Accepted Manuscript

Title: Construction of a high-efficiency cloning system using

the Golden Gate method and I-SceI endonuclease for targeted
gene replacement in Bacillus anthracis

Authors: Tiantian Wang, Dongshu Wang, Yufei Lyu, Erling

Feng, Li Zhu, Chunjie Liu, Yanchun Wang, Xiankai Liu,
Hengliang Wang

PII: S0168-1656(18)30050-6
DOI: https://doi.org/10.1016/j.jbiotec.2018.02.006
Reference: BIOTEC 8117

To appear in: Journal of Biotechnology

Received date: 24-5-2017

Revised date: 8-2-2018
Accepted date: 9-2-2018

Please cite this article as: Wang T, Wang D, Lyu Y, Feng E, Li Z, Liu C, Wang Y, Liu
X, Wang H, Construction of a high-efficiency cloning system using the Golden Gate
method and I-SceI endonuclease for targeted gene replacement in Bacillus anthracis,
Journal of Biotechnology (2010), https://doi.org/10.1016/j.jbiotec.2018.02.006

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 Running title: Constructing high-efficiency cloning and gene replacement system
in B. anthracis

Construction of a high-efficiency cloning system using the Golden

Gate method and I-SceI endonuclease for targeted gene

replacement in Bacillus anthracis

T
IP
Tiantian Wang, Dongshu Wang, Yufei Lyu, Erling Feng, Li Zhu, Chunjie Liu,

R
Yanchun Wang*, Xiankai Liu*, Hengliang Wang*

SC
State Key Laboratory of Pathogens and Biosecurity, Beijing Institute of Biotechnology, 20

Dongdajie Street, Fengtai District, Beijng 100071, China

U
N
List of the authors’ emails:
A
Tiantian Wang: con_an@126.com
Dongshu Wang: wangdongshu@foxmail.com
M

Yufei Lyu: flygogo.cool@163.com

Erling Feng: fengel@sohu.com
ED

Li Zhu: jewly54@126.com
Chunjie Liu: liucj@nic.bmi.ac.cn
PT

*Corresponding authors:
Yanchun Wang: springwyc@gmail.com
Xiankai Liu: liuxk007@163.com
E

Hengliang Wang: wanghl@nic.bmi.ac.cn

CC

Highlights

 The efficiency of using Golden Gate method to build knockout

A

vector is more than 85%.

 2.The expression of I-SceI endonuclease promoted homologous
recombination. The allelic-exchange schemes driven by I-SceI
nuclease at rates of 85%–100%.
1
  3.The Golden Gate method and the counterselectable marker
were combined innovatively.

Abstract

To investigate gene function in Bacillus anthracis, a high-efficiency cloning

system is required with an increased rate of allelic exchange. Golden Gate cloning

T
is a molecular cloning strategy allowing researchers to simultaneously and

IP
directionally assemble multiple DNA fragments to construct target plasmids using

R
type IIs restriction enzymes and T4 DNA ligase in the same reaction system. Here,

SC
a B. anthracis S-layer protein EA1 allelic exchange vector was successfully

U
constructed using the Golden Gate method. No new restriction sites were
N
introduced into this knockout vector, and seamless assembly of the DNA
A
fragments was achieved. To elevate the efficiency of homologous recombination
M

between the allelic exchange vector and chromosomal DNA, we introduced an

I-SceI site into the allelic exchange vector. The eag gene was successfully
ED

knocked out in B. anthracis using this vector. Simultaneously, the allelic exchange

vector construction method was developed into a system for generating B.

PT

anthracis allelic exchange vectors. To verify the effectiveness of this system,

E

some other allelic exchange vectors were constructed and gene replacements were
CC

performed in B. anthracis. It is speculated that this gene knockout vector

construction system and high-efficiency targeted gene replacement using I-SceI

A

endonuclease can be applied to other Bacillus spp.

Keywords: Bacillus anthracis;Golden Gate; allelic exchange; I-SceI

2
1. Introduction

DNA recombination technology emerged in the early 1970s. Restriction

T
digestion and ligation is the most commonly used cloning method in the

IP
construction of recombinant DNA molecules. This method is one of the most

basic molecular biology techniques, in which restriction endonucleases and DNA

R
SC
ligase are used to clone the gene(s) of interest into a vector. However, this method

has its limitations. Specifically, when multi-target DNA fragments are ligated,

U
they then need to be inserted into the vector step-by-step, which can be a tedious
N
and time-consuming procedure; in addition, restriction endonuclease sites are
A
retained between target DNA fragments in the final construct (Goff and Berg,
M

1976).

Approximately 20 years ago, the development of site-specific recombination

ED

cloning technology provided a new, rapid, and effective method of constructing

DNA recombinants in vitro, such as the Gateway cloning system developed by

PT

Invitrogen (Hartley et al., 2000; Walhout et al., 2000), the Creator cloning system
E

by Clontech (Sternberg et al., 1981), and the Univector cloning system by Stephen
CC

Elledge’s laboratory (Liu et al., 1998; Liu et al., 2000). With these cloning

techniques, target fragments can be simply, efficiently, and accurately fused to

A

specific vectors. However, specific recombination site sequences are retained and

cannot be eliminated from the constructs developed using this cloning method, so

extra amino acids are added into the expression products if these sites are located

in an expression sequence.

3
 In 2008, Engler et al. reported a cloning strategy mediated by the type IIs

restriction enzyme, known as the IIs cloning method or the Golden Gate cloning

method (Engler et al., 2008; Engler et al., 2014), which has been applied in many

fields. In 2011, Weber and colleagues used the Golden Gate cloning method to

create three dTALEs to activate a reporter construct (Weber et al., 2011). In 2014,

Terfrüchte and coworkers established a versatile Golden Gate cloning system for

T
genetic engineering in fungi (Terfrüchte et al., 2014), and Engler and coworkers

IP
used the Golden Gate cloning method to construct multigene constructs for plant

R
transformation (Engler et al., 2009). In 2016, Vad-Nielsen and colleagues used the

SC
Golden Gate cloning method to assemble a CRISPR gRNA expression array to

U
simultaneously target multiple genes (Vad-Nielsen et al., 2016). However, this
N
method has not yet been applied for constructing the allelic exchange vectors used
A
in Bacillus anthracis.
M

The method of gene knockout in B. anthracis is based on the principle of

homologous recombination, which usually uses a suicide plasmid to construct an

ED

allelic exchange vector carrying the upstream part of a target gene, an antibiotic

resistance cassette, and the downstream part of the target gene. The most
PT

commonly used method to construct B. anthracis homologous recombination

E

vectors is restriction digestion and ligation, which can involve multiple steps and
CC

can be tedious and time-consuming. In this study, we attempted to use the Golden

Gate cloning method to build a knockout vector in B. anthracis, thereby allowing

A

high-efficiency one-step subcloning, without the inclusion of any additional

nucleotides in the final cloned product. This greatly improves the efficiency of

producing the allelic exchange vector, and we believe that this system could also

be applied to construct allelic exchange vectors for other Bacillus spp., such as B.

4
subtilis, B. cereus, and B. thuringiensis.

For B. anthracis, gene replacements for some loci have been reported, but the

methods used to obtain the desired gene replacement strains are often

time-consuming and labor-intensive. This is owing to the lack of a

counterselection scheme to deal with the low efficiency of homologous

recombination between the target plasmids and the chromosomal DNA in B.

T
anthracis. An alternative to such counterselection schemes involves the use of the

IP
intron-encoded homing restriction enzyme I-SceI, which can cleave a unique

R
introduced site in a genome; this property has been exploited in the promotion of

SC
homologous recombination in organisms as diverse as bacteria, Drosophila, and

U
other higher eukaryotes (Choulika et al., 1995; Rong et al., 2002; Schmidt-Puchta
N
et al., 2004). In 2006, Stibitz and colleagues used the I-SceI enzyme to promote
A
allelic exchange in B. anthracis and, in 2015, this group improved this method
M

further (Janes and Stibitz, 2006; Schuch et al., 2015). In this report, we describe

further modification of this methodology using the I-SceI enzyme to facilitate

ED

homologous recombination in B. anthracis. We introduced an I-SceI site into our

allelic exchange vector as a counterselection marker. This greatly elevated the

PT

efficiency of homologous recombination in our study.

E
CC

2. Materials and methods

A

2.1. Materials

2.1.1. Bacterial strains and plasmids

Plasmid pKSV7 contains temperature-sensitive replication functions derived from

pE194ts (permissive temperature, 30°C; restrictive temperature, 37°C and above; Table

5
1). It also contains an ampicillin resistance gene for selection in Escherichia coli, and a

chloramphenicol resistance gene for selection in B. anthracis. The pMAD vector contains

a bgaB gene encoding a thermostable -galactosidase from Bacillus stearothermophilus,

which is expressed from a constitutive promoter recognized in Gram-positive and

Gram-negative bacteria. The pSET4s vector contains a spectinomycin resistance gene.

This vector could be propagated at 37°C in E. coli. The pGEM-T easy vector is a

T
linearized vector with a single 3ʹ-terminal thymidine at both ends. B. anthracis vaccine

IP
strain A16R (pXO1+, pXO2−; Table 1) was derived from strain A16 by UV irradiation.

R
SC
2.1.2. Enzymes and biochemical reagents
U
2xEs Taq Master Mix (Dye) (Beijing ComWin Biotech, Beijing, China),
N
Cobuddy High-Fidelity Rapid DNA polymerase (Beijing ComWin Biotech), Fast
A

Digest EcoRI (Thermo Fisher Scientific, Rockford, IL, USA), Fast Digest BamHI
M

(Thermo Fisher Scientific), T4 DNA ligase (Thermo Fisher Scientific), BsaI (New
ED

England Biolabs, USA), and ATP (10 mM; New England Biolabs, Beverly, MA,

USA) were used in this study.

PT

2.1.3. Primers
E
CC

The primers used in this study are listed in Table 2. Primers were designed

according to the DNA sequences of target gene fragments and plasmids, with the
A

addition of BsaI restriction sites (lowercase italic letters) and other required

restriction sites (underlined).

6
2.2. Construction of the entry vectors of pGEM-U, pGEM-D,

and pGEM-S

The sequences of eagup (947 bp) and eagdn (811 bp), that is, the upstream

T
and downstream parts of the eag gene sequence, were amplified with two pairs of

IP
specifically designed primers, eagup_F/R and eagdn_F/R, using the B. anthracis

R
strain A16R genome as a template. The spectinomycin resistance gene (spc r, 1155

SC
bp) was amplified using the primer pair spc_F/R and the pSET4s plasmid as a

template. Each of the three DNA fragments was flanked by two BsaI sites.
U
The three PCR-amplified fragments eagup, eagdn and spc were ligated to the
N
pGEM-T easy vector using T4 DNA ligase, resulting in three entry vectors
A

designated pGEM-U, pGEM-D and pGEM-S, respectively.

M
ED

2.3. Construction of the acceptor vector of pKMBKI

To successfully implement the Golden Gate cloning protocol and introduce a

PT

counterselectable marker, the shuttle plasmid pKSV7 was first modified. An

E

internal BsaI site in the pKSV7 plasmid was destroyed by the introduction of a
CC

point mutation to prevent unwanted plasmid cleavage in the subsequent digestion,

giving plasmid pKMU7.

A

The exogenous 2259-bp fragment of pclpB-bgaB was amplified by PCR

from pMAD plasmid DNA using the bgaB_F/R primer pair and the Cobuddy

high-fidelity enzyme. Furthermore, the exogenous 1315-bp fragment of Tn5

neomycin phosphotransferase (containing the kanamycin resistance gene, kanr)

7
was amplified by PCR from pKD4 plasmid using the kan_F/R primer pair. Then,

the pclpB-bgaB and kanr DNA fragments were cloned between EcoRI and BamHI

restriction sites in pKMU7 using the Seamless Cloning Kit (Beijing Taihe

Biological Technology, Beijing, China), resulting in the vector pKMBK.

The I-SceI site was amplified using the primers I-SceI_F/R and inserted into

the SalI site of pKMBK using the Gibson assembly method, resulting in the

T
acceptor vector pKMBKI (Fig. 1).

R IP
2.4. Subcloning the entry DNA fragments into the acceptor

SC
vector pKMBKI to construct the recombinant allelic

exchange vector U
N
The scheme for subcloning the entry DNA fragments into the acceptor vector
A

pKMBKI is illustrated in Fig. 2. First, the concentration of each plasmid

M

(pGEM-U, pGEM-S, pGEM-D and pKMBKI) was measured using a NanoDrop

ED

spectrophotometer. The reaction mixture comprised: 50 ng of each entry vector

and the acceptor vector together with 2.5 units of BsaI enzyme, 2.25 units of T4
PT

DNA ligase, 1 μl of CutSmart buffer, 1 μl of ATP (10 mM), and ddH 2O up to a

final volume of 10 μl. The restriction–ligation procedure was performed in

E

accordance with the following program: The mixture was incubated in a water
CC

bath at 37°C for 30 min, followed by heating at 50°C for 5 min to redigest and
A

eliminate any plasmid that might still contain a BsaI restriction site and then 80°C

for 5 min to inactivate both restriction enzyme and ligase. A 5-μl

restriction–ligation product was then added to 50 μl of DH5 chemically

competent cells, which were subsequently incubated in an ice-bath for 30 min,

8
followed by a 90-s thermal shock at 42°C, then another 2 min in the ice-bath,

followed by 1-h incubation at 30°C with shaking at a speed of 225 r/min. Then,

300 μl of bacterial culture was plated on Luria–Bertani (LB) agar plates

containing X-Gal (40 μg/ml) and kanamycin (50 μg/ml) and the plates were

incubated at 30°C overnight. Twenty white colonies were picked, streaked onto

LB agar plates, and incubated at 30°C for 10 h. Single clones were analyzed by

T
colony PCR to determine whether the three entry DNA fragments had been

IP
inserted into the acceptor vector pKMBKI using three pairs of primers: eagup_F/R,

R
eagdn_F/R and spcin_F/R. The recombinant plasmid was confirmed by DNA

SC
sequencing using an ABI Prism Model 3730XL DNA analyzer (Applied

U
Biosystems, Carlsbad, CA, USA) and the resulting plasmid was designated
N
pKMUSD (Fig. 2).
A
M

2.5. Measurement of the effects of restriction enzyme I-SceI

on the elimination kinetics of the allelic exchange vector

ED

pKMUSD
PT

For plasmid elimination, B. anthracis strain A16R (pKMUSD) was grown in

LB medium without kanamycin and A16R (pKMUSD+pSS4332) was grown in

E

LB medium with kanamycin at 42°C for 12 h. Then, the samples were diluted and
CC

plated onto LB agar with or without spectinomycin. The plates were incubated at
A

30°C until colonies appeared. Both strains were passaged up to four times in the

same medium and the plasmid elimination rates of each generation were analyzed

as described previously (Wolfson et al., 1982).

9
2.6. Deletion of the eag gene from B. anthracis

The allelic exchange vector pKMUSD extracted from DH5 was introduced

into the dam−dcm− E. coli host strain JM110 by electroporation (1.8 kV, 25 μF, 200

Ω) using a Gene Pulser II electroporator (Bio-Rad, Hercules, CA, USA) to obtain

demethylated plasmid DNA for transformation into B. anthracis strain A16R.

T
Electroporation-competent cells of B. anthracis strain A16R were prepared as

IP
described previously. Then, 5 l of plasmid DNA (1 g in water) was added to a

50-l aliquot of competent cells and pulsed (2.5 kV, 25 F, 200 ) in a 0.2-cm

R
SC
gap cuvette (Shatalin and Neyfakh, 2005). The cells were resuspended in 1 ml of

brain–heart infusion supplemented with 0.5% glycerol and incubated for 3 h at

U
30C with aeration. The recovered cells were spread onto LB agar plates
N
containing spectinomycin. Spectinomycin-resistant colonies were picked, and
A
PCR was performed using the primer pair pKSV7_F/R to confirm that
M

recombinant pKMUSD had been introduced into B. anthracis strain A16R. This

strain was then electroporated with pSS4332, which contains the gene encoding
ED

the I-SceI enzyme. The transformants were selected at 30°C on LB medium

PT

containing spectinomycin and kanamycin. Fluorescent colonies carrying pSS4332

were visible after 16–20 h.

E

B. anthracis mutants were constructed by replacing coding sequences with

CC

the spcr cassette, which confers spectinomycin resistance. Transformation and the

selection of transformants in B. anthracis were conducted as described previously.

A

A single colony of A16R (pKMUSD+pSS4332) was transferred into liquid

medium with kanamycin and incubated with shaking for 8 h at 30°C; then, the

cultures were shifted to 37°C for 8 h. After three passages in the same medium,

serial dilutions of the culture were plated onto LB agar plates containing
10
spectinomycin and kanamycin, and incubated overnight at 37°C. At least 50

colonies were screened for spectinomycin resistance and chloramphenicol

sensitivity (SpcrCms). Colonies (SpcrCms) thought to have undergone a

double-crossover recombination event were validated by PCR analysis using three

pairs of primers, eagus_F/R, eagsd_F/R and eagin_F/R, and DNA sequencing.

Strains with the desired gene replacement were passaged up to three times in the

T
absence of antibiotics to ensure the loss of pSS4332, which was confirmed by the

IP
loss of fluorescence. The eag gene deletion was further verified by western blot

R
analysis using an antibody against EA1 protein. The resulting colony was

SC
designated A16R△eag::spc.

U
N
3. Results and Discussion
A

3.1 Efficiency of Golden Gate restriction–ligation and

M

identification of the recombinant allelic exchange vector

ED

When the acceptor vector pKMBKI was introduced into E. coli DH5α cells,

all transformants were blue because of the reaction between the -galactosidase
PT

expressed by the pclpB–bgaB gene cassette on the acceptor vector pKMBKI and
E

the X-Gal on the plate. When the Golden Gate restriction–ligation product was
CC

transformed into E. coli DH5α cells, almost all of the transformants were white,

indicating that the three entry DNA fragments had been inserted into acceptor
A

vector pKMBKI in the designed order, replacing the original pclpB–bgaB gene

cassette on the acceptor vector. This indicated that this cloning system had high

efficiency and the blue/white colors enabled easy screening of the recombinant

clones.

11
 Twenty white colonies were picked up and analyzed by colony PCR using

three pairs of primers, eagup_F/R, eagdn_F/R and spcin_F/R. In these reactions,

genomic DNA of B. anthracis strain A16R and plasmid DNA of pSET4s were

used as positive control templates to amplify the DNA fragments of eagup, eagdn

and spcin. Genomic DNA of strain DH5α was used as a negative control template

in all tests. The results showed that all of the colonies except for No. 8 and No. 20

T
contained the correct eagup amplified fragment (Fig. 3A). For the spcin DNA

IP
fragment, 19 of the 20 colonies contained the correct amplified DNA bands (Fig.

R
3B). For the eagdn DNA fragment, 18 of the 20 white colonies contained the

SC
correct amplified DNA bands (Fig. 3C). In total, 17 of the 20 colonies possessed

U
the desired ligation of the three entry fragments eagup, spc and eagdn into the
N
acceptor vector pKMBKI, giving a success rate of 85% (Fig. 3).
A
To further test the restriction–ligation efficiency, two types of
M

restriction–ligase buffer, namely, ligase buffer from Promega and ligase buffer

from New England Biolabs, were also tested. The amount of T4 DNA ligase used
ED

in the reaction was also tested in the range of 1.125–2.25 units. Although

increasing the amount of T4 DNA ligase slightly enhanced the restriction–ligation

PT

efficiency, no significant difference was detected between any of the treatments,

indicating a degree of flexibility in the choice of ligase buffer and amount of

E
CC

ligase used in this restriction–ligation cloning system. Engler et al. had already

proved that incubation for 30 min was optimal for experiments in which three
A

inserts were inserted into an acceptor vector. The ratios of white/blue colonies

were 24/0 and 27/1 in their two repeated trials under these conditions (Engler et

al., 2008), which are similar to our results.

Regarding the conventional process to construct an allelic exchange vector,

12
the upstream and downstream regions of the target gene, the antibiotic resistance

gene and the acceptor vector are subjected to a series of enzyme digestion and

ligation steps to insert these three DNA fragments into the acceptor vector, which

could be time-consuming and tedious, until type IIs restriction enzymes started to

be used to design the linker of ligation. Type IIs restriction enzymes can

specifically identify recognition sites and nonspecifically cleave the DNA double

T
strand outside of the recognition sequences, resulting in 5′ or 3′ overhangs that

IP
may consist of any nucleotide. The 4-nt overhang can be designed and several

R
DNA fragments with complement overhang ends can be ligated using a ligase

SC
according to an established order into a vector without restriction enzyme

U
recognition sites. These properties have been used to develop protocols for the
N
efficient assembly of multiple DNA fragments in a single seamless ligation
A
reaction (Szybalski et al., 1991). Engler and colleagues developed the Golden
M

Gate cloning strategy based on the use of type IIs restriction enzymes, which

allowed one-step subcloning of a DNA fragment from one plasmid to another with
ED

high efficiency (Engler et al., 2008; Engler et al., 2014). Since then, the Golden

Gate cloning strategy has been applied in many microorganisms to subclone

PT

multiple DNA fragments from multiple entry plasmids into an acceptor vector in a

one-step reaction (Agmon et al., 2015; Terfrüchte et al., 2014). This prompted us
E
CC

to construct a multiple DNA fragment subcloning system to be used for allelic

exchange in B. anthracis. We attempted to modify the temperature-sensitive E.

A

coli–B. subtilis shuttle vector pKSV7 into an acceptor vector to facilitate the

screening of transformants (Smith and Youngman, 1992). First, we analyzed the

sequence of pKSV7 and found an internal BsaI site. Then, we mutated the BsaI

site using a point mutation kit, resulting in pKMU7 lacking this site. To enable

13
blue/white colony screening, a promoterless bgaB gene from Bacillus

stearothermophilus encoding a thermostable -galactosidase and a constitutive

pclpB promoter was amplified from the pMAD plasmid and inserted into the

pKMU7 plasmid (Arnaud et al., 2004). The BsaI restriction sites flanking

pclpB–bgaB would be recognized and cleaved by the BsaI restriction enzyme,

allowing them then to accept a set of entry DNA fragments designed to have

T
compatible 4-nt restriction overhangs. To enhance the restriction–ligation

IP
efficiency, the 4-nt restriction overhangs were carefully designed. Theoretically,

R
256 different overhangs can be created using a type IIs restriction endonuclease

SC
that produces a 4-nt overhang, 16 of which are palindromic sequences that were

U
excluded in our cloning strategy to reduce the possibility of self-ligation, leaving
N
240 possible 4-nt overhangs from which selection could be made. For each
A
additional site to be selected, the 4 nt of previously selected sites or their
M

complement should be excluded. In this study, we applied strict criteria and

designed a set of recombination sites in which none of the sites shared three
ED

consecutive nucleotides with any other selected site, and thereby achieved

seamless ligation without introducing any new restriction sites in a single step and
PT

a single tube.
E

To enable antibiotic screening of the transformants, the entry vector and the
CC

acceptor vector should contain different antibiotic resistance genes. In this study,

the pGEM-T easy vector used as the entry vector and the pKSV7 used as the
A

acceptor vector contained the same ampicillin resistance gene, which made it

difficult to screen the recombinant transformants. Therefore, we amplified the Tn5

neomycin phosphotransferase-encoding kanamycin resistance gene from the

pKD4 plasmid and inserted it into pKMU7 adjacent to plcpB–pgaB, resulting in

14
the acceptor vector pKMBK.

When applying this cloning system for high-efficiency gene replacement in B.

anthracis described here, the only requirement is that the upstream and

downstream regions of the target gene are inserted into the pGEM-T easy vector.

This greatly improves the efficiency of producing the allelic exchange vector. In

our study, we applied this system to construct a range of allelic exchange vectors

T
for B. anthracis, all of which were successfully implemented. Because the

IP
backbone plasmid pKSV7 is an E. coli–B. subtilis shuttle vector, we believe that

R
this system could also be applied to construct allelic exchange vectors for other

SC
Bacillus spp., such as B. subtilis, B. cereus and B. thuringiensis. We also believe

U
that, through a design with a series of flexible 4-nt overhangs, more than three
N
DNA fragments could also be easily inserted into the pKSV7.
A
M

3.2 Loss of effectiveness of the allelic exchange vector using

I-SceI endonuclease and verification of eag gene replacement

ED

with the spcr cassette

PT

We tested the elimination rate of the allelic exchange vector when I-SceI

endonuclease was expressed in the host strains. The kinetics of allelic exchange
E

vector pKMUSD elimination from B. anthracis strain A16R was analyzed

CC

according to a previously published method (Wolfson et al., 1982). The expression

A

of I-SceI endonuclease resulted in rapid elimination of allelic exchange vector

pKMUSD. The rate of elimination of pKMUSD after one passage was 97.5%

when pSS4332 was present, whereas >50% of bacteria contained vector

pKMUSD when pSS4332 was absent (Fig. 4). This suggested that allelic

exchange between the allelic exchange vector pKMUSD and the genome of the
15
host is greatly accelerated.

Next, we used this allelic exchange vector to knock out the eag gene on the

chromosome of A16R. PCR analysis was performed using two pairs of external

identification primers eagus_F/R and eagsd_F/R and one pair of internal

identification primers eagin_F/R (Table 1, Fig. 5A) to determine whether the eag

gene had been replaced by the spcr cassette. The results indicated that the eag

T
gene was present in wild-type strain A16R but not in A16Reag::spc. The two

IP
pairs of external primers specifically amplified PCR bands in A16Reag::spc (Fig.

R
5A). Sequencing data obtained using the PCR fragments amplified with the two

SC
external primers were consistent with the eag gene being replaced by the spcr gene

U
(Fig. 5A). Western blot analysis using the anti-EA1 antibody indicated that there
N
was a specific protein band in the B. anthracis A16R strain that was not present in
A
A16R△eag::spc. Taken together, these results indicated that the eag gene had been
M

replaced by the spcr gene (Fig. 5B).

In 2006, Brian K. and coworkers deleted the plcR, pagA, lef, cya and spo0A
ED

genes of B. anthracis using allelic-exchange schemes driven by I-SceI nuclease at

rates of 20%–100% (Janes and Stibitz, 2006; Tischer et al., 2006). Using the
PT

protocol developed in this study, the ideal mutants were obtained after only three
E

passages by screening 50 clones. Three independent experiments indicated that

CC

approximately 85%–100% (43/50, 45/50, 50/50) of the spcrcms clones had eag

replaced by the spcr cassette in the correct manner. We also deleted several other B.
A

anthracis genes, such as BA2380, atxA, mntA, qoxA, abrB, and sigF, using this

method, with a rate of allelic exchange similar to that in the case of eag deletion.

That our method has a higher rate of allelic exchange of the spc rcms clones may be

because we used the double-antibiotic screen.

16
 For Gram-positive bacteria, homologous recombination, transposition

mutagenesis, and group II intron-based insertion mutagenesis have been used for

gene inactivation. However, the most commonly used method of mutagenesis is

homologous recombination. This is usually a rare event, so strategies to enhance

the efficiency of homologous recombination are vital (Janes and Stibitz, 2006;

Schuch et al., 2015; Shatalin and Neyfakh, 2005). Many methods have been

T
proposed that artificially increase the efficiency of endogenous homologous

IP
recombination. The success of the λ Red recombination system for

R
recombineering in E. coli provided the basis for initial experimental designs

SC
(Datsenko and Wanner, 2000; Tischer et al., 2006). Using

U
mycobacteriophage-encoded recombination proteins, a recombineering system for
N
mycobacteria was developed based on the RecE/T system (Van Kessel and Hatfull,
A
2007). The same protocol has also been applied in lactic acid bacteria and B.
M

subtilis (Van Pijkeren and Britton, 2012; Van Pijkeren et al., 2012; Wang et al.,

2012). However, for B. anthracis, there is no effective method for transforming

ED

single-stranded DNA into competent cells and then ensuring its stability in B.

anthracis cells. Thus, it is difficult to use the RecE/T system in this case. As such,
PT

research has focused on increasing the allelic exchange efficiency between the

chromosome and the target plasmid. Plasmids usually used in B. anthracis based
E
CC

on the widely used pE194ts replicon (e.g., pMAD, pKSV7) were able to replicate

in B. anthracis, but curing was inefficient even at the highest temperature

A

tolerated by this bacterium (44C), at which more than 50% of the cells retained

the plasmid in free form. The pWV01 replicon-based plasmids also require more

than 20 passages to lose the target plasmids, which is impractical and even then

does not guarantee success (Shatalin and Neyfakh, 2005). Thus, maintenance at a

17
nonpermissive temperature is critical during the passage of cells in the absence of

antibiotic selection conferred by the plasmid. The approach to facilitating this

recombination event involves the use of site-specific cleavage, mediated by the

restriction enzyme I-SceI, of a cognate I-SceI site present in the allelic exchange

vector but not elsewhere in the bacterial chromosome. This method was originally

used in E. coli, and adapted for use in B. anthracis by Stibitz and coworkers

T
(Janes and Stibitz, 2006; Schuch et al., 2015).

IP
To address this problem, we introduced an I-SceI site into the allelic

R
exchange vector. After the allelic exchange vector was transformed into B.

SC
anthracis, a second plasmid, pSS4332, which expressed endonuclease I-SceI, was

U
introduced to digest the first plasmid resulting in a linear allelic exchange vector
N
during passaging. The linear allelic exchange vector was forced to incorporate
A
into the chromosome by growth at elevated temperatures concomitant with
M

selection for spectinomycin as a consequence of a first homologous recombination

event. The loss rate of the final allelic exchange vector improved significantly
ED

using this method. This method can also simulate gene targeting using linear DNA

fragments to some extent and increase the homologous recombination efficiency

PT

greatly. This approach will also facilitate the second recombination event by both

selecting against bacteria in which the chromosome has been damaged and
E
CC

stimulating recombination of the flanking homologous sequences to repair the

lesion. Repair of this break by a second recombination event results in loss of the
A

integrated plasmid and the desired gene replacement strains can then be obtained

by appropriate antibiotic resistance screening(Fig. 6).

In the present study, an I-SceI site, which was situated outside the two

homologous genomic sequences, was introduced into the target plasmid.

18
Recombination between the target plasmid and genomic DNA was forcibly

induced by three factors, namely, the induction of endonuclease I-SceI to initiate

site-specific cleavage, a nonpermissive temperature, and an antibiotic. Using this

strategy, eag and several other genes of B. anthracis were deleted extremely

efficiently within approximately 3 weeks with less screening work. B. anthracis, B.

weihenstephanensis, B. mycoides, B. pseudomycoides, B. cereus, and B.

T
thuringiensis all belong to the B. cereus group, which are found in diverse habitats.

IP
Although the B. cereus group is phenotypically and genetically heterogeneous

R
overall, strikingly, strains of the different constitutive species are closely related at

SC
the chromosomal level. Therefore, we speculated that the high-efficiency allelic

U
exchange driven by I-SceI endonuclease described here could be applied in other
N
species of the B. cereus group.
A
The flexibility, speed and efficiency of the multiple DNA fragment cloning
M

and gene replacement technique reported here should make it a valuable tool to

investigate gene function in B. anthracis and other species of the B. cereus group.
ED

Conclusion
PT

In this study, we constructed a high-efficiency cloning system using the

E

Golden Gate method and introduced an I-SceI site into our allelic exchange vector
CC

as a counterselection marker to elevate the homologous recombination efficiency

for targeted gene replacement in B. anthracis. The efficiency of using the Golden
A

Gate method to build a knockout vector is more than 85%. We proved that the

expression of I-SceI endonuclease greatly promoted homologous recombination.

eag and several other genes of B. anthracis were deleted using the technique

described here. That the Golden Gate method and the counterselectable marker

19
were combined innovatively greatly shortens the time required for gene knockout

in B. anthracis. It is speculated that this gene knockout vector construction system

and high-efficiency targeted gene replacement driven by I-SceI endonuclease can

be applied to other Bacillus spp.

Funding sources

T
This work was supported by the National Natural Science Foundation of
China (grant numbers 81571958, 81601739, and 81671979) and the State Major

IP
Science and Technology Special Projects of China (grant number

R
2017ZX10104002).

SC
Conflict of interest
U
N
The authors declare no conflicts of interest.
A
M

Acknowledgments
ED

We thank Liwen Bianji, Edanz Group China (www.liwenbianji.cn/ac), for

editing the English text of a draft of this manuscript.
E PT
CC
A

20
References
Agmon, N., L. Mitchell, Y. Cai, S. Ikushima, J. Chuang, A. Zheng, W. Choi, J. Martin, K. Caravelli, G.
Stracquadanio, and J. Boeke, 2015. Yeast Golden Gate (yGG) for the Efficient Assembly
of S. cerevisiae Transcription Units. ACS Synth Biol 4(7):853-859.
Arnaud, M., A. Chastanet, and M. Debarbouille, 2004. New vector for efficient allelic
replacement in naturally nontransformable, low-GC-content, gram-positive bacteria.
Appl Environ Microbiol 70(11):6887-6891.

T
Choulika, A., A. Perrin, B. Dujon, and J. Nicolas, 1995. Induction of homologous recombination
in mammalian chromosomes by using the I-SceI system of Saccharomyces cerevisiae.

IP
Mol. Cell. Biol. 15(4):1968-1973.
Cybulski, R.J., Jr., P. Sanz, F. Alem, S. Stibitz, R.L. Bull, and A.D. O'Brien, 2009. Four superoxide

R
dismutases contribute to Bacillus anthracis virulence and provide spores with
redundant protection from oxidative stress. Infect Immun 77(1):274-285.

SC
Datsenko, K., and B. Wanner, 2000. One-step inactivation of chromosomal genes in Escherichia
coli K-12 using PCR products. Proc. Natl. Acad. Sci. U.S.A. 97(12):6640-6645.
Engler, C., R. Gruetzner, R. Kandzia, and S. Marillonnet, 2009. Golden gate shuffling: a one-pot

U
DNA shuffling method based on type IIs restriction enzymes. PLoS ONE 4(5):e5553.
Engler, C., R. Kandzia, and S. Marillonnet, 2008. A one pot, one step, precision cloning method
N
with high throughput capability. PLoS ONE 3(11):e3647.
A
Engler, C., M. Youles, R. Gruetzner, T. Ehnert, S. Werner, J. Jones, N. Patron, and S. Marillonnet,
2014. A golden gate modular cloning toolbox for plants. ACS Synth Biol 3(11):839-843.
M

Goff, S., and P. Berg, 1976. Construction of hybrid viruses containing SV40 and lambda phage
DNA segments and their propagation in cultured monkey cells. Cell 9(4 PT 2):695-705.
Hartley, J., G. Temple, and M. Brasch, 2000. DNA cloning using in vitro site-specific
ED

recombination. Genome Res. 10(11):1788-1795.

Janes, B.K., and S. Stibitz, 2006. Routine markerless gene replacement in Bacillus anthracis.
Infect Immun 74(3):1949-1953.
PT

Liu, Q., M. Li, D. Leibham, D. Cortez, and S. Elledge, 1998. The univector plasmid-fusion system,
a method for rapid construction of recombinant DNA without restriction enzymes. Curr.
Biol. 8(24):1300-1309.
E

Liu, Q., M. Li, D. Liu, and S. Elledge, 2000. Rapid construction of recombinant DNA by the
CC

univector plasmid-fusion system. Meth. Enzymol. 328:530-549.

Rong, Y., S. Titen, H. Xie, M. Golic, M. Bastiani, P. Bandyopadhyay, B. Olivera, M. Brodsky, G.
Rubin, and K. Golic, 2002. Targeted mutagenesis by homologous recombination in D.
melanogaster. Genes Dev. 16(12):1568-1581.
A

Schmidt-Puchta, W., N. Orel, A. Kyryk, and H. Puchta, 2004. Intrachromosomal homologous

recombination in Arabidopsis thaliana. Methods Mol Biol 262:25-34.
Schuch, R., R.D. Plaut, and S. Stibitz, 2015. Improvements to a Markerless Allelic Exchange
System for Bacillus anthracis. PLoS ONE 10(12):e0142758.
Shatalin, K.Y., and A.A. Neyfakh, 2005. Efficient gene inactivation in Bacillus anthracis. FEMS
Microbiol Lett 245(2):315-319.
Smith, K., and P. Youngman, 1992. Use of a new integrational vector to investigate

21
 compartment-specific expression of the Bacillus subtilis spoIIM gene. Biochimie
74(7-8):705-711.
Sternberg, N., D. Hamilton, S. Austin, M. Yarmolinsky, and R. Hoess, 1981. Site-specific
recombination and its role in the life cycle of bacteriophage P1. Cold Spring Harb. Symp.
Quant. Biol. 45 Pt 1:297-309.
Szybalski, W., S. Kim, N. Hasan, and A. Podhajska, 1991. Class-IIS restriction enzymes--a review.
Gene 100:13-26.
Takamatsu, D., M. Osaki, and T. Sekizaki, 2001. Thermosensitive suicide vectors for gene
replacement in Streptococcus suis. Plasmid 46(2):140-148.
Terfrüchte, M., B. Joehnk, R. Fajardo-Somera, G. Braus, M. Riquelme, K. Schipper, and M.

T
Feldbrügge, 2014. Establishing a versatile Golden Gate cloning system for genetic

IP
engineering in fungi. Fungal Genet. Biol. 62:1-10.
Tischer, B., J. von Einem, B. Kaufer, and N. Osterrieder, 2006. Two-step red-mediated

R
recombination for versatile high-efficiency markerless DNA manipulation in Escherichia
coli. BioTechniques 40(2):191-197.

SC
Vad-Nielsen, J., L. Lin, L. Bolund, A.L. Nielsen, and Y. Luo, 2016. Golden Gate Assembly of CRISPR
gRNA expression array for simultaneously targeting multiple genes. Cell Mol Life Sci
73(22):4315-4325.

U
Van Kessel, J.C., and G.F. Hatfull, 2007. Recombineering in Mycobacterium tuberculosis. Nature
methods 4(2):147-152.
N
Van Pijkeren, J.P., and R.A. Britton, 2012. High efficiency recombineering in lactic acid bacteria.
Nucleic Acids Res 40(10):e76-e76.
A
Van Pijkeren, J.P., K.M. Neoh, D. Sirias, A.S. Findley, and R.A. Britton, 2012. Exploring
M

optimization parameters to increase ssDNA recombineering in Lactococcus lactis and

Lactobacillus reuteri. Bioengineered 3(4):209-217.
Walhout, A., G. Temple, M. Brasch, J. Hartley, M. Lorson, S. van den Heuvel, and M. Vidal, 2000.
ED

GATEWAY recombinational cloning: application to the cloning of large numbers of open

reading frames or ORFeomes. Meth. Enzymol. 328:575-592.
Wang, Y., J. Weng, R. Waseem, X. Yin, R. Zhang, and Q. Shen, 2012. Bacillus subtilis genome
PT

editing using ssDNA with short homology regions. Nucleic Acids Res 40(12):e91-e91.
Weber, E., R. Gruetzner, S. Werner, C. Engler, and S. Marillonnet, 2011. Assembly of designer TAL
effectors by Golden Gate cloning. PLoS ONE 6(5):e19722.
E

Wolfson, J., D. Hooper, M. Swartz, and G. McHugh, 1982. Antagonism of the B subunit of DNA
gyrase eliminates plasmids pBR322 and pMG110 from Escherichia coli. J. Bacteriol.
CC

152(1):338-344.
A

22
Figure captions
Fig. 1. Construction of the acceptor vector pKMBKI
A point mutation was introduced into plasmid pKSV7 to create plasmid pKMU7
lacking the BsaI site. Then, pKMU7 was digested with EcoRI and BamHI
enzymes. The pclpB-bgaB and kanr DNA fragments were amplified using the
primer pairs bgab_F/R and kan_F/R, respectively. Then, the pclpB-bgaB and kanr
DNA fragments were cloned between the EcoRI and BamHI restriction sites of

T
pKMU7 using the Seamless Cloning Kit, to create plasmid pKMBK. Finally, the

IP
I-SceI site was inserted into the SalI site of pKMBK using the Gibson assembly
method, resulting in acceptor vector pKMBKI.

R
SC
Fig. 2. Cloning strategy for constructing the recombinant allelic exchange
vector

U
Entry vectors pGEM-U, pGEM-S and pGEM-D and acceptor vector pKMBKI
were mixed together along with BsaI and ligase. The mixture was incubated in a
N
water bath at 37℃ for 30 min followed by heating at 50℃ for 5 min and then at
A
80℃ for 5 min. Next, the mixture was transformed into DH5 chemically
M

competent cells. X-Gal and kanamycin were used for the selection of correct
clones, white kanamycin-resistant colonies were picked out, and the recombinant
ED

allelic exchange vector pKMUSD was successfully identified.

Fig. 3. PCR identification of the recombinant allelic exchange vector

PT

(A) All colonies except No. 8 and No. 20 possessed the correct eagup amplified
fragment. (B) For the spcin DNA fragment, 19 of the 20 colonies possessed the
E

correct amplified DNA bands. No amplification products were obtained with

CC

colony No. 8. (C) For the eagdn DNA fragment, 18 of the 20 white colonies
possessed the correct amplified DNA bands. No amplification products were
obtained with colonies No. 6 and No. 8. Overall, the success rate of ligation of the
A

three entry fragments eagup, spc and eagdn into the acceptor vector pKMBKI was
85% (17/20 colonies). +: the positive controls, −: the negative controls.

Fig. 4. Kinetics of plasmid pKMUSD elimination from Bacillus anthracis

strain A16R
23
B. anthracis strain A16R (pKMUSD) was grown in LB medium without
kanamycin and A16R (pKMUSD+pSS4332) was grown in LB medium with
kanamycin at 42°C for 12 h. The samples were diluted and plated onto LB agar. At
each time point, the colonies were counted and scored for the elimination of the
plasmid. Plasmid elimination rate (PER) was calculated based on the following
equation: PER= (1−N/N0)×100%. Note: N0, total colonies tested; N,
plasmid-containing colonies (spcr).

T
Fig. 5. Verification of eag gene replacement by the spcr cassette in B. anthracis

IP
strain A16R

R
1: B. anthracis strain A16R△eag::spc, 2: B. anthracis strain A16R. (A) The results
indicated that the eag gene was present in wild-type strain A16R but not in

SC
A16Reag::spc because the eagin fragment was only amplified in the wild-type
strain. Furthermore, two specific PCR bands were amplified by the external

U
primers in A16Reag::spc and sequencing results were consistent with the eag
N
gene being replaced by the spcr gene. (B) Western blot analysis using anti-EA1
A
antibody revealed a specific protein band in B. anthracis strain A16R that was not
present in A16R△eag::spc.
M
ED

Fig. 6. Potential schematic of endonuclease I-SceI-mediated allelic-exchange

procedure. This protocol used the endonuclease I-SceI to cleave the target
plasmid or genomic DNA in vivo. (I) The plasmid pSS4332 was transformed into
PT

B. anthracis containing pKMUSD, which was then grown at a permissive

E

temperature. (II) pKMUSD was integrated into the chromosome in some bacterial
CC

cells while free plasmid also existed in many other cells. (III) The endonuclease
I-SceI expressed and cleaved the target sites on the integrated plasmid, broke the
chromosome or cleaved the target sites on pKMUSD, and linearized it. (IV) The
A

chromosome was damaged by endonuclease I-SceI, followed by recombination

repair resulting in loss of the integrated plasmid or the occurrence of homologous
recombination between linear plasmid DNA and chromosome. (V) Both of these
different pathways result in replacement of the targeted gene (eag) by spcr cassette.

24
The red diamond denotes the I-SceI site.

T
IP
R
SC
U
N
A
M
ED
E PT
CC
A

25
A
Table 1 Bacterial strains and plasmids used in this study

Plasmids and
Relevant genotype and characteristics Source
strains
This
Shuttle vector, temperature-sensitive, Ampr in E. coli lab(Smith and
pKSV7
and Cmr in B. anthracis Youngman,
1992)

T
This
Shuttle vector, Apr in E. coli; Emr in B. subtilis;

IP
pMAD lab(Arnaud et
containing pclpB-bgaB fragment
al., 2004)
pKMU7 Derive from pKSV7 by destroying a BsaI site in it. This work

R
pKD4 Containing kanamycin resistant gene This lab

SC
Constructed by inserting pclpB-bgaB and Kanr
pKMBK This work
fragments into pKMU7
Constructed by inserting I-SceI fragment into
pKMBKI
pKMBK
U This work
N
The production of “Golden Gate” cloning with the
upstream part of the target gene, the spectinomycin
pKMUSD This work
A
resistance cassette, and the downstream part of the
target gene inserting into the vector pKMBKI
M

pGEM-T easy Ampr in E. coli, with single 3’-T overhangs This lab
Constructed by inserting eagup fragment into
pGEM-U This work
pGEM-T easy
ED

Constructed by inserting eagdn fragment into

pGEM-D This work
pGEM-T easy
Constructed by inserting Spcr cassette fragment into
PT

pGEM-S This work

pGEM-T easy
This
pSET4s Shuttle vector,Spcr both in E. coli and B. anthracis lab(Takamats
E

u et al., 2001)
CC

expressing endonuclease I-SceI，Kanr in B. anthracis

pSS4332 This
and Ampr in E. coli，is relatively unstable in B.
lab(Cybulski
anthracis allowing rapid loss by segregation after et al., 2009)
removal of selection.
A

F-,ϕ80d/lacZ∆M15, ∆(lacZYA-argF)U169, deoR,

E. coli DH5α recA1, endA1, hsdR17(rk-,mk+),phoA,supE44λ-, This lab
thi-1, gyrA96,relA1

rpsL (Str R) thr leu thi-1 lacY galK galT ara tonA tsx
E. coli JM110 dam dcm supE44 Δ(lac-proAB) /F′[traD36 This lab
proABlacIq lacZΔM15]

B. anthracisA16R Vaccine strain; pXO1+, pXO2- This lab

26
B.anthracisA16R△ A16R with eag gene knocked out and Spcr gene
This work
eag::spc instead
Note: Ampr, ampicillin resistance; Cmr, chloramphenicol resistance; Emr, erythromycin resistance;
Spcr, spectinomycin resistance; Kanr, kanamycin resistance.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

27
Table 2 Primers used in this study

Name Sequences

bgaB_F TGACATGATTACGAATTCTGACAgagaccGTCTAGTTAATGTGTAACG

bgaB_R AGTGCTTGCGGCAGCGTGACCAAAgagaccGCATATTATGTTGCCAAC

kan_F TCACGCTGCCGCAAGCACTCAGGGCGCAAGGGCTGC

T
kan_R GTCGACTCTAGAGGATCCTGGGCGAAGAACTCCAGCAT

IP
CAAGCTTGCATGCCTGCAGGTCGACTAGGGATAACAGGGTAATGTC
I-SceI_F

R
GACTCTAGAG

SC
I-SceI_R TTCGCCCAGGATCCTCTAGAGTCGACATTACCCTGTTATCCCTAG

eagup_F TTTggtctcATGACCTTATCTGTTGATGGTGTACC

eagup_R

eagdn_F
TTTggtctcACTCCCTCCTTCAGGAATATGCTACT

TTTggtctcAATCGCGAATCACTATACACGAACA U
N
eagdn_R TTTggtctcACCAACCAATCCATGCCTGCTAT
A
spc_F TTTggtctcAGGAGCTAGTGTTCGTGAATACATG
M

spc_R TTTggtctcACGATTATGCCGATAACTAG

eagin_F CAATGACAGCAGCAATGG
ED

eagin_R TCAACATCAGCGTTACCTT

eagus_F GAAGTAGACGCAACTGATG
PT

eagus_R GCTCTTGTAACCATTCTCC
E

eagsd_F GGCTGAATCTTCTCCATTA
CC

eagsd_R CATACTAGAACTTGCACCAA

spcin_F TGATGTGAGAAGAGCCATTA
A

spcin_R CTCCAAGATAACTACGAACTG

pKSV7_F ATCGTTGTCAGAAGTAAGTTGG

pKSV7_R TGGCGTAATAGCGAAGAGG
Note: Underlined sequences indicate restriction sites; lower-case italic sequences
indicate BsaI restriction sites.

28