Subscriber access provided by READING UNIV
Article
Reprogramming one-carbon metabolic pathways to decouple Lserine catabolism from cell growth in Corynebacterium glutamicum Yun Zhang, Xiuling Shang, Shujuan Lai, Yu Zhang, Qitiao Hu, Xin Chai, Bo Wang, Shuwen Liu, and Tingyi Wen ACS Synth. Biol., Just Accepted Manuscript • DOI: 10.1021/acssynbio.7b00373 • Publication Date (Web): 09 Jan 2018 Downloaded from http://pubs.acs.org on January 11, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS 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 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
1
Reprogramming one-carbon metabolic pathways to decouple L-serine catabolism
2
from cell growth in Corynebacterium glutamicum
3 4
Yun Zhang†, Xiuling Shang†, Shujuan Lai†, Yu Zhang†,‡, Qitiao Hu†,‡, Xin Chai†, Bo
5
Wang†,‡, Shuwen Liu†, Tingyi Wen†,§,*
6 7
†CAS
8
Microbiology, Chinese Academy of Sciences, Beijing 100101, China
9
‡University
Key Laboratory of Pathogenic Microbiology and Immunology, Institute of
of Chinese Academy of Sciences, Beijing 100049, China
10
§Savaid
11
China
12
*Corresponding author: Institute of Microbiology, Chinese Academy of Sciences, 1
13
Beichen West Road, Chaoyang District, Beijing 100101, China.
14
E-mail address:
[email protected] (T. Wen).
15
Phone: +86 10 64806119. Fax: +86 10 64806157.
Medical School, University of Chinese Academy of Sciences, Beijing 100049,
16 17
Graphical Table of Contents
18 19 20 1
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
21 22
Page 2 of 43
ABSTACT L-Serine,
the principal one-carbon source for DNA biosynthesis, is difficult for
23
microorganisms to accumulate due to the coupling of
24
microbial growth. Here, we reprogrammed the one-carbon unit metabolic pathways in
25
Corynebacterium glutamicum to decouple L-serine catabolism from cell growth. In
26
silico model-based simulation showed a negative influence on glyA-encoding serine
27
hydroxymethyltransferase flux with
28
transcription resulted in increased L-serine accumulation, and a decrease in purine
29
pools, poor growth and longer cell shapes. The gcvTHP-encoded glycine cleavage
30
(Gcv) system from Escherichia coli was introduced into C. glutamicum, allowing
31
glycine-derived
32
resulted in an increased amount of one-carbon units. Gcv introduction not only
33
restored cell viability and morphology but also increased
34
Moreover, comparative proteomic analysis indicated that abundance changes of the
35
enzymes involved in one-carbon unit cycles might be responsible for maintaining
36
one-carbon unit homeostasis. Reprogramming of the one-carbon metabolic pathways
37
allowed cells to reach a comparable growth rate to accumulate 13.21 g/L L-serine by
38
fed-batch fermentation in minimal medium. This novel strategy provides new insights
39
into the regulation of cellular properties and essential metabolite accumulation by
40
introducing an extrinsic pathway.
13
L-serine
L-serine
catabolism and
productivity. Attenuation of glyA
CH2 to be assimilated into intracellular purine synthesis, which
41 42 2
ACS Paragon Plus Environment
L-serine
accumulation.
Page 3 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
43
KEYWORDS:
Corynebacterium
glutamicum,
L-serine,
serine
44
hydroxymethyltransferase, glycine cleavage system, one-carbon units, in
45
silico model-based simulation
46
3
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
47 48
Page 4 of 43
INTRODUCTION L-Serine
is a building block for the biosynthesis of amino acids, such as glycine,
49
cysteine and tryptophan,1-3 a predominant source of one-carbon (C1) units in de novo
50
biosynthesis of purines and thymidylate and an intermediate in phospholipid
51
biosynthesis.4 Due to its importance in cell growth and division in prokaryotes and
52
eukaryotes,5 L-serine is currently used in the pharmaceutical and cosmetic industries.6
53
Industrial production of L-serine depends mainly on enzymatic or cellular conversion
54
from glycine plus formaldehyde or methanol using serine hydroxymethyltransferase
55
(SHMT).7, 8 In spite of a very high theoretical yield from glucose to L-serine, it is
56
challenging for direct fermentation to produce
57
indispensability of L-serine as an essential intermediate for cell growth.
L-serine
from glucose due to the
58
Many attempts have been made to construct L-serine-producing strains derived
59
from Escherichia coli or Corynebacterium glutamicum. The traditional approach of
60
releasing the feedback inhibition of 3-phosphoglycerate dehydrogenase and
61
co-expressing L-serine synthetic genes (serA, serC, and serB) has only led to marginal
62
accumulations of
63
mechanism for the catabolism of
64
reaction
65
5,10-methylenetetrahydrofolate (5,10-methylene-THF) from 5,6,7,8-tetrahydrofolate
66
(THF). When glyA and the sdaA, sdaB and tdcG genes encoding L-serine deaminase
67
were simultaneously knocked out to block the formation of one-carbon units, an E.
68
coli strain exhibited poor cell growth.11,
that
L-serine
converts
due to the catabolism of
L-serine
L-serine
to
L-serine.
9, 10
One important
is the glyA-encoded SHMT-catalyzed
glycine
12
and
simultaneously
generates
Physiological observation showed
4
ACS Paragon Plus Environment
Page 5 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
69
abnormal shapes and sizes of E. coli ∆sdaA∆sdaB∆tdcG cells, which was attributed to
70
the inhibition of peptidoglycan synthesis and cell division due to
71
accumulation, which triggered the starvation of one-carbon units.13,
72
illustrated why serine, even at low concentrations, is toxic to E. coli, representing a
73
bottleneck in
74
bottleneck, random mutagenesis or adaptive laboratory evolution (ALE) has been
75
implemented to increase the tolerance of E. coli towards L-serine for improving
76
L-serine
L-serine
14
L-serine
This result
production by microbial fermentation. To overcome this
production.12, 15
77
In C. glutamicum, the glyA-encoded SHMT-catalyzed reaction is the only way to
78
generate 5, 10-methylene-THF.117 The transcription of glyA is activated in the
79
stationary phase by two regulators, RamB and GlyR, to supply one-carbon units for
80
cell metabolism.16 The presence of the glyA gene only led to a marginal accumulation
81
of L-serine even when serACB was overexpressed.10 In spite of the feasibility of
82
deleting the glyA gene in E. coli,12 inactivation of SHMT makes C. glutamicum
83
folate-auxotrophic, resulting in high L-serine productivity combined with reduced
84
growth at low folate concentrations.18 Strong inhibition of glyA mRNA expression
85
triggers spontaneous mutations in the genome and genetic instability in C. glutamicum,
86
which is unfavorable for industrial fermentation.10,18 To control SHMT activity by
87
coenzyme THF availability, the pabAB and pabC genes of THF synthesis were
88
deleted, resulting in dose-dependent L-serine accumulation in response to an external
89
folate supply.18 However, cell growth was significantly impaired with an increase in
90
L-serine
accumulation.18, 19 Therefore, L-serine accumulation is bound to impair cell 5
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
91 92
Page 6 of 43
growth due to interference with one-carbon unit metabolism. L-serine-derived
5, 10-methylene-THF is an important form of THF derivatives in
93
one-carbon transfer reactions. Conversion between the forms of THF derivatives
94
drives
95
10-Methylene-THF can be oxidized to form 5, 10-methyl-THF via the enzymatic
96
action of FolD and can continue to be hydrolyzed to 10-formyl-THF, or can be
97
converted
98
2’-deoxythymidine-5’-phosphate synthesis.20
99
10-methylene-THF in C. glutamicum; however, glycine is also a donor for 5,
100
10-methylene-THF generation through the glycine cleavage system in other bacteria,
101
such as E. coli and Bacillus subtilis. The glycine cleavage system (Gcv) is composed
102
of four proteins, GcvT, GcvH, GcvP and GcvL, which catalyze the oxidized cleavage
103
of glycine to CO2, ammonia, and a methylene group that is accepted by THF.21 The
104
first reaction of glycine cleavage is the decarboxylation catalyzed by GcvP to convert
105
the carboxyl carbon of glycine to CO2 and the remaining aminomethyl moiety is
106
transferred to the lipoyl prosthetic group of GcvH. Then, the aminomethyl moiety
107
attached to GcvH is subjected to GcvT-catalyzed degradation by transferring the
108
methyl group to THF for the synthesis of 5, 10-methylene-THF, which releases
109
ammonia.22 The reduced lipoic group of GcvH is reoxidated by GcvL, which
110
functions as a component of 2-oxoacid (e.g., pyruvate, 2-oxoglutarate) dehydrogenase
111
multi-enzyme, to form the disulfide.21 However, the gcv operon from E. coli contains
112
no L-protein gene, and no specific GcvL for the Gcv system has been found. In E. coli,
one-carbon
to
unit
cycles
dihydrofolate
in
cellular
to
metabolism
donate L-Serine
6
ACS Paragon Plus Environment
methyl
(Figure
1a).
groups
5,
for
is the only donor for 5,
Page 7 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
113
the Gcv system provides only a secondary mechanism for 5, 10-methylene-THF
114
formation under tight regulation to maintain a homeostatic balance between glycine
115
and one-carbon units.23, 24 Therefore, L-serine and glycine availability is essential for
116
the metabolism of one-carbon units and cell viability.
117
In our previous study, the correlation between the transient dynamics of L-serine
118
and cell growth demonstrated that L-serine pool management is fundamental for
119
sustaining the viability of C. glutamicum.4 Interestingly, enhancement of extracellular
120
glycine indicated that redundant glycine was exported to maintain intracellular
121
homeostasis. In this study, to overcome the bottleneck between L-serine accumulation
122
and one-carbon unit generation for cell growth, we present a novel engineering
123
strategy for recycling glycine by introducing Gcv systems from E. coli into C.
124
glutamicum to reprogram the one-carbon unit metabolic pathways. This strategy
125
efficiently partially decoupled cell growth with L-serine catabolism and improved
126
L-serine
accumulation.
127 128
RESULTS
129
Metabolic profile prediction of SHMT flux disturbance by iCW773
130
The relationship between SHMT flux and biomass as well as L-serine productivity
131
was investigated by constraint-based metabolic flux balance analysis using a new
132
genome-scale metabolic network model, iCW773.25 Variations in SHMT flux
133
exhibited a prominent influence on L-serine productivity. A high SHMT flux was
134
properly adjusted to supply 5, 10-methylene-THF for maximal biomass synthesis, 7
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 43
135
leading to marginal L-serine productivity (Figure 1a). Conversely, maximum L-serine
136
productivity by in silico prediction resulted in a SHMT flux close to zero (Figure 1a),
137
indicating a negative influence of SHMT flux on L-serine productivity. A low SHMT
138
flux leads to moderate biomass synthesis and
139
Furthermore, decreased SHMT flux also led to decreases in pentose phosphate
140
pathway (PPP) and tricarboxylic acid cycle (TCA) fluxes (Figure 1b). 5,
141
10-methylene-THF generated from SHMT plays a critical role in the one-carbon unit
142
cycle, which generates 10-formyl-THF for purine synthesis (Figure 1a). Therefore, a
143
low flux through the SHMT node triggered the inefficient formation of 5,
144
10-methylene-THF, resulting in decreased flux towards the purine synthesis pathway
145
(Figure 1b). The growth rate simultaneously reduced with a decrease in SHMT flux.
146
This result suggested that SHMT had a significant impact on growth rate by
147
controlling flow toward the purine biosynthetic pathway, indicating that reasonable
148
control of SHMT flux is necessary for L-serine productivity and cellular growth.
149
Attenuation of SHMT flux by replacing the promoter of glyA
L-serine
productivity (Figure 1a).
150
Guided by model-predicted metabolic flux distributions, attenuation of SHMT flux
151
was implemented by regulator depletion and promoter replacement. In our previous
152
study, deficiency of GlyR removed the transcriptional activation of the glyA gene in
153
strain SER-3 (∆sdaA∆glyRserAr).4 Based on SER-3, cysE (encoding serine
154
O-acetyltransferase) was deleted to prevent the conversion of
155
O-acetyl-L-serine,1 resulting in strain SER-9 (Table 1). Compared to the wild-type
156
strain, the mRNA level of glyA and specific enzymatic activity of SHMT decreased by 8
ACS Paragon Plus Environment
L-serine
to
Page 9 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
157
30% and 48% in glyR-deficient SER-9, respectively (Figure 2a). To further attenuate
158
the transcription of glyA, the weak promoter Phom was used as a replacement for the
159
native promoter of glyA.26 The resultant strain (SER-10) showed 50% and 48%
160
decreases in the transcriptional level of the glyA gene and the specific enzymatic
161
activity of SHMT, respectively (Figure 2a). When 24 mg/L cysteine was added for
162
shake flask cultivation,
163
SER-10 (Table S1). In contrast, the specific growth rate and maximum cell dry weight
164
of SER-10 decreased by 17% and 11% compared to those of SER-9, respectively
165
(Table S1), which was consistent with the model-predicted change in growth rate in
166
response to SHMT flux variation (Figure 1).
167
Introduction of GcvTHP to recycle glycine for one-carbon unit generation
168
L-serine
accumulation significantly improved 2-fold in
To investigate whether an increase in one-carbon unit formation could restore
169
growth
rate,
three
170
10-methylene-THF generation from E. coli were added to iCW773 to simulate
171
intracellular metabolic flux distributions. A significantly increased flux toward the
172
purine synthesis pathway was observed with a simultaneous increase in Gcv flux and
173
decrease in SHMT flux (Figure 1b). As expected, the growth rate also improved with
174
an increase in Gcv flux. Due to the deficiency of gcvTHP homologs in C.
175
glutamicum,27,
176
glutamicum chromosome.
28
gcvTHP-catalyzed
reactions
for
glycine-derived
5,
we introduced the gcvTHP system from E. coli into the C.
177
GcvTHP genes, which act as one transcriptional unit in E. coli, were designed in
178
two independent expression cassettes for integration into the SER-10 genome (Figure 9
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 43
179
2b). The co-transcribed gcvTH gene region under the control of a moderate P45
180
promoter was inserted into the deleted glyR region to construct the resultant SER-11
181
strain (Figure 2b). In the same manner, a P45-gcvP expression cassette was inserted
182
into the downstream region of the gcvTH genes to construct the resultant SER-12
183
strain (Figure 2b). Protein bands at the corresponding molecular weights of GcvT,
184
GcvH and GcvP proteins were detected by SDS-PAGE. Western blotting analysis
185
demonstrated the expressions of GcvT, GcvH and GcvP proteins in the crude cell
186
extract of SER-12 (Figure 2c). To validate that the Gcv system incorporates glycine-derived one-carbon units into
187
13
188
intracellular metabolites,
C-glycine-labeling tracing was performed by cultivating
189
SER-12 in CGX medium supplemented with [2-13C] glycine or unlabeled glycine. In
190
HPLC-Q-TOF-MS analysis, total ion chromatograms of the eluted components in the
191
intracellular extracts were obtained (Figure 3a and 3b). Purines and their derivatives,
192
adenine, guanine, adenosine monophosphate (AMP), and guanosine monophosphate
193
(GMP), were detected at the corresponding retention times (Figure 3c and 3e). In the
194
13
195
different
196
incorporation of [2-13C] glycine into intracellular purine derivatives (Figure 3d and
197
3f). The mass isotopomer distributions of AMP in the 13C-labeled cell extract showed
198
the unlabeled isotopomer at ~ 50%, whereas the remaining isotopomer represented
199
incorporation of [2-13C] glycine-derived one-carbon units. In contrast, the non-labeled
200
cell extract exhibited the natural distribution of corresponding isotopes from the
C-labeled cell extract, the labeling fractions in the m + 1 mass isotopomers of purine
derivatives
were
dominantly
detected,
10
ACS Paragon Plus Environment
demonstrating
the
Page 11 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
201
natural presence of 13C (1%). The same distribution patterns of the mass isotopomers
202
of adenine, guanine and GMP were observed for AMP, suggesting that
203
GcvTHP-mediated one-carbon unit generation from glycine plays the same role as
204
GlyA-mediated one-carbon unit generation from L-serine for purine synthesis.
205
Shake flask fermentation showed that the
L-serine
accumulation of SER-12
206
improved from 0.70 ± 0.11 mM to 1.21 ± 0.15 mM. Compared to SER-10, SER-12
207
exhibited 15% increased specific growth rate and 8% increased biomass (Table S1),
208
which validated the simulated effect of Gcv introduction on growth rate by iCW773.
209
Cellular growth and metabolite profiles in response to reprogramming
210
one-carbon metabolic pathways
211
To increase the flux towards L-serine synthesis, pWYE1151, in which the artificial
212
assembled operon consisting of serArBC genes under the control of Ptac was induced
213
to express feedback-resistant SerAr, SerB and SerC, was transformed into SER-9,
214
SER-10 and SER-12 to construct the corresponding plasmid-carrying strains.
215
Fermentation profiles showed the effects of reprogramming one-carbon metabolic
216
pathways on
217
attenuated-transcriptional glyA gene) exhibited a 41% increase in
218
accumulation, accompanied by a 26% decrease in specific growth rate and an 11%
219
reduction in biomass compared to the basic strain SER-13. SER-15 (GcvTHP
220
introduction) showed a specific growth rate and biomass that were nearly comparable
221
to strain SER-13 (Figure 4a). Remarkably, the final L-serine accumulation of strain
222
SER-15 significantly increased by 42% and 100% compared to those of strains
L-serine
accumulation and cell growth (Table S1). SER-14 (the
11
ACS Paragon Plus Environment
L-serine
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
223
SER-14 and SER-13, respectively (Figure 4b). Notably, SER-15 showed a 33%
224
increase in the yield of L-serine to glucose (Table S1). The assay of extracellular
225
byproducts showed that glycine concentrations remarkably decreased in strain
226
SER-15 (Figure 4c). The major byproducts were pyruvate-derived L-alanine and
227
L-valine,
which is similar to previously reported results.29
228
Differences were noted in intracellular metabolites related to one-carbon units.
229
SER-14 exhibited a 95% increase in its intracellular L-serine pool and a 32% decrease
230
in its glycine pool in response to the downregulated expression of glyA (Figure 4d). In
231
addition, the intracellular pools of adenine and guanine decreased by 20% and 35%,
232
respectively (Figure 4d). This result demonstrated that the reduction of
233
L-serine-derived
234
consistent with the model-predicted decrease in the purine synthetic pathway flux. By
235
introducing the Gcv system into SER-14, the L-serine and glycine pools decreased by
236
37% and 28%, whereas the adenine and guanine pools improved 1.58- and 1.69-fold
237
(Figure 4d), respectively, demonstrating that glycine supplied one-carbon units for the
238
synthesis of purine nucleotides.
239
Cell morphology restoration by improving the supply of glycine-derived one-
240
carbon units
one-carbon units interfered with the synthesis of purines, which was
241
Because L-serine is an intermediate in the biosynthesis of phospholipids, which
242
are one of the components of membranes, morphological observation was performed
243
to investigate whether one-carbon unit disturbances affect cell shape. Microscopic
244
analysis and cell size determination of C. glutamicum showed that the deletion of 12
ACS Paragon Plus Environment
Page 12 of 43
Page 13 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
245
sdaA (the only L-serine deaminase) had a negligible effect on cell length and shape
246
compared to the wild-type strain (Figure 5a). In contrast, reprogramming of the
247
one-carbon metabolic pathways affected cell morphology (Figure 5b). When glyA
248
transcription was attenuated, the average cell size of SER-14 was 18% longer than
249
that of SER-13. The average cell size of SER-15 decreased by 11% compared to
250
SER-14 and was comparable to that of SER-13 (Figure 5b), indicating that GcvTHP
251
introduction could restore cell length when its variation was triggered by attenuating
252
glyA transcription.
253
Changes in enzymes involved in one-carbon unit cycle are responsible for the
254
homeostasis of one-carbon units
255
The total soluble proteins of strains SER-13, SER-14 and SER-15 were separated
256
by two-dimensional electrophoresis to investigate proteome responses to one-carbon
257
metabolic disturbances. The proteome profiles exhibited considerable similarity in
258
terms of the quantity and distribution of proteins (Figure S1). The expression
259
abundances of twenty-one proteins involved in different cellular processes changed
260
(Table S2). In the protein profile of SER-14, down-regulation of GlyA expression was
261
consistent with the decrease in glyA mRNA expression and specific activity of GlyA
262
compared with SER-13 (Figure 2a). Remarkable differences in the abundances of
263
proteins involved in one-carbon unit metabolism were observed (Figure 6). The folD
264
gene, encoding bifunctional 5, 10-methylene-tetrahydrofolate dehydrogenase/5,
265
10-methylene-tetrahydrofolate cyclohydrolase increased 1.88-fold. Moreover, AICAR
266
transformylase/IMP cyclohydrolase (PurH), which catalyzes AICAR to form 13
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 43
267
intermediates of purine synthesis using 10-formyl-THF as a donor of formyl group,
268
also improved 1.82-fold. The abundance of SAM-dependent methyltransferase
269
(MTase) increased 1.55-fold, indicating that cells might strengthen the methyl donors
270
derived from SAM to regulate one-carbon unit metabolism. The decreased abundance
271
of bifunctional PyrR suggested that pyrimidine synthesis might be repressed to a
272
certain extent to deal with decreased one-carbon pools. Significantly, the abundances
273
of FolD, PurH and MTase in strain SER-15 were much lower than those in SER-14
274
and were comparable to those in SER-13 (Figure 6). In addition, the elongation
275
factors TufP and NusA and single-strand DNA-binding protein Ssp also exhibited
276
reduced abundances in response to glyA down-regulation, which could be related to
277
the retarded cell proliferation of SER-14 (Table S2). The increased abundances of
278
these proteins in SER-15 might contribute to increases in cell viability. Interestingly,
279
one-carbon unit disturbance incurred a response similar to the stress response, which
280
triggered increased abundances of catalase and Hsp70.
281
The effect of reprogramming
282
accumulation in fed-batch fermentation
one-carbon unit pathways
on
L-serine
283
Due to the significant increase in L-serine accumulation by Gcv introduction,
284
production performance of the SER-15 strain was investigated in a 7.5-L bioreactor
285
with minimal medium containing 40 g/L glucose and 24 mg/L cysteine. As shown in
286
Figure 7, the maximal specific growth rate of SER-15 was 0.22 h-1, and the maximal
287
cell dry weight reached 22.68 gCDW/L. The specific glucose consumption rate of
288
SER-15 was 3.48 mmol/gCDW/h. L-serine accumulated at a marginal level in the 14
ACS Paragon Plus Environment
Page 15 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
289
exponential growth phase due to the requirement of
L-serine-derived
5,
290
10-methylene-THF for growth, but its level rapidly increased in the stationary phase.
291
The glycine concentration was maintained at a relatively low level during the whole
292
fermentation. SER-15 produced up to 13.21 g/L L-serine with a yield of 0.22 mol/mol
293
glucose and a productivity of 0.24 g/L/h (Figure 7). The fermentation performance of
294
SER-15 was much better than that of SER-8 in the previous report,4 demonstrating
295
that reprogramming of the one-carbon metabolic pathways can improve
296
production by partially decoupling L-serine catabolism from cell growth.
L-serine
297 298
DISCUSSION
299
One-carbon units are incorporated into the 2nd and 8th positions of the purine ring
300
for DNA and RNA biosynthesis, and are required for the conversion of homocysteine
301
to methionine, which is the precursor of s-adenosylmethionine for the methylation of
302
DNA, RNA, proteins, and lipids.5 Because L-serine is a principal source of one-carbon
303
units, SHMT-catalyzed L-serine catabolism occupied a key position in central carbon
304
metabolism, and followed a complicated regulatory mode, such as PurR-mediated
305
repression and MetR-mediated activation in E. coli and RamB- and GlyR-mediated
306
activation in C. glutamicum.16, 30 SHMT was an abundant protein and distributed 60%
307
of the metabolic flux over L-serine in C. glutamicum;17 therefore, SHMT flux was a
308
key factor in controlling L-serine accumulation. Deletion of GlyR merely decreased
309
the flux towards
310
resulted in increased L-serine accumulation and a decrease in the intracellular purine
L-serine
to glycine by 24%.4 Attenuation of glyA transcription
15
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 43
311
pool, which agreed with the model-simulated downward flux towards the purine
312
synthesis pathway. Decreases in the specific growth rate and biomass were consistent
313
with model-simulation and were observed in a previous report.10 Similarly, the
314
engineered E. coli strain exhibited obvious
315
growth.11,
316
growth.
12
Taken together,
L-serine
L-serine
accumulation and poor cell
accumulation occurs at the expense of cell
317
The export of redundant glycine inspired us to introduce Gcv systems, which
318
enable glycine to be recycled to generate 5, 10-methylene-THF, into C. glutamicum.
319
As expected, one-carbon units derived from 13C-labeled glycine could be assimilated
320
in purine synthesis, resulting in an increase in intracellular purine pools. This result
321
was consistent with the model-predicted increase in the flux towards purine synthesis.
322
Moreover, Gcv introduction had an obvious impact on L-serine accumulation and
323
biomass. There are several explanations for the increases in L-serine accumulation and
324
biomass. One possibility is that the cleavage of glycine mediated by the Gcv system
325
in SER-15 increases the amount of 5, 10-methylene-THF compared to SER-14, which
326
triggers the cellular response to variations in one-carbon unit concentrations. It is well
327
known that L-serine is a physiological substrate for SHMT in the THF-dependent
328
reaction, and the regulation of SHMT activity depends on the concentrations of two
329
substrates (L-seine and THF) and two products (glycine and 5, 10-methylene-THF).31,
330
32
331
amount of 5, 10-methylene-THF to a certain degree. The cleavage of glycine leads to
332
a large pool of 5, 10-methylene-THF, which might feedback inhibit SHMT activity
We can assume that feedback regulation for SHMT activity is dependent on the
16
ACS Paragon Plus Environment
Page 17 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
333
and give rise to a decrease in the amount of L-serine cleavage. On the other hand, an
334
increasing amount of 5, 10-methylene-THF might decrease the demand for the
335
cleavage of L-serine to generate 5, 10-methylene-THF. We assume that an unknown
336
regulatory effect would play a role to enable cellular export of L-serine to decrease the
337
intracellular pool of
338
decrease the reaction rate of SHMT to maintain metabolic homeostasis of one-carbon
339
units. Therefore, a decrease in L-serine cleavage and an increase in L-serine export
340
result in an increase in the yield of L-serine to glucose. Similarly, the increase in the
341
yield of L-serine to biomass can be explained in the same way. The recycling of
342
glycine elevates the capacity of one-carbon unit synthesis to meet the demand of cell
343
growth, which enables cell accumulation of L-serine instead of cleavage of L-serine
344
and boosts the yield of L-serine to biomass. Consequently, the recycling of glycine
345
partially decoupled cell growth from
346
improving L-serine accumulation and cell growth at the same time.
L-serine.
The decline in the concentration of substrate will
L-serine
catabolism, which contributed to
352
Morphology analysis showed the changes in cell shapes in response to one-carbon
353
unit disturbance. Attenuation of glyA transcription together with the deletion of sdaA
354
led to an increase in the intracellular L-serine pool and longer cell shape. Similarly,
355
the inactivation of three
356
accumulation and abnormally shaped and sized E. coli.14 On the basis of the decrease
357
in the intracellular purine pool, it could be deduced that intracellular
358
accumulation might trigger the deficiency of one-carbon units, which has been
359
demonstrated to be the primary cause of defects in cell shape by inhibiting
L-serine
deaminases resulted in intracellular
17
ACS Paragon Plus Environment
L-serine
L-serine
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 18 of 43
360
peptidoglycan synthesis and cell division in E. coli.13,
14
361
system into C. glutamicum restored cell shape and size, which is consistent with
362
previous observations in the E. coli triple mutant.13 This result demonstrated that the
363
cleavage of glycine could restore the supply of one-carbon units to maintain the
364
overall shapes of cells. Alternatively, the unusual sizes and shapes of triple mutant E.
365
coli could be recovered by exogenous supplementation of SAM.13 Therefore, both
366
L-serine-derived
367
generate sufficient one-carbon units to maintain cell morphology.
Introduction of the gcv
and glycine-derived one-carbon pathways make C. glutamicum
368
Interestingly, comparative proteomic analysis indicated self-regulation of the
369
enzymes involved in the one-carbon unit cycle in response to one-carbon metabolic
370
disturbance in C. glutamicum. When the attenuation of glyA expression led to the
371
insufficient supply of 5, 10-methylene-THF, the upregulation of FolD and PurH might
372
enhance the reciprocal interconversion of THF derivatives to maintain the one-carbon
373
unit cycle. When glycine could be used as a donor to generate 5, 10-methylene-THF,
374
the abundances of FolD and PurH were regulated to their original levels. One-carbon
375
unit metabolism was interconnected with purine anabolism through the intermediate
376
10-formyl-THF, which acts as a donor of formyl groups.20 Recently, ZTP/THF
377
riboswitch, an ancient riboswitch class in bacteria, has been found to be directly
378
involved in de novo purine synthesis and the generation of one-carbon units
379
associated with folate coenzymes.33 Our results indicated that C. glutamicum might
380
respond specifically to changing levels of THF derivatives through potential
381
riboswitches to regulate the expression of FolD and PurH to maintain metabolic 18
ACS Paragon Plus Environment
Page 19 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
382
homeostasis of one-carbon units.
383 384
MATERIALS AND METHODS
385
Bacterial strains, plasmids, and cell growth. All bacterial strains and plasmids
386
used in this study are listed in Table 1. C. glutamicum ATCC 13032∆sdaA∆glyRserAr
387
(SER-3) was used as the parent strain for strain engineering.4 E. coli strain DH5α was
388
used as host cells for plasmid construction. pK18mobsacB and pXMJ19 were used for
389
genetic disruption and expression as described previously.34, 35 E. coli strains were
390
cultured in Luria-Bertani medium at 37 ºC. C. glutamicum strains were cultured in
391
brain heart infusion medium at 30 ºC. If necessary, antibiotics were added at the
392
following concentrations: 50 µg/mL kanamycin or 20 µg/mL chloramphenicol for E.
393
coli and 25 µg/mL kanamycin and 10 µg/mL chloramphenicol for C. glutamicum.
394
Construction of plasmids and strains. General DNA manipulations were
395
performed as described previously. Primers used for gene amplification and plasmid
396
construction are listed in Table S3. The pK18mobsacB derivatives that were used for
397
promoter replacement, gene deletion and chromosomal insertion were constructed in
398
E. coli (Table 1) and transformed into C. glutamicum by electroporation. For serC,
399
serB and serAr overexpression, pWYE1116 (carrying serAr) was digested with
400
PstI/XbaI, and the serC and serB-containing insert was ligated into PstI/XbaI-treated
401
pWYE1116
402
isopropylthio-β-d-galactopyranoside (IPTG) was added to the culture media to induce
403
target gene expression.
404
to
construct
pWYE1151.
When
needed,
1
mM
A markerless homologous recombination system carrying the sacB gene was used 19
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 43
405
for promoter replacement, gene disruption and chromosomal insertion of the E. coli
406
gcvTHP genes to construct engineered C. glutamicum strains.35 The upstream and
407
downstream homologous fragments of cysE were amplified using PCR with primers
408
WZ618/WZ619 and WZ620/WZ621 (Table 1). The amplified DNA fragments were
409
spliced
410
pK18mobsacB-∆cysE was transformed into SER-3 by electroporation to generate
411
cysE-deficient SER-9. For promoter replacement of the glyA gene with the weak Phom
412
promoter, a fusing fragment with the upstream region of the glyA gene, the Phom
413
promoter and the glyA gene were obtained by overlap PCR and ligated to
414
pK18mobsacB. The resulting plasmid was transformed into C. glutamicum cells by
415
electroporation.35,
416
confirmation of the chromosomal deletion were performed as described previously.35
417
Promoter replacement was verified by DNA sequencing to obtain the SER-10 strain.
418
For chromosomal insertion of the E. coli gcvTHP genes, the strong P45 promoter was
419
used to control the constitutive expression of the gcvTHP genes, which were
420
amplified using the E. coli genome as the template. To construct SER-11, the plasmid
421
pK18mobsacB-P45-gcvTH was used to insert the gcvTH genes into the glyR deletion
422
gene
423
pK18mobsacB-P45-gcvP was used to insert the gcvP genes after the integration site of
424
the gcvTH genes, resulting in SER-12 (Figure 2).
425 426
using
position
overlap
36
extension
PCR
and
ligated
into
pK18mobsacB.
Screening for the first and second recombination events and
of
SER-10
via
two
recombination
events.
Then,
Flux balance analysis. A new genome-scale metabolic network model of C. glutamicum ATCC13032
25
was used to perform all computations based on the 20
ACS Paragon Plus Environment
Page 21 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
427
COBRA toolbox v2.05 within MATLAB 2010b using the GLPK solver.37, 38 For FBA
428
analysis, simulation constraints followed the uptake rate of glucose with 4 .0
429
mmol/gDCW/h and free transport of external metabolites, such as CO2, H2O, SO3,
430
NH3, and PO4. Simulation of the impact of SHMT flux on biomass and cellular
431
metabolic flux distributions was performed with decreasing relative SHMT fluxes
432
from 1.0 to 0.2 by flux balance analysis.
433
SHMT activity assay. The cells of C. glutamicum were collected by centrifugation
434
and were resuspended with 0.02% CTAB. SHMT activity was measured by
435
β-phenylserine assay.39 The assay mixture contained 50 mM D,L-β-phenylserine, 50
436
µM PLP, 1 mM Na2EDTA, 25 mM sodium sulfate, 50 mM Hepes buffer (pH 7.5) and
437
the crude enzyme in a total volume of 0.5 mL. The rate of the appearance of the
438
product, benzaldehyde, was measured at 25 °C at 279 nm employing a molar
439
absorption coefficient of 1400 1/M/cm. D, L-β-phenylserine was used as the substrate
440
and the production of benzaldehyde was followed by its strong UV absorbance at 279
441
nm.
442
Real-time reverse transcription-PCR. The engineered C. glutamicum strains
443
were grown to the late exponential phase in CGXII minimal medium. Total RNA was
444
isolated by RNAprep Pure Cell/Bacteria Kit (Tiangen, China). Reverse transcription
445
of approximately 200 ng of RNA was performed with the specific primers listed in
446
Table 1 using the FastQuant RT Kit (Tiangen, China). Quantitative PCR was
447
performed with BRYT Green from the GoTaq qPCR master mix (Promega, USA) in a
448
20-µL mixture with a LightCycler® 96 Real-Time PCR System (Roche, Switzerland). 21
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
449
The RpoB gene of C. glutamicum was used as the reference gene to normalize the
450
glyA mRNA level. Negative controls were used in each PCR run to exclude DNA and
451
other contamination. The qPCR products were verified by melting curve analysis.
452
Data collection and analysis were facilitated using LightCycler® 96 software (Roche,
453
Switzerland), according to the 2-∆∆CT method.40
454
Western blotting. The crude proteins from C. glutamicum SER-10, SER-11 and
455
SER-12 were extracted using a lysis buffer (50 mM Tris-HCl (pH 7.5), 1 mM EDTA,
456
5% glycerol, 1 mM DTT) by ultrasonication. The supernatants were collected by
457
centrifugation and analyzed by SDS-polyacrylamide gel electrophoresis. Specific
458
detection of the GcvT, GcvH and GcvP proteins was performed by Western blotting.
459
Proteins were electrophoretically transferred to a PVDF membrane and probed with
460
three rabbit polyclonal antibodies that were raised against the purified proteins of
461
GcvT, GcvH and GcvP, respectively. The blots were visualized with a
462
peroxidase-coupled goat anti-rabbit secondary antibody and an ECL color
463
development reagent (GE, USA).
464
13
C-labeling glycine tracing. [2-13C] glycine (99% enriched) from Sigma-Aldrich 13
465
Co. LLC was used for
C-labeling tracing. C. glutamicum stains were grown
466
aerobically to the exponential growth phase in minimal media CGX containing 40 g/L
467
glucose and 24 mg/L cysteine with the addition of [2-13C] glycine (100 mg/L) and
468
unlabeled glycine. Extraction was performed by mixing preheated perchloric acid
469
(PCA, 55 °C, final concentration of 500 mM) with the sample, followed by incubation
470
at 55 °C for 10 min.41 The supernatants were collected by centrifugation (10 min, 4°C, 22
ACS Paragon Plus Environment
Page 22 of 43
Page 23 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
471
8,000 x g) and stored at -20 °C .
472
High-performance liquid chromatography-quadrupole time-of-flight/mass
473
spectrometry (HPLC-Q-TOF-MS) assay. Samples were analyzed on an Agilent
474
1260 series binary HPLC system (Agilent Technologies, Waldbronn, Germany)
475
coupled with an Agilent 6520 Accurate-Mass quadrupole time-of-flight tandem mass
476
spectrometry equipped with an electrospray ionization (ESI) source. Approximately
477
10 µL of the standard or sample solution were injected by the autosampler to the
478
Agilent 1260 HPLC using a Zorbax SB-Aq column (4.6 mm×250 mm, 5 µm; Agilent)
479
with the column temperature kept at 40 ºC. The metabolites were separated by
480
isocratic elution in the mobile phase of 0.1% formic acid in water (eluent A) and
481
methanol (95%/5%, v/v) at a constant flow rate of 500 µL/min. MS data were
482
collected in positive ionization mode with nitrogen supplied as the nebulizing and
483
drying gas. Full-scan spectra were acquired over a scan range of m/z 50-1000 at 1.03
484
spectra/s. Data were acquired and evaluated via MassHunter Workstation software.
485
Optical analysis of cells and cell size determination. C. glutamicum strains were
486
grown at 30 °C in LB medium, and cells were harvested in the exponential phase.
487
After washing the cells two times with PBS buffer, the cell suspension was diluted
488
into 0.2 of OD600 on glass slides. Microscopic investigations and cell size
489
determinations were performed with a Zeiss Axio Imager A1 Microscope equipped
490
with a transmitted light illuminator and a high-resolution Zeiss Axio Cam camera.
491
Imaging was acquired with Zeiss Axio Vision imaging software (Carl Zeiss
492
MicroImaging GmbH, Jena, Germany). Cell sizes were determined with a 23
ACS Paragon Plus Environment
ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
493
computationally efficient image analysis tool MicrobeJ on the platform ImageJ.42
494
Comparative proteomic analysis. Cells were harvested at the late exponential
495
phase by centrifugation for 10 min at 8,000 × g. The pellets were resuspended in lysis
496
buffer (8 M Urea, 4% CHAPS and 2% Bio-Lyte) and were broken by sonication on
497
ice. The supernatant was collected and protein concentration was determined
498
according to the method described by Bradford (Bradford, 1976). An aliquot of
499
supernatant containing 800 µg of proteins was analyzed by 2-DE. Approximately 450
500
µL of rehydration solution (8M urea, 2% CHAPS, 65 mM DTT, 2% IPG buffer (pH
501
4-7)) containing the sample were loaded onto IPG strips (pH 4-7; 24 cm in length; GE
502
Healthcare, USA) by active rehydration. IEF was performed in Ettan IPGphor III (GE
503
Healthcare, USA). After the equilibration of IPG strips, the second dimension of
504
SDS-PAGE was performed on a 13% polyacrylamide gel using the Ettan DALT six
505
Electrophoresis instrument (GE Healthcare, USA). The gels were stained with
506
Coomassie Blue R-250 (BioRad, Hercules, CA, USA) and scanned using an EPSON
507
Expression 11000XL at a resolution of 300 dpi. The samples from three biological
508
replicates were run in parallel. Image analysis was performed with ImageMasterTM 2D
509
Platinum 7.0 (GE Healthcare, USA). To compare proteome profiles of SER-13,
510
SER-14 and SER-15 strains, gel images in triplicates from each strain were grouped
511
into a class. Ratio was defined as the fold change of protein in quantity between two
512
classes to evaluate protein expression variations. The highest spot value of a protein in
513
one class was divided by the lowest spot value of the same protein in the other class to
514
calculate the ratio of each protein between two classes. The comparison between 24
ACS Paragon Plus Environment
Page 24 of 43
Page 25 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Synthetic Biology
515
two-class data was performed by ANOVA with a two-tailed Student’s t-test to
516
evaluate the significantly differential proteins (P