Experimentally Validated Model Enables Debottlenecking of in Vitro

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

Experimentally validated model enables debottlenecking of in vitro protein synthesis and identifies a control shift under in vivo conditions Alexander Nieß, Jurek Failmezger, Maike Kuschel, Martin Siemann-Herzberg, and Ralf Takors ACS Synth. Biol., Just Accepted Manuscript • Publication Date (Web): 19 Jun 2017 Downloaded from http://pubs.acs.org on June 20, 2017

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 30

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

Title:

2

Experimentally validated model enables debottlenecking of in

3

vitro protein synthesis and identifies a control shift under in vivo

4

conditions

5 6

Authors:

7

Alexander Nieß12; Jurek Failmezger1,2; Maike Kuschel1; Martin Siemann-Herzberg1; Ralf

8

Takors1*

9 10

1

11

* For correspondence:

12

[email protected]; Tel.: (+49) 711 685 64535; Fax: (+49) 711 685 65164

13

2

14

Jurek Failmezger performed the experiments.

Affiliation: Institute of Biochemical Engineering, University of Stuttgart, Stuttgart

Authors contributed equally to this work. Alexander Nieß performed the modeling, and

15 16

Keywords: Ribosomes, Translation Factors, Translation rate, cell-free synthetic biology

17

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

18

1

19

Cell-free (in vitro) protein systems (CFPS) provide a versatile tool that can be used to

20

investigate different aspects of the transcription-translation machinery by reducing cells to the

21

basic functions of protein formation. Recent improvements in reaction stability and lysate

22

preparation offer the potential to expand the scope of in vitro biosynthesis from a research

23

tool to a multifunctional and versatile platform for protein production and synthetic biology.

24

To date, even the best-performing CFPS are drastically slower than in vivo references. Major

25

limitations are imposed by ribosomal activities that progress in an order of magnitude slower

26

on the mRNA template. Owing to the complex nature of the ribosomal machinery,

27

conventional ‘trial and error’ experiments only provide little insights into how the desired

28

performance could be improved. By applying a DNA-sequence-oriented mechanistic model,

29

we analyzed the major differences between cell-free in vitro and in vivo protein synthesis. We

30

successfully identified major limiting elements of in vitro translation, namely the supply of

31

ternary complexes consisting of EFTu and tRNA. Additionally, we showed that diluted in

32

vitro systems suffer from reduced ribosome numbers. Based on our model, we propose a new

33

experimental design predicting 90% increased translation rates, which were well achieved in

34

experiments. Furthermore, we identified a shifting control in the translation rate, which is

35

characterized by availability of the ternary complex under in vitro conditions and the initiation

36

of translation in a living cell. Accordingly, the model can successfully be applied to

37

sensitivity analyses and experimental design.

ABSTRACT

38 39

2

40

Cell-free protein synthesis (CFPS) is a powerful technique for the fast production of

41

recombinant proteins. In contrast to traditional in vivo protein formation, CFPS offers

INTRODUCTION


Page 2 of 30

Page 3 of 30

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


42

advantages such as expression of genes encoding growth-inhibiting proteins, high-throughput

43

protein production and easy access for investigating translation conditions (1–3). Recent

44

studies have demonstrated the potential of CFPS as platform technique for genetic circuit

45

engineering (4).

46

Whereas early CFPS batch reactions suffer from short activity periods of less than 1 hour, the

47

implementation of sophisticated energy regeneration systems has allowed active protein

48

formation up to 10 hours (5), reaching final product titers of several mg/mL. Nevertheless,

49

CFPS systems achieve maximum synthesis rates of only 0.4-1 mg/mL/h (6–11), which lags

50

behind in vivo protein production rates by a factor of several hundred (12). To compete with

51

in vivo protein production, the duration of the in vitro reaction and the synthesis rates must

52

both be improved, and key factors that limit the in vitro space-time yield must be identified.

53

According to current opinion, in vivo protein synthesis is thought to be limited by translation

54

initiation rate (13,14), but it is not clear whether this is also true for in vitro protein synthesis.

55

Research groups have attempted to identify key steps that control in vitro conditions (15). For

56

example, Freischmidt et al. (16) investigated the impact of adding translation factors,

57

aminoacyl-tRNA synthetases and active ribosomes, on the total amount of protein produced.

58

Surprisingly, this did not improve the final protein concentration, and the authors concluded

59

that stoichiometry of the translation factors was optimal in their system. However, putative

60

impacts on translation rate (i.e. molar volumetric protein synthesis rate) were not addressed.

61

A detailed analysis of CFPS was performed by Underwood et al. (12). Combining polysome

62

profile analysis with protein quantification allowed researchers to quantitatively analyze the

63

cell-free system. Among others, Underwood et al. (12) clearly identified that a highly reduced

64

ribosome concentration, in combination with a low specific elongation rate of around 1.5

65

amino acids per ribosome per second, heavily influenced the overall in vitro translation rate.

66

In fact, the in vitro translation rate was estimated to be more than three orders of magnitude



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

67

lower than the in vivo translation rate under fast growth conditions. The discrepancy between

68

in vitro and in vivo elongation rates could partly be overcome by supplementing the cell-free

69

system with additional elongation factors, which increased the elongation rate by 33%. It

70

appeared that the limitation of protein growth shifted from initiation to elongation under in

71

vivo and in vitro conditions, indicating that translation is limited by the elongation rate of

72

ribosomes. There are two major mechanisms involved in elongation. The first is peptide chain

73

elongation, which is controlled by ternary complexes consisting of charged tRNA, GTP, and

74

elongation factor Tu (EFTu), and the second is translocation, which is catalyzed by elongation

75

factor G (EFG).

76

As depicted in Figure 1, the translation machinery comprises a highly complex, synergistic

77

system that consists of a large number of components. As a result, its optimization requires

78

sophisticated tools to properly investigate the parameter space. Arnold et al. (17) established a

79

mathematical model that combines all steps of transcription and translation, such as mRNA

80

synthesis and degradation, ribosomal translation, energy regeneration and inactivation kinetics

81

of certain proteins, to simulate a CFPS system. A DNA sequence oriented approach enabled

82

estimation of crucial parameters such as reaction rates, which are not readily accessible

83

experimentally. Our approach builds on previous work by performing a thorough modeling

84

and experimental study aimed at identifying key control elements of in vitro protein synthesis

85

and defining ways in which the system can be improved.


Page 4 of 30

Page 5 of 30

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


86 87

Figure 1: Overview of cell-free protein synthesis and reaction composition of S30 extract and added supplements. An additional schematic

88

overview of the model mechanisms is shown, starting with T7 RNA polymerase and subsequent sequence oriented translation as

89

well as regeneration of co-factors. Abbreviations: RNAP = RNA polymerase; IF = initiation factor; EFTu = elongation factor Tu;

90

EFTs = elongation factor Ts; EFG = elongation factor G; RF = releasing factor; NTP = nucleotide triphosphate; GTP = guanosine

91

triphosphate; GDP = guanosine diphosphate.

92 93

3

94

3.1

95

Routinely performed batch-CFPS reactions in 96-well microtiter plates always showed a three

96

phase pattern, as depicted in Figure 2. First, the transcription-translation machinery was

97

initiated and began producing mRNAs and proteins, and characteristic increases in translation

98

rate were observed. Then, a pseudo steady-state was achieved, as defined by constant reaction

99

rates. The length of phase two is dependent on the supply of substrates. Finally, phase three

RESULTS AND DISCUSSION MODEL VALIDATION



100

comprised a decrease in translation rate due to limiting components and decomposition of the

101

original enzymes.

6

10

4

8

6

4

2

Protein synthesis rate | nM sec-1

12

eGFP concentration | µM

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 30

2

0 0

25

50

125

0 150

time | min

102 103 104

Figure 2: Time courses of protein concentration (biological triplicates) and corresponding error bands. Protein synthesis rate, derived from eGFP concentration via central differences, is also shown.

105

All three phases were considered for the translation analysis. An increase in GFP

106

concentration in the observation intervals was calculated with central differences at each time

107

point (shown in Equation 1). Maximum translation rates were determined for phase two and

108

were compared to the simulation results.

( ) =

( ) − ( ) −

1

109 110

With a complete set of parameters and the experimentally determined in vitro concentrations,

111

it was possible to predict translation rates for the given cell-free reaction system (Table 1).

112

Our in silico predictions were consistent with the experimental observations of Underwood et

113

al. (12), although the latter studied the translation rate of chloramphenicol acetyltransferase


Page 7 of 30

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


114

(CAT), which has a molecular weight close to that of eGFP. On the other hand, only 61% of

115

the predicted 90% increase in translation rate were measured experimentally, even though the

116

simulations and experiments were focused on the same protein sequence.

117

Despite the fact that eGFP has the advantage of being able to utilize online monitoring via

118

fluorescence measurement, several studies identified a drawback in that only a fraction of the

119

synthesized eGFP is correctly folded and fluorescently active. Examples are given by

120

Caschera et al. (18) and Iskakova (19), who measured only 80% and 50% fluorescently active

121

GFP, respectively. Further sophisticated efforts such as the analysis of

122

incorporation could determine the degree of active eGFP. However likewise studies were not

123

performed as key parameters affecting eGFP folding (such as the temperature (19)) remained

124

constant. Consequently, a stable ratio of active to non-active eGFP was assumed for all

125

expression rates, which most likely underestimated the measured translation rates.

126

Systems modeling is always a trade-off between the detailed reflection of reality and the

127

availability of proper data sets for configuring the theoretical set-up. Because likewise

128

parameters were not available, the regeneration of tRNA is treated equally for all tRNA

129

species disregarding putative interactions with aminoacyl tRNA synthases. Moreover, the

130

presented model describes the transcription and translation of only one single mRNA species

131

per time. Another characteristic feature is its hybrid structure that is given by a black box

132

model for transcription and a detailed structured model for translation. As such, structural

133

limitations of translation are highlighted which enables the identification of key rate

134

controlling elements of the total processing from transcription to translation. Accordingly, the

135

modelling approach differed from published black box models by Stögbauer et al. (20) and

136

Chizzolini et al. (21). Notably, the model offers the opportunity for simulating in vitro and in

137

vivo conditions, for instance by including mechanisms like mRNA secondary structure.


14

C-lysine


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

Page 8 of 30

138

Summarizing, we conclude that the presented model fairly represents the mechanisms of in

139

vitro translation.

140

Table 1: Comparison of experimentally determined parameters of in vivo and in vitro translation, and simulated in vitro reaction parameters.

141

In vivo data taken from Bremer and Dennis (22)a and Underwood et al. (12)b.

in vivo Ribosome concentration (µM)

42a

Amount of actively translating ribosomes (%) Elongation rate (AA/sec)

a

Initiation rate (nM/sec) Bulk protein synthesis rate (mg/L/min)

80

18a

in vitro Underwood et al. 1.6 ± 0.1

in silico prediction

This study

1.7

1.7 ± 0.1

72 ± 4

82

n.d.

1.5 ± 0.2

1.12

n.d.

2.8·10

3b

8.2 ± 0.3

7.1

4.3 ± 1.07

3.3·10

3b

12.3 ± 0.5

10.6

6.5 ± 1.6

142 143

3.2

144

To identify possible rate-limiting factors in in vitro translation systems, we altered several

145

components of the translation machinery in silico and simulated the respective impact on

146

translation dynamics (Figure 3). Specifically, different concentrations were tested for (i) all

147

initiation factors (simultaneously) and selectively for (ii) EFTu, (iii) EFG, (iv) EFTs, (v)

148

tRNA, (vi) ribosome releasing factor (RF), (vii) T7 RNA polymerase and (viii) ribosomes. All

149

other input parameters and reaction conditions remained constant. For comparison, the

150

calculated translation rates were normalized to the reference condition with unchanged

151

parameter settings.

MODEL BASED IDENTIFICATION OF LIMITING FACTORS


Page 9 of 30

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


IF 1.25

Ribosomes

EFTu 1.00

0.75

0.50

T7

RF

152

EFTs

EFG tRNA

153

Figure 3: Sensitivity analysis of translation factors regarding translation rate. Values are derived from an in silico 50% increase in

154

concentration of each translation factor. Resulting translation rate is normalized to the non-altered rate to allow easy comparison.

155

Factors considered are initiation factors 1-3 (IF); elongation factors G (EFG), Ts (EFTs), and Tu (EFTu); ribosome releasing

156

factor (RF); T7 RNA polymerase; and tRNA. A value of one indicates no influence of this factor on the translation rate

157

Surprisingly, predicted translation rates persisted when the pool size of initiation factors,

158

EFG, EFTs, and RF increased. These findings suggest that none of the factors limited

159

translation rates under the conditions tested. In fact, increasing ribosome numbers can even

160

decrease the translation rate, as shown in Figure 3. To be precise, raising the total number of

161

ribosomes by 50% caused 53% increase of actively translating ribosomes which underlines

162

that each ‘added’ ribosome is directly incorporated in the translation machinery. This finding

163

is consistent with previous experimental observations by Freischmidt et al. (16), who detected

164

a decline in translation rate with increasing ribosome density. The more the ribosomes are

165

active, the more tRNA is bound, which decreases the amount of tRNA that can form novel

166

ternary complexes. Ternary complex concentrations of tRNA species (according to the



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

167

nomenclature of Dong et al. (24)) Ala2, Asp1, Phe, His, Lys, Asn, Pro2, Gln2, Ser3, Ser5 and

168

Thr13 are decreased by 10-40% and their corresponding codons are required by more than

169

50% of the GFP sequence. Our modeling results show that the majority of ribosomes slow

170

down translation when repetitive codons occur (e.g. His tag). Under such conditions, each

171

ribosome is bound to a specific tRNA that is also required in the next elongation step.

172

Accordingly, the specific tRNA becomes limiting, which decreases the elongation rate. This

173

slowdown will also lead to ribosomal stalling.

174

Moreover, our simulation revealed a slight decrease in translation performance due to

175

increased RNA polymerase levels. The explanation for this follows a similar rationale to that

176

previously described for ribosomes. An increase in mRNA concentration leads to an increase

177

in the number of actively translating ribosomes and a lower ribosome density on the mRNA,

178

which causes a shortage of free tRNA. In this case, the amount of freely accessible ternary

179

complexes for the species Ala2, Ser5 and His decrease by more than 20 % compared to

180

reference levels.

181

Interestingly, translation rates only increased with rising levels of tRNA or EFTu. The critical

182

role of EFTu is emphasized by the fact that EFTu is the most abundant protein in E. coli (23).

183

In vivo EFTu concentrations correlate with the concentration of total aminoacyl-tRNA (24).

184

GTP binding confers EFTu affinity to charged tRNA, the substrate for translation elongation.

185

Accordingly, the stimulating effects of EFTu and tRNA on translation basically mirror the

186

lack of substrate for protein formation.

187

As a consequence of this, there are indications that the ternary complex consisting of EFTu,

188

tRNA, and GTP is rate-limiting in cell-free translation. To investigate the synergistic impacts

189

of tRNA and EFTu, we analyzed the combinatorial effect of tRNA and EFTu on translation

190

rate. As depicted in Figure 4A, the combination of tRNA and EFTu considerably increased

191

the simulated translation rate, compared to increasing either EFTu or tRNA alone. This result


Page 10 of 30

Page 11 of 30

192

supports the hypothesis that availability of the ternary complex exerts a key level of control

193

on translation speed. Figure 4D shows that the rise in translation rate slows down as the

194

availability of ternary complexes increases. As the translation rate is a function of the amount

195

of actively translating ribosomes and the elongation rate, it is likely that at least one of these

196

will limit the process. Further studies into mean elongation rate (Fig. 4C) and amount of

197

actively translating ribosomes (Fig. 4B) revealed a contradictory trend for the number of

198

actively translating ribosomes. Whereas translation rate and specific elongation rate steadily

199

increased with tRNA and EFTu concentration, the amount of translating ribosomes decreased,

200

and as a consequence the translation rate leveled off. These findings suggest that as specific

201

elongation rates increase, a threshold is reached where initiation becomes limited. In other

202

words, as the supply of tRNA and EFTu improves, the more accurately in vitro translation

203

reflects the in vivo conditions.

204 Translation rate vTL

B

3.0 1.0 1.5 2.0 2.5

2.5

2.0

1.5

normalized EFTu concentration

normalized EFTu concentration

A

1.0

active ribosomes 3.0 0.70 0.85 0.90 0.95

2.5

2.0

1.5

1.0 1.0

1.5

2.0

2.5

3.0

1.0

normalized tRNA concentration

C

2.0

2.5

3.0

D

mean elongation rate velo

2.5

2.0

1.5

value compared to reference

3.5 1.0 1.5 2.0 2.5

1.0

translation rate mean elongation rate active ribosomes

3.0 2.5 2.0 1.5 1.0 0.5

1.0

205

1.5


3.0 normalized EFTu concentration

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


1.5

2.0

2.5


3.0

1.0

1.5

2.0

2.5

normalized ternary complex concentration


3.0


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

206

Figure 4: Influence of EFTu and tRNA concentration (normalized to reference values) and resulting normalized translation rate (A),

207

normalized amount of active ribosomes (B) (calculated as sum of actively translating ribosomes divided by overall number of 30S/50S

208

subunits) and normalized elongation rate (C) (average value over all codons). All values were normalized to values at unaltered conditions.

209

Part D shows the influence of coupled EFTu and tRNA multipliers and the resulting translation rate (solid line), mean elongation rate (dotted

210

line) and amount of active ribosomes (dashed line).

211

3.3

212

To verify that ternary complex concentration does in fact limit in vitro translation rates, we

213

doubled the EFTu and tRNA concentration by supplementing with purified components.

214

Because a ribosome concentration of 1.45 µM was detected in the samples, 8 µM EFTu and

215

13 µM tRNA were added to double the respective concentrations. An equal volume of protein

216

storage buffer was added to the reference samples. Maximum protein synthesis rates were

217

measured under both conditions. For the given experimental setup, the model predicted an

218

89% increase in the translation rate. The underlying translation system shows a 117% increase

219

in elongation rate, but the amount of actively translating ribosomes decreased by 12%.

220

Experimental measurements revealed a 62 ± 9% rise in the maximum translation rate

221

compared to the reference (see Appendix figure 1).

222

Considering the inherent drawbacks of experimental detection, we conclude that the

223

simulation results are a good prediction of the experimental readout. Thus, our hypothesis that

224

the ternary complex is a key rate-limiting target was confirmed.

225

3.4

226

Figure 4D shows that an increase in ternary complex concentration causes a proportional rise

227

in elongation rate, but reduces the rise in translation rate. Furthermore, falling fractions of

228

actively translating ribosomes were observed, which indicates the increasing impact of

229

limiting initiation. Furthermore, the influence of involved reactants was investigated using

230

metabolic control analysis. Because details of tRNA recharging via amino acylation could not

INCREASING TERNARY COMPLEX CONCENTRATIONS IN CFPS REACTIONS

TRANSLATIONAL CONTROL


Page 12 of 30

Page 13 of 30

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


231

be properly parameterized in the model, tRNA regeneration was finally regarded as not rate

232

limiting.

233

We evaluated (i) , (ii) and (iii) , which represent sensitivities with respect to

234

varying ternary complex (i), initiation factors (ii) and ribosome concentrations (iii) on the

235

translation rate at a given CT3, respectively. Variations in these impact factors were calculated

236

by estimating the partial differentials with central differences (Fig. 5). Under standard in vitro

237

conditions (normalized concentration =1), the influence of ribosomes and initiation factors

238

turned out to be negligible ( = = 0). Neither changes in ribosomes nor variations in

239

the initiation factors could improve the given translation rate. However, the translation rate

240

could be increased significantly by increasing , which is indicated by an elasticity of

241

1.

242

The stimulating impact of on translation rate diminished with increasing levels. As

243

indicated in Figure 5, control of the translation process is taken over by the amount of actively

244

translating ribosomes, and this number can be increased by two means: raising either the

245

concentration of ribosomes ( ) or the concentration of all initiation factors ( ). When

246

neither ribosomes nor initiation factors are increased, the number of actively translating

247

ribosomes becomes reduced as the ternary complex concentration increases, as depicted in

248

Figure 4D.

249

Figure 5 outlines the differences between in vitro and in vivo protein production. The first is

250

given by the standard conditions, which may benefit from the addition of ternary complexes.

251

The second, in vivo conditions, is not affected by a limiting ternary complex supply, which

252

identifies availability of initiation factors and ribosome number as key control elements of

253

translation rate.



254

Compared to the previous model (17) our hybrid approach reduced model complexity and in

255

turn accelerated computational speed. Accordingly, the investigation of elasticities as a key

256

issue of metabolic control analysis was enabled.

FCC

1.0

εIF

0.8

0.6

0.4

εT3

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

Page 14 of 30

0.2

0.0 1.0

1.5

2.0

2.5

3.0

CT3

257 258

Figure 5: Calculated elasticities and and flux control coefficient (FCC) as a function of increasing normalized ternary complex

259

concentrations. Elasticities and and FCC represent the sensitivity of translation rate with respect to varying total ternary

260

complex, initiation factor and total ribosome concentration

261

3.5

262

Owing to the fact that the translation machinery is derived from viable E. coli cells, in vivo

263

measurements of translation rates show strong potential for CFPS applications. Using the

264

dataset published by Bremer and Dennis (22), a translation rate for eGFP of almost 150 µM

265

per minute was estimated as theoretical upper limit for undiluted CFPS extracts (Equation 2).

266

However, the in vivo translation rate comprises the parallel processing of multiple different

267

mRNA species, which implies the concomitant use of a variety of codon sequences, whereas

268

CFPS process only one mRNA sequence. This effect leads to one-sided use of tRNA species

ESTIMATION OF CFPS POTENTIAL


Page 15 of 30

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


269

in CFPS, causing limited availability of distinct ternary complexes. Accordingly, the observed

270

elongation rates in CFPS are inherently limited compared to in vivo conditions. = ̅ ! "##,

271

2

with: = 33.85 )* (25)

272

̅ ! = 18 ,, -./01023 1 (22)

273

"##, = 245 ,, 367

274 275

Neglecting CFPS dilution (8 = 1), i.e. assuming in vivo-like conditions, and assuming the

276

translation rate of a single mRNA only, we found that the measured translation rates of

277

Bremer and Dennis (22) could not be reached. Instead, we found either a low number of

278

active ribosomes with a high elongation rate (Table 2, low TC), or a high number of active

279

ribosomes with a low elongation rate (Table 2, high TC). Apparently, the ‘implementation’ of

280

in vivo concentrations had increased the active ribosome levels so extensively that the

281

availability of ternary complexes becomes limiting and therefore the elongation rate decreases

282

(see chapter 3.2, increased T7 concentrations). This emphasizes the close interaction between

283

the supply of active ribosomes and ternary complexes, which creates shifting control of

284

translation under different conditions. The finding is also supported by previous studies (17)

285

who estimated the ratio of actively translating to total ribosomes to be below 5% which

286

imposed an artificial control on initiation and not on tRNA supply as it is the case in this

287

study.

288



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

289

Page 16 of 30

Table 2: Translation rates under in vivo conditions. Optimal conditions show values from the literature (22). Simulations of in vivo conditions

290

ignore dilution and investigate the impacts of ternary complexes (TC) as indicted.

9:;< == >?@

Experimentally Validated Model Enables Debottlenecking of in Vitro

Recommend Documents