INTRODUCTION
Exoelectrogenic
bacteria (EEB) can couple the reduction of metal oxide and electrodes
with their cellular metabolism and growth, and thus play an essential
role in environmental bioremediation,
energy
generation from wastes, and biogeochemical cycling of metals in
environments (Li, Yu, & He, 2014; Logan, 2009; Lovley, 1991; Rittmann,
2008; TerAvest & Ajo-Franklin, 2016). Shewanella oneidensisMR-1, a model strain of EEB, can use various extracellular electron
acceptors such as iron and manganese oxides, oxygen, sulfur species,
radionuclides, toxic metals, and so on for cell respiration (Fredrickson
et al., 2008; Nealson & Cox, 2002). However, in contrast with its
diverse electron
acceptors,S. oneidensis MR-1 can use only a narrow range of substrates,
such as lactate (Serres & Riley, 2006),
pyruvate
(Pinchuk et al., 2009) and
N-acetyl-glucosamine
(GlcNAc) (Rodionov et al., 2010), as its carbon sources which
significantly
limits its practical applications in waste treatment and environmental
bioremediation.
Carbohydrates
supply carbon sources for a variety of microbes (Chaudhuri & Lovley,
2003; Rodionov et al., 2010). Except for chitin, GlcNAc, and glycerate
(Rodionov et al., 2010), S. oneidensis MR-1 can’t utilize the
vast majority of carbohydrates including glucose. In fact, the genomic
annotations predict that S. oneidensis MR-1 has the ability to
grow on glucose or maltodextrin (Rodionov et al., 2010). However, when
some carbohydrates utilization pathways were reconstructed with genomic
annotations, the phenotypic experimental results show
they did not work in S.
oneidensis MR-1 (Rodionov et al., 2010). Unfortunately, the
genomic
annotations derived from a sequence similarity analysis could not
identify the inactivation of the reconstructed pathways caused by
transcriptional regulations, genetic mutations, or something else
(Serres & Riley, 2006). Thus, it is necessary to rapidly identify the
functions of the crucial genes involved in
the
carbohydrate metabolism in S. oneidensis MR-1. Then, robust
engineered S. oneidensis MR-1 with an increased metabolic
capacity for carbohydrates could be used to accelerate its
environmental
bioremediation. To achieve this goal,
a
rapid, highly efficient and
easily
tuned approach is highly desired for engineering S. oneidensisMR-1.
Recently, the clustered regularly interspaced short palindromic repeat
(CRISPR) related systems have been widely used for targeted genome
editing in numerous organisms including eukaryotes and prokaryotes
(Barrangou & Doudna, 2016; Choi & Lee, 2016; Luo, Leenay, & Beisel,
2016). Among these CRISPR related systems, a CRISPR/Cas9n (Cas9 D10A
nickase)-mediated tool, termed “base editor”, has been developed in
mammalian
cells and some other organisms for precise base editing (Chen et al.,
2018; Komor, Kim, Packer, Zuris, & Liu, 2016; Tong et al., 2019; Wang
et al., 2018). This approach can
rapidly
and efficiently generate mutations without double strand break, which is
required in other CRISPR/Cas9 or Cpf1-mediated genome editing systems.
As
a typical base editor, BE3 system developed in mammalian cells contains
the engineered fusions of the Cas9n(D10A), the rat cytidine deaminase
APOBEC1 (rAPOBEC1) and a uracil glycosylase inhibitor (UGI) (Komor et
al., 2016). Guided by the target binding capability of Cas9n/sgRNA
complex, cytidine deamination by the rAPOBEC1 occurs in the displaced
single-strand DNA within the Cas9n-sgRNA-targeted DNA ‘R-loop’ structure
(Figure 1a). The deamination of the target C in a C:G base pair results
in a
U:G
heteroduplex, which can be converted to a T:A base pair following DNA
replication (Komor et al., 2016). The C-to-T conversion of target C
within the coding genes can rapidly result in the introduction of
premature stop codons or point mutations in target genes. Thus, the base
editor might be an ideal tool to quickly identify the carbohydrates
metabolic pathways in S. oneidensis MR-1 and construct engineered
strains with an enhanced pollutant degradation capacity.
This work aims to develop such a rapid and robust base editing system
inS. oneidensis MR-1 and apply this tool to rapidly identify the
functions of the genes involved in
carbohydrate
metabolism, and expand the metabolic capacity of S. oneidensisMR-1 for contaminant degradation. For this purpose, the
single-plasmid-based pCBEso system was firstly constructed inS.
oneidensis MR-1. Then, the effective conversions of C-to-T within
targeted genes were confirmed and the editing efficiencies of target C
within each NC motifs (N: A, T, C, G) at different positions were
systematically evaluated. Furthermore, multiplex genome editing using
the pCBEso system was tested and the key
genes
related to GlcNAc or glucose metabolism
were
identified. Finally, an engineered strain was constructed and its
ability to degrade multiple organic pollutants was compared with the
wild type when glucose or GlcNAc was used as the sole carbon source.
2 | MATERIALS AND METHODS
2.1 | Microbial strains, plasmids, and growth
conditions
All of the strains and plasmids used in this work are listed in Table S1
(Supporting Information). E. coli strains were aerobically
cultured in Luria broth (LB) medium at 37 oC. S.
oneidensis strains were grown at
30oC in LB medium for
preparing
electrocompetent cells, or in mineral medium for degrading organic
pollutants (Min et al., 2017). If necessary, appropriate chemicals were
added
at the following concentrations: 34 µg/mL chloramphenicol for E.
coli Turbo, 50 µg/mL diaminopimelic acid and 34 µg/mL chloramphenicol
for E. coli WM3064, and
10
µg/mL chloramphenicol for S. oneidensis MR-1.
2.2 | Construction of the pCBEso plasmid and
reporter strain-MR-1/GFP-lacZ
The pCBEso plasmid was rapidly assembled as follows: the ColE1 origin of
replication and the
rAPOBEC1-Cas9n
(D10A) expression cassette were amplified from the
plasmid
pnCasPA-BEC (Chen et al., 2018). The chloramphenicol-resistance marker
(CmR) was copied from the pRE112 plasmid (Min et al., 2017). TherpsL promoter, controlling the transcription of the
rAPOBEC1-Cas9n (D10A) expression cassette, was cloned from the genome ofS. oneidensis MR-1. The trc promoter-driven sgRNA
expression cassette, I-SceI recognition site, and lacpromoter-driven I-SceI Endonuclease expression cassette were synthesized
by General Biosystems Co., Anhui, China. All these fragments were
assembled to construct the final plasmid pCBEso via
Gibson
assembly method (Gibson et al., 2009). In the sgRNA expression cassette
driven by the trc promoter, twoBsa I
sites were introduced to facilitate the construction of targeted pCBEso
plasmid via
Golden
Gate assembly (Engler, Kandzia, & Marillonnet, 2008). All primers used
in this work are listed in Table S2.
The
GFP-lacZ -fused reporter was integrated into the genome ofS. oneidensis MR-1 by homologous recombination. The plasmid
pRE112-GFP-lacZ harboring two
1
kb homologous arms flanking the gene SO_4341 and the lacpromoter-driven GFP-lacZ expression cassette was transformed intoE. coli WM3064 and then transferred into S. oneidensisMR-1 by conjugation. The positive exconjugants grown on LB agar plates
containing 10 µg/mL chloramphenicol were firstly cultured in
antibiotic-free LB medium and then plated onto LB agar plates containing
10% (wt/vol) sucrose. The chloramphenicol-sensitive and
sucrose-resistant colonies were screened by PCR for the desired mutants
with insertion of the GFP-lacZ -fused reporter.
2.3 | Base editing by the pCBEso system
S. oneidensis electrocompetent cells were prepared with the
following procedures: after cultivation in 30 mL LB medium for 16
h, S. oneidensis cells were harvested by centrifugation (5000 g,
5 min) and washed twice with 10 mL of sterile ice-cold 300 mM sucrose,
and finally resuspended with 2 mL of 300 mM sucrose.
For the base editing in S. oneidensis MR-1, two single-strand DNA
(ssDNA) oligos (24-nt) were designed as: 1)
5’-GTGGNNN…NNN
(N20) -3’; 2) 5’-AAACNNN…NNN
( inverted
repeat of N20) -3’. Then, the two ssDNA oligos were
phosphorylated
and annealed to form a double-strand (dsDNA) fragment with
cohesive
ends, which was then ligated to the Bsa I linearized pCBEso
plasmid via
Golden
Gate assembly (Engler et al., 2008). The resulting pCBEso plasmid was
transformed into E. coli Turbo. Then, S. oneidensiselectrocompetent cells of 100 µl were mixed with the targeted pCBEso
plasmid of 1 µg extracted from the corresponding E. coli Turbo,
and
subsequently
transferred into a 0.2-cm sterile electroporation cuvette. Immediately
after electroporation (1.3 kV mm-1, 100 Ω, and 25 µF),
the
mixtures were supplemented with 1 mL LB and recovered at 30oC for 2 h. The cells were collected by centrifugation
and then plated onto a LB agar plates containing 10 µg/mL
chloramphenicol. Colonies grown on the plates
were
randomly picked and amplified by PCR. Amplified products of the targeted
regions were subjected to Sanger sequencing. Similar steps were applied
in the multiplex genome editing using the pCBEso system in S.
oneidensis MR-1. After base editing, the targeted pCBEso plasmid was
cured through the I-SceI counter-selection. The edited strains were
cultured with fresh antibiotic-free LB medium containing 1 mM
isopropyl-beta-D-thiogalactopyranoside
(IPTG)
for 12 h and then plated onto the surface of LB agar plates containing 1
mM IPTG. After incubation at 30 oC for 30 h, the
plasmid-free colonies were obtained.
To
systematically
evaluate the effects of the target C position and sequence context on
the base editing efficiency, the following 7 protospacers were designed:
1)
5’-N AC1TC3GC5C6 NNN…NNN(12
nt) NGG- 3’; 2)
5’-NN AC2TC4GC6 NNN…NNN(12
nt) NGG -3’; 3)
5’-N CC1AC3TC5 NNN…NNN(13
nt) NGG -3’; 4)
5’-N GC1C2AC4TC6 NNN…NNN(12
nt) NGG -3’; 5)
5’-NN GC2C3AC5 NNN…NNN(13
nt) NGG -3’; 6)
5’-N TC1GC3C4AC6 NNN…NNN(12 nt) NGG -3’; 7)
5’-NN TC2GC4C5 NNN…NNN
(13 nt) NGG -3’. All 7 suitable protospacers randomly
distributed in nonessential genes were carefully screened from the
genome of S. oneidensis MR-1 (Figure 2a).
2.4| Phenotypic evaluations
After overnight cultivation in 3 mL LB medium, the edited
MR-1/GFP-lacZ cells were collected and washed twice by
phosphate-buffered saline (PBS) buffer, and finally resuspended in 15 mL
PBS buffer. The edited cells were measured by a Flow Cytometer
(CytExpert,
Beckman Coulter Inc., USA) and
collected
data were analyzed by CytExpert
2.0
software (CytExpert 2.0, Beckman Coulter Inc., USA). The wild-type cells
were taken as a negative (GFP-) control, while the MR-1/GFP-lacZcells were taken as a positive (GFP+) control. Meanwhile, a
fluorescence
microscope (BX51, Olympus Inc., Japan) was used to detect the
fluorescence of these cells (Cheng, Min, Liu, Li, & Yu, 2019).
In
β-galactosidase activity assay,
2
µL
overnight
culture of the positive control and edited MR-1/GFP-lacZ cells
was dripped onto the surface of a LB agar plate containing 30 µg/mL
5-bromo-4-chloro-3-indolyl β-D-galactopyranoside (X-gal) and incubated
at 30 oC for 24 h for imaging.
Ampicillin susceptibility assay: overnight cultures of the wild-type and
edited strains were serially diluted 10-fold over a range of
10-1 to 10-6 with fresh LB. Then, 2
µL of each diluent was dripped onto the surface of LB agar plates
containing 0 or 2.5 µg/mL ampicillin and incubated at 30oC for 24 h for imaging.
Growth curves of the wild-type and edited strains cultured in mineral
medium supplemented with
GlcNAc
or glucose were obtained by measuring the
optical
density at 600 nm (OD600)
using
a BioTek microplate reader (Synergy HT,
BioTek
Ins., USA).
2.5| Anaerobic biodegradation oforganic
pollutants
Anaerobic reductions of organic pollutants were carried out in serum
vials containing 30 mL mineral medium (Cheng et al., 2020; Min et al.,
2017). The serum vials were
sparged
with N2 to maintain anaerobic conditions. Methyl orange
(MO, 30 mg/L), amaranth (30 mg/L), or roxarsone (50 µM) was separately
added into the mineral medium as the electron acceptor. The experiments
for organic pollutants degradation by the wild type and the strainnagR BE in mineral medium supplemented with 20 mM
glucose were conducted as follows: the wild-type and the strainnagR BE were firstly cultured in mineral
medium
supplemented
with 20 mM lactate and glucose, respectively. After cultivation for 60 h
at 30 oC,
the
cells in mineral medium were harvested by centrifugation (5000 g, 5
min), washed twice by sterile mineral medium and immediately injected
into the serum vials containing 30 mL mineral medium (with 20 mM glucose
as the sole electron donor) to reach a final OD600 of
0.3. To examine the reduction kinetics of the organic pollutants, the
samples were taken from each serum vial at given time intervals. MO and
amaranth were quantified and analyzed by high performance liquid
chromatography (HPLC, Shimadzu Co., Japan) (Gomi et al., 2011).
Roxarsone was analyzed by using a high performance liquid
chromatography-hydride generation-atomic fluorescence spectrometry
combined with a species analysis instrument (HPLC-HG-AFS, SAP-10,
Beijing Titan Co., China), as described previously (Han et al., 2017).
It was worth noting that when organic pollutants degradation abilities
of the wild type and the strain nagR BE in mineral
medium supplemented with 20 mM GlcNAc were evaluated, both strains were
firstly cultured in mineral medium supplemented with 20 mM lactate.
2.6 | RNA extraction and quantitative reverse
transcription-PCR (qRT-PCR) analysis
The wild-type and the strain nagR BE were cultured
for 24 h in mineral medium with lactate as the sole carbon source. The
resulting cells were collected to extract total RNA by using RNAiso plus
kit (Takara Co., China). After treated with DNase I (Takara Co., China),
the purified total RNA was reverse transcribed via the PrimeScript RT
regent kit
(Takara
Co., China). qRT-PCR analysis was conducted as described previously (Liu
et al., 2017). The related primers for qRT-PCR experiments are listed in
Table S2.
3 | RESULTS
3.1 | Development of the base-editing system
pCBEso in Shewanella oneidensis MR-1
To perform efficient conversion of target C to T
inS. oneidensis MR-1,
a
single-plasmid-based, CRISPR/Cas9n(D10A)-mediated base editing system
pCBEso was designed and constructed (Figure 1). In the
“all-in-one-plasmid” system, the cytidine deaminase
rAPOBEC1
was fused to the N terminus of
the
Cas9n(D10A)
and their expression was driven by the constitutive promoter
(PrpsL ).
To
apply the pCBEso plasmid for base editing of targeted genes, only a
20-bp spacer was needed to insert into the sgRNA cassette, which was
controlled by the constitutive promoter (Ptac ).
A
Golden Gate assembly strategy was adopted to facilitate efficient
cloning of the 20-bp spacer for rapid assembly of the targeted pCBEso
plasmid (Engler et al., 2008). To cure the plasmids after editing, both
of I-SceI recognition site and lac promoter-driven I-SceI
Endonuclease expression cassette were integrated into pCBEso plasmid.
After transforming the targeted pCBEso plasmid into S. oneidensisMR-1, the base-editing system could readily convert the target C to T
and enable highly efficient gene inactivation and point mutations in the
targeted genes.
3.2 | Conversion of the target C to T by the
pCBEso system in S. oneidensis MR-1
To
find out whether the
pCBEso
system was an effective base editing tool, the coding region of the
metal-reducing/respiratory pathway was selected as a target. The
reported base editors in mammalian or Streptomycetes cells have a
less than
10-nucleotide
editable window in the protospacer region distal to the protospacer
adjacent motif (PAM) (Komor et al., 2016; Tong et al., 2019). Thus, each
cytidine
from Positions 1 to 10 within the protospacer in the PAM distal position
was investigated (the 5’ end of the protospacer was set to position 1,
Figure S1 and Table S3). Sequencing results show that none of Cs from
Positions 1, 2, 9 and 10 within the
hypothetic
base editing window was converted to
thymidines
in all 8 protospacers and the conversion of C to T was observed in the
rest positions (Figure S1). The efficiency of C-to-T editing in the
hypothetic base editing window was 33.3% to 100%. These results
suggest that the pCBEso system had a
6-nucleotide
editing window from Positions 3 to 8 within the hypothetic base editing
window (Figure 2a). Unexpectedly, two Cs at Positions 3 and 8 within the
protospacer mtr_ Ps7 were not converted to thymidines and the
target C was preceded by a G or C. Therefore, the base editing
efficiency of the pCBEso system was substantially affected by the
position of target C and the sequence context within the base editing
window.
3.3 | Characteristics ofthe
pCBEso system in vivo
In order to evaluate the effects of the target C position and sequence
context on the base editing efficiency of the pCBEso system in
vivo , a matrix was designed based on the four possible NC (N: A, T, C,
or G) combinations of the target nucleotide with all 4 nucleotides. The
target C of each NC motif occurred at all six possible positions (Figure
2a). As shown in Figure 2b,
all
the TC motifs within the base editing window achieved the highest
conversion efficiency of C-to-T, while the editing efficiencies of the
target Cs in all the GC motifs were the lowest. Overall, the editing
activity of the pCBEso system on the target Cs of all four NC motifs
follows the order TC > AC > CC >
GC. Moreover, the target Cs at Positions 3, 4, and 5 within the
6-nucleotide editing window exhibited higher editing efficiencies than
the rest.
To test the actual editing activity
of
the
pCBEso system for the functional genes in S. oneidensis MR-1, an
exogenous
GFPmut3b-lacZ -fused
reporter gene and an
endogenousblaAgene were selected as targets. The protospacer in GFP gene was selected
and assembled into the pCBEso plasmid (Figure 3a). Sanger sequencing
results of the targeted region show that the C at Position 4 within the
GFP spacer was successfully converted to T with a high editing
efficiency of 4/4, generating an expected STOP codon. The phenotypic
changes of the edited
MR-1/GFP-lacZwere further characterized by flow cytometry,
chromogenic
test with X-gal, and fluorescence microscope. As shown in Figure 3b,
99.7% of edited cells collected from all four edited strains lost
fluorescence, indicating that the highly efficient base editing and GFP
gene inactivation occurred in the strain MR-1/GFP-lacZ . The X-gal
reaction and fluorescence imaging results are consistent with the above
observations (Figures 3c and S2).
BlaA, one putative β-lactamase, was reported to confer S.
oneidensis MR-1 resistance to β-lactam (Yin et al., 2013). As shown in
Figure S3a, the
target
C at Position 8 within the blaA spacer was effectively mutated to
T with an efficiency of 3/3, producing a desired STOP codon.
Susceptibility
assay of these strains to
ampicillin
was adopted to characterize the phenotypic changes of the edited
strains, which further confirms the highly efficient base editing
efficiency of the pCBEso system in S. oneidensis MR-1. The growth
of all the three edited strains was comparable to that of the wild-type
strain on LB agar plates without ampicillin addition. However, all the
three edited strains, contrary to the wild-type strain, failed to grow
on LB agar plates when 2.5 µg/mL ampicillin was added (Figure S3b). In
general, the pCBEso system exhibited a highly efficient conversion of C
to T within the suitable protospacers of the functional genes.
3.4 | Highly efficient double-locus base
editing of S. oneidensis MR-1 by the pCBEso system with one
single plasmid
Compared with the conventional genetic manipulation tools, CRISPR-based
genome editing methods have an unparalleled advantage in multiplex
genome editing (Choi & Lee, 2016). To achieve multiplex genome editing
using the pCBEso system in S. oneidensis MR-1, two protospscers
harboring target Cs in nagR and dmsE were selected and
simultaneously assembled into the pCBEso plasmid. In the resulting
plasmid, the two sgRNAs targeting nagR and dmsE were
controlled by the same promoters and terminators (Figure 4a). Sanger
sequencing results show that both of the target Cs at Position 3 within
the nagR spacer and Position 4 within the dmsE were
converted to Ts with high efficiencies of 7/8 and 8/8, respectively
(Figure 4b). Therefore, simultaneous base editing of two genes using the
pCBEso system with a high efficiency in S. oneidensis MR-1 was
achieved.
3.5 | Identification the key genes involved in
GlcNAc or glucose metabolism by the pCBEso system in S.
oneidensis MR-1
As an available carbon source for S. oneidensis MR-1, comparative
genomic analysis reveals that GlcNAc can be transported by a permease
NagP into the cytoplasm and subsequently transformed to fructose
6-phosphate
(Fructose-6P)
via an N-acetylglucosamine (NAG) catabolic pathway (Rodionov et al.,
2010). Such a NAG pathway (from GlcNAc to Fructose-6P) consists of a
three-step biochemical conversion (Figure 5a): 1) from GlcNAc to
GlcNAc-6P (catalyzed by a GlcNAc kinase NagK); 2)
from
GlcNAc-6P to GlcN-6P
(catalyzed
by a GlcNAc-6P deacetylase NagA); and 3) from GlcN-6P to Fructose-6P
(catalyzed by a GlcN-6P deaminase NagB) (Yang et al., 2006). Moreover,
another uncharacterized membrane protein NagX was predicted to be
involved in the uptake of GlcNAc (Yang et al., 2006). To examine whether
NagX could be classified as
a
GlcNAc transporter and provide more experimental evidence on the roles
of functional genes in the NAG catabolic pathway, several single or
double-locus edited strains were
constructed by the pCBEso system
(Figures
5b and S4). The growth curves of these edited strains demonstrate thatnagK BE,nagA BE, nagP BE and the
double-locus edited strain nagX/P BE lost the
ability to grow in mineral medium with
GlcNAc
as the sole carbon source (Figure 5b). However, the wild type strain andnagX BE could grow on GlcNAc, indicating that
NagX
was not the GlcNAc transporter.
NagR, a transcriptional repressor,
governs the expression of multiple genes (e.g., nagK ,nagB , nagA , nagX , and nagP ) associated with
the GlcNAc metabolism in S. oneidensis MR-1 (Rodionov et al.,
2010; Yang et al., 2006). Interestingly, the deletion of nagRenabled S. oneidensis MR-1 to catabolize glucose. This might be
attributed to the overexpression of NagP and NagK involving in the
transport
and phosphorylation of glucose in MR-1, respectively (Chubiz & Marx,
2017). In order to confirm this hypothesis, the strainnagR BEwas constructed by using the pCBEso system (Figue S5).
Expectably, the expression levels ofnagK , nagB , nagA , nagX , and nagP innagR BEincreased substantially than those in the wild type strain when cells
were grown in mineral medium supplied with 20 mM lactate
(Figure
5c). The growth curves show that all the editednagR BEstrains could grow on glucose, while the wild-type strain couldn’t
(Figure S5b). Meanwhile, two double and one triple-locus edited strains
(nagR /P BE,nagR /X BE andnagR /X/P BE) based on the strainnagR BE were also constructed using the pCBEso
system to confirm whether NagP or NagX could transport glucose inS. oneidensis MR-1. The results suggest that NagP, rather than
NagX, was the major glucose transporter in the strainnagR BE (Figure 5d). Interestingly, the deletion
of nagK slightly inhibit the growth of the strainnagR BE on glucose, demonstrating that other
unknown kinases played key roles in glucose metabolism of the strainnagR BE.
3.6 | Degradation of organic pollutants by the
edited strain nagRBE
Firstly, several organic pollutants were selected to evaluate the
degradation
capacity of the strain nagRBE using glucose as
the sole carbon source. MO and amaranth, two typical azo dyes, could be
anaerobically reduced by S. oneidensis MR-1 when lactate was used
as the sole carbon source (Cai et al., 2012; Hong et al., 2007). HPLC
analysis shows that either MO or amaranth could be degraded by the
edited strain nagR BE when glucose was used as the
sole carbon source (Figure S6). Moreover, the strainnagR BEgrown on glucose exhibited a faster MO or amaranth degradation than the
wild type strain (Figure 6). The MO or amaranth degradation capacity by
the wild type strain in mineral medium
supplemented
with glucose as the sole carbon source might be associated with the
accumulation of carbon source in the inoculated culture. Meanwhile, the
MO and amaranth degradation rates were also compared by calculating the
first-order rate constants (k ) (Figure 6). The k values of
MO and amaranth degradation by the strain nagR BE(0.493 h-1 and 0.306 h-1) were 11.7
and 7.3 times higher than that of the wild type strain (0.042
h-1 and 0.042 h-1), respectively.
Previous studies demonstrated that the
roxarsone,
an organoarsenic compound, could be anaerobically degraded by S.
oneidensis MR-1 when lactate was used as the sole carbon source (Han et
al., 2017). HPLC analysis demonstrates that HAPA(V) was the main
reduction product of roxarsone by the strainnagR BE when glucose was used as the sole carbon
source (Figure S6). Moreover, the k value of roxarsone reduction
bynagR BE(0.044 h-1) was 2.4-folds higher than that of the wild
type strain (0.018 h-1) (Figure 6c). Overall, base
editing endowed the strain nagR BE with the
ability for pollutant degradation using glucose as the sole carbon
source.
The increased expression levels of the genes associated with the GlcNAc
metabolism in
the
strain nagR BE imply that such an edited strain
might have a faster utilization rate for GlcNAc and thus exhibit a
higher pollutant degradation capability than S. oneidensis MR-1
when
GlcNAc was used as the sole carbon source. Expectably, the engineered
strain nagR BE exhibited higher degradation rates
for all three organic pollutants than the wild type strain (Figure 7).
Meanwhile, the k values of MO, amaranth, and roxarsone
degradation by the strain nagR BE (0.711
h-1,
0.469 h-1, and 0.089 h-1) were
27.3-, 12.7-, and 6.4-folds higher than that of the wild type strain
(0.026 h-1, 0.037 h-1, and 0.014
h-1), respectively. Taken together, the base editing
of nagR in S. oneidensis MR-1 not only expanded its carbon
source utilization spectra, but also accelerated its degradation
efficiencies for organic pollutants when GlcNAc was used as the sole
carbon source.
4 | DISCUSSION
CRISPR/Cas9-mediated genome editing systems have been widely used in
microbes (Choi & Lee, 2016; Hong, Zhang, Cui, Wang, & Wang, 2018;
Jakociunas, Jensen, & Keasling, 2016; Wang, Dong, Wang, Tao, & Wang,
2017). Basically, these systems introduce a double-strand DNA break at a
target position and precise genome editing is achieved by homologous
recombination with a donor DNA template in the self-repairing process of
double-strand DNA break (Barrangou & Doudna, 2016; Choi & Lee, 2016).
However, the efficiency of homologous recombination for Cas9-mediated
double-strand DNA break repair in most of microbes is limited, which may
cause substantial cell death. Since cytidine deamination restrictedly
occurs in the displaced single-strand DNA, the
pCBEso
system developed in S. oneidensis MR-1 can rapidly mediate the
highly efficient conversion of target C to T without requiring
double-strand DNA break or a donor DNA template. In the
“all-in-one-plasmid”
system, a Golden Gate assembly strategy is adopted to facilitate
efficient cloning of the 20-bp spacer for rapid assembly of the targeted
base editing plasmid.
Thus,
the engineered strain can be successfully constructed by the pCBEso
system
within
four days.
In
addition, multiplex genome editing of two or more loci in S.
oneidensis MR-1 can be accomplished by the pCBEso system. All these
features enable the pCBEso systemas a rapid, highly efficient and
readily tuned approach for the identification of gene functions inS. oneidensis MR-1 and the robust construction of engineered
strains.
Efforts have been made to broaden the spectra of carbohydrates as
available carbon sources of S. oneidensis MR-1. For example, the
expose to glucose enabled few S. oneidensis MR-1 cells to use
glucose as the sole carbon source through an unknown mutation (Howard,
Hamdan, Lizewski, & Ringeisen, 2012). A similar evolution strategy was
used to construct a
xylose-utilizing
strain by activating the silent xylose-related pathway (Sekar, Shin, &
DiChristina, 2016). An engineered Escherichia coli capable of
utilizing xylose was introduced as a fermenter to synthesize lactate,
which could be used byS.
oneidensis MR-1 as carbon source (Yang et al., 2015). Recently, the
synthetic biology strategies were adopted to engineer S.
oneidensis MR-1 to respectively utilize glycerol, glucose, or xylose as
the sole carbon source via heterologous expressions of modules
correlated to carbohydrate metabolism (Choi et al., 2014; Flynn, Ross,
Hunt, Bond, & Gralnick, 2010; Li et al., 2017). Although these
approaches are feasible, they either suffer from a low operation
efficiency, or are labor intensive and time consuming. In our work, to
fully exploit its intrinsic and latent carbohydrate metabolic capacity
in S. oneidensis MR-1, the functions of the key genes involved in
GlcNAc or glucose metabolism were quickly identified and validated by
the pCBSso system. Moreover, an engineered S. oneidensis MR-1
capable of utilizing glucose was rapidly constructed, which exhibited a
higher capability of biodegrading multiple organic pollutants than the
wild type strain using glucose or GlcNAc as the sole carbon source.
Despite of the above advantages, one drawback of this editing tool is
the target coverage. In few non-essential genes of S. oneidensisMR-1, there are no editable Cs by the pCBEso system.
To
expand the target scope, several C-to-T base editors using Cas9 variants
with different PAMs (such as VQR-BE3 and EQR-BE3) can be introduced inS. oneidensis MR-1 (Kim et al., 2017). Another possible solution
to this problem is that Cas9n (D10A) with NGG in pCBEso system can be
replaced by other CRISPR effectors with diverse PAMs like Cpf1 with
BTTV, CasX with TTCN, and so on (Burstein et al., 2017; Yang, Gao,
Rajashankar, & Patel, 2016; Zetsche et al., 2015). Editing preference
of the pCBEso system (poor conversion of target C within the GC motifs)
is another shortcoming. Evolved cytosine base editors like
evoAPOBEC1-BE4max and evoFERNY with
highly efficient editing of target C within the GC context has been
developed to overcome the editing preference of conventional cytosine
base editors (Thuronyi et al., 2019). Recently, to
write
new genetic information at targeted sites in cells, a
revolutionary
genome editing tool (prime editing), targeting by a specific editing
guide RNA, was developed by fusing a Cas9 variant with the engineered
reverse transcriptase (Anzalone et al., 2019). Overall, these approaches
would further expand the scopes and capabilities of base editing inS. oneidensis MR-1.
5 | CONCLUSIONS
In this work, we develop a rapid, highly efficient and readily tuned
base editing system-pCBEso inS.
oneidensis MR-1. Unlike other genome editing tools, the pCBEso system
can achieve efficient genome editing without requiring a double-strand
DNA break or repair templates. Such an effective editing system is
applied to identify the key genes in GlcNAc or glucose utilization inS. oneidensis MR-1. Moreover, we rapidly construct an engineered
strain for more efficient biodegradation of organic pollutants than the
wild type when glucose or GlcNAc was used as the sole carbon source. In
addition to the gene inactivation by introducing a premature stop codon,
the pCBEso system
has
a great potential in repairing the undesired point mutation, engineering
proteins by replacing key amino acid residues in vivo , and
enhancing electron transfer pathways by multiplex genome editing inS. oneidensis MR-1 and other EEB. Such an efficient base editing
system will promote the application of EEB in environmental remediation.