A toolbox for genetic manipulation in intestinal Clostridium symbiosum

Abstract

Gut microbes are closely related with human health, but remain much to learn. Clostridium symbiosum is a conditionally pathogenic human gut bacterium and regarded as a potential biomarker for early diagnosis of intestinal tumors. However, the absence of an efficient toolbox that allows diverse genetic manipulations of this bacterium limits its in-depth studies. Here, we obtained the complete genome sequence of C. symbiosum ATCC 14940, a representative strain of C. symbiosum. On this basis, we further developed a series of genetic manipulation methods for this bacterium. Firstly, following the identification of a functional replicon pBP1 in C. symbiosum ATCC 14940, a highly efficient conjugative DNA transfer method was established, enabling the rapid introduction of exogenous plasmids into cells. Next, we constructed a dual-plasmid CRISPR/Cas12a system for genome editing in this bacterium, reaching over 60 % repression for most of the chosen genes as well as efficient deletion (>90 %) of three target genes. Finally, this toolbox was used for the identification of crucial functional genes, involving growth, synthesis of important metabolites, and virulence of C. symbiosum ATCC 14940. Our work has effectively established and optimized genome editing methods in intestinal C. symbiosum, thereby providing strong support for further basic and application research in this bacterium.

1. Introduction

The human intestinal tract is not only an important digestive organ but also a favorable habitat for microorganisms. On average, the intestine of a healthy adult is colonized by over 1000 microbial species [1], which can affect host health via different mechanisms. Gut microbes have been known to play vital roles in physiological activities of host, such as inflammation, barrier homeostasis, metabolism, and hematopoiesis [2]. Thus, the roles of gut microbes in human's physiological processes and disease occurrence and development have been attracting extensive attention.

Firmicutes are the most major component of the gut microbiota, in which 95 % belongs to clostridia [3]. Specific to intestinal Clostridium, multiple species have been found to be associated with human diseases such as bacteremia, anorexia, diabetes, autism spectrum disorders, and obesity. C. symbiosum has been reported as an opportunistic pathogen causing bacteremia [[4], [5], [6]]. Currently, research indicates that C. symbiosum may be associated with the development of inflammatory bowel disease (IBD) [7] and coronary artery disease [8]. Of note, recent studies revealed significantly increased abundance of C. symbiosum in the stools of patients with colonic adenomas or early and advanced colorectal cancer, suggesting that this bacterium can serve as a tumor biomarker for the early diagnosis of colorectal cancer [9].

A comprehensive and thorough understanding of causal relationship between C. symbiosum and the development of colorectal cancer will require efficient genetic manipulation of this bacterium. The first step towards this goal has been made recently with three atypical C. symbiosum strains (C. symbiosum DSM 29356, WAL-14163, and WAL-14673) tested, in which gene overexpression and CRISPRi-based gene repression were achieved in WAL-14163 [10]. This work strongly supports a continued development and optimization of the toolbox in C. symbiosum, enabling its expansion to more important strains.

In this study, the first complete genome of the representative C. symbiosum strain, ATCC 14940, is provided. Subsequently, a CRISPR-Cas system employing Cas12a was constructed, achieving efficient gene deletion and expressional regulation in the ATCC 14940 strain. Furthermore, a series of promoters with wide strength range in this bacterium were identified, enabling the controlled overexpression of target genes. Based on this comprehensive toolbox, multiple genes associated with crucial characteristics of C. symbiosum ATCC 14940 were identified. Together, this work effectively expands and optimizes the existing toolbox in intestinal C. symbiosum strains.

2. Materials and methods

2.1. Strains, cell lines, media, and reagents

The E. coli strains TOP10, HB101, S17-1, and their derived strains were cultured in Luria-Bertani (LB) broth or on LB agar plate at 37 °C supplemented with chloramphenicol (12.5 μg/mL) or erythromycin (500 μg/mL) when needed. The E. coli strain TOP10 was used for plasmid construction. HB101 and S17-1 were used as the conjugation donors. C. symbiosum ATCC 14940 was anaerobically cultivated at 37 °C using the Brain-Heart Infusion medium (37 g/L brain heart infusion, 5 g/L yeast extract, 5 g/L K₂HPO₄, 0.5 g/L l-cysteine hydrochloride, 0.2 % (v/w) resazurin,5 mg/L hemin, 1 mg/L menadione). When required, antibiotics were added to growth media at the working concentrations: thiamphenicol (10 μg/mL), erythromycin (1 μg/mL), polymyxin B (50 μg/mL).

HCT116 human colorectal adenocarcinoma (CRC) cell line was obtained from American Type Culture Collection (ATCC). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (Gibco BRL, Grand Island, NY, USA) supplemented with 10 % fetal bovine serum (FBS, YOUSHI, Wuhan, China) and incubated at 37 °C under a humidified incubator (37 °C, 5 % CO₂). Cell passage was treated with trypsin-EDTA (Gibco, Burlington, Canada). For the co-culture of C. symbiosum and HCT116 cells, approximately 4 × 10⁵ HCT116 cells were grown in a 6-well plate and then infected with C. symbiosum for 4 h (multiplicity of infection = 100) under anaerobic condition at 37 °C. Then, the cells were washed with PBS and cultured in complete medium prepared for subsequent experiments.

KOD plus Neo and KOD FX DNA polymerase (Toyobo, Osaka, Japan) were used for PCR amplification. Plasmid construction was performed using the ClonExpress MultiS One Step Cloning Kit (Vazyme Biotech Co., Ltd., Nanjing, China) or restriction enzymes (Thermo Fisher Scientific, Vilnius, Lithuania) and ligase (Takara, Dalian, China). Extraction of plasmid and DNA purification were performed by kits (Axygen Biotechnology Company Limited, Hangzhou, China). All the primers used in this study were synthesized by BioSune (Biosune, Shanghai, China).

2.2. Plasmid construction

The plasmids and primers used in this study are listed in Tables S1 and S2. The starting plasmids used for constructing the other plasmids in this study were pMTL82151 and pMTL82254.

The plasmid pMTL-P₁₃₃₉-ccpA used for gene overexpression in C. symbiosum ATCC 14940 was constructed as follows: the DNA fragment of the promoter P₁₃₃₉ was obtained via PCR amplification using the primers P₁₃₃₉–F/P₁₃₃₉–R and the genomic DNA of C. acetobutylicum ATCC 824 as template. The ccpA gene was obtained via PCR amplification using the primers ccpA-F/ccpA-R and the genomic DNA of C. symbiosum ATCC 14940 as the template. Subsequently, the DNA fragment (P₁₃₃₉-ccpA) was yielded via overlapping PCR using the primers P₁₃₃₉–F/ccpA-R, and then linked with the linear plasmid pMTL82151 vector (digested with BamHI and SacI) using ClonExpress II One Step cloning kit (Vazyme, Nanjing, China), generating the target plasmid pMTL-P₁₃₃₉-ccpA. The pMTL-P₁₃₃₉-ackA plasmid was constructed via the same steps except that gene sequences were changed accordingly.

The plasmid pMTLddcas12a was generated as follows: the DNA fragments corresponding to the promoter P₀₁₄₄₀ and P_thl were obtained via PCR amplification using the primers P₀₁₄₄₀–F/P₀₁₄₄₀–R and P_thl-F/P_thl-R, and the genomic DNA of C. ljungdahlii and C. acetobutylicum as the template, respectively. The ddCas12a was codon-optimized based on the genomic sequences of C. beijerinckii NCIMB 8052 and C. acetobutylicum ATCC 824. The gene fragment of ddCas12a was obtained via PCR amplification using the primers ddcas12a-F/ddcas12a-R. The P₀₁₄₄₀ fragment was assembled with the P_thl fragment via overlapping PCR using the primers P₀₁₄₄₀–F/P_thl-R. The obtained two DNA fragments, FnddCas12a and P_thl-P₀₁₄₄₀, were further linked with the linear pMTL82151 (digested with restriction enzymes BamHI and NotI) using the ClonExpress MultiS One Step Cloning Kit, generating the target plasmid pMTLddcas12a. The pMTLddcas12a-birAi-TS plasmid was constructed as follows: a DNA fragment containing the crRNA spacer and repeat sequence was first obtained by PCR amplification using the primers birAT-crRNA-F/birAT-crRNA-R. Then, this DNA fragment was used as a template for PCR amplification using the primers universal-F/universal-R. The generated DNA fragment birAT-crRNA, which contained the crRNA, terminator, and repeat sequence, was assembled with the linear plasmid pMTLddcas12a (digested with restriction enzymes BamHI and NcoI) using the ClonExpress II One Step cloning kit, yielding the target plasmid pMTLddcas12a-birAi-TS. The pMTLddcas12a-birAi-NTS and pMTLddcas12a-ackAi plasmids were constructed via the same steps except that the crRNA expression cassettes were changed accordingly.

The plasmid pMTLcas12a was constructed as follows: The DNA fragment cas12a-M was obtained via PCR amplification using the primers M-F/M-R and the plasmid pMTLddcas12a as the template with the introduction of a point mutation (A1006E) in the Cas12a protein. The obtained DNA fragment was assembled with the linear plasmid pMTL-ddcas12a (digested with restriction enzyme SwaI) using the ClonExpress II One Step cloning kit, generating the target plasmid, pMTLcas12a.

The plasmid pMTLcas12a-ΔccpA was constructed as follows: the two homologous arms flanking the open reading frame of ccpA (obtained by PCR amplification using the primers ccpA-up-F/ccpA-up-R and ccpA-down-F/ccpA-down-R) and the crRNA (obtained by PCR amplification using the primers ccpA-crRNA-F/ccpA-crRNA-R) were assembled by overlapping PCR, yielding a large DNA fragment ccpAcrRNA-ccpAup-ccpAdown. Thereafter, this DNA fragment was assembled with the linear plasmid pMTLcas12a (digested with XhoI and BamHI) using the ClonExpress II One Step cloning kit, generating the target plasmid pMTLcas12a-ΔccpA. The pMTLcas12a-Δ02043 plasmid was constructed via the same steps.

The plasmid pMTL-P₀₁₄₄₀-cas12a was constructed as follows: the DNA fragment P₀₁₄₄₀-cas12a was obtained via PCR amplification using the primers cas12a-F/P₀₁₄₄₀-BamHI-R and the pMTLcas12a plasmid as the template. The obtained DNA fragment P₀₁₄₄₀-cas12a was linked with the linear pMTL82254 (digested with BamHI and NotI) using the ClonExpress II One Step cloning kit, generating the target plasmid pMTL-P₀₁₄₄₀-cas12a. The derivative plasmids, i.e., pMTL-P_fba2-cas12a, pMTL-P_gpm-cas12a, and pMTL-P_thl-cas12a, were constructed via the same steps.

The plasmid pMTL-P_thl-ccpAcrRNA was constructed as follows: the DNA fragment P_thl-ccpAcrRNA-ccpAup-ccpAdown was obtained via PCR amplification using the primers P_thl-crRNA-F/ccpAdown-SalI-R and plasmid pMTLcas12a-ΔccpA as the template. Then, the obtained DNA fragment P_thl-ccpAcrRNA-ccpAup-ccpAdown was linked with the linear plasmid pMTL82151 (digested with BamHI and SacI) using ClonExpress II One Step cloning kit, generating the target plasmid pMTL-P_thl-ccpAcrRNA. The derivative plasmids, i.e., pMTL-P_fba2-ccpAcrRNA, pMTL-P_gpm-ccpAcrRNA, and pMTL-P₀₁₄₄₀-ccpAcrRNA, were constructed via the same steps.

The plasmid pMTL82151-P₁₃₃₉-LacZ was constructed as follows: the DNA fragment corresponding to LacZ was obtained via PCR amplification using the primers LacZ-F/LacZ-R and the pIMP1-P_thl-LacZ plasmid [11] as the template. The DNA fragment corresponding to the P₁₃₃₉ promoter was obtained via PCR amplification using the primers P₁₃₃₉-lacZ-F/P₁₃₃₉-lacZ-R and the genomic DNA of C. acetobutylicum ATCC 824 as the template. These two DNA fragments were assembled via overlapping PCR using the primers P₁₃₃₉-lacZ-F/LacZ-R. The obtained P₁₃₃₉-lacZ fragment was further linked with the linear plasmid pMTL82151 (digested with BamHI and SacI) using ClonExpress II One Step cloning kit, generating the target plasmid pMTL82151-P₁₃₃₉-LacZ. The other plasmids for assessing promoter strength (pMTL82151-P_thl-LacZ, pMTL82151-P₀₁₄₄₀-LacZ, pMTL82151-P_fba2-LacZ, pMTL82151-P_gpm-LacZ, pMTL82151-P_pfkA2-LacZ, pMTL82151-P_pgi-LacZ, pMTL82151-P_eno-LacZ, pMTL82151-P_pyk-LacZ, pMTL82151-P_pgk-LacZ, pMTL82151-P_gap-LacZ, and pMTL82151-P_fbaA-LacZ) were constructed via the same steps.

2.3. RNA purification and real-time quantitative PCR (RT-qPCR)

C. symbiosum cells were collected when they reached an optical density (OD₆₀₀) of 1.0 and then stored at −80 °C. Total RNA extraction and RT-qPCR were conducted as reported previously [12]. The 16S rDNA was employed as the internal control. The primers used for RT-qPCR are listed in Supplementary Table S2.

2.4. C. symbiosum conjugation

The culture of C. symbiosum ATCC 14940 was anaerobically cultured at 37 °C in BHI medium until reaching OD₆₀₀ of 0.4–1.2. The donor strain of E. coli (containing the shuttle plasmid) was cultivated in LB broth supplemented with appropriate antibiotics until reaching OD₆₀₀ of 0.4–1.2. Subsequently, 5 mL of the E. coli culture was subjected to centrifugation (at 3,000 × g for 10 min). The obtained cell pellets were gently resuspended in 1 mL of PBS buffer and subjected to centrifugation again. The obtained cell pellets were then gently suspended in the C. symbiosum ATCC 14940 culture (200 μL). The mixed culture was spotted onto the BHI agar plates and incubated at 37 °C for 2−8 h. The colonies on the agar plates were collected and resuspended in 1 mL of PBS, and then plated onto the BHI agar plates containing the appropriate antibiotics (50 μg/mL polymyxin B, 10 μg/mL thiamphenicol, or 1 μg/mL erythromycin). Colonies appeared after 3−4 d of incubation at 37 °C.

2.5. Fermentation

The C. symbiosum strains were cultivated anaerobically in BHI medium containing the appropriate antibiotic. In brief, 250 μL of the frozen stock was inoculated into BHI medium (5 mL) and then incubated anaerobically at 37 °C. When the OD₆₀₀ of grown cells reached 1.0, 0.5 mL of the inoculum was transferred into 9.5 mL of BHI medium for cultivation at 37 °C. The culture was sampled every 2 h. All the procedures were performed in an anaerobic chamber (Whitley A35 Anaerobic Workstation, Don Whitley Scientific Limited, Bingley, West Yorkshire, UK).

2.6. Analytical methods

The measurements of cell growth and acetate concentration were performed according to previously reported methods [11]. Briefly, cell growth was determined based on the absorbance of the culture at A₆₀₀ (OD₆₀₀) using a spectrophotometer (U-1800, Hitachi, Japan). The concentration of acetate was determined as described previously using a gas chromatography (7890 A, Agilent, Wilmington, DE, USA) equipped with a flame ionization detector and a capillary column (EC-Wax, Alltech, Lexington, KY, USA) [11].

2.7. Cell proliferation assays

Cell proliferation was evaluated using a Cell Counting Kit-8 (CCK-8, DOJINDO, Kumamoto, Japan) according to the manufacturer's instructions. In brief, the HCT116 cells were seeded into 96-well culture plates at a density of 3,000 cells per well and cultivated in a CO₂ (5 %, v/v) incubator at 37 °C overnight. 10 μl of CCK-8 solution was added to each well every 24 h for another 2h incubation at 37 °C. The absorbance was measured at 450 nm.

2.8. β-Galactosidase assays

C. symbiosum cells were grown at 37 °C under anaerobic condition and harvested by centrifugation (6,000 × g for 10 min) when the grown cells reached the optical density (OD₆₀₀) of 1.0. Cells were resuspended in 500 μl of B-PER reagent (Thermo Scientific Pierce, USA) with vigorous vortexing for 3 min. The resulting cell lysates were heated at 60 °C for 30 min to remove heat unstable proteins and then subjected to centrifugation (12,000 × g for 30 min). The supernatants were used for the β-galactosidase assays as previously reported [13].

2.9. DNA sequencing and genome assembly

The C. symbiosum cells were collected when the optical density (OD₆₀₀) reached 1.0 by centrifugation (10,000 × g for 10 min) and then immediately frozen in liquid nitrogen. Next, genomic DNA was extracted by using a DNA extraction kit (Shandong Sparkjade Biotechnology Co., Ltd., China).

Genome sequencing was carried out using the Pacific Biosciences (PacBio) RS II sequencing platform. In brief, genomic DNA was sheared into ∼10 kb fragments using a Covaris g-TUBE shearing device (Covaris, Woburn, Massachusetts, USA). DNA fragments were purified, end-repaired, and ligated to SMRTParkbell hairpin adapters using the PacBio SMRTbell library preparation kit (Pacific Biosciences, Menlo Park, China, USA). SMRTbell DNA libraries were constructed according to the manufacturer's protocol (Pacific Biosciences, Menlo Park, CA, USA). The library quality analysis and quantification were determined using the Qubit 2.0 Fluorometer (Bio-Medlab, China) and the Agilent 2100 Bioanalyzer system (Agilent Technologies, SantaClara, CA, USA). The SMRT sequencing was accomplished using the PacBioRSII system (Pacific Biosciences, Menlo Park, CA, USA) according to the standard protocol. The obtained continuous PacBio long-reads generated from SMRT sequencing runs were adopted for de novo assembly using the program Spades (version 3.14.1), yielding one circular chromosome.

2.10. Gene prediction and functional annotation

RepeatMasker (v4.0.6) was used to identify and annotate repeats [42]. The tandem repeats were further identified using Tandem Repeats Finder software (v4.07b) [43] with the following parameters: Match = 2, Mismatch = 7, Delta = 7, PM = 80, PI = 10, Minscore = 50.

Augustus (v3.2.1) [44] was used for the de novo gene prediction. For homology-based prediction, protein sequences of C. symbiosum ATCC 14940 were downloaded from the NCBI database, and then aligned to the repeats of a soft-masked genome using TBlastN (v2.2.29)[45] with a cut-off e-value of 1e−5. The results from the de novo and homology prediction were integrated using EVidenceModeler (v1.1.1) [46]. The annotated genes were also subjected to a BLAST analysis in the Virulence Factor Database (VFDB; http://www.mgc.ac.cn/VFs/main.htm) using the Diamond software with the default parameters to identify virulence genes [14].

3. Results and discussion

3.1. Complete genome sequence of C. symbiosum ATCC 14940

C. symbiosum ATCC 14940 is the model strain of intestinal C. symbiosum. Its draft genome sequence has been presented by the National Institutes of Health (NIH) (GenBank accession No.514700). However, the complete genome sequence of this strain has not yet been reported so far. To fill this gap and obtain the detailed genomic information, we re-sequenced this stain and finished de novo assembly. The sequencing data showed a circular chromosome of 5,022,615 bp with a G + C content of 47.8 % (Fig. 1 and Table 1). A total of 4,496 coding sequences were predicted, including 4,419 protein-coding sequences, 12 rRNA and 65 tRNA genes (Genome Warehouse in National Genomics Data Center: GWHBEHX01000000). No significant difference was observed in the genome size and GC content between C. symbiosum ATCC14940 and the C. symbiosum LT0011 (a reference genome providing by NCBI) (Table S4).

Fig. 1.

Circular genome map for Clostridium symbiosum ATCC 14940. From outside to inside: circle 1, CDSs on the forward strand colored according to COG category; circle 2, CDSs, rRNA and tRNA on forward strand; circle 3, CDSs, rRNA, and tRNA on the reverse strand; circle 4, CDSs on the reverse strand colored according to COG category; circle 5, GC content; circle 6, GC skew.

Table 1.

The genome features of Clostridium symbiosum ATCC 14940.

Features	Results
Sequencing technology	PacBio, Illumina
Genome topology	Circular
Total size (bp)	5,022,615
GC content (%)	47.80
Contig N50 (bp)	977,160
Scaffold number	1
Protein-coding genes	4,419
tRNA genes	65
rRNA genes	12

3.2. Efficient DNA transformation for C. symbiosum

It is known that bacteria process a series of defense systems that provide protection against foreign DNA, in which the restriction-modification (R-M) system normally plays a major role [15]. During the conjugative DNA transfer, plasmid DNA is transferred in a single-stranded form; thus, it is likely to be minimally affected by the R-M system [16]. Therefore, we attempted to establish the DNA transformation method of C. symbiosum ATCC 14940 via conjugational transfer using E. coli as the donor strain.

We firstly assessed the resistance of C. symbiosum ATCC 14940 to different antibiotics and found that the supplementation of 10 μg/mL thiamphenicol or 1 μg/mL erythromycin in the medium could completely inhibit cell growth (Table 2). Thus, the Clostridium perfringens catP gene that can give Clostridium strains the resistance to thiamphenicol [17] was used in the following plasmid construction. Next, four pMTL80000 plasmids, i.e., pMTL82151, pMTL83151, pMTL84151, and pMTL85151, which contain the abovementioned catP gene as well as different clostridial replicons (pBP1, pCB102, pCD6, and pIM13, respectively) [18], were tested in conjugative plasmid transfer of C. symbiosum ATCC 14940 (Fig. 2A). Here, two types of the E. coli strains were adopted as the donor strains, namely a standard strain S17-1 and a methylation-deficient strain HB101 (pRK2013). Concurrently, polymyxin B, which can inhibit the growth of these E. coli stains but has no effect on C. symbiosum ATCC 14940, was used for the selection of transconjugants. The four plasmids were firstly transferred into the E. coli strain S17-1 and strain HB101 (details shown in Materials and Methods). Encouragingly, both the E. coli S17-1 and HB101 strain that contained the plasmid pMTL82151 generated transconjugants, in which S17-1 produced an average of 30 transconjugants per conjugation (equivalent to a conjugation efficiency of 8 × 10⁻⁸) while HB101 only produced an average of 12 transconjugants (equivalent to a conjugation efficiency of 3 × 10⁻⁸), suggesting that E. coli S17-1 is a better conjugal donor strain (Fig. 2B).

Table 2.

Antibiotic sensitivity test of C. symbiosum.

Antibiotics	Concentration (mg/L)	Sensitivity
Thiamphenicol	1	+
	5	–
	10	–
	20	–
Erythromycin	0.1	–
	0.5	–
	10	–
Spectinomycin	50	+++
	100	+
	200	–
Ampicillin	50	–
	100	–
Kanamycin	50	++
	100	+
	150	–
Streptomycin	50	+
	100	–
Apramycin	50	++
	100	+
	150	–
Tetracycline	5	–
	10	–
Polymyxin B	100	+++
	200	+++

−: High sensitivity; +: moderate sensitivity; ++: low sensitivity; +++: insensitive.

Fig. 2.

Development of a conjugative DNA transfer system for C. symbiosum ATCC 14940. (A) Schematic diagram of bacterial conjugation. (B) Effect of the donor strain and shuttle plasmid on conjugation efficiency. (C) Effect of co-incubation time on conjugation efficiency. (D) Effect of the growth state (OD₆₀₀) of the donor bacterium (E. coli S17-1) on conjugation efficiency. (E) Effect of the growth state of the recipient bacteria (C. symbiosum ATCC 14940) on conjugation efficiency. Conjugation efficiency was calculated as transconjugant CFU/total C. symbiosum CFU. Data are represented as mean ± SD (n = 3).

Next, the conjugation transfer process was further optimized. We firstly adjusted the co-culturing time of E. coli S17-1 and C. symbiosum ATCC 14940. The combined cultures were incubated stationarily at 37 °C for 2, 4, 6, or 8 h. As shown in Fig. 2C, the 6-h incubation offered higher conjugation transfer efficiency over the other three time-intervals. We further examined the influence of the growth state of the donor and recipient strains on the conjugation efficiency. The highest plasmid transfer efficiency was obtained from the culture of E. coli S17-1 and C. symbiosum ATCC 14940 grown to OD₆₀₀ of 1.0 and 0.4, respectively (Fig. 2D and E). Finally, a robust conjugation protocol was presented, reaching 5 × 10³ transconjugants per milliliter of recipient cells (equal to a conjugation efficiency of 2 × 10⁻⁵) for C. symbiosum ATCC 14940.

3.3. Gene overexpression in C. symbiosum

Having established an efficient conjugative plasmid transfer method for C. symbiosum ATCC 14940, we attempted to test gene overexpression in this bacterium. Three heterologous strong promoters, P_thl and P₁₃₃₉ from C. acetobutylicum ATCC 824 [13,19] and P₀₁₄₄₀ from C. ljungdahlii DSM 13528 [20], were chosen for an initial test. These promoters were placed upstream of a reporter gene lacZ in the pMTL82151 plasmid for strength evaluation by measuring β-galactosidase activity. As shown in Fig. 3A, all the promoters generated high β-galactosidase activities, demonstrating that they are active in C. symbiosum ATCC 14940.

Fig. 3.

Establishment of the gene expression system in C. symbiosum ATCC 14940. (A) Determination of the β-galactosidase activities that represent the strength of the P₁₃₃₉, P₀₁₄₄₀, and P_thl promoters. Control: no promoter driving the expression of lacZ. (B) A schematic diagram of the pMTL-P₁₃₃₉-ccpA plasmid for the overexpression of ccpA. P₁₃₃₉: the reported promoter from C. acetobutylicum ATCC 824; traJ: encoding an oriT-recognizing protein for initiation of transfer DNA replication; catP: chloramphenicol resistance gene; pBP1: the Gram-positive replicon; TCD0164 and TFdx: terminators. (C) Comparison of the transcriptional levels of ccpA in ccpA-overexpressing and wild-type strains. (D) The specific growth rates of the wild-type and ccpA-overexpressing strains. Control: the wild-type strain contained the plasmid pMTL82151; WT-ccpA: the ccpA-overexpressing strain. Data are represented as mean ± SD (n = 3). Statistical analysis was performed by a two-tailed Student's t-test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, no significance.

Subsequently, these promoters were further used to overexpress some genes that are potentially associated with specific phonotypic outcomes of C. symbiosum. Here, CcpA (catabolite control protein A), a global transcriptional regulator involved in the regulation of cell growth and substrate metabolism in Gram-positive bacteria including Clostridium species [21,22], was chosen for testing. According to the genome annotation, the gene (CSYM_04033) that encodes CcpA in C. symbiosum was found and then cloned into the pMTL82151 plasmid under the control of the promoter P₁₃₃₉. The yielding plasmid was introduced into C. symbiosum through conjugative transfer for gene overexpression (Fig. 3B). As expected, a significantly increased transcriptional level of ccpA was detected in the ccpA-overexpressing strain compared with the control strain (Fig. 3C), indicating that this gene was sufficiently expressed with the promoter P₁₃₃₉. Interestingly, the growth rate of the ccpA-overexpressing strain was lower than that of the control strain (Fig. 3D), which was different from the findings in the industrial C. acetobutylicum [12], in which ccpA overexpression could promote cell growth. Therefore, this result indicates the functional diversity of CcpA in different bacteria.

3.4. CRISPR interference (CRISPRi) in C. symbiosum using the DNase-deactivated Cas12a (ddCas12a)

Next, we attempted to further develop a CRISPRi system in C. symbiosum ATCC 14940, aiming to achieve the efficient repression of target genes. Compared with Cas9, Cas12a has a less toxic effect and requires only a single RNA molecule, the crRNA, which presents a more minimalistic system [23,24], and thus, Cas12a rather than Cas9 was employed in the construction of the CRISPRi system. Using the pMTL82151 plasmid as the backbone, a DNase-deactivated Cas12a from Francisella novicida (ddFnCas12a) was inserted into the plasmid under the control of the abovementioned promoter P₀₁₄₄₀ (Fig. 4A). A previously reported functional gene, birA, was adopted as the target to test our CRISPRi system. BirA is known to be a bifunctional protein acting as both a biotin ligase and a repressor on the biotin synthesis operon in bacteria [25,26].

Fig. 4.

Development of the CRISPRi system in C. symbiosum ATCC 14940. (A) Schematic diagram showing the workflow of birA repression using the CRISPRi system. P₀₁₄₄₀ and P_thl: the promoters for the expression of ddCas12a and crRNA, respectively; RNAP: RNA polymerase. (B) The transcriptional changes of birA after introducing the CRISPRi plasmid into C. symbiosum ATCC 14940. (C) The effect of the birA repression on the growth of C. symbiosum ATCC 14940. Control: the strain containing the plasmid carrying ddCas12a but no crRNA; birAi-TS: the strain harboring the CRISPRi plasmid with crRNA targeting the template strand of birA; birAi-NTS: the strain harboring the CRISPRi plasmid with crRNA targeting the non-template strand of birA. Data are represented as mean ± SD (n = 3). Statistical analysis was performed by a two-tailed Student's t-test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, no significance.

To target the birA gene (CSYM_01417) in C. symbiosum ATCC 14940, two crRNA sequences complementary to different regions (one on the template DNA strand and another on the non-template DNA strand) of birA, were designed and then separately inserted into the plasmid under the control the abovementioned promoter P_thl (Fig. 4A). As expected, after introducing the plasmid into the strain, both the two ddFnCas12a-crRNA constructs could efficiently inhibit the expression of birA, reaching 80 % and 60 % repression for the template and non-template DNA strand, respectively (Fig. 4B). Meanwhile, both of the resulting two strains exhibited slower growth rate compared to the control strain (Fig. 4C), indicating the impact of birA repression on cell metabolism. This phenotypic change was similar with the previous finding in C. ljungdahlii that the inactivation of birA could impair the cell growth [27], which was largely due to the damage of BirA's biotin protein ligase activity on the biotinylation of acetyl-CoA carboxylase (ACC) and pyruvate carboxylase (PYC), two crucial enzymes for the basic metabolism in cells [27]. However, it should be noted that the repression degree on birA based on the above ddFnCas12a-crRNA construct was still incomplete, requiring a further promotion (Fig. 4B). A feasible strategy is to use a dual-crRNA construct targeting the same gene, which could lead to stronger transcriptional inhibition [28]. According to the previous reports, the efficiency of ddCpf1 (ddCas12a) in CRISPRi-based gene repression was higher using the crRNAs targeting the template strand relative to those targeting the non-template strand [29]. Our study also had a similar finding. In addition, the repression efficiency was found to be higher when the targeted site of crRNA was closer to the transcription start site [30]. Therefore, we designed the crRNAs that targeted the template DNA strand within 100 bp relative to the translation initiation code “ATG”.

3.5. Construction and optimization of CRISPR-Cas12a-based gene deletion in C. symbiosum

After achieving CRISPRi, we attempted to further develop gene deletion in C. symbiosum ATCC 14940 using the CRISPR–Cas12a system. An expression cassette containing the FnCas12a protein driven by the P₀₁₄₄₀ promoter and a crRNA sequence (targeting the ccpA gene mentioned above) driven by the P_thl promoter was integrated into the plasmid. Meanwhile, two homologous arms (1,000 bp for both) flanking ccpA for repairing double strand break were also inserted into the plasmid (Fig. 5A). The yielding plasmid pMTLcas12a-ΔccpA was introduced into cells, and the transformant colonies occurred on the agar plates were used for PCR screening to examine the desired deletion events. Encouragingly, of the total 32 colonies, evidence for ccpA deletion was obtained for five cases, in which one colony harbored pure gene-deleted cells while the other four colonies harbored mixed wild-type and gene-deleted cells (Fig. 5B and Table 3). The deletion events were further confirmed by sequencing (Fig. 5C). Furthermore, we investigated the influence of ccpA deletion on the growth of C. symbiosum ATCC 14940, and found that the mutant strain (ΔccpA) exhibited an improved growth compared with the control strain (Fig. 5D); when ccpA was re-introduced into the mutant strain via an expression plasmid, the growth of the resulting strain (ΔccpA-P₁₃₃₉-ccpA) was partially restored (Fig. 5D). These results confirmed that the growth change caused by ccpA deletion was reliable.

Fig. 5.

CRISPR-Cas12a-basded gene deletion in C. symbiosum ATCC 14940 using the single-plasmid system. (A) Schematic diagram of the gene deletion system. P₀₁₄₄₀: the reported promoter from C. ljungdahlii DSM13528; P_thl: the reported promoter from C. acetobutylicum ATCC 824; RHA: right homologous arm; LHA: left homologous arm. (B) PCR screening of the ccpA deletion events. (C) Confirmation of the ccpA deletion by sequencing. (D) The growth of the ccpA-deleted, ΔccpA-P₁₃₃₉-ccpA, and control strains. Control: the wild-type strain containing the plasmid pMTL82151; ΔccpA: the ccpA-deleted strain containing the plasmid pMTL82151; ΔccpA-P₁₃₃₉-ccpA: the ccpA-deleted strain complemented with the plasmid-carried ccpA expression. Data are represented as mean ± SD (n = 3).

Table 3.

Efficiency of ccpA deletion in C. symbiosum.

Plasmids	Elements	Results (M/P/W/T)a	Efficiencyb (%)	Conjugation efficiency
Control	Cas12a + crRNA	none	–	–
pMTLcas-ΔccpA	Cas12a + crRNA + Homologous arms	4/1/27/32	15.6 %	2.48 ± 0.28 × 10−8

^aM: number of colonies that harbored mixed wild-type and genedeleted cells; P: number of colonies that harbored pure gene-deleted cells; W: number of colonies that harbored pure wild-type cells; T: total number of colonies used for PCR screening.

^bEfficiency: probability of deletion events occurring, calculated as (M + P)/T × 100 %.

Although we have achieved CRISPR-Cas12a-based gene deletion in C. symbiosum ATCC 14940, the efficiency of the conjugative transfer and gene deletion was unacceptable (Table 3), thereby requiring a further promotion. Considering that the conjugative plasmid transfer efficiency is usually associated with the size of plasmid [31,32], we thus allocated the elements of the above CRISPR-Cas construct (Cas12a, crRNA, and homologous arms) to two plasmids: (i) pMTL-cas12a carrying Cas12a; (ii) pMTL-crRNA carrying crRNA and homologous arms (Fig. 6A). Additionally, for efficient expression of Cas12a and crRNA, multiple endogenous promoters in C. symbiosum ATCC 14940 were identified and subjected to activity assay (Fig. 6B). Among them, two stronger promoters (P_fba2 and P_gpm), together with the aforementioned P_thl and P₀₁₄₄₀ promoters, were chosen for driving the expression of Cas12 or crRNA, generating 16 pairs of plasmids (Fig. 6A). These plasmids were then transferred into C. symbiosum ATCC 14940, yielding the transformed cells that harbored eight different plasmid pairs (the other eight plasmid pairs did not give transformed cells due to unknown factors). A number of transformed cells were subjected to PCR screening to detect the desired double-crossover recombination events. The results showed that the deletion efficiencies of these dual-plasmid systems for the ccpA gene varied greatly (ranging from 30 % to 90 %) with the highest efficiency obtained by the combination of pMTL-P_fba2-cas12a and pMTL-P_fba2-crRNA (Fig. 6C). Interestingly, P_fba2 was not the strongest within the abovementioned four promoters; thus, it seems that an appropriate expression level of the Cas12a protein and crRNA was crucial for the editing efficiency in C. symbiosum. Obviously, compared with one plasmid that carried all the genetic elements (Fig. 6C), the dual-plasmid system exhibited higher editing efficiency. Actually, the advantage of the dual-plasmid system over single plasmid in CRISPR-Cas12a-based genome editing has also been found in some other Clostridium species, mainly attributing to the smaller plasmid size of the former [[33], [34], [35]].

Fig. 6.

CRISPR-Cas12a-basded gene deletion in C. symbiosum ATCC 14940 using a dual-plasmid system. (A) Adjustment of the expression of Cas12a and crRNA with different promoters. (B) The strength of the tested endogenous promoters in C. symbiosum ATCC 14940. The LacZ activity was used to indicate promoter strength. Control: no promoter driving the lacZ expression. (C) Editing efficiencies of the transformed cells that harbored different plasmid pairs. Data are represented as mean ± SD (n = 3).

3.6. Identification of bacterial virulence factors via the genetic toolbox

The establishment of an efficient genetic toolbox enabled in-depth studies to be carried out in C. symbiosum. Short-chain fatty acids are crucial metabolites of gut microbial for maintaining intestinal homeostasis [36]. For example, acetate produced by intestinal bacteria can induce apoptosis by inhibiting cell proliferation, thereby inhibiting the occurrence and progression of colorectal cancer [37,38]. To investigate the influence of acetic acid synthesis on C. symbiosum, the ackA gene (CSYM_02545), encoding acetate kinase (Fig. 7A), was repressed through the CRISPRi system (Fig. S1), yielding the strain (ackAi) to check phenotypic changes. As shown in Fig. 7B and C, the ackAi strain exhibited improved growth compared with the control strain, although the acetic acid formation decreased after ack repression. This finding is interesting, because it is quite different from the previous report regarding other Clostridium species, in which the disruption of acetic synthesis normally impaired cell growth [11]. We also overexpressed ackA in the wild-type strain, and the resulting strain (WT-ackA) showed increased acetic acid formation but impaired growth (Fig. 7E and F), which were just opposite to the above results from ackA repression (Fig. 7B and C). Additionally, the ackAi strain showed higher promotion effect on the proliferation of colon cancer cells compared with the control strain (Fig. 7D), indicating a crucial role of acetate in inhibiting cell proliferation. Taken together, all these findings revealed the phenotypic changes generated by altering ackA expression in C. symbiosum ATCC 14940.

Fig. 7.

The influence of the downregulation of ack and the deletion of CSYM_02043/CSYM_03550 on the performance of C. symbiosum ATCC 14940. (A) Schematic diagram of the acetate synthesis pathway. (B–D) The changes in growth, acetate synthesis, and virulence of C. symbiosum ATCC 14940 after the repression of ackA. Control: the wild-type strain harboring the plasmid that carried ddCas12a but no crRNA; ackAi: the wild-type strain containing the CRISPRi plasmid pMTLddcas12a-ackAi. (E, F) The changes in the growth and acetate synthesis after ackA overexpression. Control: the wild-type strain containing the plasmid pMTL82151; WT-ackA: the wild-type strain containing the pMTL82151-P₁₃₃₉-ackA plasmid. (G, H) The changes in the growth and virulence of C. symbiosum ATCC 14940 after CSYM_02043 deletion. Control: the wild-type strain; Δ02043: the CSYM_02043-deleted strain. (I, J) The changes in the growth and virulence of C. symbiosum ATCC 14940 after CSYM_03550 deletion. Control: the wild-type strain; Δ03550: the CSYM_03550-deleted mutant. Data are represented as mean ± SD (n = 3). Statistical analysis was performed by a two-tailed Student's t-test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, no significance.

As aforementioned, C. symbiosum is a potential pathogenic bacterium related with colorectal cancer [9], in which virulence genes may play crucial roles. Based on the genomic data, a large number of putative virulence genes were found in C. symbiosum ATCC 14940 (Fig. S2). Among them, CSYM_02043 and CSYM_03550, two potentially crucial virulence genes encoding a glucose-1-phosphate thymidylyltransferase (catalyzing the conversion of glucose-1-phosphate to dTDP-d-glucose) and a rearrangement hotspot (RHS) repeats-containing protein, respectively, were picked out and subjected to the deletion. It has been known that peptidoglycan containing l-rhamnose in enteric pathogens can serve as an antigen to activate the immune system of intestinal epithelial cells, thereby leading to a series of inflammatory reactions if the host's immune response is excessive [39]. Additionally, RHS repeat-containing proteins are known for their diverse functions and roles in intercellular communication and cell-cell contact [40,41]. Through the abovementioned dual-plasmid CRISPR-Cas12a editing system, the CSYM_02043 and CSYM_03550-deleted strains were rapidly obtained with 97 % and 95 % deletion efficiency, respectively (Figs. S3 and S4), further confirming the utility of this editing system in C. symbiosum. The following phenotypic analyses showed that the deletion of CSYM_02043 and CSYM_03550 had the opposite effect on cell growth (Fig. 7G and I), but both led to the reduced promotion effect on the proliferation of colorectal cancer cell HCT116 compared with the wild type strain (Fig. 7H and J). This finding suggests that these two gene are closely associated with the growth and pathogenesis of C. symbiosum ATCC 14940.

4. Conclusions

Efficient genome editing of enteric pathogens is essential for the comprehensive discovery of genes associated with important physiological processes as well as pathogenesis in these organisms. In this study, we established a toolbox in C. symbiosum ATCC 14940, a conditionally pathogenic bacterium, achieving efficient gene overexpression and CRISPR-Cas12a-based gene deletion and repression. Based on this toolbox, multiple genes involved in synthesis of important metabolites and pathogenesis were identified, exemplifying the utility of this toolbox. The tools developed in this study fill a major gap in the existing genetic tools for intestinal C. symbiosum, providing opportunities for understanding and engineering this important intestinal bacterium.

CRediT authorship contribution statement

Pengjie Yang: Investigation, Validation, Writing – original draft. Jinzhong Tian: Methodology. Lu Zhang: Resources. Hui Zhang: Formal analysis. Gaohua Yang: Investigation. Yimeng Ren: Resources. Jingyuan Fang: Conceptualization, Resources. Yang Gu: Conceptualization, Writing – review & editing, Supervision, Project administration. Weihong Jiang: Conceptualization, Writing – original draft, Supervision, Project administration.

Declaration of competing interest

The authors declare that they have no competing interests.

Appendix A. Supplementary data

The following is the Supplementary data to this article.