Novel T7-like expression systems used for Halomonas

Abstract

To engineer non-model organisms, suitable genetic parts must be available. However, biological parts are often host strain sensitive. It is therefore necessary to develop genetic parts that are functional regardless of host strains. Here we report several novel phage-derived expression systems used for transcriptional control in non-model bacteria. Novel T7-like RNA polymerase-promoter pairs were obtained by mining phage genomes, followed by in vivo characterization in non-model strains Halomonas spp TD01 and Pseudomonas entomophila. Three expression systems, namely, MmP1, VP4, and K1F, were developed displaying orthogonality (crosstalk<0.7%), tight regulation (3085-fold induction), and high efficiency (2.5-fold of Ptac) in Halomonas sp. TD01, a chassis strain with a high industrial value. The expression under the corresponding T7-like promoter libraries persisted with striking correlations (R² >0.94) between Escherichia coli and Halomonas sp. TD01, implying suitability of broad-host range. Three Halomonas TD strains were then constructed based upon these expression systems that enabled interchangeable and controllable gene expression. One of the strains termed Halomonas TD-MmP1 was used to express the cell-elongation cassette (minCD genes) and polyhydroxybutyrate (PHB) biosynthetic pathway, resulting in a 100-fold increase in cell lengths and high levels of PHB production (up to 92% of cell dry weight), respectively. We envision these T7-like expression systems to benefit metabolic engineering in other non-model organisms.

Full text

Original Research Article
Novel T7-like expres
Q2
sion systems used for Halomonas
Han Zhao
a,1
, Haoqian M. Zhang
b,c,1
, Xiangbin Chen
a
, Teng Li
a,c
, Qiong Wu
a
, Qi Ouyang
b,
n
,
Guo-Qiang Chen
a,
n
Q1
a
Center for Synthetic and Systems Biology, School of Life Sciences, Tsinghua-Peking Center for Life Sciences, Tsinghua University, Beijing 100084, China
b
Peking-Tsinghua Joint Centre for Life Sciences, Center for Quantitative Biology, Peking University, Beijing 100871, China
c
BluePHA Co., Ltd., Beijing 100084, China
article info
Article history:
Received 17 September 2016
Received in revised form
21 October 2016
Accepted 21 November 2016
Keywords:
Part mining
Expression system
Morphological engineering
PHB
Synthetic biology
Non-model microorganism
abstract
To engineer non-model organisms, suitable genetic parts must be available. However, biological parts are
often host strain sensitive. It is therefore necessary to develop genetic parts that are functional regardless
of host strains. Here we report several novel phage-derived expression systems used for transcriptional
control in non-model bacteria. Novel T7-like RNA polymerase-promoter pairs were obtained by mining
phage genomes, followed by in vivo characterization in non-model strains Halomonas spp TD01 and
Pseudomonas entomophila. Three expression systems, namely, MmP1, VP4, and K1F, were developed
displaying orthogonality (crosstalko0.7%), tight regulation (3085-fold induction), and high efficiency
(2.5-fold of P
tac
)inHalomonas sp. TD01, a chassis strain with a high industrial value. The expression
under the corresponding T7-like promoter libraries persisted with striking correlations (R
2
40.94) be-
tween Escherichia coli and Halomonas sp. TD01, implying suitability of broad-host range. Three Halomonas
TD strains were then constructed based upon these expression systems that enabled interchangeable and
controllable gene expression. One of the strains termed Halomonas TD-MmP1 was used to express the
cell-elongation cassette (minCD genes) and polyhydroxybutyrate (PHB) biosynthetic pathway, resulting
in a 100-fold increase in cell lengths and high levels of PHB production (up to 92% of cell dry weight),
respectively. We envision these T7-like expression systems to benefit metabolic engineering in other
non-model organisms.
&2016 International Metabolic Engineering Society. Published by Elsevier Inc.
1. Introduction
Engineering of microorganisms has enabled the production of
various chemicals including platform molecules (Steen et al., 2008;
Yang et al., 2016a), amino acids (Yu et al., 2015), biofuels (Lee et al.,
2015), pharmaceuticals (Breitling and Takano, 2015) and bioma-
terials (Yang et al., 2016b). Synthetic biology, as a new metho-
dology for microorganism engineering (Breitling and Takano,
2016;Jensen and Keasling, 2014;Stephanopoulos, 2012), allows
controls of biological processes in a more efficient, reliable, and
predictable manner (Brophy and Voigt, 2014;Gu et al., 2016;
Rajkumar et al., 2016;Venturelli et al., 2016). Advances in en-
gineering genetic control circuits, especially those with intriguing
spatial and temporal control functions, have been made in several
model organisms especially Escherichia coli (Bonnet et al., 2012;
Daniel et al., 2013;Ke et al., 2016;Kotula et al., 2014;Nielsen et al.,
2016;Wang et al., 2014). However, in metabolic engineering
practices, non-model organisms are often used (Hammar et al.,
2015;Tran and Charles, 2016); these include soil bacteria in the
plant rhizosphere, probiotics in the human gut, and industrial
bacterial strains in open or closed bioreactors (Cardinale and Ar-
kin, 2012;Venturelli et al., 2016).
Non-model bacteria usually preserve some unique pathways
for their unique products. Their values could be further improved
by introducing new pathways and adding new genetic control (Li
et al., 2016a). However, when a control circuit is transferred from
one species/strain to another, substantial efforts are usually re-
quired to rebuild the circuit to preserve its functions, and un-
expected failures often occur (Cardinale and Arkin, 2012;Ne-
vozhay et al., 2013;Prindle et al., 2012;Venturelli et al., 2016). This
is largely because biological parts in a circuit are often influenced
by host strains, and altered activity of biological parts frequently
causes failures in process control (Brophy and Voigt, 2014;Kit-
tleson et al., 2012). For example, when a well-built transcriptional
circuit, an AND gate (Anderson et al., 2007), was transferred from
E. coli strain MC1061 to E. coli DH10B, the gate function un-
expectedly failed (Moser et al., 2012). Another example is that TetR
homologues as transcriptional regulators sourced from their na-
tive hosts resulted in growth defects and failed in its regulatory
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/ymben
Metabolic Engineering
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
http://dx.doi.org/10.1016/j.ymben.2016.11.007
1096-7176/&2016 International Metabolic Engineering Society. Published by Elsevier Inc.
n
Corresponding authors.
E-mail addresses: qi@pku.edu.cn (Q. Ouyang),
chengq@mail.tsinghua.edu.cn (G.-Q. Chen).
1
These authors contributed equally to the paper as first authors.
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i
Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎

functions in E. coli (Stanton et al., 2014). These observations in-
dicate that novel biological parts that are operational regardless of
chassis strains are highly in demand; with these parts, engineering
genetic control in non-model bacteria would be much more effi-
cient and reliable.
In this study, we used Halomonas sp. TD01 as a representative
industrial strain to address the development of broad-host biolo-
gical parts for genetic control in non-model microorganisms. Ha-
lomonas sp. TD01 is a halophile able to grow in highly saline
medium under unsterile conditions (Tan et al., 2011). It can utilize
glucose as a single substrate to intracellularly produce large
amounts of polyhydroxyalkanoates (PHAs), a family of degradable
bioplastics with diverse structures and properties (Fu et al., 2014;
Tan et al., 2014). A problem in the engineering of this strain is the
lack of robust, high-performance genetic parts for implementing
biological process control. Previous studies developed several
constitutive and inducible promoters applicable in Halomonas
(Li et al., 2016b;Tan et al., 2014). However, the number and per-
formance of these transcriptional parts are rather limited. For in-
stance, P
porin
, the strongest promoter developed thus far for
Halomonas spp., is constitutive and often aggravated genetic in-
stability. Meanwhile for an inducible gene expression in Halomo-
nas, there is only one available promoter, namely, IPTG-dependent
P
trc
. Therefore, new genetic parts with a wide host range, tight
regulation, and high efficiency represent a much-needed and en-
abling step toward fully realizing the potential of Halomonas and
other non-model microorganisms.
In our initial attempts to develop inducible expression system
for efficient transcriptional control in Halomonas sp. TD01, the
conventional T7 expression system was used (Elroy-Stein and
Moss, 1990;Kushwaha and Salis, 2015;Studier and Moffatt, 1986;
Temme et al., 2012). However, the T7 system inexplicably failed
despite multiple troubleshooting attempts. Instead of continued
debugging, we turned to part mining (Martinez-Garcia et al.,
2015b;Nielsen et al., 2013;Rhodius et al., 2013;Stanton et al.,
2014) to source novel T7-like expression systems from a phage
genome database. Computational and in vivo experimental results
revealed six T7-like RNA polymerase-promoter pairs with cross-
species activities (three species, E. coli S17-1, Halomonas sp. TD01,
and Pseudomonas entomophila LAC31). Based upon these six pairs,
three expression systems with a high efficiency, mutual ortho-
gonality, strict regulation, genetic stability, and variable tran-
scriptional levels in Halomonas were developed. The subsequent
cross-species evaluation showed that the functional characteristics
of these expression systems persisted with strikingly high corre-
lations when the systems were transferred from E. coli to Halo-
monas sp. TD01, highlighting their potentials for broad-host-range
applications. Subsequently, the components of expression systems
including RNA polymerases (RNAPs) were integrated into the
chromosome of Halomonas sp. TD01 to stabilize the T7-like func-
tion, thus yielding three Halomonas platform strains. Further, a
platform strain harboring a representative T7-like system, MmP1,
was used to control the bacterial cell shapes (Tan et al., 2014) and
polyhydroxybutyrate (PHB) biosynthetic pathway (Pohlmann
et al., 2006), respectively. Results showed that the MmP1 system
was extraordinarily efficient and robust to the host context. To-
gether, these T7-like novel expression systems and the idea of part
mining provide intriguing design flexibility for transcriptional
control of biological process in Halomonas and enable the transfer
of transcriptional control from a model bacterium to an in-
dustrially interesting one with reliability and predictability.
2. Material and methods
2.1. Strains, media, and chemicals
E. coli strain S17-1 was used for molecular cloning and plasmid
propagation throughout this study. E. coli strain S17-1, Halomonas
sp. TD01, and Pseudomonas entomophila LAC31 were used for the
characterization of T7-like RNAP-promoter pairs. Luria–Bertani
(LB) medium (g/L: 10 tryptone, 5 yeast extract, and 10 NaCl) was
used for culturing E. coli and P. entomophila, while LB medium
supplemented with 60 g/L NaCl (termed “60LB”hereafter) was
employed for Halomonas sp. TD01. For PHA production, LB and
60LB media with 20 g/L glucose (called “LBG”and “60LBG”here-
after) were used. Antibiotics, including (mg/L) 25 chlor-
amphenicol, 100 spectinomycin, or 100 ampicillin, were added as
needed. The inducer for P
tac
promoter was isopropyl
β
-D-1-thio-
galactopyranoside (IPTG). All chemicals were purchased from
Sigma-Aldrich unless otherwise indicated.
2.2. Computational methods for part mining
The protein sequence of T7 RNAP was used to search the NCBI
database with the tblastn program (https://blast.ncbi.nlm.nih.gov/
Blast.cgi) for bacteriophages carrying homologues. The program
was run against the NCBI genome (chromosome) database with
default algorithm parameters. For the identification of T7-family
RNAP-specific promoters, PHIRE (phage in silico regulatory ele-
ments) software package (Lavigne et al., 2004;Temme et al., 2012)
was used. The parameter settings were string length (L), 20; de-
generacy (D), 4; dominanNum, 4; and window size (W), 30. The
window size used was larger than the default because we in-
tended to a priori identify conserved sequence motifs across the
promoter region whenever possible. WebLogo (Crooks et al., 2004)
was used to generate the sequence logos for each set of putative
promoters. The putative transcription start site (þ1) and the 7
to 12 bp region of phage promoters were manually identified.
2.3. Plasmid construction and coarse characterization
T7-family RNAPs and promoters (P
T7-like
)werede novo syn-
thesized at GenScript (Nanjing, China), cloned into RNAP-module
plasmid (p15A origin, Amp
þ
) and promoter-module plasmid
(pSC101 origin, Cm
þ
) respectively, via the Golden Gate Assembly
method (see Table 1 for plasmid properties and Supplementary
Figure 1 for plasmid maps). For coarse characterization, each RNAP
module and its corresponding promoter module (superfolder
green fluorescent protein, sfGFP, as the reporter gene) were as-
sembled together on the plasmid backbone pSEVA 321 using the
Gibson Assembly method to create an RNAP-promoter-p321
plasmid series and then were transferred into E. coli, Halomonas
sp. TD01, and Pseudomonas entomophila LAC31.
For the orthogonality test, the chosen RNAP modules were
exhaustively assembled with the corresponding P
T7-like
-p321
modules (3 RNAPs 3 promoters ¼9 combinations in detail) on
pSEVA 321. Orthogonality was characterized using the same
method as that for the coarse characterization described above.
The concentration of IPTG was 1 mM.
Flat-bottom 96-well plates (Corning) covered with air perme-
able sealing films (Corning, BF-400-S) were used for the char-
acterization studies. For coarse characterization, E. coli strain S17-1,
Halomonas sp. TD01, and Pseudomonas entomophila LAC31 har-
boring T7-like systems were first inoculated into a plate contain-
ing LB or 60LB medium, respectively, for overnight cultivation
(37 °C for E. coli and Halomonas while 30 °C for Pseudomonas en-
tomophila, 1000 rpm, mB100-40 Thermo Shaker, Aosheng). Then,
cell cultures were diluted 500-fold using fresh LB medium in a
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎2
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

new plate, supplemented with IPTG of appropriate concentrations,
and cultivated for 18 h. Subsequently, an appropriate aliquot of
each cell culture (1
μ
L for E. coli; 0.5
μ
L for Halomonas sp. TD01)
was transferred to another plate containing 200
μ
L/well phos-
phate-buffered saline (PBS) for analysis by flow cytometry.
2.4. Flow cytometry
Bacterial fluorescence was detected using an LSRII flow cyto-
meter (BD Biosciences) with appropriate voltage settings (FSC 440;
SSC 260; FITC 480) and gated by forward and side scatter. At least
20,000 cells were recorded for each sample. Cytometry data were
processed using FlowJo (v7.6) to obtain the geometric mean of
fluorescence.
In this study, all fluorescence level of modified strains was
determined by subtracting that from wild type strains.
2.5. Conjugation methods
Plasmids were transferred from E. coli S17-1 to Halomonas sp.
TD01 and Pseudomonas entomophila LAC31 using an optimized
conjugation protocol. In brief, the donor cells, E. coli S17-1 har-
boring the plasmid, and the recipient cells, Halomonas sp. TD01 or
P. entomophila LAC31, were cultivated in LB (for E. coli and P. en-
tomophila) or 60LB (for Halomonas sp. TD01), respectively, with the
relevant antibiotics to an OD of 0.8. Cells were harvested (2500 g,
4°C, 3 min) and washed twice with PBS, and then mixed in a ratio
of 1:1. Next, the mixture was directly spread on an agar plate
containing 60LB with antibiotics, and incubated at 37 °C for 48 h.
2.6. Construction of platform strains
The RNAP module of K1F, VP4, or MmP1 was integrated into the
chromosome of Halomonas sp. TD01, respectively. Integration was
carried out using the conventional method (Fu et al., 2014). In
brief, the integration site was targeted by homologous re-
combination sequences. Homologous recombination sequences
(H1 and H2) were introduced into the suicide plasmid via PCR
primers (Supplementary Table 1A), after which the RNAP module
was inserted between H1 and H2. The suicide plasmids carrying
RNAP modules were conjugated to Halomonas sp. TD01 and
recombined with the chromosome to produce TD-Suicide-Co
strains (1st step recombination). A helper plasmid was then con-
jugated to the TD-Suicide-Co strains to trigger 2nd step re-
combination to produce the daughter cells with RNAP modules
integrated (RNAP positive) or not (wild-type Halomonas sp. TD01).
The positive clones were confirmed using PCR primers (Supple-
mentary Table 1A) that mapped to the chromosomal integration
site. The recombinant Halomonas sp. TD01 strains were accord-
ingly named TD-K1F, TD-VP4, and TD-MmP1, and used as platform
strains for the following experiments.
2.7. Optimization of T7-like expression systems
The promoter module of K1F, VP4, or MmP1 was cloned into
pSEVA321, respectively, to obtain a PT
7-like
-p321 plasmid series
using the Gibson Assembly method. Lac operator (lacO) was in-
serted immediately downstream the phage promoter. To do this,
lacO sequence was designed in primers and the primers were used
to amplify the P
T7-like
-p321 plasmid series. The linear PCR products
were purified and recyclized using T4 ligase and polynucleotides
kinase (PNK) to obtain the P
T7-like-LacO
-p321 plasmid series. These
plasmids were conjugated to the corresponding platform strains
(TD-K1F, TD-VP4, and TD-MmP1). At the same time, the same
P
T7-like-LacO
modules were integrated into the chromosome of the
corresponding Halomonas sp. TD01 platform strains, respectively,
using the same integration process as described above. In this
process, a new chromosomal integration site was targeted via the
corresponding homologous recombination sequences (L1 and L2)
and primers (Supplementary Table 1B). The fold induction and
genetic stability of these resultant Halomonas sp. TD01 strains
were then quantified using the characterization method outlined
above.
2.8. Construction and characterization of phage promoter libraries
A saturated random sequence corresponding to the 1to4
region of the phage promoter was designed in primers for T7-like
systems K1F, VP4, and MmP1, respectively. PCR amplifications
were performed to introduce the saturated random mutations into
each phage promoter using the P
T7-like-LacO
-p321 plasmid series as
the template. Linear PCR products were recyclized via ligation
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Table 1
Strains and plasmids used in this study.
Strain Description Reference
Escherichia coli S17-1 A vector donor in conjugation, harbors the tra genes of plasmid RP4 in the chromosome; proA, thi-1 (Simon et al., 1983)
Halomonas TD01 Halomonas TD wild type, isolated from a salt lake in China (Tan et al.,2011)
P. entomophila LAC31 P. entomophila L48 mutant(nfadB nfadA nPSEEN 0664 nPSEENccdc 4635 nPSEEN 4636 nphaC1-phaZ-phaC2)Chung et al., 2013)
Plasmid
pRE112–6ISceI pRE112 derivate with six ISceI recognition sites, a suicide vector for gene knock out in Halomonas TD strain Fu et al., 2014
pBBR1 MCS1-ISceI pBBR1-MCS1 derivate expressing ISceI endonuclease, helper plasmid Fu et al., 2014
pSEVA321 p321, RK2 replication origin, oriT, an expression vector in Halomonas TD strain, Cm Silva-Rocha et al.,
2013
RNAP-promoter-p321 series p321 derivates, with 6 phages' RNAP module and cognate promoter module on it, respectively This study
pRE112-H1H2 pRE112–6ISceI derivate, with homologous sequences H1 and H2 on it, which targets integration sites for RNAP This study
pRE112-L1L2 pRE112–6ISceI derivate, with homologous sequences L1 and L2 on it, which targets integration sites for pro-
moter modules
This study
pRE112-RNAP-Suicide series pRE112-H1H2 derivate, with different RNAP between homologous sequence H1 and H2 This study
pRE112-P
T7-like-lacO
-Suicide series pRE112-L1L2 derivate, with different promoter modules between homologous sequence L1 and L2 This study
P
T7-like
-p321 plasmid series p321 derivate, with 3 T7-like sytems' promoter modules on it. Respectively This study
P
T7-like-LacO
-p321 plasmid series P
T7-like
-p321 plasmid series derivate, with another lacO just after promoter sequences This study
P
MmP1-LacO
-minCD-p321 P
MmP1-LacO
-p321 plasmid derivate, by substituting sfgfp with minCD This study
P
MmP1-LacO
-phaCAB-p321 P
MmP1-LacO
-p321 plasmid derivate, by substituting sfgfp with phaCAB This study
Ptrc-
LacO
-minCD-p321 p321 derivate, with P
trc
promoter expressing minCD. Control of P
MmP1-LacO
-minCD-p321 This study
P
Re
-PhaCAB-p321 p321 derivate, with P
Re
promoter expressing phaCAB. Control of P
MmP1-LacO
-phaCAB-p321 This study
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎ 3
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

using T4 ligase and polynucleotides kinase (PNK) to obtain pro-
moter libraries for K1F, VP4, and MmP1 in the P
T7-like-LacO
-p321
plasmid series. Subsequently, 94 mutants from each promoter li-
brary were characterized in E. coli S17-1 harboring the corre-
sponding RNAP-module plasmids, and then approximately 10 re-
presentative mutants were selected and conjugated to the corre-
sponding Halomonas sp. TD01 platform strains (TD-K1F, TD-VP4, or
TD-MmP1). The resultant E. coli and Halomonas sp. TD01 strains
harboring promoter mutants were then characterized as described
above using flow cytometry.
2.9. Morphological engineering of Halomonas sp. TD01 using T7-like
expression system MmP1
The minCD genes were PCR-amplified from the genome of
Halomonas sp. TD01 using primers listed in Supplementary
Table 1C. The coding sequence of sfgfp in the P
MmP1-LacO
-p321
plasmid was substituted with minCD using Gibson Assembly
method (Supplementary Fig. 2) to generate plasmid P
MmP1-LacO
-
minCD-p321. P
MmP1-LacO
-minCD-p321 was first transferred to the
Halomonas sp. TD01 platform strain TD-MmP1, and then the
P
MmP1-LacO
-minCD module was integrated into the TD-MmP1
chromosome. The resulting strains were first inoculated into 60LB
medium for overnight growth, and then 1000-fold diluted into a
shake flask containing fresh 60LBG medium for further incubation.
IPTG was added when the OD
600
reached 0.4. Bright-field micro-
scopy was performed 18 h later with an Olympus IX83 microscope.
Photos were captured and cell lengths were calculated using
CellSens Standard 1.9 software (Olympus). Morphological ob-
servations of the PHA granules were conducted using ultra-thin-
section transmission electron microscopy (TEM, H-7650B instru-
ment, Hitachi Ltd.) at 120 kV.
2.10. PHB production by different species equipped with expression
system MmP1
The phaCAB gene cluster was PCR amplified from the Ralstonia
eutropha H16 genome with primers listed in Supplementary
Table 1C and used to construct plasmid P
MmP1-LacO
-phaCAB-p321
(Supplementary Fig. 2). P
MmP1-LacO
-phaCAB-p321 was transformed
into E. coli S17-1 carrying the MmP1 RNAP-module plasmid and
then conjugated to strain TD-MmP1. LBG and 60LBG media were
used for E. coli and Halomonas sp. TD01, respectively. IPTG (1 mM)
was added for full induction when the OD
600
reached 0.6. For PHB
production at different levels of induction, IPTG concentrations in
media of (mg/L) 0.02, 1, 2, 10, and 20 were used, respectively. After
the addition of IPTG, all cultivations in shake flasks continued for
24 h. PHB content was analyzed using gas chromatography (GC,
GC-2014, Shimadzu) (Tan et al., 2011).
2.11. PHB production in fermentors
P
MmP1-LacO
-PhaCAB cluster was integrated into the chromosome
of Halomonas TD-MmP1. The resultant strain named TD-HIGH was
used for a 48 h-fermentation study. Wild-type Halomonas sp. TD01
was used as a control. A 10-L fermentor was employed with a
working volume of 3 L 60-MMG medium at 37 °C. The initial
medium contained 30 g/L glucose, and the glucose concentration
was maintained at 20 g/L by continuous addition of an aqueous
solution of 300 g/L glucose. Dissolved oxygen (DO) was provided
by injecting filtered air with a flow rate of 10 L/min. The pH was
automatically controlled at 9.0 by addition of an aqueous solution
of 5 M NaOH. Culture broths were sampled every 12 h for studying
cell dry weight (CDW) and PHB content, respectively.
3. Results
3.1. Study of a T7-based expression system in Halomonas sp. TD01
In our initial effort, T7 RNAP under the control of a lacI-regu-
lated P
tac
promoter was integrated into the chromosome of Halo-
monas sp. TD01 (TD-T7, Fig. 1A). A T7 promoter driving the ex-
pression of superfolder gfp (sfgfp) was placed on the Standard
European Vector Architecture (SEVA) (Martinez-Garcia et al.,
2015a;Silva-Rocha et al., 2013) plasmid backbones pSEVA321
(p321, RK2 origin, low copy number), and transferred into Halo-
monas TD-T7 via conjugation. Unexpectedly, the green fluores-
cence of resultant strains was negligible, regardless of the added
concentrations of IPTG induction (Fig. 1B, middle; Fig. 1C). To ex-
clude plasmid loss as an issue, the promoter module was in-
tegrated into the chromosome, the fluorescence was undetectable
again (Fig. 1B, right; Fig. 1C). In contrast, a P
tac
promoter driving
sfgfp on pSEVA321 led to an obvious fluorescence, validating the
plasmid suitability (Fig. 1B, upper). Significant GFP expression was
observed when the plasmid of T7 promoter module was trans-
formed into E. coli S17-1 carrying the lacI-P
tac
-T7RNAP module,
indicating that the T7-based expression system functioned well
(Fig. 1B, left). Therefore, we substituted the coding sequence of T7
RNAP on the TD-T7 chromosome with sfgfp, thus yielding the
strain: TD-sfgfp. Under induction with IPTG, the fluorescence of
GFP was detected, showing that the lacI-P
tac
components (re-
sponsible for regulating T7 RNAP) and sfgfp (as a reporter) were
both functional (Fig. 1C). This revealed that the failure of T7-based
system was attributed to the T7 components themselves. We fur-
ther examined the RNA levels of T7 RNAP and sfgfp using reverse
transcription PCR (RT-PCR). Results showed that TD-gfp's sfgfp
mRNA level was normal, thus revealing the lacI-P
tac
module to be
functional. However, T7 RNAP mRNA was hardly detectable in
Halomonas strain TD-T7 (Fig. 1D), indicating possible degradation
of T7 RNAP with unknown reasons. Moreover, we noted that, de-
spite seriously influenced by RNAP deficiency, the P
T7
driving GFP
level was positively correlated with the existence of the T7 pro-
moter, implying that T7 promoter was functional. Altogether, these
data demonstrated the failure of the conventional T7-based ex-
pression system in Halomonas sp. TD01, which could be attributed
to the inability of this strain to stably maintain T7 RNAP mRNA.
3.2. Genomic mining of T7-like RNAP-promoter pairs
Since the mechanism on how host context affects the perfor-
mance of biological parts/circuits is poorly understood (Brophy
and Voigt, 2014;Cardinale and Arkin, 2012;Cardinale et al., 2013),
efforts to resolve such an mRNA degradation issue tend to be in-
efficient and ad hoc. Therefore, we attempted to adopt a more
rational approach, namely, part mining (Martinez-Garcia et al.,
2015b;Nielsen et al., 2013)tofind novel and efficient expression
systems that function in a broad-host range. This would provide a
general way for engineering of non-model bacteria.
The T7 RNA polymerase protein sequence (GenBank AAA32569.
1) was used to screen the NCBI Genomes (chromosome) database
for bacteriophage genomes carrying T7-family RNAPs (Fig. 2A).
This yielded 75 phages with fully sequenced genomes in the top
100 hits; 25 negative hits were bacterial genomes carrying T7
RNAP or T7-family prophages, such as E. coli strain BL21 (DE3), and
were thereby discarded (Supplementary Table 2). Next, two bio-
logical constraints were taken to exclude those phages whose T7
RNAP homologues were less likely to function: first, the E-value
should be lower than e
100
; second, the optimum temperature for
growth of the phage host should be 430 °C. These constraints
decreased the number of phages to 46 (Supplementary Table 2). To
identify putative T7-family promoters for each phage, PHIRE
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎4
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

software (Lavigne et al., 2004) was used to perform an algorithmic
string-based search on the phage genome sequences to identify
conserved sequence motifs based on sequence similarity. Three
phages, Yersinia phage Yepe2, Yep-phi, and Caulobacter phage
Percy, were eliminated because of their complicated genome
structures, making it difficult for T7-family promoters to be re-
liably identified. For other phages, at least 10 promoters were
identified for each T7-family RNAP, resulting in 43 promoter sets
(Supplementary Table 3). Next, WebLogo (Crooks et al., 2004)was
exploited to extract the consensus sequences from each promoter
set.
Previous studies showed that transcription at T7-family pro-
moters usually start at a guanine followed by a purine-rich se-
quence motif, and that the binding specificity between T7-family
RNAPs and promoters is mediated by DNA positions 7to12
relative to the transcription start site (Bandwar et al., 2002;
Cheetham et al., 1999;Rong et al., 1998)(Fig. 2B). Accordingly, the
putative transcription start sites (þ1) and RNAP-binding regions
of 41 consensus promoters were manually determined; those of
Enterobacter phage E-2 and Escherichia phage P483 could not be
identified based on existing knowledge and therefore, were dis-
carded. Each T7-famlily RNAP-promoter pair can be regarded as
one T7-like system set. These 41 sets were clustered into 15 sub-
families based on the RNAP-binding regions of promoters. For each
subfamily, the RNAP-binding region was identical (see Fig. 2C for
subfamily representatives).
3.3. Characterization of T7-like systems in different bacterial strains
Due to the high cost of commercial DNA synthesis, we selected
eight T7-like subfamily representatives (gh-1, K30, K1F, ICP3, K1-5,
SP6, MmP1, and VP4) for de novo DNA synthesis (Fig. 3A). The
coding sequences of T7-like RNAPs were codon-optimized, and the
corresponding T7-like promoters were used to design the con-
sensus promoter sequences. Similar to the T7 system, RNAPs of
novel T7-like systems were placed under the control of a lacI-
regulated P
tac
promoter in an RNAP-module plasmid (p15A origin),
and the corresponding promoters were cloned to prefixsfgfp in the
promoter-module plasmid (pSC101 origin). The cloning process for
the RNAP-module plasmid of ICP3 and SP6 repeatedly failed,
probably because of cytotoxicity similar to that seen with T7 RNAP
(Studier and Moffatt, 1986;Temme et al., 2012). Subsequently, two
modules for each of the remaining six T7-like systems were
combined into pSEVA321 for the following characterization
(Fig. 3B).
Three representative strains, Escherichia coli S17-1, Halomonas
sp. TD01, and Pseudomonas entomophila LAC31, were selected for
this study. E. coli is a model bacterium for a lot of studies, Halo-
monas sp. TD01 is a halophile bacterium with potential industrial
value, especially for open and continuous fermentation (Li et al.,
2014). P. entomophila LAC31 is a metabolically versatile soil bac-
terium that is closely related to Pseudomonas putida; it is often
used as the producer of medium-chain-length poly-3-hydro-
xyalkanoic acids (Chung et al., 2013). The broad-host-range pro-
moter, P
tac
, was used for the calibration of the promoter activity.
In E. coli, all six T7-like systems exhibited significant fold
changes (410-fold) for GFP expression when cells were induced
with 1 mM IPTG (Fig. 3C, upper panel). The GFP expression levels
of four systems (K1-5, K1F, VP4, and MmP1) were comparable to
that of P
tac
, although leakage expressions were significant. When
the systems were transferred into Halomonas sp. TD01 and P. en-
tomophila LAC31, the system activity was generally proportional to
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Fig. 1. T7-based expression system failed in Halomonas sp.TD01. (A): Circuit design of T7 expression system. The RNAP module was integrated into the chromosome of
Halomonas sp. TD01 (strain TD-T7) while the P
T7
promoter module was placed on pSEVA321. (B): GFP fluorescence from the T7 expression system using different genetic
permutations. TGreen Transilluminator (Tiangen Biotech Co. Ltd.) emitting blue light (wavelength of 470 nm) was used for GFP visualization. Middle: Strain TD-T7 with P
T7
module on pSEVA321. Right: Strain TD-T7 with P
T7
module in chromosome. Upper: Strain TD-T7 with P
T7
module replaced by P
tac
. Left: E. coli S17-1 harboring both modules
on pSEVA321. IPTG: 1 mM. (C): Upper: The coding sequence of RNAP was substituted by sfgfp (strain TD-gfp). Lower: GFP fluorescence level of the strain TD-gfp, compared
with other strains shown in (B). GFP þ: Cells with GFP fluorescence 4100 a.u. (D): Results of RT-PCR for the strains TD-T7 and TD-sfgfp. The coding sequence of lacI in RNAP
module and 16 S rRNA of Halomonas sp. TD01 were used as the positive controls.
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎ 5
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

that of E. coli (Figs. 3D-E, upper panel). One exception was K30, for
which the GFP level dramatically decreased in Halomonas sp. TD01.
Fold changes in response to IPTG induction varied a lot along the
system identity in these two non-model bacteria (Figs. 3D-E, up).
The gh-1, K1-5 and MmP1 exhibited 74-, 19- and 89-fold induction
in Halomonas, and 11-, 13- and 21-fold in P. entomophila LAC31,
respectively. For K1F and VP4, fold changes were negligible due to
significant leakage expression in both non-model bacteria. Based
on the maximal expression level as the criterion, we chose three
T7-like systems (K1F, VP4, MmP1) for further development as they
were comparable to P
tac,
or stronger in transcriptional activity
across all three bacterial strains (gh-1, 1/8 of P
tac,
in Halomonas;
K1-5, 1/4 of P
tac,
in Halomonas and Pseudomonas).
The orthogonality of the three best-performing T7-like systems
were examined by exhaustively combining the RNAP and pro-
moter modules (Figs. 2C-E, lower panel). Results showed that
these three systems were highly orthogonal to each other, with
specific GFP level 110-, 142- and 68-fold stronger than that from
crosstalk (o0.9%, o0.7%, and o1.5% of specific activation),
respectively.
3.4. Development of novel T7-like expression systems
As to our target strain Halomonas sp. TD01, high levels of
leakage expression at phage promoters indicated that simple re-
pression of RNAP expression using lacO at the P
tac
promoter was
not enough for tight control. It was notable that constitutive, high-
level leakage expression of circuit/part components severely re-
duced the evolutionary fitness of cells, thereby causing genetic
instability (Sleight et al., 2010;Sleight and Sauro, 2013), especially
in non-model bacteria (Jones, 2014;Yin et al., 2014), in which
exogenous DNA tends to be less genetically stable. Therefore, op-
timization of T7-like system performance must be conducted for
their application as expression systems in Halomonas. The RNAP
modules of T7-like K1F, VP4, and MmP1 were integrated into the
Halomonas sp. TD01 chromosome to create three platform strains,
TD-K1F, TD-VP4, and TD-MmP1. One more lacO was introduced
immediately after the transcription start site (þ1) of P
T7-like
(producing P
T7-like-LacO
) in order to investigate whether a high-fold
induction could be achieved with these platform strains (Fig. 4A).
Flow cytometry was used to record the fluorescence of GFP-
positive cells. As expected, when the P
T7-like-LacO
-sfgfp module was
encoded on plasmid pSEVA321 (P
T7-like-LacO
-p321 plasmid series),
1153-, 164-, and 3085-fold induction strengths were observed for
T7-like K1F, VP4, and MmP1, respectively. After integrating the
P
T7-like-LacO
-sfgfp modules into Halomonas sp. TD01 chromosome,
the inductions of these 3 modules were 813-, 275-, and 2441-fold,
respectively (Figs. 4B–D). When a promoter module without lacO
(P
T7-like
) was encoded on plasmids, the fraction of GFP-positive
cells was only 8%, 37%, and 24% for K1F, VP4 and MmP1, respec-
tively, indicating leakage expressions resulted in genetic instability
(Fig. 4E). When lacO was added (P
T7-like-LacO
, on plasmid), a sig-
nificant increase in genetic stability was observed (63%, 73%, and
67% of cells with GFP expression for K1F, VP4 and MmP1), re-
spectively. This could be further improved by integrating the
promoter module into the chromosome, a method with which the
fraction of GFP-positive cells increased to 499% for all three
systems (Fig. 4E). These data together demonstrated that 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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Fig. 2. Part mining to source novel T7-like RNAP-promoter pairs from a phage genome database. (A): Part mining workflow, with bioinformatic computation and biological
constraints combined. (B): Schematic of T7-family promoters. TSS, transcription start site ( þ1). The T7 promoter is shown as an example. (C): Representatives for each of the
15 T7-like subfamilies revealed by part mining. Predicted RNAP binding regions are shown in the IUPAC nucleotide code. The promoter sequence logos were created using
WebLogo (Crooks et al., 2004), and the nucleotides in the black box are putative transcription start sites for each T7-like promoter. See detailed results for each step of part
mining in Supplementary Tables 2–3.
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎6
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Fig. 3. Characterization of T7-like systems in three non-model bacterial species. (A): In vivo characterization. (B): Schematic of genetic circuits design for the characterization. Left
panel, RNA polymerase (RNAP) module; right panel, phage promoter (P
T7-like
) module. ICP3 and SP6 failed in molecular cloning, probably because of cytotoxicity. (C-E): Tran-
scriptional activity and orthogonality of T7-like systems in E. coli S17-1 (C), Halomonas sp. TD01 (D), and Pseudomonas entomophila LAC31 (E), indicated by GFP fluorescence. The
RNAP and P
T7-like
modules were combined on plasmid pSEVA321. 1 mM IPTG was added to the þIPTG group. Error bars represent standard deviations; n ¼3.
Fig. 4. Optimization of three T7-like RNAP-promoter pairs for tight regulation and genetic stability. (A): An additional lacO was introduced immediately downstream the
transcription start site ( þ1) of the T7-like promoter. TSS, transcription start site. (B–D): Fold induction of K1F (B), VP4 (C), and MmP1 (D), respectively, using different genetic
permutations (P
T7-like,
P
T7-like-LacO
, and P
T7-like-lacO
-chromosome). þIPTG: 1 mM IPTG added at OD ¼0.4. (E): Fractions of GFP-positive cells in overnight batch cultures
expressing different T7-like systems. Raw flow cytometry data is shown in Supplementary Figure 3. Error bars represent standard deviations; n ¼3.
n
and
nn
indicate po0.05
and po0.01, respectively.
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎ 7
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

optimized T7-like systems were genetically stable, highly efficient,
and tightly-regulated expression systems.
3.5. Characterization of dose-response functions and cross-species
comparison
The unexpected failure of the T7 expression system in Halo-
monas sp. TD01 motivated us to systematically evaluate the per-
formance of these novel T7-like expression systems in intended
hosts. For each system, the dose-response curves (input, IPTG
concentration; output, GFP fluorescence) were determined in E.
coli S17-1 expressing each RNAP module and the corresponding
Halomonas sp. TD01 platform strains (e.g. TD-K1F), with the pro-
moter module encoded on plasmid pSEVA321 or integrated into
the chromosome of the platform strain (Fig. 5, left panel). Full
induction occurred at 10 mg/L IPTG, which was far lower than the
commonly used concentration (200 mg/L). The host context ef-
fect was evaluated by plotting the dose-dependent response of the
plasmid-carrying promoter module in E. coli against that in Halo-
monas sp. TD01. Results showed that this effect was negligible,
with R
2
¼0.96, 0.97, and 0.99 for expression systems K1F, VP4, and
MmP1, respectively, suggesting that the T7-like components were
not interfered by the host cells (Fig. 5, middle panel). The vector
effect (plasmid or chromosome in Halomonas) was slightly
stronger, as indicated by a slightly smaller R
2
(0.90, 0.97, and 0.98,
respectively) when system performances of the plasmid vector
were plotted against those of the chromosome vector (Fig. 5, right
panel). These data demonstrated these novel expression systems
were proportionally activated in different hosts and they were
likely to act as genetic controllers in non-model industrial strain
Halomonas.
3.6. Construction of promoter libraries to achieve various tran-
scriptional levels
The application of these novel expression systems requires
promoters of different transcriptional strengths for each phage
RNAP, especially when multiple circuit/pathway components are
being fine-tuned. Previous studies showed that the region from
1bpto4 bp in the T7 promoter has a remarkable influence on
the transcription initiation rate, yet it has little effect on the
binding constant or specificity of RNAP (Bandwar et al., 2002;
Cheetham et al., 1999;Temme et al., 2012). Accordingly, saturation
mutagenesis was performed in the corresponding region of each
T7-like promoter (Fig. 6A), and the resulting number of mutants
for each T7-like promoter was calculated to be 256. For each T7-
like system, 94 mutants were selected, verified by sequen-
cing, quantified using flow cytometry, and then, mutants with
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Fig. 5. Cross-context comparison of dose-response functions of novel T7-like expression systems K1F, VP4 and MmP1. Left panel: Dose-response functions in different
contexts. IPTG was added when OD ¼0.4. Middle panel: Cross-species comparison of plasmid-carried T7-like systems. The dose-response functions in E. coli were plotted
against those in Halomonas. Right panel: Cross-vector comparison of T7-like systems in Halomonas. The dose-response functions of plasmid-carried T7-like systems were
plotted against those carried by chromosome. The size of error bars is smaller than that of symbols. Error bars represent standard deviations; n ¼3.
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎8
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

representative promoter strengths were selected and character-
ized in E. coli S17-1 and Halomonas sp. TD01. Results showed that,
within each promoter library, promoter strengths varied by 2–4
orders of magnitude, regardless of the bacterial chassis, and the
strengths of promoters from all three libraries were systematically
lower in Halomonas sp. TD01 than in E. coli (Fig. 6B, upper panel).
Remarkably, the strengths of promoters in E. coli S17-1 were highly
correlated with those in Halomonas sp. TD01, with R
2
¼0.94, 0.99,
and 0.97 for K1F, VP4, and MmP1, respectively (Fig. 6B, lower pa-
nel). This demonstrated that the RNAP-promoter interactions of
these expression systems were not affected by these two strains,
allowing various levels of genetic control in Halomonas.
3.7. Morphological engineering of Halomonas sp. TD01 using T7-like
MmP1 system
After systematic characterization of these systems, we wanted
to determine whether they could be practically used. The data
above showed that T7-like MmP1 system displayed the highest
expression level (450000 a.u) and the most strict induction
control (43000 folds) in Halomonas sp. TD01. It has been reported
that the polymerization of tubulin-like protein FtsZ is essential for
the bacterial division and that the overexpression of minCD, the
inhibitors of FtsZ protein, would turn normal rod-shaped bacteria
into long filamentary ones (Tan et al., 2014). Therefore, to de-
monstrate the application of MmP1 system, the cassette minCD
was selected to conduct morphological engineering of Halomonas
(Fig. 7A). Genes minCD were cloned downstream of P
MmP1-lacO
to
form P
MmP1-lacO
-minCD, and then this module was inserted into
plasmid pSEVA321. An inducible promoter Ptrc-
lacO
, used for
minCD expression in a previous study (Tan et al., 2014) was used as
the control (Ptrc-
LacO
-minCD). After the induction with 1 mM IPTG
when OD
600
reached 0.4, cells were cultured overnight. The
average length of cells using the control promoter was 4.5 mm,
whereas that of cells carrying P
MmP1-lacO
-minCD (MmP1-group)
was 13 mm, which tripled that of control (Fig. 7B). Notably, ex-
tremely long cells (over 100 mm) were observed in the MmP1
group but not in the control group (Supplementary Figure 4). This
suggested that the cell length resulting from the use of T7-like
MmP1 could be far more than 13 mm if the P
MmP1-lacO
-minCD
module could be stably maintained. As expected, when the
P
MmP1-lacO
-minCD module was integrated into the chromosome, an
average length of 102 mm was observed (Fig. 7B). Moreover, the
non-induced cells showed nearly the same shapes as the wild-
type; in contrast, the induced cells appeared as long fibers having
prominent PHB nodes (Fig. 7C, upper panel). We further analyzed
the cell morphology using transmission electron microscopy
(TEM). The intracellular PHB granules in all groups were directly
observed (Fig. 7C, lower panel), and the IPTGþgroup showed an
unobstructed PHB distribution throughout the entire bacterial
cytoplasmic spaces, indicating that the filamentary cells had no
division rings or cell walls/membranes inside. These data together
demonstrated the remarkable stability, high efficiency, and control
stringency of T7-like system MmP1.
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Fig. 6. Promoter libraries offered a wide range of transcriptional levels and exhibited resistance to host context. (A): Saturation mutagenesis in P
T7-like-lacO
modules. Resultant
mutants were randomly selected, characterized using flow cytometry, and transferred into Halomonas platform strains for further characterization. (B): Comparison of
transcriptional activity of promoter mutants in E. coli and Halomonas. Upper panel, raw data for promoter mutants. Lower panel, transcriptional activity of promoter mutants
measured in E. coli was plotted against that measured in Halomonas. IPTG: 1 mM. Error bars represent standard deviations; n ¼3.
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎ 9
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

3.8. Cross-species transfer of biosynthesis pathways using T7-like
MmP1 system
Next, the MmP1 system was investigated for expressing bio-
synthesis pathways in both E. coli and Halomonas. The PHB bio-
synthetic pathway was selected for proof of concept. PHB synth-
esis requires acetyl-CoA as a precursor (Fig. 8A). Overexpression of
phaA is crucial for PHB accumulation competing against the TCA
cycle. Therefore, the PHB synthesis operon phaCAB was amplified
from the Ralstonia eutropha H16 genome (Pohlmann et al., 2006)
and inserted downstream of P
MmP1-lacO
to generate plasmid
P
MmP1-lacO
-phaCAB-p321 (Method in Supplementary Figure 2). At
the same time, phaCAB equipped with its native R. eutropha pro-
moter (P
Re
) was used to construct the plasmid P
Re
-phaCAB-p321 as
a control. Both plasmids were transferred into E. coli S17-1 har-
boring the MmP1-RNAP module and the strain TD-MmP1, re-
spectively. Results showed that the MmP1 system with full in-
duction (1 mM IPTG) led to an obviously higher cell dry weight
(CDW) and PHB content (PHB%) in both E. coli and TD-MmP1 than
those from P
Re
(Fig. 8B–C). Dose-response curves for plasmid
P
MmP1-lacO
-phaCAB-p321, with IPTG concentration as the input and
PHB content as the output, were then determined in both E. coli
S17-1 and TD-MmP1 (Fig. 8D), and a consistent saturation point
(2 mg/L) was observed. As expected, linear regression of the two
dose response curves showed that MmP1-driven PHB biosynthesis
activity was consistent across species (Fig. 8E, R
2
¼0.92), indicat-
ing the robustness of T7-like MmP1 for pathway expression in
different hosts. Moreover, it should be noted that the MmP1-ex-
pressed phaCAB pathway was much more sensitive (saturated by
2 mg/L IPTG) than MmP1-expressed sfgfp (10 mg/L), indicating a
cascade effect from enzymatic reactions, which should be taken
into consideration in future applications.
3.9. Ultra-high-level production of PHB in Halomonas using MmP1
system
Because the T7-like expression systems exhibited extraordinarily
strong transcriptional activity, it was valuable to discover whether an
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Fig. 7. Morphology engineering ofHalomonasusing MmP1 system. (A):minCD genes (cell elongation cassette) under the control of T7-like MmP1 system were used to
elongate Halomonas cells. Two different genetic permutations were used, as shown in (B). (B): Results of Halomonas cell elongation using different genetic permutations of
MmP1 system. Cell length was measured using CellSens Standard 1.9 software (Olympus). P
trc
, the promoter used for Halomonas cell elongation in previous studies. The error
bars represent standard deviations; n ¼20 0;
n
and
nn
indicate po0.05 and p o0.01, respectively. (C): Cell shapes with and without the integration of P
MmP1-lacO
-minCD
module into the chromosome of strain TD-MmP1. wt, wild-type Halomonas sp. TD01; MmP1, strain TD-MmP1 with P
MmP1-lacO
-minCD module in the chromosome; þIPTG,
1 mM IPTG added at OD ¼0.4. Upper panel, optical bright-field microscopy. Photos were captured at 18 h after induction. White arrows indicate PHB nodes in cell fibers.
Lower panel, transmission electron microscopy. Cells were fixed at 18 h after induction. Each photo represents a cross-section of cells.
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎10
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

ultra-high-level PHB accumulation could be achieved in Halomonas
using the MmP1 system. The P
MmP1-LacO
-phaCAB module was in-
tegrated into the chromosome of Halomonas TD-MmP1, resulting in a
new strain named TD-HIGH; the wild-type Halomonas sp. TD01 was
used as the control. A 48 h fermentation study was carried out, with
10 mg/L IPTG added at 4 h after inoculation. At 12 h, both strains had
similar CDW and PHB%; at 24 h, 36 h, and 48 h, however, TD-HIGH
displayed an obviously higher PHB% than the control (56% vs 42%, 82%
vs 60%, and 92% vs 71%, respectively), despite that the CDWs of these
two strains were still similar (Fig. 8F). At the end (48 h), TD-HIGH
produced 69 g/L (75 g/L *92%) of PHB, which was 36% higher than the
wild-type Halomonas TD01 (50 g/L). This PHB content (92%) not only
broke the record of PHB production in Halomonas (Tan et a l., 2 011), but
also validated the feasibility for introducing more intended genetic
components into the platform strains to achieve platform upgrading.
4. Discussion and conclusion
In this study, we attempted to address the issue of lacking
biological parts that are operational regardless of chassis context
for the genetic engineering of non-model bacteria. Particularly,
one halophilic bacterium with high industrial interest, Halomonas
sp. TD01, was used as the object. We first tried to transfer the
conventional T7 expression system (T7 RNAP and its cognate
promoter) into this bacterium. The motivation arose from avid
desires for high-performance gene expression tools in non-model
bacteria, and results showed that T7-based expression systems
failed due to the poorly understood interactions between genetic
parts/circuits and host cells. This observation is quite re-
presentative because the host interference often results in un-
predictable changes in the activity of biological parts or even
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
Fig. 8. T7-like MmP1 system enabled predictable cross-species transfer and ultra-high-level expression of PHB biosynthetic pathway. (A) Schematic of PHB biosynthetic
pathway. The corresponding phaCAB cluster was cloned from the genome of Ralstonia eutropha H16 and was used to construct the P
MmP1-lacO
-phaCAB-p321 plasmid. The
native promoter, P
Re
(Pohlmann et al., 2006), was used as a control (P
Re
-phaCAB-p321). (B): MmP1-based PHB synthesis in E. coli using heterologous PHB biosynthetic
pathway. (C): MmP1-based PHB synthesis in Halomonas. Dotted line, background PHB content. The concentration of IPTG was 1 mM. (D): Dose-dependent PHB synthesis in E.
coli and Halomonas under the control of MmP1 system. Left vertical axis, PHB content in E. coli; right vertical axis, PHB content in Halomonas. (E): Comparison of dose-
dependent PHB synthesis in E. coli and Halomonas. Error bars represent standard deviations; n ¼3.
n
and
nn
indicate po0.05 and p o0.01, respectively. (F): Fermentation
study of TD-HIGH and wild-type Halomonas sp. TD01. WT, wild-type Halomonas sp. TD01; HIGH, strain TD-HIGH. 10 mg/L of IPTG was added at 4 h after inoculation.
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎ 11
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

failures in the function of genetic control circuits (Brophy and
Voigt, 2014;Cardinale and Arkin, 2012;Kittleson et al., 2012).
Deciphering the interactions between genetic parts/circuits and
host cells is always challenging (Cardinale and Arkin, 2012). As a
result, there has not been a reliable approach to directly solve the
host-interference problems.
Instead of continual troubleshooting of the T7 expression sys-
tem in Halomonas sp. TD01 without rational guidance, another
approach, namely part mining (Martinez-Garcia et al., 2015b;
Nielsen et al., 2013), was used to source novel T7-like systems
from the natural phage genomes. Resultant candidates were ex-
perimentally characterized in bacterial strains with significantly
different genetic context, and the best-performing ones were then
subjected to further genetic optimization. As expected, three novel
T7-like systems were developed to function as transcriptional
control systems that offer high efficiency, strict regulation, genetic
stability, cross-species activity and mutual orthogonality
(Figs. 3 and 4). These results indicate that part mining via bioin-
formatic computation and experimental validation (Martinez-
Garcia et al., 2015b;Nielsen et al., 2013) to source biological parts
from natural resources not only expanded the part repertoire but
also increased the probability of finding biological parts with
minimal interference from hosts.
It was found that the T7-like systems functioned well in P.
entomophila, Halomonas and E. coli with some differences (Fig. 3C-
E). For example, gh-1, a T7-like system from Pseudomonas putida,
exhibited rather weak transcriptional activity in E. coli and Halo-
monas yet became strikingly strong in P. entomophila. This might
be attributed to the fact that P. entomophila is closely related to P.
putida both genetically and biochemically. While, MmP1, in con-
trast to K1F and VP4 that suffered from some leaked expressions in
Halomonas and Pseudomonas, showed strong induction and ex-
pression in all three bacterial strains, implying its broad suitability
across species (Figs. 3D-E, Figs. 7 and 8). These observations al-
together demonstrated that different biological parts have differ-
ent effective host ranges. For some biological parts (e.g., gh-1
system), they are strongly affected by the host context, thus having
a relatively narrow host range. Other parts (e.g., MmP1 system),
however, appeared to be robust to the change in the genetic and
biochemical contexts, indicating that they are the better choices
for non-model bacteria.
In our study, a “platform strain”referred to recombinant Halo-
monas sp. TD01 with a phage RNAP integrated into the chromo-
some. Each T7-like system consisted of two modules (RNAP
module and promoter module) that required each other to gain a
function (Fig. 3B), just like “battery”and “switch”. With “battery”
(RNAP) integrated into the chromosome, genetic circuits with the
corresponding “switch”(promoter) could be flexibly designed and
introduced into the bacteria for target functions. Meanwhile, with
genetic stability guaranteed by double lacO (Fig. 4A), the expres-
sion systems exhibited a striking consistency not only under dif-
ferent regulatory conditions (plasmid vs chromosome, IPTG con-
centration) (Fig. 5) but also with diverse innate characteristics
(promoter libraries) (Fig. 6). These data from practical cases in
Halomonas TD-MmP1 altogether demonstrated the reliability of
the platform strain for heterologous gene expression.
Chromosome-based expression, as a method to improve the
genetic stability of heterologous circuits/pathways in metabolic
engineering practices, is often impaired by the shortage of DNA
copy number (Yin et al., 2014). Our T7-like expression systems
overcame this negative impact because of their ultra-high tran-
scriptional activity. One evidence is that the expression of minCD
genes using MmP1 system on the chromosome was not only high-
level but also genetically stable (Figs. 7B-C). Therefore, the chro-
mosome-carried T7-like systems are genetic tools that are both
efficient and robust for pathway overexpression, which was
validated by the case of overexpressing phaCAB operon in Halo-
monas sp. TD01 (Fig. 8). Besides, on the basis of the platform
strains Halomonas TD-K1F, TD-VP4, and TD-MmP1, more rounds of
chromosomal integration of intended genetic components could
be conducted to continuously upgrade the pathway/circuit design,
as demonstrated by the case of developing Halomonas TD-HIGH
(Fig. 8F).
In summary, the lack of high-performance biological parts for
transcriptional control of gene expression in non-model bacteria was
addressed in this study by applying part mining. Novel T7-like ex-
pression systems that offer wide dynamic range, tight regulation,
high efficiency, and low crosstalk were successfully developed and
tested in E. coli, Pseudomonas entomophila and Halomonas TD01.
Based on these T7-like systems, Halomonas sp. TD01 was engineered
to serve as platform strains for the efficient expression of biosyn-
thetic pathways. This provides design flexibility and wide scope for
metabolic engineering of Halomonas. This study demonstrated a
good example of applying synthetic biology approach such as the
part mining method to find a new solution for non-model organisms.
We envision this strategy to impact the engineering and applications
of many other non-model microorganisms.
Acknowledgement
This study was supported by the Natural Science Foundation of
China (grant numbers 31430003 and 31270146 to GQC, 11074009,
11434001 and 31470818 to QO), the National High Technology Re-
search and Development Program 863 (grant number 2012AA02A702
to QO), and the State Basic Science Foundation 973 program (grant
numbers 2012CB725201 to GQC). We are grateful for the donation of
pSEVA plasmids by Professor Victor de Lorenzo of CSIC/Spain.
Appendix A. Supporting information
Supplementary data associated with this article can be found in
theonlineversionathttp://dx.doi.org/ 10.1016/j.ymben.2016.11.007.
References
Anderson, J.C., Voigt, C.A., Arkin, A.P., 2007. Environmental signal integration by a
modular AND gate. Mol. Syst. Biol. 3, 133.
Bandwar, R.P., Jia, Y., Stano, N.M., Patel, S.S., 2002. Kinetic and thermodynamic basis
of promoter strength: multiple steps of transcription initiation by T7 RNA
polymerase are modulated by the promoter sequence. Biochemistry 41,
3586–3595.
Bonnet, J., Subsoontorn, P., Endy, D., 2012. Rewritable digital data storage in live
cells via engineered control of recombination directionality. Proc. Natl. Acad.
Sci. USA 109, 8884–8889.
Breitling, R., Takano, E., 2015. Synthetic biology advances for pharmaceutical pro-
duction. Curr. Opin. Biotechnol. 35, 46–51.
Breitling, R., Takano, E., 2016. Synthetic Biology of Natural Products. Cold Spring
Harb Perspect Biol.
Brophy, J.A., Voigt, C.A., 2014. Principles of genetic circuit design. Nat. Methods 11,
508–520.
Cardinale, S., Arkin, A.P., 2012. Contextualizing context for synthetic biology–
identifying causes of failure of synthetic biological systems. Biotechnol. J. 7,
856–866.
Cardinale, S., Joachimiak, M.P., Arkin, A.P., 2013. Effects of genetic variation on the E.
coli host-circuit interface. Cell Rep. 4, 231–237.
Cheetham, G.M., Jeruzalmi, D., Steitz, T.A., 1999. Structural basis for initiation of
transcription from an RNA polymerase-promoter complex. Nature 399, 80–83.
Chung, A.L., Zeng, G.D., Jin, H.L., Wu, Q., Chen, J.C., Chen, G.Q., 2013. Production of
medium-chain-length 3-hydroxyalkanoic acids by beta-oxidation and phaC
operon deleted Pseudomonas entomophila harboring thioesterase gene. Metab.
Eng. 17, 23–29.
Crooks, G.E., Hon, G., Chandonia, J.M., Brenner, S.E., 20 04. WebLogo: a sequence
logo generator. Genome Res. 14, 1188–1190.
Daniel, R., Rubens, J.R., Sarpeshkar, R., Lu, T.K., 2013. Synthetic analog computation
in living cells. Nature 497, 619–623.
Elroy-Stein, O., Moss, B., 1990. Cytoplasmic expression system based on constitutive
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎12
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i

synthesis of bacteriophage T7 RNA polymerase in mammalian cells. Proc. Natl.
Acad. Sci. USA 87, 6743–6747.
Fu, X.Z., Tan, D., Aibaidula, G., Wu, Q., Chen, J.C., Chen, G.Q., 2014. Development of
Halomonas TD01 as a host for open production of chemicals. Metab. Eng. 23,
78–91.
Gu, P., Su, T., Wang, Q., Liang, Q., Qi, Q., 2016. Tunable switch mediated shikimate
biosynthesis in an engineered non-auxotrophic Escherichia coli. Sci. Rep. 6,
29745.
Hammar, P., Angermayr, S.A., Sjostrom, S.L., van der Meer, J., Hellingwerf, K.J.,
Hudson, E.P., Joensson, H.N., 2015. Single-cell screening of photosynthetic
growth and lactate production by cyanobacteria. Biotechnol. Biofuels. 8, 193.
Jensen, M.K., Keasling, J.D., 2014. Recent applications of synthetic biology tools for
yeast metabolic engineering. FEMS Yeast Res.
Jones, P.R., 2014. Genetic instability in cyanobacteria - an elephant in the room?
Front Bioeng. Biotechnol. 2, 12.
Ke, C., Yang, X., Rao, H., Zeng, W., Hu, M., Tao, Y., Huang, J., 2016. Whole-cell con-
version of l-glutamic acid into gamma-aminobutyric acid by metabolically
engineered Escherichia coli. Springerplus 5, 591.
Kittleson, J.T., Wu, G.C., Anderson, J.C., 2012. Successes and failures in modular
genetic engineering. Curr. Opin. Chem. Biol. 16, 329–336.
Kotula, J.W., Kerns, S.J., Shaket, L.A., Siraj, L., Collins, J.J., Way, J.C., Silver, P.A., 2014.
Programmable bacteria detect and record an environmental signal in the
mammalian gut. Proc. Natl. Acad. Sci. USA 111, 4838–4843.
Kushwaha, M., Salis, H.M., 2015. A portable expression resource for engineering
cross-species genetic circuits and pathways. Nat. Commun. 6, 7832.
Lavigne, R., Sun, W.D., Volckaert, G., 2004. PHIRE, a deterministic approach to reveal
regulatory elements in bacteriophage genomes. Bioinformatics 20, 629–635.
Lee, S.Y., Kim, H.M., Cheon, S., 2015. Metabolic engineering for the production of
hydrocarbon fuels. Curr. Opin. Biotechnol. 33, 15–22.
Li, T., Chen, X.B., Chen, J.C., Wu, Q., Chen, G.Q., 2014. Open and continuous fer-
mentation: products, conditions and bioprocess economy. Biotechnol. J. 9,
1503–1511.
Li, T., Guo, Y.Y., Qiao, G.Q., Chen, G.Q., 2016a. Microbial synthesis of 5-aminolevu-
linic acid and its coproduction with polyhydroxybutyrate. ACS Synth. Biol.
Li, T., Li, T., Ji, W., Wang, Q., Zhang, H., Chen, G.Q., Lou, C., Ouyang, Q., 2016b. En-
gineering of core promoter regions enables the construction of constitutive and
inducible promoters in Halomonas sp. Biotechnol. J. 11, 219–227.
Martinez-Garcia, E., Aparicio, T., Goni-Moreno, A., Fraile, S., de Lorenzo, V., 2015a.
SEVA 2.0: an update of the Standard European Vector Architecture for de-/re-
construction of bacterial functionalities. Nucleic Acids Res 43, D1183–D1189.
Martinez-Garcia, E., Benedetti, I., Hueso, A., De Lorenzo, V., 2015b. Mining en-
vironmental plasmids for synthetic biology parts and devices. Microbiol Spectr.
3 (PLAS-0033-2014).
Moser, F., Broers, N.J., Hartmans, S., Tamsir, A., Kerkman, R., Roubos, J.A., Bovenberg,
R., Voigt, C.A., 2012. Genetic circuit performance under conditions relevant for
industrial bioreactors. ACS Synth. Biol. 1, 555–564.
Nevozhay, D., Zal, T., Balazsi, G., 2013. Transferring a synthetic gene circuit from
yeast to mammalian cells. Nat. Commun. 4, 1451.
Nielsen, A.A., Segall-Shapiro, T.H., Voigt, C.A., 2013. Advances in genetic circuit
design: novel biochemistries, deep part mining, and precision gene expression.
Curr. Opin. Chem. Biol. 17, 878–892.
Nielsen, A.A., Der, B.S., Shin, J., Vaidyanathan, P., Paralanov, V., Strychalski, E.A., Ross,
D., Densmore, D., Voigt, C.A., 2016. Genetic circuit design automation. Science,
352.
Pohlmann, A., Fricke, W.F., Reinecke, F., Kusian, B., Liesegang, H., Cramm, R., Eitinger,
T., Ewering, C., Potter, M., Schwartz, E., Strittmatter, A., Voss, I., Gottschalk, G.,
Steinbuchel, A., Friedrich, B., Bowien, B., 2006. Genome sequence of the bio-
plastic-producing "Knallgas" bacterium Ralstonia eutropha H16. Nat. Bio-
technol. 24, 1257–1262.
Prindle, A., Selimkhanov, J., Danino, T., Samayoa, P., Goldberg, A., Bhatia, S.N., Hasty,
J., 2012. Genetic circuits in Salmonella typhimurium. ACS Synth. Biol. 1,
458–464.
Rajkumar, A.S., Liu, G., Bergenholm, D., Arsovska, D., Kristensen, M., Nielsen, J.,
Jensen, M.K., Keasling, J.D., 2016. Engineering of synthetic, stress-responsive
yeast promoters. Nucleic Acids Res 44, e136.
Rhodius, V.A., Segall-Shapiro, T.H., Sharon, B.D., Ghodasara, A., Orlova, E., Tabakh, H.,
Burkhardt, D.H., Clancy, K., Peterson, T.C., Gross, C.A., Voigt, C.A., 2013. Design of
orthogonal genetic switches based on a crosstalk map of sigmas, anti-sigmas,
and promoters. Mol. Syst. Biol. 9, 702.
Rong, M., He, B., McAllister, W.T., Durbin, R.K., 1998. Promoter specificity de-
terminants of T7 RNA polymerase. Proc. Natl. Acad. Sci. USA 95, 515–519.
Silva-Rocha, R., Martinez-Garcia, E., Calles, B., Chavarria, M., Arce-Rodriguez, A., de
Las Heras, A., Paez-Espino, A.D., Durante-Rodriguez, G., Kim, J., Nikel, P.I., Pla-
tero, R., de Lorenzo, V., 2013. The standard European vector architecture
(SEVA): a coherent platform for the analysis and deployment of complex pro-
karyotic phenotypes. Nucleic Acids Res 41, D666–D675.
Sleight, S.C., Sauro, H.M., 2013. Visualization of evolutionary stability dynamics and
competitive fitness of Escherichia coli engineered with randomized multigene
circuits. ACS Synth. Biol. 2, 519–528.
Sleight, S.C., Bartley, B.A., Lieviant, J.A., Sauro, H.M., 2010. Designing and en-
gineering evolutionary robust genetic circuits. J. Biol. Eng. 4, 12.
Stanton, B.C., Nielsen, A.A., Tamsir, A., Clancy, K., Peterson, T., Voigt, C.A., 2014.
Genomic mining of prokaryotic repressors for orthogonal logic gates. Nat.
Chem. Biol. 10, 99–105.
Steen, E.J., Chan, R., Prasad, N., Myers, S., Petzold, C.J., Redding, A., Ouellet, M.,
Keasling, J.D., 2008. Metabolic engineering of Saccharomyces cerevisiae for the
production of n-butanol. Micro. Cell Fact. 7, 36.
Stephanopoulos, G., 2012. Synthetic biology and metabolic engineering. ACS Synth.
Biol. 1, 514–525.
Studier, F.W., Moffatt, B.A., 1986. Use of bacteriophage T7 RNA polymerase to direct
selective high-level expression of cloned genes. J. Mol. Biol. 189, 113–130.
Tan, D., Xue, Y.S., Aibaidula, G., Chen, G.Q., 2011. Unsterile and continuous pro-
duction of polyhydroxybutyrate by Halomonas TD01. Bioresour. Technol. 102,
8130–8136.
Tan, D., Wu, Q., Chen, J.C., Chen, G.Q., 2014. Engineering Halomonas TD01 for the
low-cost production of polyhydroxyalkanoates. Metab. Eng. 26, 34–47.
Temme, K., Hill, R., Segall-Shapiro, T.H., Moser, F., Voigt, C.A., 2012. Modular control
of multiple pathways using engineered orthogonal T7 polymerases. Nucleic
Acids Res 40, 8773–8781.
Tran, T.T., Charles, T.C., 2016. Genome-engineered Sinorhizobium meliloti for the
production of poly(lactic-co-3-hydroxybutyric) acid copolymer. Can. J. Micro-
biol 62, 130–138.
Venturelli, O.S., Egbert, R.G., Arkin, A.P., 2016. Towards engineering biological sys-
tems in a broader Context. J. Mol. Biol. 428, 928–944.
Wang, B., Barahona, M., Buck, M., 2014. Engineering modular and tunable genetic
amplifiers for scaling transcriptional signals in cascaded gene networks. Nucleic
Acids Res 42, 9484–9492.
Yang, C., Gao, X., Jiang, Y., Sun, B., Gao, F., Yang, S., 2016a. Synergy between me-
thylerythritol phosphate pathway and mevalonate pathway for isoprene pro-
duction in Escherichia coli. Metab. Eng. 37, 79–91.
Yang, J.E., Kim, J.W., Oh, Y.H., Choi, S.Y., Lee, H., Park, A.R., Shin, J., Park, S.J., Lee, S.Y.,
2016b. Biosynthesis of poly(2-hydroxyisovalerate-co-lactate) by metabolically
engineered Escherichia coli. Biotechnol. J.
Yin, J., Fu, X.Z., Wu, Q., Chen, J.C., Chen, G.Q., 2014. Development of an enhanced
chromosomal expression system based on porin synthesis operon for halophile
Halomonas sp. Appl Microbiol Biotechnol. 98, 8987–8997.
Yu, X., Jin, H., Liu, W., Wang, Q., Qi, Q., 2015. Engineering corynebacterium gluta-
micum to produce 5-aminolevulinic acid from glucose. Micro. Cell Fact. 14, 183.
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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
H. Zhao et al. / Metabolic Engineering ∎(∎∎∎∎)∎∎∎–∎∎∎ 13
Please cite this article as: Zhao, H., et al., Novel T7-like expression systems used for Halomonas. Metab. Eng. (2016), http://dx.doi.org/
10.1016/j.ymben.2016.11.007i