Streamlined assembly of cloning and genome editing vectors for genus Clostridium

Summary Reported herein is a new set of vectors designed to streamline molecular cloning and genome editing by exploiting modern cloning methods. The new vectors build on the existing pMTL8000 vectors that have been a staple of Clostridium research for more than a decade. The introduction of two pairs of type IIS restriction sites flanking an insulated multiple cloning site in both a cloning vector and a CRISPR-Cas9 gene editing vector enables plasmid construction in a “one-pot” reaction, avoiding the more laborious steps of conventional cloning. A synthetic lacZα expression cassette introduced between the cloning sites enables visual detection of background colonies. In addition, distinct selection markers on each vector permit selection of the desired clones according to antibiotic resistance. An example of strain development using the new vectors is demonstrated.


INTRODUCTION
Interest in species of genus Clostridium has increased considerably over the last 20 years. While early research efforts focused on pathogenic species, more recent studies have highlighted the ubiquity of clostridia, both in the environment and in mammalian hosts. 1,2 In addition, a rapidly expanding list of species relevant to industrial production of solvents, biofuels, and chemical precursors has attracted attention, aided by increasing demand for fossil fuel alternatives. 3,4 This enthusiasm is reflected in the literature, with publications containing ''Clostridium'' in their title increasing from 300 per year in the 2000s to a pre-pandemic peak of 1300 per year.
The community of researchers that study Clostridium species have relied heavily on the pMTL8000 vectors, a series of modular plasmids designed to be easily adapted to the species of interest and the needs of the experiment. These vectors have been a major facilitator for research in this field, enabling gene knock-outs and complementation studies as well as overexpression of native or heterologous proteins, among other uses. Published in 2009, the pMTL8000 vectors functioned as both cloning and expression vectors that could be shuttled between E. coli cloning strains and the Clostridium species of interest. Conserved restriction sites between functional modules, consisting of a gram-positive replicon, selection marker, gramnegative replicon, and ''application-specific'' module enabled a simple method for adapting the vector to different clostridia and different purposes. These vectors have spawned several mutagenesis systems, including intron-directed gene disruption, 5 a mariner transposon system, [6][7][8][9] an allelic exchange system, [10][11][12] and, most recently, multiple versions of a CRISPR-Cas system. [13][14][15][16][17][18] In parallel to these advances, a number of new, commercially available cloning methods have emerged, including Gibson assembly and its derivatives 19,20 and golden gate assembly (GGA). 21 Common to all of these techniques are a simplified cloning process, typically involving fewer steps and fewer reagents, and the possibility to clone without ''scars'' (restriction enzyme recognition sequences). GGA has a number of distinct advantages over other techniques. Remarkably complex assemblies are possible using this method. A recent publication demonstrated that precise selection of fusion sites to maximize ligation fidelity enabled assembly of a 40 kb phage genome from 52 fragments. 22 On a practical level, only two enzymes (a restriction enzyme and DNA ligase) are required for GGA, and the inherent simplicity and efficiency of the reaction make this method more accessible for students. Plasmid construction using GGA does require an upfront effort to develop suitable vectors, that is, addition of the desired type IIS restriction sites in the correct orientation and removal of any existing recognition sequences outside of the multiple cloning site (MCS) (domestication). An example of this is the start-stop system, which uses 3-base cutting type IIS restriction enzymes to construct units from promoters, UTRs, coding sequences, and terminators. By Pairs of type IIS restriction sites were added to the original vector via primer tails in two sequential steps (A). The BsaI sites are flanked by characterized terminators to ensure insulated expression. The BsmBI sites flank the terminators so that the entire insulated expression cassette can be transferred to the second vector. BsaI sites can be incorporated into amplicons that are to be assembled into vector pGG222L via primer tails (B). The orientation of the BsaI sites is important to ensure correct Golden Gate assembly. iScience Article sites of their forerunners for easy exchange of backbone modules. The central change in both vectors is the creation of a cloning site that is compatible with BsaI-based GGA and blue-white screening using Xgal. The introduction of additional type IIS sites, flanking the cloning site, enables transfer of ''subcloned'' expression cassettes from cloning vector to gene editing vector in a second GGA reaction. As demonstrated in this paper, the gene editing vector can be ''pre-targeted'' to a known integration site, for which a guide RNA (gRNA) and repair cassette have been validated previously. The results presented here demonstrate that these vectors can support a fast, cost-effective, and user-friendly method for construction of expression and gene editing plasmids for use in Clostridium species.

Vector construction
A number of modifications were made to vector pMTL82221 to support the function of the final vector, pGG222L. The 5 0 MCS terminator was replaced by the validated tyrosyl-tRNA synthetase (tyrS) terminator from Bacillus subtilis. 24 To enable blue-white screening of transformants (blue representing undigested vector), a synthetic lacZa coding sequence was added in between the tyrS and C. pasteurianum ferredoxin (fdx) terminators, flanked by BsaI recognition sites. Up-and downstream of the terminators, an additional pair of type IIS restriction sites (BsmBI) were incorporated. The BsaI sites are oriented so that the recognition sequence remains on the empty MCS after cleavage, while the BsmBI sites are oriented to ensure that the recognition sequence remains on the vector backbone. The four-base sticky ends that result from cleavage at these locations were selected based on several factors: avoiding conflict with other cut sites on the vector, the availability of sequences in the original vector, and maximizing ligation efficiency and fidelity Figure 2. Construction of the gene editing vector and amplicon primer design Five fragments were created by PCR using the previously reported pPME-101-g1 vector as template (A). Sequences were introduced or eliminated through primer tails and selective primer annealing, respectively. The Shine-Dalgarno sequence at the junction between the promoter and coding sequencing of the gRNA and the sequence downstream of the Cas9 gene, corresponding to the M13-R primer annealing site and the CD0164 terminator, were eliminated. A validated terminator was introduced downstream of the Cas9 gene, and the insulated lacZa cloning site with flanking BsmBI sites was added between the SLS homology arms. The sequence and orientation of the type IIS restriction sites are shown (B). Primers used to produce amplicons to be cloned at the BsaI sites must contain the 5 0 sequences shown (B).

OPEN ACCESS
iScience 26, 107484, August 18, 2023 3 iScience Article (determined using the NEB Ligase Fidelity tool -https://ligasefidelity.neb.com/). The sequence of the final vector with annotations is available from NCBI GenBank (GenBank: OQ302270). A detailed description of the cloning process is provided in the STAR Methods section. Figure 1 is a schematic representation of the process and shows the arrangement and orientation of the cloning sites in the final vector. 5' primer sequences for cloning fragments at the BsaI cloning sites are indicated.
The gene editing vector was derived from the plasmid reported previously. 25 Alongside the addition of an MCS within the repair cassette, several modifications were made to the CRISPR-Cas9 vector through primer design that excluded sequences or added them through non-annealing primer tails. The previous terminator and M13R primer site downstream of the Cas9 gene were removed and replaced with the validated terminator from the Lactococcus lactis pepN gene. The BsmBI restriction site present in the native Cas9 coding sequence was removed via the synonymous codon change, CGT to AGA, and the Shine-Dalgarno sequence of the C. acetobutylicum arabinose sugar-proton symporter promoter (ParaE) was deleted. The full name of the resultant vector is pCas9n312-g7SLS-GGL. The sequence of the final vector with annotations is available from NCBI GenBank (GenBank: OQ302271). A detailed description of the cloning steps to produce pGG222L and pCas9n312-g7SLS-GGL is provided in the STAR Methods section. Figures 1 and 2 are schematic representations of the cloning process for pGG222L and pCas9n312-g7SLS-GGL, respectively, and display the arrangement and orientation of the cloning sites in the final vectors. 5' primer sequences for cloning fragments at the BsaI cloning sites of each vector are indicated.

Assembly reactions
An overview of the workflow for assembly using the cloning and CRISPR-Cas9 vectors is given in Figure 3. To evaluate the efficiency of assembly reactions using the new vectors, a previously validated expression cassette was selected. Clostridium-directed enzyme prodrug therapy requires the heterologous expression of a nitroreductase (NTR) enzyme by tumor-colonizing C. sporogenes. This experimental treatment of cancer has been demonstrated previously in a nude mouse xenograft tumor model. 26 To recreate this strain, the native sequence for the C. sporogenes fdx promoter (Pfdx) and the codon-optimized Neisseria meningitidis NTR gene (reported previously) were cloned by PCR. Cloning using the type IIS restriction sites in the new vectors was sufficiently efficient to enable easy detection of correct clones. Plasmids assembled with the promoter and coding sequence fragments produced plates with an abundance of colonies that were predominantly white ( Figure 4A). In contrast, control plates contained significantly fewer colonies, which were predominantly blue ( Figure 4B). The relatively high number of blue colonies on control plates suggests that, in the absence of alternative clonable DNA, the MCS re-ligates to the vector backbone. Blue and white colonies from each plate were counted, and the efficiency of assembly was calculated (Table 1). Colony PCR and gel electrophoresis of white iScience Article colonies (8) indicated that the screened colonies contained the correctly assembled plasmid ( Figure 4C). Sanger sequencing of plasmid from a positive clone revealed correct assembly without mutation. One clone, denoted pGG222-Pfdx-NTR, was taken forward.
The aim of the second assembly reaction is to exchange the lacZa gene of the pCas9n312-g7SLS-GGL vector for the cloned expression cassette of pGG222-Pfdx-NTR. As the BsaI restriction sites of the pGG222L vector are excised during the first assembly reaction, the second reaction relies on the BsmBI sites that flank the MCS. Plating the resultant transformation on chloramphenicol/X-gal plates enables exclusion of two forms of background: undigested pGG222-Pfdx-NTR (chloramphenicol-sensitive) and undigested pCas9n312-g7SLS-GGL (chloramphenicol-resistant blue colonies). This reaction is virtually identical to the first assembly, with the exception that the fragment to be cloned is precloned on the pGG222L plasmid rather than linear PCR products. Transformation and plating revealed a similar result to the first assembly: predominantly white colonies on the cloning plate and a reduced number of colonies on the control plate that were primarily blue ( Figure 5A and 5B). After 24 h incubation, the colonies were noticeably smaller than those of the first assembly. In the authors' experience this is typical, possibly due to the larger size of the vector and the constitutive expression of the Cas9 gene. A PCR screen of randomly selected white colonies indicated that assembly was correct ( Figure 5C). The sequence of one colony was confirmed by Sanger sequencing and taken forward for genome integration.

Genome integration
Genome integration of the NTR expression cassette was initially attempted with the original gRNA, used previously to delete the streptolysin S (SLS) operon homologue from C. sporogenes. 25 Recombination was not detected, neither from the first conjugation selection plate nor following multiple subcultures in liquid with selection (data not shown). To improve the cutting performance of the Cas9 enzyme, the SLS operon was analyzed, using the Benchling CRISPR Guide RNA Design Tool, 27 for target sequences that would cut more efficiently. The

DISCUSSION
Vectors are essential for the study and exploitation of prokaryotic and eukaryotic organisms. While cloning vectors for E. coli have been in use since the 1970s, stable vectors for other species have taken considerably longer to develop. The publication and distribution of the pMTL8000 modular shuttle vector in 2009 were a major catalyst for the study of clostridia, a genus of bacteria that has become increasingly relevant. This publication introduces an update to these vectors that enables the use of the Golden Gate cloning method to assemble fragments directly into the plasmid. In addition, a second CRISPR-Cas9n gene editing vector has been developed to be compatible with the first vector, accelerating the creation of ''knock-in'' mutants. The primary objective of these changes was to make cloning and genome engineering in clostridia more accessible, with both the experienced postdoc and early PhD student in mind.
By relying solely on GGA for plasmid construction, this system can be utilized with only three enzymes: BsaI, BsmBI, and T4 DNA ligase. The gel and PCR cleanup steps required in conventional cloning are not needed, further reducing costs. While this removes the need to stock a collection of enzymes and kits, it does shift the burden to primer design. Web-based tools, such as the NEBridge Golden Gate Assembly Tool (https://goldengate.neb.com/), can assist in primer design. If the end user choses to design primers manually, particular attention must be given to the orientation of the recognition sequence to ensure that it is excluded from the fragment post-cut. It should be noted that in order for cloned fragments to be passed from the cloning vector to the gene editing vector during the assembly reaction, the BsmBI sites have opposite orientations in each vector: in the cloning vector, the recognition sites remain on the backbone, while on the gene editing vector they remain on the MCS fragment. This ensures that undesirable products (cloning vector and empty gene editing vector) are digested, while the intended assembly is not. DNA fragments can be cloned using the BsaI sites, either directly into the gene editing vector (white, Cm-resistant colonies) or via the cloning vector (white, Em-resistant). If the latter is used, the BsmBI sites must be used to transfer from cloning to gene editing vector. Both methods result in the same final expression cassette.
A number of changes were made to the parent vectors based on rational design, with the intention of improving their performance and utility. A synthetic lacZa replaces the MCS in both vectors to enable blue-white screening. The synthetic sequence was designed to remove any restriction sites that are present in the vector backbones to avoid issues in downstream cloning steps. The decision not to alter the promoter sequence of the lacZa does result in two M13R sequences in the MCS. However, this is only relevant for ''empty vector'' clones, which will appear blue on X-gal plates.
For the CRISPR-Cas9 vector, the Shine-Dalgarno sequence of the ParaE promoter, present in previous versions of this vector, was deleted to avoid unwanted translation of the guide RNA, and a characterized terminator was introduced downstream of the Cas9 gene to prevent transcription readthrough into the backbone. The Rhoindependent terminator of the Lactococcus lactis pepN gene was selected due to its documented efficiency in C. acetobuylicum and its small size, 24 which enabled simple cloning as primer tails. The length of the sequence between the translation stop codon and the terminator secondary structure has been shown to affect termination efficiency. 28 To promote efficient termination, the new terminator was incorporated 30 bp downstream of the Cas9 stop codon. The M13R primer annealing site, present downstream of the Cas9 gene in the iScience Article previous vector, was not cloned to avoid an additional primer annealing site in the same vector. By cloning Cas9 in two fragments, it was possible to remove the single BsmBI restriction site in the native coding sequence. This was achieved by changing the codon encoding arginine from CGT to AGA, a preferred codon in cluster I clostridia according to recent codon usage data. 29 In both vectors particular attention was given to creating a transcriptionally ''insulated'' MCS. Due to their strong secondary structure, rho-independent terminators can be challenging to clone. The hairpin can be split between adjacent fragments as primer tails during subcloning, 30 but this is difficult to do at the junction with the vector. The terminator downstream of the MCS, inherited from the pMTL82121 vector, has been validated previously, 31 but the upstream terminator, derived from open reading frame (ORF) CD0164 of C. difficile 630, has not. This was exchanged for the validated terminator from the tyrS gene of Bacillus subtilis. 24 Vector assembly was demonstrated using the C. sporogenes fdx gene promoter and Neisseria meningitidis NTR gene. 26 While GGA appeared to be highly efficient in both vectors, genome integration proved to be very inefficient. Initial attempts using a gRNA sequence that was previously used to delete the SLS locus of C. sporogenes failed to produce integrants. A new screen of the SLS locus yielded a new gRNA target sequence, which was cloned into the gene editing vector. Using the new construct, a single integrant was obtained from a screen of 32 colonies. This result highlights the sharp decrease in the efficiency of genome editing when ''cargo'' is added to the repair cassette of this system.
The motivation for developing these vectors was to support more efficient construction of plasmid-or genome-based recombinant strains and to enable standardized cloning of expression cassettes within research groups. The simplicity of GGA compared to conventional cloning opens up the possibility to clone, transform, and conjugate in microplates, as demonstrated in an equivalent system for Bacteroides species. 32 This must be designed carefully in order to downscale volumes while preserving reactions iScience Article kinetics. However, the creation of these vectors and the demonstration of a 96-well workflow in Bacteroides (which also relies on conjugative transfer from E. coli) make it a considerably easier task. The introduction of these vectors, which are openly available to all researchers, will enable high-throughput experiments that can yield large datasets, and it is our hope that this will accelerate the study of Clostridium biology. In addition, the design of these vectors provides a concept that can be adopted and adapted to any other species, promoting a smarter, more refined method of strain development.

Limitations of the study
The vectors in their current format can be used in bacteria that support the gram-positive or gram-negative replicons present on the vector backbone. The segregational stability of the gram-positive replicons, along with several others, has been reported previously. Depending on the species of Clostridium in which the vectors will be used, the replicon will need to be changed according to the reported data. 33,34 For the CRISPR-Cas9 vector, users should avoid using very stable replicons, which may lead to issues with plasmid loss following genome editing. The CRISPR-Cas9 vector is currently designed to target the SLS operon of C. sporogenes NCIMB 10696. The current homology arms and gRNA must be changed in order to target genome loci in other Clostridium species. The MCS can be retained using the MCS-intBsmBI-F/MCS-intBsmBI-R primer pair (See Table S1) and cloned conventionally using the BsmBI sites. Figure 6. PCR screening the pCas9n312-g7SLS-Pfdx-NTR transconjugants for Pfdx-NTR integration at the SLS locus Primers were designed to anneal to sequences flanking the sites of homologous recombination (SLS-Flank-F and SLS-Flank-R: see Table S1). WT colonies containing the 8.5kb SLS operon do not produce a product. C. sporogenes-NT (''DSLS''), in which the SLS operon has been deleted, produces a 1.8kb product and is provided here as a control for the PCR reaction. The predicted size for Pfdx-NTR integration is 2.8kb. Integration was detected for one colony (number 9), which was subsequently confirmed by sanger sequencing. The major limitation of the current vectors is the apparent reduced potency of the Cas9 nickase, in contrast to the original, unaltered spCas9. In the current vector design, both the Cas9n gene and the gRNA are expressed constitutively. The promoter driving the Cas9 gene, Pthl, was recently shown to produce very high levels of gusA reporter activity in E. coli, but low levels in C. sporogenes and C. butyricum. 35 This suggests that the truncation of the Cas9 gene to a nickase, as reported previously, 13 is likely to have occurred during cloning in E. coli, possibly due to the inherent toxicity of the Cas9 protein. 36 Utilizing the true spCas9 in the gene editing vector is likely to require an inducible promoter for the Cas9 gene, and possibly the gRNA too.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

ACKNOWLEDGMENTS
This project received funding from KWF-Alpe d'HuZes (n 13056/2021). The authors would like to thank Erik Beuken of Maastricht UMC+ for supporting this work with constructive discussions and by sharing lab equipment.

AUTHOR CONTRIBUTIONS
TB designed and constructed the vectors and wrote the manuscript. PH, YZ, and AK participated in construction and validation of the vectors and reviewed the manuscript.