., Kessler, B., Vonrhein, C., Vonck, J., Oesterhelt, D., and Grininger, M. (2008) Inhibition of the fungal fatty acid synthase type I multienzyme complex, Proc Natl Acad Sci U S A. 105, 12803-12808.

432 433

[7] Wakil, S. J. (1989) Fatty Acid Synthase, A Proficient Multifunctional Enzyme, 28, 4523-4530.

434 435

[8] Hertweck, C. (2009) The Biosynthetic Logic of Polyketide Diversity, Angew. Chem. Int. Ed. 48, 4688–4716.

436 437

[9] Weissman, K. J. (2015) The structural biology of biosynthetic megaenzymes, Nat. Chem. Biol. 11, 660-670.

438 439 440

[10] Tang, Y., Kim, C. Y., Mathews, I. I., Cane, D. E., and Khosla, C. (2006) The 2.7Å crystal structure of a 194-kDa homodimeric fragment of the 6-deoxyerythronolide B synthase, Proc. Natl. Acad. Sci. U.S.A. 103, 11124-11129.

441 442 443

[11] Herbst, D. A., Jakob, R. P., Zähringer, F., and Maier, T. (2016) Mycocerosic acid synthase exemplifies the architecture of reducing polyketide synthases, Nature 531, 533-537.


17

Page 19 of 34 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


444 445 446 447

[12] Keatinge-Clay, A. T., Shelat, A. A., Savage, D. F., Tsai, S.-C., Miercke, L. J. W., O'Connell III, J. D., Khosla, C., and Stroud, R. M. (2003) Catalysis, Specificity, and ACP Docking Site of Streptomyces coelicolor Malonyl-CoA:ACP Transacylase., Structure 11, 147-154.

448 449 450

[13] Dutta, S., Whicher, J. R., Hansen, D. A., Hale, W. A., Chemler, J. A., Congdon, G. R., Narayan, A. R., Håkansson, K., Sherman, D. H., Smith, J. L., and Skiniotis, G. (2014) Structure of a modular polyketide synthase., Nature 510, 512-517.

451 452

[14] Smith, S., and Tsai, S.-C. (2007) The type I fatty acid and polyketide synthases: a tale of two megasynthases., Nat. Prod. Rep. 24, 1041-1072.

453 454 455

[15] Smith, S., and Stern, A. (1983) The Effect of Aromatic CoA Esters on Fatty Acid Synthetase: Biosynthesis of ω-Phenyl Fatty Acids, Arch. Biochem. Biophys. 222, 259-265.

456 457 458 459

[16] Buckner, J. S., Kolattukudy, P. E., and Rogers, L. (1978) Synthesis of Multimethyl-Branched Fatty Acids by Avian and Mammalian Fatty Acid Synthetase and Its Regulation by Malonyl-CoA Decarboxylase in the Uropygial Gland., Arch. Biochem. Biophys. 186, 152-163.

460 461 462

[17] Seyama, Y., Otsuka, H., Kawaguchi, A., and Yamakawa, T. (1981) Fatty Acid Synthetase from the Harderian Gland of Guinea Pig: Biosynthesis of MethylBranched Fatty Acids, J. Biochem. 90, 789-798.

463 464 465

[18] Pirson, W., Schuhmann, L., and Lynen, F. (1973) The Specificity of Yeast Fatty Acid Synthetase with Respect to the "Priming" Substrate. Decanoyl-CoA and Derivatives as "Primers" of Fatty Acid Synthesis in vitro, Eur. J. Biochem. 36, 16-24.

466 467 468

[19] Sumper, M., Oesterhelt, D., Riepertinger, C., and Lynen, F. (1969) Synthesis of various carboxylic acids by the multienzyme complex of fatty acid synthesis from yeast, and clarification of their structure, Eur. J. Biochem. 10, 377-387.

469 470 471

[20] Dutton, C. J., Gibson, S. P., Goudie, A. C., Holdom, K. S., Pacey, M. S., Ruddock, J. C., Bu'Lock, J. D., and Richards, M. K. (1991) Novel avermectins produced by mutational biosynthesis., The Journal of Antibiotics 44, 357-365.

472 473 474

[21] Wang, F., Wang, Y., Ji, J., Zhou, Z., Yu, J., Zhu, H., Su, Z., Zhang, L., and Zheng, J. (2015) Structural and Functional Analysis of the Loading Acyltransferase from Avermectin Modular Polyketide Synthase, ACS Chem. Biol. 10, 1017-1025.

475 476 477 478

[22] Bhatt, A., Molle, V., Besra, G. S., Jacobs, W. R. J., and Kremer, L. (2007) The Mycobacterium tuberculosis FAS-II condensing enzymes: their role in mycolic acid biosynthesis, acid-fastness, pathogenesis and in future drug development., Mol. Microbiol. 64, 1442-1454.

479 480 481 482

[23] Sundermann, U., Bravo-Rodriguez, K., Klopries, S., Kushnir, S., Gomez, H., Sanchez-Garcia, E., and Schulz, F. (2013) Enzyme-Directed Mutasynthesis: A Combined Experimental and Theoretical Approach to Substrate Recognition of a Polyketide Synthase, ACS Chem. Biol. 8, 443-450.

483 484 485 486

[24] Bravo-Rodriguez, K., Ismail-Ali, A. F., Klopries, S., Kushnir, S., Ismail, S., Fansa, E. K., Wittinghofer, A., Schulz, F., and Sanchez-Garcia, E. (2014) Predicted Incorporation of Non-native Substrates by a Polyketide Synthase Yields Bioactive Natural Product Derivatives, ChemBioChem 15, 1991-1997.


18


Page 20 of 34

487 488 489 490

[25] Gregory, M. A., Petkovic, H., Lill, R. E., Moss, S. J., Wilkinson, B., Gaisser, S., Leadlay, P. F., and Sheridan, R. M. (2005) Mutasynthesis of Rapamycin Analogues through the Manipulation of a Gene Governing Starter Unit Biosynthesis, Angew. Chem. Int. Ed. 44, 4757-4760.

491 492 493

[26] Molnos, J., Gardiner, R., Dale, G. E., and Lange, R. (2003) A continuous coupled enzyme assay for bacterial malonyl–CoA:acyl carrier protein transacylase (FabD), Anal. Biochem. 319, 171-176.

494 495 496

[27] Dunn, B. J., Cane, D. E., and Khosla, C. (2013) Mechanism and Specificity of an Acyltransferase Domain from a Modular Polyketide Synthase, Biochemistry 52, 1839-1841.

497 498

[28] Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) Basic local alignment search tool., J. Mol. Biol. 215, 403-410.

499 500 501

[29] Cleland, W. W. (1963) The kinetics of enzyme-catalyzed reactions with two or more substrates or products, Biochimica et Biophysica Acta (BBA) - Specialized Section on Enzymological Subjects 67, 188-196.

502 503 504

[30] Rangan, V. S., and Smith, S. (1996) Expression in Escherichia coli and Refolding of the Malonyl-/Acetyltransferase Domain of the Multifunctional Animal Fatty Acid Synthase, J. Biol. Chem. 271, 31749-31755.

505 506 507

[31] Misra, A., Sharma, S. K., Surolia, N., and Surolia, A. (2007) Self-Acylation Properties of Type II Fatty Acid Biosynthesis Acyl Carrier Protein, Chem. Biol. 14, 775-783.

508 509 510 511

[32] Wang, Y.-Y., Bai, L.-F., Ran, X.-X., Jiang, X.-H., Wu, H., Zhang, W., Jin, M.-Y., Li, Y.-Q., and Jiang, H. (2015) Biochemical Characterization of a Malonyl-Specific Acyltransferase Domain of FK506 Biosynthetic Polyketide Synthase., Protein Pept. Lett. 22, 2-7.

512 513 514

[33] Liew, C. W., Nilsson, M., Chen, M. W., Sun, H., Cornvik, T., Liang, Z. X., and Lescar, J. (2012) Crystal Structure of the Acyltransferase Domain of the Iterative Polyketide Synthase in Enediyne Biosynthesis, J. Biol. Chem. 287, 23203-23215.

515 516

[34] Maier, T., Leibundgut, M., and Ban, N. (2008) The Crystal Structure of a Mammalian Fatty Acid Synthase, Science 321, 1315-1322.

517 518 519

[35] Pappenberger, G., Benz, J., Gsell, B., Hennig, M., Ruf, A., Stihle, M., Thoma, R., and Rudolph, M. G. (2010) Structure of the Human Fatty Acid Synthase KS–MAT Didomain as a Framework for Inhibitor Design, J. Mol. Biol. 397, 508-519.

520 521 522

[36] Krissinel, E., and Henrick, K. (2004) Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions, Acta Crystallogr. D Biol. Crystallogr. 60, 2256-2268.

523 524 525 526

[37] Bunkoczi, G., Misquitta, S., Wu, X., Lee, W. H., Rojkova, A., Kochan, G., Kavanagh, K. L., Oppermann, U., and Smith, S. (2009) Structural Basis for Different Specificities of Acyltransferases Associated with the Human Cytosolic and Mitochondrial Fatty Acid Synthases, Chem. Biol. 16, 667-675.

527 528

[38] Nardini, M., and Dijkstra, B. W. (1999) α/β Hydrolase fold enzymes: the family keeps growing, Curr. Opin. Struct. Biol. 9, 732-737.


19

Page 21 of 34 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


529 530

[39] Hol, W. G. (1985) Effects of the α-helix dipole upon the functioning and structure of proteins and peptides., Adv. Biophys. 19, 133-165.

531 532 533

[40] Oefner, C., Schulz, H., D'Arcy, A., and Dale, G. E. (2006) Mapping the active site of Escherichia coli malonyl-CoA-acyl carrier protein transacylase (FabD) by protein crystallography, Acta Crystallogr. D Biol. Crystallogr. 62, 613-618.

534 535 536 537

[41] Rangan, V. S., and Smith, S. (1997) Alteration of the Substrate Specificity of the Malonyl-CoA/Acetyl-CoA: Acyl Carrier Protein S-Acyltransferase Domain of the Multifunctional Fatty Acid Synthase by Mutation of a Single Arginine Residue, J. Biol. Chem. 272, 11975-11978.

538 539 540

[42] Gajewski, J., Buelens, F., Serdjukow, S., Janßen, M., Cortina, N., Grubmüller, H., and Grininger, M. (2017) Engineering fatty acid synthases for directed polyketide production, Nat. Chem. Biol. 13, 363–365.

541 542

[43] Warshel, A., Aqvist, J., and Creighton, S. (1989) Enzymes work by solvation substitution rather than by desolvation, Proc. Natl. Acad. Sci. U. S. A. 86, 5820-5824.

543 544

[44] Helfrich, E. J. N., and Piel, J. (2016) Biosynthesis of polyketides by trans-AT polyketide synthases, Nat. Prod. Rep. 33, 231-316.

545 546 547

[45] Zhu, Z., Zhou, Y. J., Krivoruchko, A., Grininger, M., Zhao, Z. K., and Nielsen, J. (2017) Expanding the product portfolio of fungal type I fatty acid synthases, Nat. Chem. Biol. 13, 360-362.

548 549 550

[46] Robbins, T., Liu, Y.-C., Cane, D. E., and Khosla, C. (2016) Structure and mechanism of assembly line polyketide synthases, Curr. Opin. Struct. Biol. 41, 1018.

551 552 553 554

[47] Peter, D. M., Schada von Borzyskowski, L., Kiefer, P., Christen, P., Vorholt, J. A., and Erb, T. J. (2015) Screening and Engineering the Synthetic Potential of Carboxylating Reductases from Central Metabolism and Polyketide Biosynthesis, Angew. Chem. Int. Ed. Engl. 54, 13457-13461.

555 556 557

[48] Mohamed, M. F., and Hollfelder, F. (2013) Efficient, crosswise catalytic promiscuity among enzymes that catalyze phosphoryl transfer, Biochim. Biophys. Acta 1834, 417-424.

558 559

[49] Vagin, A., and Teplyakov, A. (1997) MOLREP: an Automated Program for Molecular Replacement, J. Appl. Crystallogr. 30, 1022-1025.

560 561

[50] Emsley, P., Lohkamp, B., Scott, W. G., and Cowtan, K. (2010) Features and development of Coot, Acta Crystallogr. D Biol. Crystallogr. 66, 486-501.

562 563 564

[51] Grininger, M. (2014) Perspectives on the evolution, assembly and conformational dynamics of fatty acid synthase type I (FAS I) systems, Curr. Opin. Struct. Biol. 25, 49-56.

565 566 567 568

[52] Afonine, P. V., Moriarty, N. W., Mustyakimov, M., Sobolev, O. V., Terwilliger, T. C., Turk, D., Urzhumtsev, A., and Adams, P. D. (2015) FEM: feature-enhanced map, Acta Crystallogr. D Biol. Crystallogr. 71, 646-666.

569


20


Page 22 of 34

570 571 572

Figure 1: Animal fatty acid synthesis and MAT-mediated transacylation.

573

(A) Compartmentalized synthesis by FAS centered around the loading function of the

574

MAT domain. The elongation cycle is divided into a condensation reaction,

575

performed by KS, and modification of the β-carbon, catalyzed by the domains KR,

576

DH and ER. The flexible attachment of ACP via linkers is abstracted by the zigzag

577

lines for the elongation cycle. Abbreviations as introduced in the text. Adapted from

578

51

579

CoA-ester is transferred to the enzyme to form an acyl-enzyme intermediate (MAT-X;

580

active site serine-acylated), with the concomitant release of the free CoA (ping). The

581

ACP then interacts with the loaded enzyme intermediate forming the acylated ACP

582

and regenerating the native enzyme (pong). The left branch represents the common

583

side reaction of hydrolysis, which can occur when water is utilized as a nucleophile

584

attacking the acyl-enzyme intermediate.

. (B) Scheme of the ping-pong bi-bi mechanism of MAT. The acyl-moiety of the

585


21

Page 23 of 34 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


586 587

Figure 2. Overall structure of murine KS-MAT didomain.

588

(A) Superposition of the two biological dimers of KS-MAT onto the porcine FAS KS-

589

MAT dimer. For clarity, the full X-ray structural model of porcine FAS, excluding the

590

domains ACP and TE that were not traceable in electron density,34 is shown as inset.

591

Porcine FAS KS-MAT is shown in grey, and the two biological dimers in orange

592

(chain A and B) and cyan (chain C and D). Subdomains and folds are indicated by

593

labels and coloring. (B) Superposition of the four KS-MAT modules in the asymmetric

594

unit aligned via the KS domains (residue range 1–407; upper panel) and of the MAT

595

(including LD/LD2) aligned via their α/β-hydrolase domain (residue ranges 488–615

596

and 685–806; lower panel). Domains and folds are color coded as depicted in the

597

attached cartoon. Cofactor and residues are shown for chain D (malonyl-loaded

598

state). (C) Binding site of murine MAT in its apo-state (chain B), in complex with

599

malonyl-CoA (chain C) and in its acyl-enzyme state (chain A and D). Electron density

600

of an unbiased feature-enhanced map of the active site regions is contoured at 1σ.52

601

View 70° rotated (y-axis) as compared to Figure 2B.

602


22


Page 24 of 34

603 604

Figure 3. MAT binding site architecture.

605

(A) Acyl-CoA binding cavity as comprised by the α/β-hydrolase and the ferredoxin-

606

like domain. The binding cavity is shown in mesh representation (left panel) and with

607

surfaces colored in electrostatic potential (right). The cavity was drawn with default

608

settings of PyMOL (vacuum electrostatics function; The PyMOL Molecular Graphics

609

System, Schrödinger, LLC.). Figures of the LD-MAT substructure in the upper panel

610

are provided for overview. (B) Comparison of the binding site of the acyl-enzyme

611

state (chain D) with the apo-state (chain A, orange) of human MAT. Distances

612

indicate that N738 is the only residue in H-bond distance in the apo state.

613

Superposition has been performed with α/β-hydrolase domains (residue ranges 488–

614

615 and 685–806; pdb: 3hhd).35 Residue numbers are identical for both proteins. H-

615

bonding to the guanidinium group is indicated by dashed lines. (C) Malonyl-loaded

616

enzyme states of MAT (chain D, residues in yellow) in superposition with malonyl

617

transferases FabD (pdb: 2g2z, grey, H-bonds in blue dashed lines),40 and AT of

618

DynE8 (pdb: 4amp, white, H-bonds in grey).33 Superposition indicates side-on vs.

619

end-on bidentate coordination of arginine and the carboxyl group of serine bound

620

malonyl. H-bond networks of active site arginines are different in FabD (right top) and

621

in AT of DynE8 (right bottom) to MAT (see Figure 3B).

622


23

Page 25 of 34 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


623 624

Figure 4: Double-reciprocal plots of the Michaelis-Menten constants. Reciprocal

625

Km (top panel) and reciprocal initial rates kcat (bottom) are presented as functions of

626

reciprocal ACP concentrations. Data collected on malonyl-CoA are shown in black,

627

acetyl-CoA in grey, and linear fits are accordingly. Error bars reflect the standard

628

deviation between three biological replicates. A biological replicate was defined by

629

three independently expressed MAT preparations and different batches of purified,

630

high quality ACP (see Figure S3), mainly from bicistronic expression with Sfp. To

631

establish the fidelity of the assay, every measurement was performed in technical

632

triplicates. Tables underneath graphs summarize absolute kinetic constants

633

calculated from the double-reciprocal plot. Michaelis-Menten constants for ACP were

634

calculated from both plots and averaged (245 µM and 245 µM for malonyl-transfer;

635

284 µM and 246 µM for acetyl-transfer).

636 637


24


Page 26 of 34

638 639 640

Figure 5: MAT as tool in PKS engineering.

641

(A) Domain organization of murine FAS and constructs engineered in this study; FAS

642

(1), KS-MAT (2), LD-MAT (3), MAT-MBP (4), LD-MAT-ACP-FAS (5), MAT-ACP-FAS

643

(6). (B) Coomassie-stained SDS-PAGE gel (NuPAGE 4–12 % Bis-Tris) of purified

644

constructs after tandem affinity chromatography and SEC. Calculated sizes of

645

constructs (1), (2), (3) and (4) are 277, 97, 49 and 81 kDa, respectively. (C) SEC of

646

purified constructs with absorption normalized to the highest peak. Peaks (1), (2)

647

(2m)), (4) and (3) correspond to apparent molecular weights of 601, 231 (monomer:

648

118), 85 and 48 kDa, respectively. Calibration is shown in inset; thyroglobulin (T),

649

aldolase (A), ovalbumin (O) and ribonuclease (R). (D) Partial activity of MAT in the

650

αKGDH-assay monitored at fixed substrate concentrations of 20 µM CoA-ester and

651

60 µM ACP. Malonyl- and acetyl-turnovers are shown as black and grey bars,

652

respectively. All data were collected in technical triplets. Error reflects standard

653

deviation from four biological replicates. (E) In-gel fluorescence for monitoring the

654

degree of phosphopantetheinylation in co-expressions with Sfp (lower panel) and

655

Coomassie-stained SDS-PAGE gel (upper panel). Star (*) indicates activated

656

(phosphopantetheinylated) protein received in co-expressions with Sfp. Calculated

657

sizes of constructs (1), (2), (5) and (6) are 277, 338 and 323 kDa, respectively.

658

Proteins were incubated with 10 mM MgCl2, 10 µM purified Sfp and 1 µM CoA 647

659

for 1 h at 37 °C. SDS-PAGE gel (NuPAGE 4–12 % Bis-Tris) of fluorescently labeled

660

constructs is shown after fluorescence detection and Commassie-staining. For

661

monitoring phosphopantetheinylation, KS-MAT (not containing ACP) was used as a

662

negative, and FAS, expressed as apo-protein and in vitro phosphopantetheinylated,

663

was used as positive control. (F) Preparative SEC of purified constructs after co-

664

expression with Sfp with absorption normalized to the highest peak to isolate the

665

dimeric fraction of each protein. Peaks (1*), (5*) and (6*) correspond to apparent


25

Page 27 of 34 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


666

molecular weights of 697 kDa, 784 kDa and 727 kDa, respectively. Peak shoulders

667

indicate the presence of monomers in all samples, which are in temperature and

668

buffer dependent equilibrium with the dimer (for analytic SEC, see also Figure S8).

669

(G) FAS activity addressed by the specific oxidation of NADPH in 50 mM potassium

670

phosphate. The consumption of NADPH was monitored fluorometrically. FAS without

671

Sfp co-expression shows no activity.

672


26


Page 28 of 34

673

Table 1: Kinetic analysis of the transacylation reaction with unusual CoA-esters at a

674

fixed acceptor concentration of 60 µM ACP. app

(µM)

app

–1

kcat/Km (M 1 × 106)

s

–

Km

Malonyl-CoA

1.28

±

0.13

15.6

±

1.4

12.2

± 1.7

9.8

±

1.7

kcat

(s )

–1

Substrate

Hydrolysis rate 10–3 (s-–1)

Methylmalonyl-CoA

0.62

±

0.07

4.0

±

0.8

6.5

± 1.4

6.0

±

1.2

Acetyl-CoA

1.63

±

0.10

13.7

±

2.1

8.4

± 1.4

9.3

±

0.8

Butyryl-CoA

1.64

±

0.16

10.8

±

0.3

6.6

± 0.7

8.5

±

1.7

Octanoyl-CoA*

0.73

±

0.28

4.1

±

0.3

5.6

± 2.2

6.1

±

1.0

Palmitoyl-CoA**

4.8

±

1.1

2.5

±

0.5

0.5

± 0.2

16.0

±

1.4

Acetoacetyl-CoA*

0.86

±

0.03

7.3

±

0.4

8.5

± 0.5

13.0

±

2.5

Hydroxybutyryl-CoA

1.72

±

0.39

6.1

±

1.1

3.6

± 1.0

20.4

±

2.7

Crotonyl-CoA

1.36

±

0.13

6.0

±

1.1

4.4

± 0.9

24.0

±

2.0

Phenylacetyl-CoA*

0.96

±

0.10

6.0

±

0.2

6.2

± 0.7

21.3

±

2.3

Methylbutyryl-CoA

2.06

±

0.36

7.2

±

0.9

3.5

± 0.7

10.4

±

1.3

Succinyl-CoA

2.10

±

0.17

13.4

0.6

6.4

± 0.6

28.7

±

2.7

R606A mutant

Kmapp

(µM)

kcat

app

± –1

(s )

kcat/Km (M 6 × 10 )

–1

s

–

1

Hydrolysis rate –3 –1 10 (s )

Malonyl-CoA

18

±

5.0

5

±

1.7

0.3

± 0.1

14.7

±

2.3

Acetyl-CoA

3.0

±

0.56

13

±

2.1

4

± 1.1

15.2

±

2.1

675

All parameters were determined in technical and biological (independent expressed

676

and purified MAT preparations) triplets. (*) and (**) MAT at 2 nM and 4 nM,

677

respectively, were used, because of low turnover rates.

678

Hydrolysis rate was determined in 6 repetitions.

679


27

Page 29 of 34 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


680


28


Figure 1: Animal fatty acid synthesis and MAT-mediated transacylation. (A) Compartmentalized synthesis by FAS centered around the loading function of the MAT domain. The elongation cycle is divided into a condensation reaction, performed by KS, and modification of the β-carbon, catalyzed by the domains KR, DH and ER. The flexible attachment of ACP via linkers is abstracted by the zigzag lines for the elongation cycle. Abbreviations as introduced in the text. Adapted from 51. (B) Scheme of the ping-pong bi-bi mechanism of MAT. The acyl-moiety of the CoA-ester is transferred to the enzyme to form an acyl-enzyme intermediate (MAT-X; active site serine-acylated), with the concomitant release of the free CoA (ping). The ACP then interacts with the loaded enzyme intermediate forming the acylated ACP and regenerating the native enzyme (pong). The left branch represents the common side reaction of hydrolysis, which can occur when water is utilized as a nucleophile attacking the acyl-enzyme intermediate.

140x54mm (300 x 300 DPI)


Page 30 of 34

Page 31 of 34 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 2. Overall structure of murine KS-MAT didomain. (A) Superposition of the two biological dimers of KS-MAT onto the porcine FAS KS-MAT dimer. For clarity, the full X-ray structural model of porcine FAS, excluding the domains ACP and TE that were not traceable in electron density,34 is shown as inset. Porcine FAS KS-MAT is shown in grey, and the two biological dimers in orange (chain A and B) and cyan (chain C and D). Subdomains and folds are indicated by labels and coloring. (B) Superposition of the four KS-MAT modules in the asymmetric unit aligned via the KS domains (residue range 1–407; upper panel) and of the MAT (including LD/LD2) aligned via their α/β-hydrolase domain (residue ranges 488–615 and 685–806; lower panel). Domains and folds are color coded as depicted in the attached cartoon. Cofactor and residues are shown for chain D (malonyl-loaded state). (C) Binding site of murine MAT in its apo-state (chain B), in complex with malonyl-CoA (chain C) and in its acyl-enzyme state (chain A and D). Electron density of an unbiased feature-enhanced map of the active site regions is contoured at 1σ.52 View 70° rotated (y-axis) as compared to Figure 2B.

140x120mm (300 x 300 DPI)



Figure 3. MAT binding site architecture. (A) Acyl-CoA binding cavity as comprised by the α/β-hydrolase and the ferredoxin-like domain. The binding cavity is shown in mesh representation (left panel) and with surfaces colored in electrostatic potential (right). The cavity was drawn with default settings of PyMOL (vacuum electrostatics function; The PyMOL Molecular Graphics System, Schrödinger, LLC.). Figures of the LD-MAT substructure in the upper panel are provided for overview. (B) Comparison of the binding site of the acyl-enzyme state (chain D) with the apostate (chain A, orange) of human MAT. Distances indicate that N738 is the only residue in H-bond distance in the apo state. Superposition has been performed with α/β-hydrolase domains (residue ranges 488–615 and 685–806; pdb: 3hhd).35 Residue numbers are identical for both proteins. H-bonding to the guanidinium group is indicated by dashed lines. (C) Malonyl-loaded enzyme states of MAT (chain D, residues in yellow) in superposition with malonyl transferases FabD (pdb: 2g2z, grey, H-bonds in blue dashed lines),40 and AT of DynE8 (pdb: 4amp, white, H-bonds in grey).33 Superposition indicates side-on vs. end-on bidentate coordination of arginine and the carboxyl group of serine bound malonyl. H-bond networks of active site arginines are different in FabD (right top) and in AT of DynE8 (right bottom) to MAT (see Figure 3B).

140x141mm (300 x 300 DPI)


Page 32 of 34

Page 33 of 34 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: Double-reciprocal plots of the Michaelis-Menten constants. Reciprocal Km (top panel) and reciprocal initial rates kcat (bottom) are presented as functions of reciprocal ACP concentrations. Data collected on malonyl-CoA are shown in black, acetyl-CoA in grey, and linear fits are accordingly. Error bars reflect the standard deviation between three biological replicates. A biological replicate was defined by three independently expressed MAT preparations and different batches of purified, high quality ACP (see Figure S3), mainly from bicistronic expression with Sfp. To establish the fidelity of the assay, every measurement was performed in technical triplicates. Tables underneath graphs summarize absolute kinetic constants calculated from the double-reciprocal plot. Michaelis-Menten constants for ACP were calculated from both plots and averaged (245 µM and 245 µM for malonyl-transfer; 284 µM and 246 µM for acetyl-transfer). 119x214mm (300 x 300 DPI)



Figure 5: MAT as tool in PKS engineering. (A) Domain organization of murine FAS and constructs engineered in this study; FAS (1), KS-MAT (2), LDMAT (3), MAT-MBP (4), LD-MAT-ACP-FAS (5), MAT-ACP-FAS (6). (B) Coomassie-stained SDS-PAGE gel (NuPAGE 4–12 % Bis-Tris) of purified constructs after tandem affinity chromatography and SEC. Calculated sizes of constructs (1), (2), (3) and (4) are 277, 97, 49 and 81 kDa, respectively. (C) SEC of purified constructs with absorption normalized to the highest peak. Peaks (1), (2) (2m)), (4) and (3) correspond to apparent molecular weights of 601, 231 (monomer: 118), 85 and 48 kDa, respectively. Calibration is shown in inset; thyroglobulin (T), aldolase (A), ovalbumin (O) and ribonuclease (R). (D) Partial activity of MAT in the αKGDH-assay monitored at fixed substrate concentrations of 20 µM CoA-ester and 60 µM ACP. Malonyland acetyl-turnovers are shown as black and grey bars, respectively. All data were collected in technical triplets. Error reflects standard deviation from four biological replicates. (E) In-gel fluorescence for monitoring the degree of phosphopantetheinylation in co-expressions with Sfp (lower panel) and Coomassiestained SDS-PAGE gel (upper panel). Star (*) indicates activated (phosphopantetheinylated) protein received in co-expressions with Sfp. Calculated sizes of constructs (1), (2), (5) and (6) are 277, 338 and 323 kDa, respectively. Proteins were incubated with 10 mM MgCl2, 10 µM purified Sfp and 1 µM CoA 647 for


Page 34 of 34

Page 35 of 34 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 h at 37 °C. SDS-PAGE gel (NuPAGE 4–12 % Bis-Tris) of fluorescently labeled constructs is shown after fluorescence detection and Commassie-staining. For monitoring phosphopantetheinylation, KS-MAT (not containing ACP) was used as a negative, and FAS, expressed as apo-protein and in vitro phosphopantetheinylated, was used as positive control. (F) Preparative SEC of purified constructs after coexpression with Sfp with absorption normalized to the highest peak to isolate the dimeric fraction of each protein. Peaks (1*), (5*) and (6*) correspond to apparent molecular weights of 697 kDa, 784 kDa and 727 kDa, respectively. Peak shoulders indicate the presence of monomers in all samples, which are in temperature and buffer dependent equilibrium with the dimer (for analytic SEC, see also Figure S8). (G) FAS activity addressed by the specific oxidation of NADPH in 50 mM potassium phosphate. The consumption of NADPH was monitored fluorometrically. FAS without Sfp co-expression shows no activity.

140x160mm (300 x 300 DPI)


Characterization of the Polyspecific Transferase of Murine Type I Fatty

Recommend Documents