ACS Synthetic Biology
Letter
individual intact cells was analyzed through a flow cytometer
and Figure 1D presents the resultant histogram plots of the
induced cells fluorescence after 19 h. Compared to the control
plasmid causing a change in the DNA bending properties for
CynR protein.
Decreasing the length between the promoters largely
reduced the leaky expression under no induction while
induction (Figure S2). For all subsequent work, pCyn-v2-GFP
design was chosen to keep the promoters at optimal distance
from each other. The nearly 626-fold increase in GFP
fluorescence observed was now promising to be able to utilize
the pCyn-v2-GFP engineered design for autonomously sensing
azide ions. Specifically, there was a significant difference (two-
tailed t test P < 0.001 for 3 biological replicates per test group)
in the fluorescence observed for pCyn-v2-GFP transformed
cell lysate after 4 h induction time in the presence of 1 mM
sodium azide (105 179 ± 3802; mean ± s.d.) versus uninduced
cells (168 ± 17). To determine the minimal azide amount
required to induce GFP protein expression, we induced cells
transformed with the pCyn-v2-GFP plasmid with varying
concentrations of azide ions. The maximum amount of azide
had a negative influence on bacterial cell growth (Figure S3).
GFP fluorescence of cell lysate showed a strong correlation in
protein expression as a function of inducer dosage. GFP
expression increased until 1 mM azide concentration induction
after which there was reduction in the GFP signal for 5 mM
inducer concentration likely due to cellular toxicity of azide
(Figure 2E). The lowest tested azide concentration (10 μM)
showed marginal GFP expression indicating that the detection
limit for this synthetic promoter would have to be greater than
10 μM. The bacterial growth phase during the time of
induction also played an important role and we observed that
induction at early exponential phase yielded higher expressed
CynR protein acts as a repressor for the promoter by causing
a bend at the −35 site hindering the binding of RNA
polymerase. The cynR protein consists of two domains: an
inducer binding domain and a DNA binding domain. The
inducer binding domain binds to the inducer (cyanate or
azide) which decreases the bend at the promoter site to
facilitate transcription. For regulating protein expression levels,
along with the inducer amounts the basal level of cynR protein
present is also critical. We therefore further engineered the
cynR constitutive promoter to adjust the background level of
repressor protein expressed. Four different constructs (v5, v6,
v7, and v8) were created by introducing mutations at the
promoter −10 site, regions between −10 and −35 site, and
between RBS and translation start site (Figure 2C and SI
Figure S5A). The expression strength of these modified
promoters was tested in the wild type (BW2113) and several
additional knockout strains (sCB1, sKS3, sKS4) by monitoring
GFP fluorescence in cell lysates. The knockout strains sKS3
and sKS4 were generated to remove the native cynX gene and
cyn operon (SI Table S1), respectively, to minimize the export
of azide ions using native transporter proteins and reduce
interference from the endogenous cyn operon. The pCyn-v8-
GFP construct, which had reduced efficiency at −10 site but
optimal length between RBS and translation start site, gave
about 120−160 fold higher reporter GFP fluorescence as
compared to the native promoter, with the maximum fold
increase observed in ΔcynX (sKS3), ΔcynR, and ΔcynTSX
(sKS4) knockout strains (Figure 2D). Specifically, there was a
significant difference (two-tailed t test P = 0.006 for 3
biological replicates per test group) in the fluorescence
(
uninduced cells), the azide induced cell lysate had a nearly 20-
fold increase in GFP fluorescence while cyanate did not result
in any significant GFP expression. Specifically, there was a
significant difference (two-tailed t test P < 0.001 for 2
biological replicates per test group) in the fluorescence
observed for pCyn-v1-GFP transformed cell lysate after 19 h
induction time in the presence of 1 mM sodium azide (3840 +
1
35; mean ± s.d.) versus uninduced cells (169 ± 38). The
absence of GFP expression in cyanate induced cells was likely
due to the breakdown of cyanate by the endogenous cyanate
hydratase enzyme encoded by the native E. coli BW25113-Wt
genome. Hence, we next tested the induction capacity of the
synthetic promoter using a ΔcynS knockout strain (BW25113-
sCB1) to clearly show GFP expression upon induction by
cyanate (SI Figure S1). However, the amount of protein
expressed based on total cell lysate GFP fluorescence was very
low suggesting that the native cyn promoter strength was quite
poor. This is not surprising since the native cyn promoter
regulates associated CynTSX proteins expression required to
overcome cyanate toxicity and that seldom requires high
protein expression yields. The native operator region also
contains suboptimal −10 sequence and Shine−Delgarno (SD)
sequences (or ribosomal binding site; RBS) which could play a
role in poor expression strength.
To improve the native promoter strength for enabling higher
inducible protein expression levels, the native cyn promoter
was next engineered to contain consensus −10 and SD
sequences. The cynR gene, which is negatively regulated by cyn
promoter, was placed under control of an independent
34
constitutive promoter (see SI Text for sequence informa-
tion). Additionally to making these modifications, the
promoters were separated by inserting a random DNA spacer
sequence of 100 bp (pCyn-v2-GFP) and 1000 bp (pCyn-v4-
GFP) to avoid any interference and steric hindrances between
the two promoters (Figure 2B and SI Figure S2). A control
construct that contained no spacer sequence (pCyn-v3-GFP)
was also generated. The modified constructs were individually
transformed into E. coli cells and tested for GFP expression
using sodium azide as inducer. All three engineered constructs
showed a greater than 50-fold increase in the fluorescence
signal in the cell lysate, indicating a significant increase in GFP
version 1 promoter (Figure 2D and SI Figure S2B).
Specifically, there was a significant difference (two-tailed t
test P = 0.002 for 3 biological replicates per test group) in the
fluorescence observed for pCyn-v2-GFP transformed cell lysate
after 2 h induction time in the presence of 1 mM sodium azide
(32 068 ± 2170; mean ± s.d.) versus pCyn-v1-GFP trans-
formed cell lysate (610 ± 80). The pCyn-v4-GFP construct
with longest spacer region showed around 37% increased
fluorescence when compared to the no spacer pCyn-v3-GFP
control. Specifically, there was a significant difference (two-
tailed t test P = 0.001 for 3 biological replicates per test group)
in the fluorescence observed for pCyn-v4-GFP transformed
cell lysate after 2 h induction time in the presence of 1 mM
sodium azide (65 468 ± 932; mean ± s.d.) versus pCyn-v3-
GFP transformed cells (47 765 ± 28). However, significant
leaky GFP expression was seen for the no induction controls of
v4 construct (SI Figure S2C). The extra-long spacer sequences
could have potentially allowed undesirable interactions in the
6
85
ACS Synth. Biol. 2021, 10, 682−689