A Novel Approach for Gene Expression Optimization through Native

Subscriber access provided by Fudan University

Article

A Novel Approach for Gene Expression Optimization through Native Promoter and 5’ UTR Combinations Based on RNAseq, Ribo-seq, and TSS-seq of Streptomyces coelicolor Jeong Sang Yi, Min Woo Kim, Minsuk Kim, Yujin Jeong, Eun-Jung Kim, Byung-Kwan Cho, and Byung-Gee Kim ACS Synth. Biol., Just Accepted Manuscript • DOI: 10.1021/acssynbio.6b00263 • Publication Date (Web): 14 Dec 2016 Downloaded from http://pubs.acs.org on December 15, 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.

ACS Synthetic 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 41

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 Synthetic Biology

1

A Novel Approach for Gene Expression Optimization through

2

Native Promoter and 5’ UTR Combinations Based on RNA-

3

seq, Ribo-seq, and TSS-seq of Streptomyces coelicolor

4 5

Jeong Sang Yi1, Min Woo Kim1, Minsuk Kim1, Yujin Jeong2, Eun-Jung Kim1, Byung-

6

Kwan Cho2, 3,*, and Byung-Gee Kim1, 4,*

7 1

8

School of Chemical and Biological Engineering, Institute of Molecular Biology and

9

Genetics, and Bioengineering Institute, Seoul National University, Seoul 151-742,

10

Republic of Korea

11

2

Department of Biological Sciences and KI for the BioCentury, Korea Advanced Institute

12

of Science and Technology, Daejeon 34141, Republic of Korea 3

13 14 15

Intelligent Synthetic Biology Center, Daejeon 34141, Republic of Korea

4

Interdisciplinary Program for Biochemical Engineering and Biotechnology, Seoul National University, Seoul 151-742, Republic of Korea

16 17 18 19 20 21

*Correspondence and requests for materials should be addressed to B.-G.K. (email:

22

[email protected]) or B.-K.C. (email: [email protected])

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

23

Table of Contents

24

25

A Novel Approach for Promoter and 5’ UTR Constructions

26

Based on RNA-seq, Ribo-seq, and TSS-seq of Streptomyces

27

coelicolor

28 29

Jeong Sang Yi1, Min Woo Kim1, Minsuk Kim1, Yujin Jeong2, Eun-Jung Kim1, Byung-

30

Kwan Cho2, 3,*, and Byung-Gee Kim1, 4,*

31

32 33


Page 2 of 41

Page 3 of 41

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

34


Abstract

35

Streptomycetes are Gram-positive mycelial bacteria, which synthesize a wide

36

range of natural products including over two-thirds of the currently available antibiotics.

37

However, metabolic engineering in Streptomyces species to overproduce a vast of

38

natural products are hampered by limited number of genetic tools. Here, two promoters

39

and four 5’ UTR sequences showing constant strengths were selected based upon

40

multi-omics datasets from Streptomyces coelicolor M145, including RNA-seq, Ribo-seq,

41

and TSS-seq, for controllable transcription and translation. Total eight sets of promoter /

42

5’ UTR combinations, with minimal interferences of promoters on translation, were

43

constructed using the transcription start site information, and evaluated with GusA

44

system. Expression of GusA could be controlled to various strengths in three different

45

media, in a range of 0.03 to 2.4 folds, compared to that of the control, ermE*P / Shine-

46

Dalgarno sequence. This method was applied to engineer three previously reported

47

promoters to enhance gene expressions. The expressions of ActII-ORF4 and MetK

48

were also tuned for actinorhodin overproductions in S. coelicolor as examples. In

49

summary, we provide a novel approach and tool for optimizations of gene expressions

50

in Streptomyces coelicolor.

51 52 53

Keywords: Promoter engineering, Ribosome Binding Site, Metabolic engineering, Streptomyces coelicolor, Antibiotics



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 41

54

Productions of antibiotics from the soil-dwelling microorganism Streptomyces1 are in

55

great demands for pharmaceutical industries owing to surge of multidrug resistances of

56

bacteria.2 In order to produce such antibiotics at commercial levels, various genetic

57

tools for fine control of gene expression of biosynthetic pathways and their regulatory

58

network are in great demands3, 4 due to limited numbers of gene expression systems

59

available for Streptomyces sp.. Currently, although other systems such as kasOp and

60

SF14p are available,5 ermE* promoter (ermE*P)6 from Saccharopolyspora erythraeus is

61

the most widely used strong and constitutive promoter. Another example is a

62

thiostrepton inducible promoter, tipA, with limited induction levels.7,

63

developments of new scalable and predictable gene expression systems are in urgent

64

needs.

8

As a result,

65

To this end, several constitutive promoters were recently evaluated and

66

developed based on the RNA-seq results from Streptomyces coelicolor M11469 and

67

Streptomyces albus J1074.10 It is reasonable to construct libraries of promoters with

68

RNA-seq, because the RNA-seq is direct measurement of mRNA transcript levels, and

69

sensitive with broader dynamic range, compared to those from microarray.11 However,

70

RNA-seq data are not enough to design the gene expression system for fine control of

71

gene expression, because actual amounts of protein synthesized depend on the

72

efficiency with which an mRNA is translated, referred as the translation efficiency (TE).

73

TE is determined by secondary structure of mRNA and ribosome binding site (RBS)

74

affinity for ribosomes.12 If TE is low, despite the abundance of mRNA, amounts of

75

protein synthesized are low accordingly.13 Moreover, strengths of RBS along with that of

76

promoters determine protein folding kinetics and solubility.14 Protein mis-folding caused


Page 5 of 41

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


77

by unmatched transcriptional and translational rates of genes may lead to inclusion

78

bodies or nonfunctional proteins.15 There have been many studies to solve such

79

problems in translations, e.g. chaperones16 and media optimization to comply with

80

codon usage of target protein,17 but ideally, promoter and RBS combinations should be

81

optimized for the best pairs with predictable and controllable target gene expression.

82

Along with RNA-seq, quantitative analysis of protein translation is acquired by

83

ribosome profiling (Ribo-seq), which analyzes the ribosome protected mRNA fragments

84

(RPFs) by deep sequencing.18, 19 TE is then calculated by dividing the RPF levels by the

85

corresponding mRNA transcript levels for each gene. Accordingly, TE of all available

86

mRNA of a species can be calculated.20 It suggests that we can select native 5’ UTR

87

sequences from the pools of thousands of mRNAs with known TE, and design synthetic

88

5’ UTRs to precisely control target gene translation instead of synthetic RBS’s.

89

In this work, a novel method for screening and designing pairs of promoters and

90

5’ UTR regions was suggested considering both transcription and translation. Recent

91

studies mainly searched separately for promoters and RBSs based on RNA-seq data

92

and 5’ UTR designing software, because multi-omics datasets like Ribo-seq and TSS-

93

seq were not available to Streptomyces sp.. They are not satisfactory enough to meet

94

our needs to provide strong or optimal gene expression of target genes in actinobacteria.

95

Using TSS information from the very first Ribo-seq and TSS-seq datasets from S.

96

coelicolor, promoters and 5’ UTRs were constructed so that they would maintain their

97

native TSS, and promoter sequences would not interfere and cross talk with 5’ UTR

98

sequences. Total eight sets of promoter / 5’ UTR combinations, consisted of two

99

promoters and four 5’ UTRs with varying transcriptional and translational strengths,



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 41

100

were constructed because Streptomyces sp. lack such systems that can control

101

strengths of gene expressions.7, 8

102

The promoters and 5’ UTRs were evaluated with β-D-glucuronidase (GusA),

103

since GusA assay is proven to be effective in several different Streptomyces sp., and is

104

as simple as color assays from β–galactosidase and X-gal.21, 22 And then, they were

105

successfully applied to enhance strengths of gene expressions from previously reported

106

promoters for Streptomyces sp., PSCO5768, PSCO4658, and PSCO3410.9 The promoter / 5’

107

UTR sets were also applied to differentially overexpress ActII-ORF4 (SCO5085) and

108

MetK (SCO1476) in S. coelicolor, two well-known targets with positive effects on the

109

polyketide antibiotic, actinorhodin (ACT), production.5,

110

expression was successfully predicted and controlled by the selection of promoters and

111

5’ UTRs from multi-omics datasets, i.e. RNA-seq, Ribo-seq, and TSS-seq. This study

112

proposes a new tool, which can be usefully applied to synthetic biology and metabolic

113

engineering for secondary metabolite productions in Streptomyces strains.

23,

24

In summary, gene

114 115

Results and Discussion

116

Promoter and 5’ UTR target selections based on Multi-omic data

117

RNA-seq and Ribo-seq data acquired at 4 different growth phases from R5- complex

118

medium cultures, and TSS-seq data measured under various growth conditions, in our

119

previous report,25 were used to select promoters and 5’ UTRs from S. coelicolor

120

genome. The data were obtained from Gene Expression Omnibus (GEO) database

121

using accession number of GSE69350. The multi-omics datasets could explain

122

important biological features about S. coelicolor in a great detail. It included TSS


Page 7 of 41

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


123

information, such as numbers of leaderless genes, and lengths of 5’ UTR sequences.

124

Various aspects of transcription and translation are also studied, which included

125

information such as transcriptional patterns, correlations between transcription and

126

translation, transition of metabolism from primary to secondary, etc. The datasets

127

comprised of such comprehensive information, which enabled us to make meaningful

128

selections and constructions of promoters and 5’ UTRs.

129

Genes having both constitutive transcription and stable translation were targeted

130

for selection of promoters and 5’ UTR candidates (Figure 1). Reads per Kilobase per

131

Million mapped reads (RPKM) values, calculated by total reads of a gene divided both

132

by its length in kilobase and total million mapped reads of the genome, which represent

133

quantitative transcription level of the genes, varied greatly from 0 to 45000. Among them,

134

genes with RPKM values under 600 were eliminated as they can be seen as under-

135

expressed. At this stage, the size of candidate pool was reduced down to 311 genes out

136

of 7846 genes. Among them, only the genes with their RPKM level changes less than

137

20%, or 0.2 folds, of their average, were considered to be constitutive. The cutoff values

138

for the promoter selections were up to 10 folds stricter than other previous reports.9, 10 In

139

this way, our promoter candidates were more reliable for constitutive transcription due to

140

the datasets acquired at more growth time points, and very thorough and harsh cutoff

141

conditions. There are no limits to the selection of promoter targets. For example,

142

housekeeping genes such as hrdB can be a good candidate for promoters. But

143

according to our study, hrdB gene did not belong to our 35 candidates. hrdB had an

144

average RPKM value of 513 ± 172, where the fluctuation was about 34% (0.34 folds) of

145

the average. Resulting 35 genes were further screened for 5’ UTR targets.



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

146

TE varied greatly from 0.20 to 29.09, with an average of 1.09 for the entire

147

genes. TE values may fluctuate depending on various conditions such as growth

148

phases.26 Thus, TE should also be as constitutive as possible throughout the life cycle

149

of the cells, for predictable and stable protein expression. The genes with their TE value

150

changes over 45%, calculated by standard deviation of TE divided by the average TE

151

value throughout the four time point sets, were removed from the target list. The genes

152

without definite transcription start site, such as leaderless or undetected based on TSS-

153

seq, were removed at this point. Final 23 genes were selected, which are listed in Table

154

1 and in Table S1 for their RPKM and TE values at different growth phases.

155

Total two promoters and four 5’ UTR were selected from the 23 candidates.

156

There are so many possible combinations of promoters and 5’ UTRs to test, even from

157

23 selected candidate genes. We selected promoters and 5’ UTR that have definitive

158

differences in their gene expression because it is hard to distinguish the effects if target

159

combinations have small differences in RPKM and TE. Cold shock protein scoF2

160

(SCO4505, P4505), the strongest of the 23 candidates, and succinyl-CoA synthetase

161

sucC (SCO4808, P4808), showing 1/10 strength of scoF2, were chosen as strong and

162

weak promoters, respectively. A putative transcriptional regulator clgR (SCO5755, R5755)

163

with the largest TE was selected as the 5’ UTR with the highest TE value. For 5’ UTR

164

with smaller TE values, terD (SCO0641, R0641), and two putative proteins, SCO3083

165

(R3083) and SCO2078 (R2078), were selected as shown in Table 2, and their specific

166

sequences are listed in Table S2. In this way, the eight combinations could represent

167

most of the constitutive promoters and 5’ UTRs.

168


Page 8 of 41

Page 9 of 41

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


169

Construction and validation of promoters and 5’ UTRs

170

TE is highly dependent on secondary structures of mRNA, so that there should be

171

minimized interferences from its promoter sequences or terminal sequences of

172

upstream genes, e.g. where transcription starts or where ribosomes bind, to maintain

173

stable and observed TE values.27 Thus, translation rates of the 23 candidate genes,

174

including the four selected 5’ UTRs (SCO5755, SCO0641, SCO3083, and SCO2078),

175

were predicted beforehand with RBS calculator28 for the guidance of designing 5’ UTR

176

regions. The resulting calculations were compared with experimental TE values from

177

Ribo-seq25. Unfortunately, however, there were no correlation between the predicted

178

translational strengths and experimentally acquired TE values (Figure S1), owing to a

179

couple of reasons.

180

RBS calculator predicts the translational strength by calculating changes in

181

Gibbs free energy, ∆Gtotal, of mRNA with 5 different variables, which are ∆G’s of i)

182

mRNA-rRNA interactions, ii) start codon – initiating tRNA anticodon loop hybridization, iii)

183

unfolding mRNA subsequence, iv) unfolding standby site, and v) ∆G loss due to the

184

space between 16S rRNA binding site and start codon. Of all, the unfolding kinetics of

185

standby site is considered as the most important factors affecting translation rate,29

186

because there are high possibility of having increased numbers of the standby sites as

187

the lengths of 5’ UTR region become longer. S. coelicolor mRNAs often contain very

188

wide ranges of 5’ UTR lengths, from leaderless to over 456 nucleotides long.

189

Determined by multi-omics datasets, 413 transcripts from total 2705 known coding

190

sequences were found to be longer than 150 nucleotides.25 It was longer than RBS

191

calculator and / or UTR designer30 could make a plausible predictions. Thus



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 41

192

experimental TE values were more likely to be different from the predictions made. In

193

addition, 5’ UTR designing methods were specific to open reading frames. As a result,

194

designing our four 5’ UTRs was done following a different method. Instead of designing

195

5’ UTR region based on modeling, complete 5’ UTR sequences of the four 5’ UTR

196

targets were cloned. (Figure 2).

197

One of the previous studies on promoter and RBS engineering used an insulator

198

between a promoter and RBS, such as RIboJ, to minimize promoter sequences

199

interfering with mRNA sequences structures, and altering TE.31 Instead of insulators,

200

promoters and 5’ UTRs were designed to have their natural TSS in this study. For

201

example, P4505R5755 would start its transcription on the same nucleotide as that of wild

202

type 5’ UTR of SCO5755. In this way, the eight promoter / 5’ UTR combinations would

203

have the same 5’ UTR structures with WT transcripts from the four 5’ UTR targets. As a

204

result, Gibbs free energy of mRNA unfolding would become the only variable altering TE

205

from the experimental TE values. The mRNA unfolding energy is also fixed to the gene

206

coding sequences. This makes the eight promoter / 5’ UTR be universally applicable to

207

optimize gene expression independently, without the needs of designing gene specific 5’

208

UTRs.

209

To validate whether or not the constructed promoter / 5’ UTR sets are functional,

210

evaluation was carried out with GusA assay system.21, 22 Because most of Streptomyces

211

species have their own native catechol 2, 3-monooxygenase (xylE) gene, such as

212

SAV1615 from Streptomyces avermitilis, assays with xylE show higher background

213

noises with less sensitivities than that of GusA. For the same reason, GusA system was

214

also more sensitive to the enhanced green fluorescent protein reporter system for the


Page 11 of 41

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


215

measurement of promoter strengths.21 The expression systems constructed following a

216

conventional cloning method were also evaluated to examine effectiveness of the eight

217

promoter / 5’ UTR combinations constructed with TSS information. GusA was expressed

218

with the strong P4505 promoter that included three randomly chosen RBS sequences

219

(Table 2), in the order of R5189, R3854, and R5061 with a decreasing order of TE. About 30

220

bp upstream regions of the randomly chosen genes were inserted without considering

221

their TSSs.

222

Examining ermE*P and the four 5’ UTR combinations were unconsidered due to

223

unavailable RPKM value of ermE* promoter, and uncertainty of TSS in pIBR25 plasmid.

224

Even though ermE*P is one of the most widely used promoter, it is not completely

225

characterized.5 Based on a sequencing data of ermE* promoter in pIBR25 plasmid

226

(data not shown), it was composed of two ermE* sequences, complete sequences of

227

ermE2 and partial sequences of ermE1. As a result, accurate evaluation of 5’ UTRs

228

were not possible because TSS information of ermE promoter in pIBR25 was not

229

available. TSS information was very important in the constructions of 5’ UTRs in this

230

study. But as a reference, GusA expressed with Streptomyces Shine-Dalgarno (RSD)

231

sequence (AGCAACGGAGGTACGGAC)32 containing ermE* promoter was used.

232

Strengths of the eight promoter / 5’ UTR combinations were ranked based on the RPKM

233

and TE values, such as stronger GusA expression with increasing order of (RPKM * TE)

234

values.

235

Despite the differences in the observed values of TE from Ribo-seq data, when

236

5’ UTRs were constructed with conventional methods, there were no correlations

237

between their TE values and GusA activities (Figure 3a). All the three expression levels



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 41

238

were about the same as that of the control, PermE* / RSD. On the other hand, when the

239

expression systems were constructed in consideration of TSS, GusA assay results

240

matched well with the predicted orders of the combinations (Figure 3b). The four 5’

241

UTRs were functional as predicted, indicating that there were minimal interference of

242

promoter sequences on translations. If there were cross-talks between the promoters

243

and 5’ UTRs, GusA expression will not comply with predicted GusA activities, when a

244

promoter was changed from P4505 to P4808 for the four 5’ UTRs.

245

This result confirmed that our design method could be successfully manipulated

246

according to the order of the prediction made by (RPKM * TE) values. In addition, when

247

expressed by the two strongest promoter / 5’ UTR combinations, P4505R5755 and

248

P4505R0641 respectively, GusA expression could be increased by about 2 folds and 1.3

249

folds compared to that of PermE* / RSD. GusA gene expression could also be manipulated

250

in a range from 0.03 to 2.4 folds of ermE* promoter and SD sequence expression.

251

GusA expressions in cultures from different media were also evaluated to

252

examine whether the applications were limited to R5- complex media or not. YEME and

253

Supplemented Minimal Media (SMM) were selected for their varying constituents.

254

Expressions of GusA in YEME media were almost identical to those from R5- complex

255

media, with two exceptions (Figure 4a). GusA expressions from both P4505R5755 and

256

P4505R0641 were increased by about 3 folds, which correspond to about 8.8 folds and 4.5

257

folds stronger than that of PermE* / RSD respectively. Gene expressions from the eight

258

promoter / 5’ UTR sets covered about 0.16 to 8.8 folds of the control. In the case of

259

SMM, overall gene expressions including PermE* / RSD were about halved. The eight

260

promoter / 5’ UTR sets covered gene expressions from about 0.10 folds to 6.4 folds of


Page 13 of 41

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


261

PermE* / RSD expression (Figure 4b). In both cases, overall gene expressions were

262

different from those in R5- media, but the eight combinations were fully functional, and

263

could express GusA in various degrees.

264

Although the GusA expression of P4505R5755 was increased by at least about 2.4

265

to 8.8 times compared to that of the control in various conditions, the fold changes in the

266

expression levels were not as high as the (RPKM * TE) values predicted. However, the

267

expressions from all the three different media are proximally close to the prediction, and

268

the eight promoter / 5’ UTR systems were shown to be functional for optimizing gene

269

expressions with varying strengths. Again, since a different gene was translated from

270

the given four 5’ UTRs, their expression levels could be altered to a certain extent. The

271

other explanation would be competitions of mRNAs for ribosomes, and resulting

272

changes in overall translatome. There could be other possible explanations such as

273

mRNAs competing for ribosomes33 or tRNA availability,34 but GusA was successfully

274

expressed with different levels following predicted orders.

275

To examine whether the promoters and 5’ UTRs were functional in other strains

276

as well, the eight plasmids carrying promoters / 5’ UTRs and gusA gene were

277

transformed into S. venezuelae and S. avermitilis. While 5’ UTRs seemed to be as

278

functional as in S. coelicolor, P4505 was not functional in both S. venezuelae (Figure S2a)

279

and S. avermitilis (Figure S2b). The result could have been due to the differences in

280

growth rates and gene regulations. As growth rates of the three Streptomyces species

281

are different greatly from one another,5 the regulations of a cold shock protein,

282

SCO4505, may be also different. The mechanisms of cold shock response are not well

283

characterized in all the three species that it is hard to draw out definite explanations.



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

284

Another possibility is that construction of expression systems using TSS information be

285

very specific to organisms owing to strain specific nature of RNA-seq, Ribo-seq, and

286

TSS-seq datasets.

Page 14 of 41

287 288

Replacements of 5’ UTR sequences from the previously reported

289

promoters, PSCO5768, PSCO4658, and PSCO3410

290

Not all promoters would be functional in different types of Streptomyces sp. or culture

291

conditions as seen with P4505 that was only functional in S. coelicolor. To examine

292

whether or not our method could be applied to solve such problems, three previously

293

published promoters were engineered to demonstrate effectiveness of gene expression

294

enhancement using our method, comparing the results to a method used to construct

295

original PSCO5768, PSCO4658, and PSCO3410.

296

PSCO5768, PSCO4658, and PSCO3410 were known to be used in various Streptomyces

297

sp.,9 but based on our RNA-seq and Ribo-seq information, their usages were also

298

limited to SMM, maltose-yeast extract-malt extract media (MYM), and tryptic soy broth

299

(TSB). Average RPKM values of SCO5768, SCO4658, and SCO3410 promoters are all

300

under 100 when grown in R5- complex media (Table S3). Compared to the gene

301

expression profiles acquired from SMM, MYM, and TSB cultures, the strengths of the

302

three SCO5768, SCO4658, and SCO3410 promoters were reduced dramatically.

303

Moreover, PSCO5768, PSCO4658, and PSCO3410 contain their own 5’ UTR nucleotide

304

sequences from -1 bp to -500 bp of each start codon. As a result, they are still under the

305

influence of original 5’ UTRs.

306

To enhance the gene expressions from the three promoters, each of the original


Page 15 of 41

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


307

5’ UTR sequences, from TSS to the 3’ end of 5’ UTRs, were replaced with R5755

308

following our method, using the TSS information. Since the three promoters are not

309

silent in the cultures from R5- complex media, 5’ UTR engineering was effective enough

310

to enhance gene expression from PSCO5768, PSCO4658, and PSCO3410. When the original 5’

311

UTRs were replaced with R5755, the expression levels of GusA from PSCO5768, PSCO4658,

312

and PSCO3410 were increased by 11.29, 8.61, and 3.97 folds, respectively (Figure 5) in

313

R5- complex media. Strengths of the SCO5768, SCO4658, and SCO3410 promoters

314

were so weak that their expression levels were not as strong as from that of the control,

315

PermERSD. However, the three promoters were successfully engineered to show more

316

effective gene expressions of interests under various culture conditions.

317 318

Applications of promoter and 5’ UTR combinations for actinorhodin

319

productions

320

To apply the promoter / 5’ UTR combinations to overproduce a polyketide secondary

321

metabolite as examples, ActII-orf4 and MetK were overexpressed with a strong,

322

P4505R5755, and a weak, P4808R3083, combinations, respectively. ActII-orf4 is a well-known

323

regulator36 showing positive effects on the production of ACT. Therefore, overexpression

324

of ActII-orf4 is expected to increase the production of ACT.

325

Overexpression of ActII-orf4 under its native promoter in a single plasmid was

326

excluded from the experiments because this study was focused on constitutive

327

promoters and 5’ UTR with constitutive TE. According to RNA-seq data, it was also

328

confirmed that promoter of ActII-orf4 is not constitutive. Transcription of ActII-orf4

329

usually started when the cell growth reached near stationary phase, suggesting that it is



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

330

Page 16 of 41

not constitutive, but only effective in stationary phase.

331

First, growths of actII-orf4 overexpressing mutants were slightly faster than that

332

of control, WT containing blank pIBR25 plasmid, but there were no overall differences

333

(Figure 6a). ACT production using P4808R3083 combination was about the same as that of

334

the control. However, in the case of P4505R5755 combination, about 2.6 folds higher ACT

335

production was observed, which corresponds to about 39% increase in ACT production

336

compared to that of ermE*P overexpressed ActII-orf4 with its own 5’UTR region (Figure

337

6b).

338

For MetK, however, fine tuning of expression optimization was required for the

339

maximum production of ACT. MetK is an enzyme that produces S-adenosylmethionine

340

(SAM) from ATP and methionine, and is an essential gene for cell growth.36 Because

341

SAM was known to activate various positive regulators of secondary metabolism in

342

many different Streptomyces sp.,37 MetK was often overexpressed. Appropriate SAM

343

concentrations for polyketide antibiotics overproduction vary according to different

344

Streptomyces strains in the ranges of 10 µM to 1 mM of external feeding.38 In the case

345

of S. coelicolor, SAM was known to up-regulate ActII-orf4, and appropriate SAM

346

concentration for the best up-regulation was about 2 µM of external treatment, which

347

was very low compared to the other Streptomyces strains.39

348

When MetK was overexpressed with either a strong or weak expression system,

349

the lag phase in growth of the weakly expressed MetK mutant using P4808R3083

350

combination was about a day longer than the other two strains, but the growth was

351

recovered near the stationary phase (Figure 6c). ACT production was the highest with

352

the weak P4808R3083 system, which is three folds increase compared to that of the control.


Page 17 of 41

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


353

In the case of the strong P4505R5755 overexpression system, ACT production was

354

increased by 1.87 folds (Figure 6d). Overexpression of MetK is different from feeding

355

SAM externally, but it is quite agreeable with the previous reports saying that weakly

356

overexpressing MetK resulted in higher ACT productions.

357

Sporulation abilities of the

two MetK overexpressing mutants were also

358

compared because the effect of MetK on delaying sporulation is also known in

359

Streptomyces strains.24, 40 When R5- complex agar media were used, the MetK mutant

360

using P4808R3083 combination started to sporulate on day two just as in WT, whereas,

361

P4505R5755-metK mutant began its sporulation around day three (Figure S3). The

362

strongly expressed MetK mutant with P4505R5755 resulted in incomplete sporulation,

363

similar to the result from previous reports.24

364 365

Conclusions

366

In summary, promoters and 5’ UTRs were screened, designed, and constructed based

367

on three deep sequencing datasets. Often there are limitations in promoter evaluation

368

and screening based only on RNA-seq data, because protein expression is also

369

dependent on TE, and there are almost no well characterized RBS or 5’ UTR

370

sequences available, especially for Streptomyce strains. Total eight promoter / 5’ UTR

371

combinations were designed where candidate genes for promoter and 5’ UTRs were

372

selected based on RNA-seq, Ribo-seq, and TSS-seq. Using TSS information, the eight

373

combinations were constructed, so that there were minimal interferences between

374

promoters and 5’ UTRs.

375

When the eight promoter / 5’ UTR sets were evaluated with GusA, GusA was



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

Page 18 of 41

376

expressed in the ranges of 0.003 GU to 0.23 GU, similar to the order of the predictions

377

made. GusA expressions from the eight combination sets covered ranges from about

378

0.03 folds to 2.4 folds of the reference, ermE*P / RSD combination, which can represent

379

most of the constitutive promoter / 5’ UTR combinations available in this species. The

380

eight combinations were also functional in two other media, YEME and SMM. These

381

expression systems were successfully applied to differentially express ActII-orf4 and

382

MetK as demonstrations of its utilities. It is unfortunate for P4808 only working in S.

383

coelicolor. But as shown with PSCO5768, PSCO4658, and PSCO3410, a working promoter from

384

literature or RNA-seq data for species of interests could be found in place of P4808.

385

It is expected that the expression system developed herein could reduce the gap

386

between genetic design with metabolic modeling41, 42 and experimental implementation,

387

providing a useful tool in metabolic engineering. This study extended and enhanced

388

previous promoter searching methods from RNA-seq, and designed promoters and 5’

389

UTRs that have minimal interferences between transcription and translations. As long

390

as there are available RNA-seq, Ribo-seq, and TSS-seq data, our method can be

391

applied in many different species. Unfortunately, one of the two promoters from S.

392

coelicolor was not fully functional in S. venezuelae and S. avermitilis, suggesting that

393

our method is still somewhat strain-specific. However, as shown with replacing 5’ UTRs

394

of PSCO5768, PSCO4658, and PSCO3410, improvements in the system would be expected

395

depending on the types of promoter used along with 5’ UTRs, as the trends of

396

translations from the four 5’ UTRs were closely maintained within the three species. It

397

would provide an insight to develop general methodology for constructions of various

398

protein expression systems in other bacterial species with simple gene cloning


Page 19 of 41

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

399


manipulations.

400 401

Methods

402

Bacterial strains, plasmids, and culture conditions

403

All the bacterial strains used, and plasmids constructed throughout the study are listed

404

in Table 3. Luria-Bertani (LB) broth (Becton Dickinson, USA) was used for E. coli strains

405

culture. R5- complex medium containing 103 g sucrose, 10.12 g MgCl26H2O, 10 g

406

glucose, 5.73 g TES buffer (pH 7.2), 5 g yeast extract, 0.25 g K2SO4, 0.1 g Difco

407

casamino acids, 7 mL of 1 N NaOH, and 1 mL of a trace element solution (10 mg

408

(NH4)6Mo7O244H2O, 20 mg CuCl22H2O, 20 mg Na2B4O710H2O, 20 mg MnCl24H2O,

409

80 mg ZnCl2, and 400 mg FeCl36H2O in 1 L distilled water) in 1 L of distilled water was

410

used for culture of Streptomyces strains. For evaluations of GusA expressed in different

411

media, YEME, containing yeast extract, malt extract, peptone, and glucose, and SMM,

412

containing glucose and casamino acid as C and N sources, were made by following

413

standard protocols.45

414

Appropriate antibiotics, 0.1 mg/mL of ampicillin, 50 µg/mL of apramycin, or 8

415

µg/mL of thiostrepton, were added to culture broths when necessary. E. coli strains were

416

cultured in 37 ºC, and Streptomyces strains were cultured in 30 ºC, both with 200 rpm

417

shaking. For Streptomyces, 150 mg of wet weighted cell, from initial 50 mL of seed

418

cultures after 24 hr, were inoculated to fresh 50 mL R5- media.

419

R5- agar media (R5- complex media with 22 g/L agar) were used for solid culture

420

and sporulation of Streptomyces strains. The cultures were carried out at 30 ºC without

421

shaking. Appropriate antibiotics were also added to the agar media when needed.



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 41

422 423

Strain construction

424

Streptomyces codon optimized E. coli originated gusA (JW1609) was PCR amplified

425

using LA-taq DNA polymerase (Takara, Japan) with XbaI cut site containing the primer 1

426

and HindIII cut site containing primer 2 as tabulated in Table S4. It was cloned into

427

XbaI/HindIII (Thermo, USA) sites of pIBR25 vector to give pIBRGUS. Since gusA

428

contained SacI cut site, it was necessary to replace SacI cut site with that of BglII to

429

replace ermE*P. pIBRGUS was PCR amplified using the primer 1 and BglII cut site

430

containing the primer 3. The PCR fragment containing pIBRGUS sequences without

431

ermE*P was ligated with fragments of target promoter and RBS combinations.

432

For a reference, ermE*P was PCR amplified with the primer 4, and SD

433

sequence containing the primer 5, and inserted into pIBRGUS to give pIBRSDGUS. To

434

construct P4505 with R5189, R3854, and R5061 in conventional method, the primers 6 and 7-

435

9 were used before GusA was inserted into the plasmid. P4505, and P4808 were PCR

436

amplified from TSS to -300 of each SCO4505 and SCO4808 with the primer 10 and 12-

437

14, and the primer 11 and 12-14, respectively. Each primer sets contained the three 5’

438

UTR sequences, R5755, R0641, and R3083, all from TSS to -1 of start codon, in reverse

439

primers. R2078 was inserted by fusion using the primer sets, 15, 16, and 5 for P4505R2078,

440

and 17-19 for P4808R2078.

441

Promoters of SCO5768, SCO4658, and SCO3410 were PCR amplified with

442

primer sets 20 and 21, 22 and 23, and 24 and 25, respectively, and the PCR amplified

443

fragments were inserted into pIBRGUS to give the three plasmids, pPSCO5768GUS,

444

pPSCO4658GUS, and pPSCO3410GUS. To construct pPSCO5768R57GUS, pPSCO4658R57GUS,


Page 21 of 41

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


445

and pPSCO3410R57GUS, SCO5768, SCO4658, and SCO3410 promoters were PCR

446

amplified using primer sets 26 and 27, 28 and 29, and 30 and 31, respectively.

447

actII-orf4 (SCO5085) and metK (SCO1476) of S. coelicolor were individually

448

cloned into XbaI/HindIII sites of the promoter sets using the primers sets 32 and 33, and

449

34 and 35, respectively. For ermE*P expressed actII-orf4, the primer 33 and 36 were

450

used.

451

E. coli DH5α was used for multiplication of plasmids, and E. coli JM110 was

452

used to acquire non-methylated DNA plasmids. Standard procedures were used to

453

introduce the non-methylated plasmids into S. coelicolor, S. venezuelae, and S.

454

avermitilis.45

455 456

RNA-seq, Ribo-seq, and TSS-seq data

457

Multi-omics datasets used throughout the study are from the previously published

458

study,25 where raw data were deposited in the Gene Expression Omnibus archive with

459

accession code GSE69350. It includes ssRNA-seq and Ribo-seq data, acquired at mid-

460

exponential phase, transition phase, late exponential phase, and stationary phase, and

461

dRNA-seq data.

462 463

GusA Activity measurements of the promoter / 5’ UTR combinations

464

On day 4 of 50 mL cell cultures, 2 mL sample was harvested for GusA activity

465

measurement. After removal of supernatants with centrifugation at 13500 rpm for 2 min,

466

cell pellets were reconstituted and underwent lysis with 1 mL of lysis buffer containing,

467

50 mM phosphate buffer with pH 7.0, 5 mM dithiothreitol (DTT), 0.1% Triton-X, and 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 22 of 41

468

mg/mL lysozyme, at 37 ºC for 15 min. 5 µL of a 0.2 M substrate, p-nitrophenyl-β-D-

469

glucuronide, was added to 0.5 mL of 8 folds diluted lysates for P4808 sets. And the other

470

0.5 mL of the lysates without the substrate was diluted by 8 folds, and control was used

471

to normalize the data. Reaction was carried out at 37 ºC for 5 min, and 0.5 mL of

472

Na2CO3 was added to terminate the reaction. UV absorbance at 415 nm was measured

473

with Multiskan spectrum (Thermo scientific, USA).46 Values of the UV absorption,

474

normalized by the controls, was divided by reaction time (min) multiplied with dried cell

475

weight of used cells (mg) to calculated Gus Unit (GU) [A415/(min·mg)]. Assay data were

476

averaged out with triplicate experimental values.

477 478

Actinorhodin measurements

479

To measure ACT production of each mutant, 1 mL of culture sample was harvested

480

every day. 4N KOH solution was added to the harvested sample, final 1N KOH

481

concentrations, followed by 200 rpm shaking at room temperature for 5 min. The

482

supernatant was acquired by centrifugation at 13500 rpm for 5 min at room temperature.

483

Using Multiskan spectrum (Thermo scientific, USA), absorbance at 640 nm of the

484

supernatant was measured.40

485 486

Supporting Information

487

Table S1: List of RPKM and TE values of the 23 target genes. Table S2: Nucleotide

488

sequences of promoters and 5’UTR. Table S3: RPKM and TE values from previously

489

reported promoters. Table S4: List of primers for promoter and 5’ UTR plasmid

490

constructions. Figure S1: Comparison of experimental TE from RNA-seq and Ribo-seq


Page 23 of 41

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


491

versus predicted translational strengths calculated by RBS calculator. Figure S2: GusA

492

activities of the eight promoter / 5’ UTR combinations in S. venezuelae and S. avermitilis.

493

Figure S3: Changes in sporulation abilities of Streptomyces coelicolor strains

494

differentially overexpressing MetK.

495 496

Author Contributions

497

J.S.Y., B.-K.C., and B.-G.K. designed the study. J.S.Y., M.W.K., Y.J., and E.K. performed

498

the experiments. J.S.Y., M.W.K., M.K., Y.J., B.-K.C., and B.-G.K. analyzed the data.

499

J.S.Y., B.-K.C., and B.-G.K. wrote the manuscript.

500 501 502

Acknowledgement This research was supported by the National Research Foundation of Korea

503

(NRF) funded by the Ministry of

504

2013R1A2A2A01069197 to B.-G.K.) and (NRF-2014K1A3A1A20034749 to B.-G.K.).

505

This work was also supported by the Intelligent Synthetic Biology Center of the Global

506

Frontier Project (2011-0031957 to B.-K.C.) and the Basic Science Research Program

507

(NRF-2015R1A2A2A01008006 to B.-K.C.) through the National Research Foundation of

508

Korea (NRF) funded by the Ministry of Science, ICT, and Future Planning (MISP).

Science, ICT & Future Planning (NRF-



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 41

509

References

510

(1) Berdy, J. (2005) Bioactive microbial metabolites, J. Antibiot. 58, 1-26.

511

(2) Pokrovskaya, V., and Baasov, T. (2010) Dual-acting hybrid antibiotics: a promising

512

strategy to combat bacterial resistance, Expert Opin. Drug Discovery 5, 883-902.

513

(3) Birch, A., Leiser, A., and Robinson, J. A. (1993) Cloning, sequencing, and expression

514

of the gene encoding methylmalonyl-coenzyme A mutase from Streptomyces

515

cinnamonensis, J. Bacteriol. 175, 3511-3519.

516

(4) Coze, F., Gilard, F., Tcherkez, G., Virolle, M. J., and Guyonvarch, A. (2013) Carbon-

517

flux distribution within Streptomyces coelicolor metabolism: a comparison between the

518

actinorhodin-producing strain M145 and its non-producing derivative M1146, PLoS One

519

8, e84151.

520

(5) Wang, W., Li, X., Wang, J., Xiang, S., Feng, X., and Yang, K. (2013) An engineered

521

strong promoter for streptomycetes, Appl. Environ. Microbiol. 79, 4484-4492.

522

(6) Bibb, M. J., Janssen, G. R., and Ward, J. M. (1985) Cloning and analysis of the

523

promoter region of the erythromycin resistance gene (ermE) of Streptomyces

524

erythraeus, Gene 38, 215-226.

525

(7) Takano, E., White, J., Thompson, C. J., and Bibb, M. J. (1995) Construction of

526

thiostrepton-inducible, high-copy-number expression vectors for use in Streptomyces

527

spp, Gene 166, 133-137.

528

(8) Murakami, T., Holt, T. G., and Thompson, C. J. (1989) Thiostrepton-induced gene

529

expression in Streptomyces lividans, J. Bacteriol. 171, 1459-1466.

530

(9) Li, S., Wang, J., Li, X., Yin, S., Wang, W., and Yang, K. (2015) Genome-wide

531

identification and evaluation of constitutive promoters in streptomycetes, Microb. Cell


Page 25 of 41

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


532

Fact. 14, 172.

533

(10) Luo, Y., Zhang, L., Barton, K. W., and Zhao, H. (2015) Systematic Identification of a

534

Panel of Strong Constitutive Promoters from Streptomyces albus, ACS Synth. Biol. 4,

535

1001-1010.

536

(11) Marioni, J. C., Mason, C. E., Mane, S. M., Stephens, M., and Gilad, Y. (2008) RNA-

537

seq: an assessment of technical reproducibility and comparison with gene expression

538

arrays, Genome Res. 18, 1509-1517.

539

(12) Yin, J., Bao, L., Tian, H., Gao, X., and Yao, W. (2015) Quantitative relationship

540

between the mRNA secondary structure of translational initiation region and the

541

expression level of heterologous protein in Escherichia coli, J. Ind. Microbiol. Biotechnol.

542

43, 97-102.

543

(13) Taylor, R. C., Webb Robertson, B. J., Markillie, L. M., Serres, M. H., Linggi, B. E.,

544

Aldrich, J. T., Hill, E. A., Romine, M. F., Lipton, M. S., and Wiley, H. S. (2013) Changes

545

in translational efficiency is a dominant regulatory mechanism in the environmental

546

response of bacteria, Integr. Biol. 5, 1393-1406.

547

(14) Baneyx, F., and Mujacic, M. (2004) Recombinant protein folding and misfolding in

548

Escherichia coli, Nat. Biotechnol. 22, 1399-1408.

549

(15) Zhu, S., Gong, C., Ren, L., Li, X., Song, D., and Zheng, G. (2013) A simple and

550

effective strategy for solving the problem of inclusion bodies in recombinant protein

551

technology: His-tag deletions enhance soluble expression, Appl. Microbiol. Biotechnol.

552

97, 837-845.

553

(16) Bukau, B., and Horwich, A. L. (1998) The Hsp70 and Hsp60 chaperone machines,

554

Cell 92, 351-366.



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 41

555

(17) Perl, A., Kless, H., Blumenthal, A., Galili, G., and Galun, E. (1992) Improvement of

556

plant regeneration and GUS expression in scutellar wheat calli by optimization of culture

557

conditions and DNA-microprojectile delivery procedures, Mol. Gen. Genet. 235, 279-284.

558

(18) Ingolia, N. T., Ghaemmaghami, S., Newman, J. R., and Weissman, J. S. (2009)

559

Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome

560

profiling, Science 324, 218-223.

561

(19) Ingolia, N. T. (2014) Ribosome profiling: new views of translation, from single

562

codons to genome scale, Nat. Rev. Genet. 15, 205-213.

563

(20) Gerashchenko, M. V., Lobanov, A. V., and Gladyshev, V. N. (2012) Genome-wide

564

ribosome profiling reveals complex translational regulation in response to oxidative

565

stress, Proc. Natl. Acad. Sci. U. S. A. 109, 17394-17399.

566

(21) Horbal, L., Fedorenko, V., and Luzhetskyy, A. (2014) Novel and tightly regulated

567

resorcinol and cumate-inducible expression systems for Streptomyces and other

568

actinobacteria, App. Microbiol. Biotechnol. 98, 8641-8655.

569

(22) Myronovskyi, M., Welle, E., Fedorenko, V., and Luzhetskyy, A. (2011) Beta-

570

glucuronidase as a sensitive and versatile reporter in actinomycetes, Appl. Environ.

571

Microbiol. 77, 5370-5383.

572

(23) Kim, D. Y., Hwang, Y. I., and Choi, S. U. (2011) Cloning of metK from Actinoplanes

573

teichomyceticus ATCC31121 and effect of its high expression on antibiotic production, J.

574

Microbiol. Biotechnol. 21, 1294-1298.

575

(24) Kim, D.-J., Huh, J.-H., Yang, Y.-Y., Kang, C.-M., Lee, I.-H., Hyun, C.-G., Hong, S.-K.,

576

and Suh, J.-W. (2003) Accumulation of S-Adenosyl-l-Methionine Enhances Production

577

of Actinorhodin but Inhibits Sporulation in Streptomyces lividans TK23, J. Bacteriol. 185,


Page 27 of 41

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


578

592-600.

579

(25) Jeong, Y., Kim, J.-N., Kim, M. W., Bucca, G., Cho, S., Yoon, Y. J., Kim, B.-G., Roe,

580

J.-H., Kim, S. C., Smith, C. P., and Cho, B.-K. (2016) The dynamic transcriptional and

581

translational landscape of the model antibiotic producer Streptomyces coelicolor A3(2),

582

Nat. Commun. 7.

583

(26) Gingold, H., and Pilpel, Y. (2011) Determinants of translation efficiency and

584

accuracy, Mol. Syst. Biol. 7, 481.

585

(27) Bashor, C. J., and Collins, J. J. (2012) Insulating gene circuits from context by RNA

586

processing, Nat. Biotechnol. 30, 1061-1062.

587

(28) Salis, H. M., Mirsky, E. A., and Voigt, C. A. (2009) Automated design of synthetic

588

ribosome binding sites to control protein expression, Nat. Biotechnol. 27, 946-950.

589

(29) Espah Borujeni, A., Channarasappa, A. S., and Salis, H. M. (2014) Translation rate

590

is controlled by coupled trade-offs between site accessibility, selective RNA unfolding

591

and sliding at upstream standby sites, Nucleic Acids Res. 42, 2646-2659.

592

(30) Seo, S. W., Yang, J. S., Kim, I., Yang, J., Min, B. E., Kim, S., and Jung, G. Y. (2013)

593

Predictive design of mRNA translation initiation region to control prokaryotic translation

594

efficiency, Metab. Eng. 15, 67-74.

595

(31) Bai, C., Zhang, Y., Zhao, X., Hu, Y., Xiang, S., Miao, J., Lou, C., and Zhang, L.

596

(2015) Exploiting a precise design of universal synthetic modular regulatory elements to

597

unlock the microbial natural products in Streptomyces, Proc. Natl. Acad. Sci. U. S. A.

598

112, 12181-12186.

599

(32) Herai, S., Hashimoto, Y., Higashibata, H., Maseda, H., Ikeda, H., Omura, S., and

600

Kobayashi, M. (2004) Hyper-inducible expression system for streptomycetes, Proc. Natl.



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 41

601

Acad. Sci. U. S. A. 101, 14031-14035.

602

(33) Mehra, A., and Hatzimanikatis, V. (2006) An algorithmic framework for genome-

603

wide modeling and analysis of translation networks, Biophys. J. 90, 1136-1146.

604

(34) Varenne, S., Buc, J., Lloubes, R., and Lazdunski, C. (1984) Translation is a non-

605

uniform process. Effect of tRNA availability on the rate of elongation of nascent

606

polypeptide chains, J. Mol. Biol. 180, 549-576.

607

(35) Arias, P., Fernandez-Moreno, M. A., and Malpartida, F. (1999) Characterization of

608

the pathway-specific positive transcriptional regulator for actinorhodin biosynthesis in

609

Streptomyces coelicolor A3(2) as a DNA-binding protein, J. Bacteriol. 181, 6958-6968.

610

(36) Wei, Y., and Newman, E. B. (2002) Studies on the role of the metK gene product of

611

Escherichia coli K-12, Mol. Microbiol. 43, 1651-1656.

612

(37) Zhao, X. Q., Gust, B., and Heide, L. (2010) S-Adenosylmethionine (SAM) and

613

antibiotic biosynthesis: effect of external addition of SAM and of overexpression of SAM

614

biosynthesis genes on novobiocin production in Streptomyces, Arch. Microbiol. 192,

615

289-297.

616

(38) Huh, J. H., Kim, D. J., Zhao, X. Q., Li, M., Jo, Y. Y., Yoon, T. M., Shin, S. K., Yong, J.

617

H., Ryu, Y. W., Yang, Y. Y., and Suh, J. W. (2004) Widespread activation of antibiotic

618

biosynthesis by S-adenosylmethionine in streptomycetes, FEMS Microbiol. Lett. 238,

619

439-447.

620

(39) Park, H. S., Shin, S. K., Yang, Y. Y., Kwon, H. J., and Suh, J. W. (2005)

621

Accumulation of S-adenosylmethionine induced oligopeptide transporters including BldK

622

to regulate differentiation events in Streptomyces coelicolor M145, FEMS Microbiol. Lett.

623

249, 199-206.


Page 29 of 41

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


624

(40) Ochi, K., and Freese, E. (1982) A decrease in S-adenosylmethionine synthetase

625

activity increases the probability of spontaneous sporulation, J. Bacteriol. 152, 400-410.

626

(41) Kim M., Yi, J. S., Kim J., Kim J.-N., Kim M. W., Kim B.-G. (2014) Reconstruction of

627

a high-quality metabolic model enables the identification of gene overexpression targets

628

for enhanced antibiotics production in Streptomyces coelicolor A3(2), Biotechnol. J. 9,

629

1185-1194.

630

(42) Kim M., Yi, J. S., Lakshmanan, M., Lee D. Y., Kim B.-G. (2016) Transcriptomics-

631

based strain optimization tool for designing secondary metabolite overproducing strains

632

of Streptomyces coelicolor, Biotechnol. Bioeng. 113(3), 651-660.

633

(43) Kim, J. N., Yi, J. S., Lee, B. R., Kim, E. J., Kim, M. W., Song, Y., Cho, B. K., and

634

Kim, B. G. (2012) A versatile PCR-based tandem epitope tagging system for

635

Streptomyces coelicolor genome, Biochem. Biophys. Res. Commun. 424, 22-27.

636

(44) Thuy, M. L., Kharel, M. K., Lamichhane, R., Lee, H. C., Suh, J. W., Liou, K., and

637

Sohng, J. K. (2005) Expression of 2-deoxy-scyllo-inosose synthase (kanA) from

638

kanamycin gene cluster in Streptomyces lividans, Biotechnol. Lett. 27, 465-470.

639

(45) Kieser T, Bibb MJ, Buttner MJ, Chater KF, Hopwood DA. (2000) Practical

640

streptomyces genetics. John Innes Foundation Norwich, Norwich, England.

641

(46) Wang, W., Ji, J., Li, X., Wang, J., Li, S., Pan, G., Fan, K., and Yang, K. (2014)

642

Angucyclines as signals modulate the behaviors of Streptomyces coelicolor, Proc. Natl.

643

Acad. Sci. U. S. A. 111, 5688-5693.



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

644

Table 1. List of 23 selected targets for promoter and 5’ UTR sequences Gene SCO0247 SCO0641 SCO1478 SCO1489 SCO1517 SCO1936 SCO2078 SCO2082 SCO2314 SCO2617 SCO3083 SCO3187 SCO3404 SCO3409 SCO4505 SCO4662 SCO4808 SCO4809 SCO5357 SCO5627 SCO5755 SCO6467 SCO6468

Hypothetical protein Tellurium resistance protein RNA polymerase DNA binding protein Putative secreted protein Transaldolase Putative mebmbrane protein Cell division protein Possible integral membrane protein ATP-dependent Clp protease Possible integral membrane protein Hypothetical protein Cell division protease Putative inorganic pyrophosphatase Cold shock protein Elongation factor Suc-CoA synthetase beta subunit Suc-CoA synthetase alpha subunit Transcription termination factor Ribosome recycling factor Possible transcriptional regulator Phosphatidylserine synthase Phosphatidylserine decarboxylase

645


Page 30 of 41

Page 31 of 41

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

646


Table 2. List of genes for the promoter / 5’ UTR constructions Gene

Average RPKM

Average TE

TSS

Promoter SCO4505

Cold shock protein, scoF2

20058

0.88

4926107

SCO4808

Succinyl-CoA synthetase beta subunit, sucC

2422

1.40

5234868

SCO5755

Possible transcriptional regulator, clgR

604

5.63

6293028

SCO0641

Tellurium resistance protein, terD

2361

1.84

681905

SCO3083

Putative integral membrane protein

2190

1.18

3377207

SCO2078

Putative membrane protein

2015

0.51

2232419

5’ UTR

Random 5’ UTR (without TSS considerations) SCO5189

Hypothetical protein

791

14.32

5647134

SCO3854

Septation inhibitor protein, crgA

343

3.49

4239443

SCO5061

GTP-dependent nucleic acid-binding protein, engD

236

1.39

Undetected

647



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

648

Page 32 of 41

Table 3. Strains and plasmids used in this study Strains or Plasmids

Descriptions

References

Streptomyces strains Streptomyces coelicolor A3 (2) M145 Streptomyces venezuelae ATCC15439 Streptomyces avermitilis ATCC31267

Wild type (WT) strain

43

WT strain

ATCC

WT strain

ATCC

E. coli strains ThermoFish er Stratagene

DH5α

A host for general gene manipulations

JM110

A strain for demethylation of plasmids

pIBR25

This work

pP45R51GUS

Amp and tsr marker; Streptomyces vector with SCP2* origin pIBR25 containing SD RBS sequence and gusA, cloned into PstI/HindIII sites pIBRGUS containing P4505R5189 in SacI/pstI sites instead of PermE*

pP45R38GUS

pIBRGUS containing P4505R3854 in SacI/pstI sites instead of PermE*

This work

pP45R50GUS

pIBRGUS containing P4505R5061 in SacI/pstI sites instead of PermE*

This work

pP45R57GUS

pIBRGUS containing P4505R5755 in BglII/pstI sites instead of PermE*

This work

pP45R06GUS


This work

pP45R30GUS


This work

pP45R20GUS


This work

pP48R57GUS


This work

pP48R06GUS


This work

pP48R30GUS


This work

pP48R20GUS


This work

pPSCO5768GUS

pIBRGUS containing PSCO5768 in BglII/pstI sites instead of PermE*

This work

Plamids

pIBRSDGUS

pPSCO5768R57GUS pPSCO4658GUS pPSCO4658R57GUS

R

44

This work

pIBRGUS containing PSCO5768R5755 in BglII/pstI sites instead of PermE*

This work


This work


This work


This work


This work

pIBR25 containing actII-orf4

This work

pP45R57actIIorf4

pIBRactIIorf4 containing P4505R5755 in BglII/XbaI sites instead of PermE*

This work

pP48R30actIIorf4

pIBRactIIorf4 containing P4808R3083 in BglII/XbaI sites instead of PermE*

This work

pIBR25 containing metK

This work

pP45R57metK

pIBRmetK containing P4505R5755 in BglII/XbaI sites instead of PermE*

This work

pP48R30metK

pIBRmetK containing P4808R3083 in BglII/XbaI sites instead of PermE*

This work

pPSCO3410GUS pPSCO3410R57GUS pIBRactIIorf4

pIBRmetK

649 650


Page 33 of 41

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


651

Figure 1. Scheme for selection of target promoter and 5’ UTR genes from RNA-seq,

652

Ribo-seq, and TSS-seq.

653 654

Figure 2. Design of promoter / 5’ UTR fragment construction. 5’ UTR fragments were

655

synthesized into reverse primers for PCR amplification of promoter targets. The PCR

656

fragments were then inserted into pIBR25 in place of ermE*P.

657 658

Figure 3. GusA evaluations from cell lysates of Streptomyces coelicolor M145 strains,

659

having PermERSD as a reference. (A) GusA assays of the reference and P4505 expression

660

with three randomly chosen, and conventionally constructed RBS sequences. (B) GusA

661

assay of the eight promoter / 5’ UTR combinations. The results are first, in order of

662

decreasing promoter strength, and second, in order or decreasing TE.

663 664

Figure 4. GusA evaluations from S. coelicolor strains carrying the eight promoter / 5’

665

UTR expression systems when they were grown in (A) YEME and (B) SMM. Expression

666

of GusA could be successfully manipulated both in YEME media and SMM, just as in

667

that from R5- complex media.

668 669

Figure 5. Evaluations of the original PSCO5768, PSCO4658, and PSCO3410, (blue bars) and

670

PSCO5768, PSCO4658, and PSCO3410 containing 5’ UTR of R5755 (yellow bars).

671 672

Figure 6. Growth and actinorhodin production curve. For all cases, Streptomyces

673

coelicolor M145 with empty pIBR25 plasmid was used as a reference. (A) Growth and



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 41

674

(B) actinorhodin productions from actII-orf4 overexpressing Streptomyces coelicolor

675

M145 mutants. (C) Growth and (D) actinorhodin productions of metK overexpressing

676

Streptomyces coelicolor M145 mutants.


Page 35 of 41

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


677 678

Figure 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

679 680

Figure 2.


Page 36 of 41

Page 37 of 41

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


681

682 683

Figure 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

684

685 686

Figure 4.

687


Page 38 of 41

Page 39 of 41

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


688 689

Figure 5.

690



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

691

692


Page 40 of 41

Page 41 of 41

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


693

694 695 696

Figure 6.


A Novel Approach for Gene Expression Optimization through Native

Recommend Documents