Improved CRISPR/Cas9 Tools for the Rapid Metabolic Engineering of Clostridium acetobutylicum

Abstract

Clustered regularly interspaced short palindromic repeats (CRISPR)/Cas (CRISPR-associated proteins)9 tools have revolutionized biology—several highly efficient tools have been constructed that have resulted in the ability to quickly engineer model bacteria, for example, Escherichia coli. However, the use of CRISPR/Cas9 tools has lagged behind in non-model bacteria, hampering engineering efforts. Here, we developed improved CRISPR/Cas9 tools to enable efficient rapid metabolic engineering of the industrially relevant bacterium Clostridium acetobutylicum. Previous efforts to implement a CRISPR/Cas9 system in C. acetobutylicum have been hampered by the lack of tightly controlled inducible systems along with large plasmids resulting in low transformation efficiencies. We successfully integrated the cas9 gene from Streptococcus pyogenes into the genome under control of the xylose inducible system from Clostridium difficile, which we then showed resulted in a tightly controlled system. We then optimized the length of the editing cassette, resulting in a small editing plasmid, which also contained the upp gene in order to rapidly lose the plasmid using the upp/5-fluorouracil counter-selection system. We used this system to perform individual and sequential deletions of ldhA and the ptb-buk operon.

1. Introduction

Multiple genetic engineering tools have previously been developed for use in Clostridium acetobutylicum; the first tools developed were those based on the group II intron gene inactivation system [1,2]. Several different tools have also been developed for in-frame gene deletion and insertion on the basis of allele-coupled exchange; these use either replicative or non-replicative plasmids [3,4,5,6,7]. However, they are limited by time-consuming screening steps, as in all cases the chance of finding the intended modification is at best 50%; yet, in practice, this can be significantly lower. Additionally, with the exception of Foulquier et al. [3], the deletion of multiple genes at the same time is not possible.

CRISPR (clustered regularly interspaced short palindromic repeats)/Cas (CRISPR-associated proteins) tools have recently shown great potential for the genetic engineering of multiple organisms [8]; however, CRISPR/Cas systems are currently not widely used for the engineering of C. acetobutylicum. Initial efforts at using the CRISPR/Cas9 system from Streptococcus pyogenes in C. acetobutylicum and other related clostridia encountered issues of toxicity when trying to constitutively express cas9 [9]. To solve this issue, researchers have used several strategies—in most cases, inducible systems have been used to control the expression of cas9, however, this has had limited success as current inducible systems are not sufficiently tightly controlled resulting in low transformation efficiencies [10,11,12]. A recent paper showed that a riboswitch-based system was able to efficiently control the expression of cas9 in several different clostridia; however, this system was not tested in C. acetobutylicum [13].

Cas9 nickase, in which Cas9 has been mutated so that it can only cut one strand of DNA, has been shown to be less toxic and has been used successfully in several clostridial species; however, generally lower efficiencies have been reported [14,15]. Additionally, although the CRISPR/Cas9 system from S. pyogenes is the most widely used, other CRISPR systems have also been used, for example, CRISPR/Cpf1. CRISPR/Cpf1 systems have been reported to be less toxic, although a tightly regulated inducible system would still be required [16,17,18,19]. Some clostridia also possess endogenous CRISPR systems and these have been successfully exploited in several clostridia [20,21,22]; however, analysis of C. acetobutylicum’s genome showed no putative CRISPR systems. Finally, in a recent paper, an anti-Cas9 protein was used to inhibit the activity of Cas9 in C. acetobutylicum, as the anhydrotetracycline inducible promoter used was unable to sufficiently control the expression of cas9 [12].

Additionally, the size of the plasmids used is also an issue; this is as having the cas9 gene plus the editing template generally results in large plasmids over 10 kb, resulting in low transformation efficiencies. This issue can be avoided by using a two-plasmid system, in which one plasmid contains the cas9 gene, and another plasmid which expresses the guide RNA (gRNA) and contains the editing template [11]. However, it is sometimes advantageous to overexpress one gene (on a plasmid) and delete another gene, resulting in both all in one plasmid systems and two plasmids systems being inconvenient.

In conclusion, an ideal CRISPR/Cas9 system would have the cas9 gene under the control of a very tightly controlled inducible promoter and allow high transformation efficiencies of the editing plasmid containing the gRNA and editing template. A Cas9 cassette was constructed consisting of the cas9 gene from S. pyogenes under the control of the tightly controlled xylose-inducible promoter from Clostridium difficile; this cassette was successfully integrated into the genome. We secondly evaluated the expression of the guide RNA and determined the minimum length of homology arms required; this allowed us to obtain very high transformation efficiencies for our editing plasmid. We managed to achieve up to 100% editing efficiency for multiple different genes.

To speed up the editing process, we also used the upp/5-fluorouracil counter-selection system to quickly remove the editing plasmid. Using this method, we were able to quickly delete individually and sequentially ldhA and the ptb-buk operon; this was achieved significantly faster than using currently available techniques. Finally, we demonstrated that the Cas9 cassette can be easily removed from the genome by expressing a gRNA targeting cas9 with the relevant homology arms, restoring the region to the wild type, creating a strain with only the desired modifications.

2. Results and Discussion

It is clear that an ideal CRISPR/Cas9 system would need to be under the control of a tightly controlled inducible promoter; previous studies clearly showed that current inducible systems used in C. acetobutylicum to control the expression of cas9 were insufficient, with the use of either NCas9 or an anti-Cas9 protein required [10,12].

A literature search was undertaken to search for other inducible systems that have previously been shown to work in clostridia. One potential candidate was the xylose-inducible system from C. difficile [23]. Importantly, this system was also shown to be able to efficiently control the expression of cas9 when used in Clostridium beijerinckii [24].

2.1. Integration of Cas9 into the Genome

We decided to integrate the Cas9 cassette into the genome in order to avoid the use of both large all-in-one plasmid systems and two-plasmid systems. The pyrE locus was chosen as it has previously been used for the expression of genes in C. acetobutylicum [7]. The wild-type cas9 gene from S. pyogenes was cloned under the control of the xylose inducible system between ≈1 kb homology arms. A gRNA targeting the intergenic region between pyrE and hydA (which is not present in the plasmid) was added outside of the homology region (Figure 1A). The gRNA was placed under the control of the j23119 promoter, which has previously been used for expression of the gRNA in C. acetobutylicum and other clostridial species [10,13], creating plasmid pINT_Cas9 (see Figure 1A).

This plasmid pINT_Cas9 was then transformed into C. acetobutylicum MGCΔcac1502, which is able to be transformed without any previous plasmid methylation [5]; cas9 was then subsequently induced by plating transformants on Clostridium growth medium (CGM) plates containing xylose (see Section 4). Colonies growing on xylose were subsequently re-streaked twice on plates containing Reinforced Clostridial Medium (RCM) without any antibiotic in order to lose the plasmid. Once plasmid loss was confirmed by colonies being unable to grow on RCM plates containing thiamphenicol, PCR screening of four independent clones showed that the Cas9 cassette was correctly integrated, indicating 100% editing efficiency (see Figure 1B). Correct integration of the Cas9 cassette was later confirmed by DNA sequencing. The resultant strain was called CAS1 (Δcac1502, pyrE::Cas9). The Cas9 cassette has subsequently been integrated into several other strains in our lab, and each time integration has been achieved with 100% efficiency.

Subsequently, MGCΔcac1502 and CAS1 were grown in liquid CGM medium with either glucose or xylose as the sole carbon source. A Western blot using an antibody raised against the Cas9 protein was subsequently performed on crude cell extracts. For MGCΔcac1502, as expected, no expression of cas9 was detected on either glucose or xylose. On the other hand, for CAS1, no expression of cas9 was detected when grown on glucose, while a high level of expression was detected when grown on xylose, implying tight control of the system (Figure 2).

2.2. Determination of the Efficiency of the CRISPR/Cas9 System

The upp gene was then selected to test the efficiency of the CRISPR/Cas9 system; several vectors were constructed with the gRNA under control of either the j23119 or the miniPthl promoter [11], with and without 1000 bp homology arms, creating pGRNAminiPthl Δupp_HA1000, pGRNAminiPthlΔupp, pGRNAJ23119ΔuppHA1000, and pGRNAJ23119Δupp.

These plasmids were then transformed into CAS1, along with pCons2-1 as a control. A very high transformation efficiency was seen for all plasmids (see

Table 1. Transformation efficiencies of strain CAS1 with pCons2-1, pGRNAminiPthl Δupp_HA1000, pGRNAminiPthlΔupp, pGRNAJ23119ΔuppHA1000, and pGRNAJ23119Δupp. Values are expressed in number of transformants per microgram DNA. Mean values and standard deviations from three independent experiments are given. A total of 0.5 μg of DNA was used in each experiment.

Plasmid	pCons2-1	pGRNAminiPthl Δupp_HA1000	pGRNAminiPthlΔupp	pGRNAJ23119ΔuppHA1000	pGRNAJ23119Δupp
CFU/μg of DNA	5.13 (±0.51) × 103	5.45 (±0.67) × 103	7.81 (±0.84) × 103	7.53 (±0.58) × 103	7.66 (±0.72) × 103

), implying no toxicity related to the expression of the gRNA, again indicating that cas9 is tightly controlled. This is in contrast to the low transformation efficiencies seen using previous systems, presumably due to background expression of cas9 [12].

Individual colonies were then grown in 1 mL of CGM until an optical density (OD) of 0.6 and plated on CGM xylose plates containing the relevant antibiotic. We achieved 100% efficiency with both the j23199 and miniPthl promoter (48 clones tested for each—see Supplementary Figures S1 and S2); however, some clones appeared on the plate containing only the gRNA (15 Colony forming units (CFU)/mL compared to 2 × 10³ CFU/mL with homology arms). This is likely due to escape mutations in either the gRNA or Cas9 cassette. Several clones were then re-streaked twice on RCM without antibiotics in order to lose the editing plasmid. Once plasmid loss was confirmed by colonies being unable to grow on RCM plates containing thiamphenicol, several clones were saved creating strain CAS2 (Δcac1502, pyrE::cas9, Δupp).

For clostridia, normally 1000 bp homology arms have been used; however, it is possible that shorter homology arms could be used. The main advantage of short homology arms is that it becomes economically viable to order synthetic DNA containing the gRNA and homology arms, which would significantly streamline the cloning process, as currently one step of cloning needs to be done to clone the homology arms and a second step to add the gRNA sequence. Additionally, if shorter homology arms could be used, the plasmid would be significantly smaller, which could be helpful for multiplex gene editing, as it would reduce the total size of the plasmid.

Plasmids were constructed which contained 500, 250, or 100 bp homology arms with the gRNA targeting the upp gene under the control of the miniPthl promoter. These plasmids were subsequently transformed into CAS1. After induction, a similar number of colonies were obtained for both 1000 bp and 500 bp homology arms (2 × 10³ CFU/mL); however, significantly fewer clones were observed after induction for plasmids containing 250 and 100 bp homology arms (1 × 10² CFU/mL).

The 500 bp homology arms had almost the same efficiency as 1000 bp with 96% efficiency, with 45 correct clones and two clones that were a mixed population of wild type and Δupp (see Figure S3). For 250 bp, 91% efficiency was achieved, with 37 correct clones, 1 clone that was a mixed population, and 3 clones that were wild type (see Figure S4). On the other hand, for 100 bp homology arms, 0% efficiency was achieved (47/47, see Figure S5). This implies that for simple deletions and insertions, 500 bp homology arms can be used (see

Table 2. Comparing the editing efficiency of the CRISPR/Cas9 system targeting the upp gene using both the J23119 or MinPthl Promoter and either 1000, 500, 250, or 100 bp homology arms.

Promoter	J23119	MiniPthl
Size of homology arms	1000 bp	1000 bp	500 bp	250 bp	100 bp
Editing efficiency	100%	100%	96%	91%	0%

).

2.3. Deletion of ldhA and the ptb-buk Operon

The efficiency of the CRISPR/Cas9 system was then tested on two other genes: ldhA and the ptb-buk operon. These genes were chosen as a clean deletion of the ptb-buk operon has never been performed in C. acetobutylicum; deletion of both ptb-buk and ldhA should result in increased solvent yields, as lactate and butyrate should not be produced as side products.

Plasmids targeting ldhA and the ptb-buk operon were constructed. The gRNA targeting ldhA was placed under the control of the j23119 promoter while the gRNA for the ptb-buk operon was placed under the control of the miniPthl promoter. Additionally, to assist with the loss of the plasmid, the upp gene was included on the plasmid backbone; this meant that after selection on xylose the plasmid could then be easily lost by plating on plates containing 5-fluorouracil [5]. This meant it took around 7 days to delete each gene and have a strain ready for characterization or another round of genetic manipulation (see Figure 3).

The plasmids pGRNAΔldhA and pGRNAΔptb-buk were transformed into CAS2, ldhA was deleted with 100% editing efficiency (14/14 clones tested, see Figure S6), and the ptb-buk operon was deleted with 93% efficiency (13/14 clones tested, see Figure S7). We found that 100% of colonies screened were sensitive to thiamphenicol after growth on plates containing 5-fluorouracil, creating the strains CAS2ΔldhA and CAS2Δptb-buk.

The plasmid pGRNAΔptb-buk was then transformed into CAS2ΔldhA, and deletion of the ptb-buk operon was again achieved with 93% efficiency (13/14 clones tested). We found 100% of colonies screened to be sensitive to thiamphenicol after growth on plates containing 5-fluorouracil, creating the strain CAS2ΔldhAΔptb-buk. A growth curve and solvent analysis was then performed for strains MGCΔcac1502, CAS2, CAS2ΔldhA, CAS2Δptb-buk, and CAS2ΔldhAΔptb-buk (Figure 4).

Analysis of the growth curve and solvent profile of MGCΔcac1502 and CAS2 showed a very similar profile, indicating integrating the Cas9 cassette into the genome had no effect on growth or solvent formation again implying no or very low background expression of cas9.

Deletion of ldhA resulted in a strain with a very similar growth and solvent profile to both MGCΔcac1502 and CAS2 with the only difference being very low levels of lactate being produced (3 vs 7 mM) (Figure 4).

Both CAS2Δptb-buk and CAS2ΔldhAΔptb-buk grew slower compared to the control strains (Figure 4A). CAS2Δptb-buk produced a molar ratio of 4.5:1.5:4 of butanol, acetone, and ethanol compared to the normal ratio of 6:3:1. CAS2Δptb-buk also produced slightly higher levels of acetate and lactate, and as expected, also produced very low levels of butyrate.

CAS2ΔldhAΔptb-buk produced a molar ratio of 3:1:7 of butanol, acetone, and ethanol. The strain also produced high levels of acetate, implying the strain was not able to properly re-uptake acetate, explaining the low level of acetone produced. As expected, CAS2ΔldhAΔptb-buk also produced very low levels of butyrate and lactate (Figure 4B).

This is the first clean deletion of the ptb-buk operon with all other mutants targeting either buk or ptb. Previous mutants were created using either a disruptive non-replicative plasmid/group II intron gene inactivation system to delete either buk or ptb; these deletion systems can often disrupt the expression of neighboring genes. Previous efforts to create a clean deletion of buk resulted in spontaneous partial deletions of ptb, implying that a Δbuk mutant is not viable [25].

Various different phenotypes have been described in different Δbuk/Δptb mutants. G.N. Bennett et al. [26] inactivated buk using a disruptive non-replicative plasmid—the resultant strain produced higher levels of butanol and acetate compared to the wild type, with a decrease in butyrate production. In a subsequent paper, ptb was disrupted using a group II intron gene inactivation system—this resulted in a strain that in pH-uncontrolled fermentations produced high levels of acetate and lactate while producing no acetone or butyrate and low levels of butanol. However, in pH-controlled fermentations, the strain produced high levels of ethanol and acetone [27]. Finally, P. Soucaille et al. [25] attempted a clean deletion of buk; however, spontaneous mutations in ptb were identified. The resultant strain showed increased butanol production compared to the wild type, with no increase in ethanol or acetone production. It is possible that this difference in phenotype is due to some residual activity of ptb or other secondary mutations. Detailed genetic analysis and characterization of these strains would be useful to determine the reason for these differing phenotypes, however, are outside the scope of this study.

2.4. Removal of Cas9

Growth curve and solvent analysis showed that integration of the cas9 gene under the control of the inducible promoter does not affect either growth or solvent production indicating that removal of the Cas9 cassette would not be routinely required. However, in certain cases it would be desirable to have a clean strain; we showed that the Cas9 cassette could easily be removed using plasmid pGRNAΔcas9, consisting of a gRNA targeting the Cas9 cassette along with the relevant homology arms cloned into an appropriate vector. The Cas9 cassette was successfully removed from strain CAS2 using this plasmid with 100% efficiency (14/14 clones tested, see Figure S8), restoring the pyrE locus to wild type.

3. Conclusions

In this paper, we were able to achieve high-efficiency genetic engineering of C. acetobutylicum by using a tightly controlled xylose inducible system. We showed that the Cas9 cassette could be quickly integrated into the genome and that this had no effect on growth or solvent formation when grown on glucose. We achieved very high editing efficiency of three different genes (upp, ldhA, and the ptb-buk operon) as well as the integration and subsequent removal of the Cas9 cassette. A table summarizing the editing efficiency is shown in

Table 3. Summarizing the efficiency of genetic modifications performed using the CRISPR/Cas9 system.

Gene	Integration of Cas9 Cassette	Deletion of upp	Deletion of ldhA	Deletion of ptb-buk Operon	Removal of Cas9 Cassette
Efficiency	100%	100%	100%	93%	100%
size ofdeletion/insertion	5872 bp	630 bp	942 bp	2001 bp	5872 bp

.

This system offers significant improvements on previously developed CRISPR/Cas9 systems for C. acetobutylicum, mainly the use of the tightly controlled xylose-inducible promoter from C. difficile, resulting in no background expression of cas9, which was not the case for the previously described lactose and anhydrotetracycline systems. Additionally, we have optimized the size of the homology arms and shown that the editing plasmid can quickly be lost using the upp/5-fluorouracil system. Importantly, integration of the Cas9 cassette into the genome means only one editing plasmid is necessary, important in C. acetobutylicum, as only two antibiotic-resistant markers are available.

We envisage that this system will significantly decrease the time needed to genetically modify C. acetobutylicum, which should result in more ambitious metabolic engineering projects being attempted. With most papers only performing one or two deletions/modifications, while to our knowledge the highest number of genes modified has been eight [28].

4. Materials and Methods

4.1. Growth Conditions

C. acetobutylicum ATCC 824 was grown anaerobically at 37 °C in either a synthetic medium (MS), in Clostridium growth medium (CGM), or Reinforced Clostridial Medium (RCM), as described previously [3]. Glucose or xylose was added to the media at a final concentration of 5%. Solid media was obtained by adding 1.5% agar to the liquid media. Media was supplemented, when required, with the appropriate antibiotic at the following concentration: thiamphenicol at 15 µg/mL. 5-Fluorouracil (5-FU) was purchased from Sigma, and stock solutions were prepared in DMSO. Transformation of C. acetobutylicum was carried out as described previously [3].

Escherichia coli TOP10 was grown at 37 °C in Luria–Bertani medium, supplemented when required, with the appropriate antibiotic at the following concentration: chloramphenicol (25 µg/mL).

4.2. DNA Isolation and Manipulation

Total genomic DNA from C. acetobutylicum ATCC 824 was isolated using GenElute Bacterial Genomic DNA Kits (Sigma, Munich, Germany). Plasmid DNA was extracted from E. coli with the NucleoSpin Plasmid (Macherey-Nagel, Düren, Germany). DNA restriction enzymes, T4 DNA ligase, Phusion polymerase, and NEBuilder HiFi DNA Assembly were obtained from NEB (New England BioLabs, Ipswich, MA, USA). DNA fragments were purified from agarose gels with the QIAquick gel purification kit (Qiagen, Hilden, Germany). PCR was carried out using chromosomal DNA as a template using Phusion polymerase (NEB, Ipswich, MA, USA). The detailed construction of each plasmid is described in the Supplementary Materials along with a list of primers used (Supplementary Table S2), and the gRNA sequences (Table S1). A list of plasmids used is shown in

Table 4. List of plasmids.

Plasmids	Relevant Characteristics	Source or Reference
pCons2-1	Cmr, repL	[5]
pCons::upp	Cmr, repL, upp	[5]
pINT_cas9	Cmr, repL, pyrE::pxyl_cas9	This study
pGRNAminiPthl Δupp_HA1000	Cmr, repL, miniPthlgRNAupp, Δupp	This study
pGRNAminiPthlΔupp	Cmr, repL, miniPthlgRNAupp,	This study
pGRNAJ23119ΔuppHA1000	Cmr, repL, J23119PthlgRNAupp	This study
pGRNAJ23119Δupp	Cmr, repL, J23119PthlgRNAupp, Δupp	This study
pGRNAminiPthl Δupp_HA500	Cmr, repL, miniPthlgRNAupp, Δupp	This study
pGRNAminiPthl Δupp_HA250	Cmr, repL, miniPthlgRNAupp, Δupp	This study
pGRNAminiPthl Δupp_HA100	Cmr, repL, miniPthlgRNAupp, Δupp	This study
pGRNAΔldhA	Cmr, repL, upp, J23119gRNAldhA, ΔldhA	This study
pGRNAΔptb-buk	Cmr, repL, upp, miniPthlgRNAbuk, Δbuk-ptb	This study
pGRNAΔcas9	Cmr, repL, upp, J23119gRNAcas9, Δcas9	This study

.

4.3. Analytical Methods

Cell growth was monitored by measuring optical density at 600 nm (OD600). Solvent and acid production, as well as glucose consumption in cell-free supernatant samples, were determined on the basis of high-performance liquid chromatography (HPLC) [29] using H₂SO₄ at 0.5 mM as mobile phase.

4.4. Western Blot

To determine the expression of cas9, we grew the relevant strains in CGM liquid with either glucose or xylose to an OD of 1.0. A total of 10 mL of cell culture was then centrifuged at 6500 RPM (Revolutions per minute of rotor), 4 °C for 10 min. The cell pellet was then stored at –20 °C for 16 h. The cell pellet was then resuspended to an OD of 10 in Phosphate Buffered Saline (PBS). SDS-PAGE and Western blot were then performed using a Horseradish-peroxidase (HRP)-conjugated antibody raised against Cas9 from S. pyogenes (Abcam -ab202580).

4.5. Construction of Strains

After transformants were obtained on RCM media containing the relevant antibiotic, we inoculated several colonies into 1 mL of liquid CGM-glucose, and after several hours of incubation, serial dilutions were performed and 100 uL of culture was spread on solid CGM-xylose plates. After around 48 h of incubation, colonies were screened by colony PCR using relevant primers (see Table S2).

In the case of constructs in which the upp gene was not present on the plasmid backbone, we re-streaked individual colonies on RCM plates to obtain single colonies. Several colonies were then screened by streaking on RCM plates with and without the relevant antibiotic; colonies that grew without antibiotic but did not grow in the presence of the antibiotic were presumed to have lost the plasmid.

In the case of constructs in which the upp gene was present on the plasmid backbone, individual colonies were inoculated in 1 mL of CGM-glucose, and after several hours of incubation, 100 uL of culture was spread on a plate containing CGM-glucose supplemented with 2-(N-morpholino)ethanesulfonic acid _(MES) (15 g/L) and 5-FU at 0.1 mM. Colonies were then screened by streaking on RCM plates with and without the relevant antibiotic; colonies that grew without antibiotic but did not grow in the presence of the antibiotic were presumed to have lost the plasmid. The list of strains used is shown in

Table 5. List of strains.

Strain	Relevant Characteristics	Source or Reference
Bacterial strains
E. coli TOP10		Invitrogen
C. acetobutylicum
MGCΔcac1502	ΔCA_C1502	[5]
CAS1	ΔCA_C1502, pyrE::pXyl_cas9	This study
CAS2	ΔCA_C1502, pyrE::pXyl_cas9, ΔCA_C2879	This study
CAS2ΔldhA	ΔCA_C1502, pyrE::pxyl_cas9, ΔCA_C2879, ΔCA_C0267	This study
CAS2Δptb-buk	ΔCA_C1502, pyrE::pxyl_cas9, ΔCA_C2879, ΔCA_C3075, ΔCA_C3076	This study
CAS2ΔldhAΔptb-buk	ΔCA_C1502, pyrE::pxyl_cas9, ΔCA_C2879, ΔCA_C0267, ΔCA_C3075, ΔCA_C3076	This study

.