3.1 CRISPR-Cpf1 mediated ldhA knock-out recombination inC. glutamicum.
Cpf1 from Francisella novicida is a putative class 2 CRISPR
effector, similar to the role of Cas9 in Streptococcus
pyogenes .[34] However, unlike Cas9, Cpf1 cleaves
target DNA with a sole single RNA-guided endonuclease and is not coupled
to a trans-activating CRISPR RNA (tracrRNA). Additionally, Cpf1
recognizes two or three thymine residues (5’-TTN-3’ or 5’-TTTN-3’) known
as the T-rich protospacer-adjacent motif (PAM) and cleaves the
phosphodiester bonds between the 23rd and
24th base in the annealing strand and between the
18th and 19th base from the PAM on
the non-annealing strand.[15, 34] Although the
CRISPR-Cas9 system was confirmed to be a simple and precise tool for
genome editing in several microbes, the system does not work well inC . glutamicum . Therefore, the Cpf1 system, adopted in this
study, enabled single-strand DNA recombination, endogenous gene
deletions, and exogenous gene insertions in C . glutamicum .[15]
To obtain a recombinant C . glutamicum strain with both a
gene deletion (lactate dehydrogenase 1, ldhA ) and a gene
insertion (kanamycin resistance gene, Knr), the
all-in-one, pJYS3_ ΔcrtYF vector was modified to obtain the
pJYS3_Amp_MCS vector (Fig. 1). The CRISPR-Cas9 genome editing system
has been used previously to modify C. glutamicum . For example,
Peng evaluated three factors affecting the recombination
efficiency.[13] The authors reported that the
recombination efficiency in a gene deletion process was much higher when
the length of each homologous arm exceeded 0.1kb, whereas the
heterogeneous gene insertion performance improved when the length of a
single arm exceeded 0.3 kb. Furthermore, a sgRNA (single guide RNA)
including a 20 bp target DNA showed different genome editing
efficiencies suggesting that the positions and ratios of the nucleotides
in the target DNA sequence must be considered. Therefore, the
pJYS3_Amp_DT vector was constructed by incorporating double target DNA
sets of the template and non-template strands of the ldhA gene
into the pJYS3_Amp_MCS vector. The lengths of the left and right
homologous arms were 974 and 942 bp, respectively, and they were
subcloned into the pCold vector. A kanamycin resistance gene (1.1 kb)
was also sub-cloned between the two homologous arms. The homologous arms
and kanamycin resistance gene set (LdhAp-Knr-LdhAt)
were introduced into multi-cloning site one of pJYS3_Amp_DT, which
resulted in the pJYS3_Amp_DT_(LdhAp-Knr-LdhAt)
vector (Fig. 1C). The vector was transformed into C. glutamicumat 46 °C for 6–12 min. The transformants appeared on
Knr LB solid media and we confirmed that the first
homologous recombination had occurred with a 100% efficiency rate (Fig.
2 and additional file 1: Fig. S1). One colony was picked and suspended
in 1 mL BHIS media. After the second heat shock at 46 °C for 6–12 min,
appropriately diluted transformant solution was spread on
Knr LB media. Twelve colonies out of several
transformants were picked and we confirmed that the second homologous
recombination occurred with a 50% recombination efficiency (Fig. 2B).
Among the transformants, the production of succinic and lactic acids was
assessed using line six, a knock-out mutant of the ldhA gene
(ΔldhA-6 ), under semi-anaerobic conditions with pure glucose. As
expected, the lactic acid production of C. glutamicum was
completely blocked (Fig. 2C).