Towards improved butanol production through targeted genetic modification of Clostridium pasteurianum

Declining fossil fuel reserves, coupled with environmental concerns over their continued extraction and exploitation have led to strenuous efforts to identify renewable routes to energy and fuels. One attractive option is to convert glycerol, a by-product of the biodiesel industry, into n-butanol, an industrially important chemical and potential liquid transportation fuel, using Clostridium pasteurianum. Under certain growth conditions this Clostridium species has been shown to predominantly produce n-butanol, together with ethanol and 1,3-propanediol, when grown on glycerol. Further increases in the yields of n-butanol produced by C. pasteurianum could be accomplished through rational metabolic engineering of the strain. Accordingly, in the current report we have developed and exemplified a robust tool kit for the metabolic engineering of C. pasteurianum and used the system to make the first reported in-frame deletion mutants of pivotal genes involved in solvent production, namely hydA (hydrogenase), rex (Redox response regulator) and dhaBCE (glycerol dehydratase). We were, for the first time in C. pasteurianum, able to eliminate 1,3-propanediol synthesis and demonstrate its production was essential for growth on glycerol as a carbon source. Inactivation of both rex and hydA resulted in increased n-butanol titres, representing the first steps towards improving the utilisation of C. pasteurianum as a chassis for the industrial production of this important chemical.

were inoculated from RCMTm15 plates into 1 ml 2x YTGTm15 broth, grown for 12 h and used to re-inoculate another 1 ml 2x YTGTm15 broth with 1% (v/v) inoculum. The fresh culture was incubated for another 12 h before it was washed three times with PBS, re-suspended in 1 ml PBS and used to inoculate 1 ml of un-supplemented 2x YTG broth at a 1% (v/v) inoculum to start the segregational stability assay. For the assay, the plasmid harbouring cells were subcultured every 12 h in un-supplemented 2x YTG broth using a 1% (v/v) inoculum of the preceding culture. Every 24 h twice 100 µl of the well growing culture were serial diluted in PBS until 10 -8 . Of each dilution 10 µl were spotted twice onto RCM and RCMTm15 agar plates and incubated for 24 h before CFUs were enumerated. The segregational stability was calculated as percent stability per generation using the formula n √R. Thereby, R represents the portion of cells which still contain the plasmid at the last time point of the experiment, n the number of generations at this time point grown without selection.

Construction of the ACE vectors
ACE pyrE KO pMTL-KS01: To generate ACE pyrE truncation plasmids on the basis of pMTL-JH12 (Heap et al., 2012) an internal 300-bp fragment of the C. pasteurianum pyrE gene (CLPA c2685, Poehlein et al., 2015) lacking the first 34 nt and an 1,200-bp sequence immediately downstream of the pyrE gene were PCR-amplified from genomic DNA (Table 1) DNA using the oligonucleotides KS001_SHA12_Fw and KS001_SHA12_Rev or KS002_LHA_Fw and KS002_LHA_Rev (Table S1). The 5' SbfI and 3' NotI restriction sites added to the shorter left homology arm (LHA, 300 bp) and the 5' NheI and 3' AscI restriction sites added to the longer right homology arm (RHA, 1,200 bp) were used to clone the PCR fragments into plasmid pMTL85151 between the SbfI and AscI restriction sites. The resultant plasmid was designated pMTL-KS01, the pyrE KO ACE vector (Table 1).
The pyrE ACE complementation vector, pMTL-KS12: A transcriptional terminator sequence (that of the C. tetani E88 glutamyl-tRNA synthetase gene, TCtet gluRS) was inserted proximal to the LHA, through its synthesis as a 77-bp SbfI DNA fragment and insertion into the unique SbfI site of pMTL-KS01 to yield plasmid pMTL-KS02. A Quik Change II Site-Directed Mutagenesis Kit (Agilent Technologies, Stockport, UK) and the oligonucleotides QE_T1Sbf_Fw and QE_T1Sbf_Rev (Table S1) was then used to silence the SbfI recognition site located upstream of TCtet gluRS resulting in the plasmid pMTL-KS03 (Table 1). As the 42bp C. pasteurianum ferredoxin terminator (TCpa fdx, Table 1) of plasmid pMTL85151 was lost during the insertion of the RHA, steps were taken to restore its presence. This was accomplished by synthesising a contiguous NheI/AscI fragment comprising the LHA and TCpa fdx and inserting it between the NheI and AscI sites of pMTL-KS03 to generate pMTL-KS04 (Table 1). To enable an easy replacement of the RHA without the loss of the TCpa fdx, an AsiSI recognition site was introduced between the LHA of pMTL-KS04 and the TCpa fdx and the by site-directed mutagenesis and the use of the oligonucleotides QE_AsiSI add_Fw and QE_AsiSI add_Rev (Table S1). The generated plasmid was named pMTL-KS10 (Table 1).
To make the pyrE repair vector pMTL-KS12, a 548-bp fragment of pyrE representing the LHA and lacking only the first 34 nt of pyre is to be swapped with the 300bp LHA fragment of pMTL-KS10. oligonucleotides KS001_SHA12_Fw and KS001_SHA14_Rev were used to PCR amplified the desired DNA fragment from genomic DNA. These oligonucleotides added 5' a SbfI and 3' a NotI recognition site to the final PCR product. Following, the modified SbfI and NotI equipped PCR fragment was cloned into appropriately digested pMTL-KS10. The result plasmid is pMTL-KS12. The only difference between pMTL-KS10 and pMTL-KS12 is the LHA it carries.

Construction of the Allelic Exchange KO plasmids
Typically, target specific AE cassettes were cloned into pMTL-KS15 using SbfI and NheI. To enable an alternative cloning strategy in case the AE cassette contains the relative abundant 6bp NheI recognition site, an 8-bp AsiSI recognition side was introduce between the NheI recognition site and the terminator TCpa fdx employing the oligonucleotides QE_ AsiSIintro_AA_Fw and QE_ AsiSIintro_AA_Rev (Table S1) and the Quik Change II Site-Directed Mutagenesis Kit (Agilent Technologies, Stockport, UK) according the manufacturer's recommendations. The resulting plasmid was called pMTL-KS16 (Table 1). For both, pMTL-KS15 as well as pMTL-KS16, the insertion of the AE cassette by SbfI and NheI or SbfI and AsiSI lead to a removal of the gratuitous 300-bp pyrE SHA-lacZα ORF/MCS-'cargo' sequence origination from the plasmid pMTL85151-TCtet gluRS-SHA.
AE cassettes for the markerless and in-frame deletion of genes were generated by SOE PCR using the BIO-X-ACTTM Short Mix according to the manufacturer's recommendation (Bioline Reagents, London, UK). A 700-bp region immediately upstream and downstream of the gene of interest including its start and stop codon were amplified employing genomic DNA and target specific oligonucleotide pairs, Fw1 and Rev1 for the amplification of the upstream region and Fw2 and Rev1 for the amplification of the downstream region. Typically oligonucleotides Rev1 and Fw1 are characterised by a 25-to 35-bp long overlapping complementary sequence enabling the later fusion of the two PCR fragments in a third PCR reaction. Oligonucleotides Fw1 and Rev2 equipped the particular PCR fragment with a 5' SbfI and a 3' NheI restriction site. Following the generation of the two separate PCR fragments, self-same were gel purified and used in equimolar amounts in a third PCR reaction employing the oligonucleotides Fw1 and Rev2. The final PCR product was digested using SbfI and NheI, purified and cloned into the equally treated AE plasmid pMTL-KS15 resulting in a target specific AE vector. All generated AE plasmids were confirmed by analytic restriction digest and Sanger sequencing.
For the generation of the spo0A KO cassette the oligonucleotides KS013_spo0A_Fw1 and KS013_spo0A_Rev1 and KS013_spo0A_Fw2 and KS013_spo0A_Rev2 were employed (Table S1). The final AE vector was called pMTL-KS15::spo0A (Table 1).
For the generation of the hdyA KO cassette the oligonucleotides KS014_hyd_Fw1 and KS014_hyd_Rev1 and KS014_hyd_Fw2 and KS014_hyd_Rev2 were employed (Table S1). The final AE vector was called pMTL-KS15::hydA (Table 1).
For the generation of the rex KO cassette the oligonucleotides KS015_rex_Fw1 and KS015_rex _Rev1 and KS015_rex _Fw2 and KS015_rex _Rev2 were employed (Table S1). The final AE vector was called pMTL-KS15::hydA (Table 1).

Complementation plasmid
To verify the observed phenotype of formerly generated KO strains, self-same strains were in trans complemented employing ACE and the genomic pyrE locus. Typically plasmids for the integration of the homologous and heterologous DNA were constructed as described below.
First, the open reading frame encoding the target gene was amplified from genomic DNA using BIO-X-ACT TM Short Mix (Bioline Reagents, London, UK) and gene specific primers which added a 5' NotI and a 3' NheI recognition site to the PCR fragment. NotI and NheI were used to clone the PCR fragment into the equally digested ACE plasmid. Generally pMTL-KS12 (Table 1) was employed. All generated complementation vectors were verified by a diagnostic restriction digest and Sanger sequencing. The homologous or heterologous genes was cloned either with its own promoter or with a heterologous promoter inserted into the plasmid pMTL-KS12 by NotI and NdeI restriction digest.
For the specific task of generating a spo0A complementation vector, the indigenous C. pasteurianum spo0A gene (CLPA c19180, Poehlein et al., 2015) as well as a 267-bp sequence immediately upstream of the spo0A start codon predicted to comprise the promoter region of spo0A (SoftBerry BPROM software, Solovyev and Salamov, 2011) were amplified from genomic DNA using the oligonucleotides KS004_spo0A_compl_Fw and KS004_spo0A_compl_Rev (Table S1). Subsequently, the 1095-bp fragment was cloned into pMTL-KS12 and confirmed as described above. The complementation plasmid pMTL-KS12::spo0A was used to generate the C. pasteurianum spo0A complementation (repair) strain CRG5518 as described in Methods.
The plasmid for complementation of the hydrogenase hyd was constructed by amplifying a 2118-bp fragment comprising the 393-bp sequence upstream of hyd and the hyd gene with primers KS011_hyd_compl_Fw/ KS011_hyd_compl_Rev (Table S1) which was cloned into the NotI/NheI recognition sites of pMTL-KS12 leading to plasmid pMTL-KS12::hyd. The plasmid was used to generate strain C. pasteurianum DSM525-H1::hydA* (CRG5530) ( Table  1).
Finally, the dhaBCE complementation plasmid C. pasteurianum DSM525-H1::dhaBCE* was constructed by amplifying a 2984-bp fragment comprising the 294-bp sequence upstream of dhaB and the dhaBCE genes with primers KS012_dhaBCE_compl_Fw/ KS012_dhaBCE_compl_Rev (Table S1)and cloning the NotI/NheI recognition sites of pMTL-KS12. The plasmid was used to generate strain C. pasteurianum DSM525-H1::dhaBCE* (CRG5536) ( Table 1). Fig. S1. PCR of double-crossovers after allelic exchange knock out of target genes. Ratios of successful knock outs over wild type revertants are given for each target.   (Minton et al., 2016). The genome was published by Poehlein et al. (2015) and restriction modification systems were identified. Plasmid DNA was protected against restriction by in vitro methylation with E. coli carrying plasmid pMTL-CR1 expressing the M.BepI methylase. Due to low transformation efficiency in the wild type a screen was done for hypertransformable mutants (not in the original roadmap, indicated by oval). 5-fluorouracil (FOA) sensitivity was assayed based on which a pyrE truncation strain (pyrE΄) was produced with the ACE plasmid pMTL-KS01. This strain allowed the use of allelic exchange technology to make knock-outs of genes of interest (Gene X) guided by homology arms (HAs) up-and downstream of the gene. Knock-out mutants were complemented at the pyrE locus by repairing the pyrE allele and supplying the gene with its native promoter downstream of pyrE using plasmids pMTL-KS12::Gene X*. The pyrE allele was repaired to wild type without additional cargo with plasmid pMTL-AGH12.