Non-homologous end-joining-deficient filamentous fungal strains mitigate the impact of off-target mutations during the application of CRISPR/Cas9

Abstract

CRISPR/Cas9 genome editing technology has been implemented in almost all living organisms. Its editing precision appears to be very high and therefore could represent a big change from conventional genetic engineering approaches. However, guide RNA binding to nucleotides similar to the target site could result in undesired off-target mutations. Despite this, evaluating whether mutations occur is rarely performed in genome editing studies. In this study, we generated CRISPR/Cas9-derived filamentous fungal strains and analyzed them for the occurrence of mutations, and to which extent genome stability affects their occurrence. As a test case, we deleted the (hemi-)cellulolytic regulator-encoding gene xlnR in two Aspergillus niger strains: a wild type (WT) and a non-homologous end-joining (NHEJ)-deficient strain ΔkusA. Initial phenotypic analysis suggested a much higher prevalence of mutations in the WT compared to NHEJ-deficient strains, which was confirmed and quantified by whole-genome sequencing analysis. Our results clearly demonstrate that CRISPR/Cas9 applied to an NHEJ-deficient strain is an efficient strategy to avoid unwanted mutations. IMPORTANCE Filamentous fungi are commonly used biofactories for the production of industrially relevant proteins and metabolites. Often, fungal biofactories undergo genetic development (genetic engineering, genome editing, etc.) aimed at improving production yields. In this context, CRISPR/Cas9 has gained much attention as a genome editing strategy due to its simplicity, versatility, and precision. However, despite the high level of accuracy reported for CRISPR/Cas9, in some cases unintentional cleavages in non-targeted loci-known as off-target mutations-could arise. While biosafety should be a central feature of emerging biotechnologies to minimize unintended consequences, few studies quantitatively evaluate the risk of off-target mutations. This study demonstrates that the use of non-homologous end-joining-deficient fungal strains drastically reduces the number of unintended genomic mutations, ensuring that CRISPR/Cas9 can be safely applied for strain development.

Keywords: Aspergillus niger; CRISPR-Cas9; biosafety; risk assessment; ΔkusA.

Fig 1

Phenotypic characterization of A. niger ΔxlnR strains in wild-type (A) and ΔkusA (B) genetic backgrounds. Strains were grown on 1% xylan-containing minimal medium plates for 4 days at 30°C.

Fig 2

Average number of random mutations caused by the transformation process. Average number of SNVs, insertions and deletions of the wild-type (WT, blue bars), and the ΔkusA strain (green bars) that were transformed with the ANEp8-pyrG plasmid in which cas9 is not present. Error bars represent the standard deviation (SD) of the replicate samples. Asterisks represent statistical significance for each of the different mutations (t test, P < 0.05). Note the break at Y-axis.

Fig 3

Assessment of the number of random mutations caused by the presence of Cas9. Average number of SNVs (A), insertions (B), and deletions (C) in the wild-type (WT, blue bars) and ΔkusA strains (green bars) detected after transformation in the absence (patterned bars) or presence (solid bars) of Cas9 endonuclease. Error bars represent the standard deviation (SD) of the replicate samples. No difference in significance was found in any analysis (P < 0.05). Note the break in Y-axis of panel A.

Fig 4

Mutations arising in the wild type (WT) after xlnR disruption with CRISPR/Cas9 depending on the DNA repair mechanism. Average number of SNPs, insertions, and deletions in the WT ΔxlnR strains that underwent NHEJ (dark blue bars) or HR (light blue bars). Error bars represent the standard deviation (SD) of the replicate samples. No difference in significance was found in any analysis (P < 0.05). Note the break in Y-axis.

Fig 5

Comparison of the number of mutations in xlnR deletion strains in the two different genetic backgrounds. Average number of SNVs (A), insertions (B), and deletions (C) in the wild-type strains (WT, blue bars) and in ΔkusA (green bars) which received the cas9-plasmid ANEp8-pyrG_cas9 without any functional guide (solid bars, control) and the strains that were transformed with the ANEp8-cas9-pyrG-xlnR plasmid (patterned bars). Error bars represent the standard deviation (SD) of the replicate samples. Difference in significance is represented by *P < 0.05. Note the break in Y-axis of panel A.

Fig 6

Distribution of the different types of genomic variants detected in each of the sequenced A. niger strains.