Mutation of the Vinv 5′ UTR regulatory region reduces acrylamide levels in processed potato to reach EU food-safety standards

The recent prohibition of Chlorpropham (CIPC) in the EU (Commission Implementing Regulation(EU) 2019/989) is prompting the potato processing industry to search for alternative and safer anti-sprouting approaches. Storage at cold temperature (i.e. 4 °C) has emerged as a valuable option for long term storage of potato without the use of CIPC. However, most commercial potato varieties accumulate high levels of reducing sugars (RS) during cold storage, a phenomenon called cold-induced sweetening (CIS). During high temperature processing of potatoes into products such as crisps and French fries, the RS react with asparagine and peptides to produce the neurotoxin acrylamide, whose presence is evidenced by a brown-to-black coloration of the processed products (Bhaskar et al., 2010). The challenges of potato storage are graphically depicted in Figure 1a.

Because of the difficulty in breeding CIS-resistant potato varieties to replace the ones that are CIS-susceptible, New Genomic Techniques (NGTs) are emerging as useful approaches to rapidly introgress the CIS-resistant trait into commercial varieties used by the processing industry. Despite the versatility of CRISPR-base approaches to target any selected sequence in a plant genome, the technology has so far been mainly used to target protein-coding sequences in plants.

In the present work, we exploited editing of a 5' UTR sequence to engineer CIS resistance in an industry-preferred potato variety. Vacuolar invertase (VInv) has been identified as a key enzyme for conversion of sucrose into RS. Previous studies have demonstrated that silencing of the VInv gene is a suitable approach to lower the accumulation of RS upon cold storage of potato (Bhaskar et al., 2010; Zhu et al., 2016).

We designed a short guide RNA (sgRNA) targeting the 5' UTR region of the VInv gene. Its activity was assessed in an in vitro cleavage assay using an amplicon from potato DNA as template (Figure S1). We next transformed potato variety Lady Rosetta (LaRo), a CIS-susceptible variety used in the crisp industry, with the construct PC2300-sgRNA1–pcoCas9–eGFP. Based on the PAM sequence, the selected sgRNA was expected to target two of the four 5' UTR allelic sequences. Thirty-two independent potato lines were generated. Eleven lines showed the desired genetic profile in a PCR–restriction enzyme (RE) assay and their mini-tubers (T0 tubers) were vegetatively propagated to produce T1 tubers. T1 tubers of wild-type LaRo and Verdi (CIS-resistant) varieties were also produced for use as controls. Further analyses were performed on the T1 tubers.

Illumina sequencing of the target region revealed four lines carrying one adenine insertion between positions −34 and −35 in the two editable alleles (P4, P6, P24 and P26), one transformed line without edits (P1), and six lines with different types and percentages of deletions in the two editable alleles (Figure 1b).

Quantification of RS (glucose + fructose) in the selected lines after storage at 4 °C for 1 month revealed that the four lines with the single A-insertion contained significantly lower levels of RS compared with the non-edited control line (P1), and the other lines displaying various editing profiles of the 5' UTR (Figure 1c). Lines P6 and P26 had significantly lower levels of RS relative to the control CIS-resistant potato variety Verdi. There were no statistically significant differences in the levels of RS amongst the wild-type LaRo potatoes, the non-edited line (P1) and the edited lines. In line with previous observations (Bhaskar et al., 2010; Shumbe et al., 2020), sucrose contents in all samples inversely correlated with the contents of RS (Figure 1c). These results indicate that the single A-insertion between positions −34 and −35 of the 5' UTR of VInv gene was sufficient to engineer CIS resistance.

We assayed VInv enzymatic activity in selected lines and control varieties (Figure 1d) and observed a similar trend to that of RS (Figure 1c) and that of the transcript levels of VInv assayed in selected CIS-contracting T1 tubers after 1-month storage at 4 °C (Figure S2a). These results indicate that a single A-insertion in the 5' UTR region cause a consistent and significant down-regulation of the VInv transcripts and a reduction of VInv activity upon cold storage. After 1-month storage at RT, no statistically significant difference was observed in the levels of RS between all the lines while sucrose levels appeared lower in the selected CIS-resistant lines and Verdi (Figure S2b). Furthermore, the levels of RS in the CIS-resistant lines and variety (P4, P24, P26 and Verdi) were consistently significantly lower than the levels in the CIS-susceptible lines and variety after two generations of propagation (T2) and after 1-month storage at 4 °C (Figure S2c).

Finally, cold stored T1 tubers of selected CIS-contrasting lines and control varieties (Figure S3) were processed into crisps. Crisps from CIS-susceptible lines (P1, P21 and P28) and the control LaRo variety displayed brown-to-black colour, whereas those from the CIS-resistant A-insertion lines (P4, P6 and P26) and the control Verdi variety consistently appeared pale yellow (Figure 1e). The acrylamide contents from crisps of the CIS-resistant lines and variety were well below the benchmark value of 750 μg/kg set by the European Union. Conversely the levels in the CIS-susceptible lines and variety were higher than the benchmark value (Figure 1f).

Taken together, our findings demonstrate that a single adenine insertion in the 5' UTR region of 50% of the LaRo VInv alleles is sufficient to alter the CIS phenotype of the LaRo variety, and thereby to reduce the acrylamide content in the processed products to an acceptable level within EU requirements. VInv is known to regulate several aspects of growth and development in plants (Wang et al., 2010; Wang and Ruan, 2016); however, no abnormal phenotype was observed for the CIS-contrasting lines (Figure S3), nor any significant differences in the mean tuber size (Figure S2d).

Because of its key function in the initiation of transcription, the 5' UTR region offers an alternative to the traditional exon targeting for genome editing (Si et al., 2020). While trait engineering usually relies on targeting all allelic coding sequences in potato, our study demonstrates that a stable CIS-resistance trait can be achieved by differential allele targeting of the VInv 5' UTR. In some specific cases, differential allele targeting might also represent a more suitable approach to reduce the possible pleiotropic effects associated with silencing or knocking out of all alleles in polyploid crop species such as potato (Andersson et al., 2017). With the newly proposed regulation on genome editing in the European Union (Vanderschuren et al., 2023), our results provide a timely solution to convert commercial CIS-susceptible potato varieties widely used by the European potato industry into CIS-resistant varieties for the subsequent production of safe and healthy crisps and French fries.