Transporter editing in cassava indicates local production of cyanogenic glucosides in, and export from, cassava roots

Cassava (Manihot esculenta) is a globally important staple crop, particularly in sub‐Saharan Africa. Its widespread consumption may be attributed to its hardiness. Relative to other staple crops such as maize, wheat and rice, cassava produces more energy per unit area in periods of drought and in marginal soils (Amelework et al., 2021). Furthermore, cassava roots can be stored below ground for extended periods of time prior to harvest, enabling greater management flexibility for producers (Amelework et al., 2021). Thus, cassava cultivation can safeguard against food insecurity, especially as climate change imposes severe threats to agricultural productivity.

Despite its global importance, cassava accumulates human‐toxic metabolites in the form of cyanogenic glucosides (CGs), chemical precursors to cyanide, which must be removed prior to safe human consumption (Ernesto et al., 2002). During periods of environmental or sociopolitical stress, the risk of improper cassava processing increases (Ernesto et al., 2002). Chronic cyanide exposure as a result of insufficient processing can result in damage to the central nervous system and, in severe cases, paralysis (Ernesto et al., 2002). To mitigate the human health impacts of CGs, researchers have developed strategies to attenuate their accumulation in cassava. CG levels have been successfully reduced by gene editing to knockout CG biosynthesis (Gomez et al., 2023; Juma et al., 2022).

Some evidence indicates that acyanogenic varieties generated through gene editing suffered from greater herbivory, which may have implications for yield (Juma et al., 2022).

To substantiate the relationship between CGs and yield, we analysed publicly available data from CassavaBase (Fernandez‐Pozo et al., 2015) (Data S1). Data aggregated from Nigerian research trials show an association of greater CG levels with increased fresh storage root weight (Figure 1a, Figure S1). This trend remained consistent when the data were aggregated by two individual field sites in Nigeria (Figure S1).

Figure 1.

Local synthesis and export of cyanogenic glucosides from cassava tuberous roots. (a) Violin plots of cassava fresh storage root weight in kg compared to measured hydrogen cyanide (HCN) picrate level. The number of replicates is denoted above each plot violin. (b) The gene model of MeCGTR1 with the location of the CRISPR‐Cas9 guide RNA indicated in blue. The unique alleles generated by CRISPR‐Cas9 are shown in red. Event 1 contains a homozygous single base‐pair insertion edit, and Event 2 is biallelic with equal frequencies of a one base‐pair insertion and a four base‐pair deletion (c) A representative cassava plant image. Measurements of mg cyanide per kg of tissue from picrate assay data comparing wild type to cgtr1 on (d) top leaves (e) bottom leaves (f) stems (g) roots. (h) Measurements of mg HCN/kg tissue from picrate assay on girdled stems of wild type and cgtr1. Individual plots for measurements above and below the girdle incision point in each genotype are presented. (i) Ratios of mg HCN/kg tissues below the girdle incision zone: above girdle incision point for wild type and cgtr1. Plots in this figure d–i are box‐and‐whisker plots where the centre horizontal indicates the median; the upper and lower edges of the box are the upper and lower quartiles; and whiskers extend to the maximum and minimum values. Individual biological replicates are presented as points. P < 0.1 denoted by *, P < 0.05 by **, P < 0.001 denoted by ***, P < 0.0001 denoted by ****. (d, e, g) Wild Type n = 15, cgtr1 n = 15 (f) Wild Type n = 4, cgtr1 n = 3 (h, i) Wild Type n = 8, cgtr1 n = 5.

Previous literature has suggested that CGs are synthesized in the shoot apex and transported to roots (Jørgensen et al., 2005). Interruption of this transport could lower levels of CGs in the roots. However, evidence of de novo biosynthesis of CGs in roots of cassava raises questions regarding the extent to which basipetal transport or de novo biosynthesis supplies cassava roots with CGs (Du et al., 1995; Jørgensen et al., 2005). Jørgensen et al. validated a putative, high‐affinity transporter of CGs from the cassava genome through heterologous expression in Xenopus laevis oocytes. Only a single queried transcript, hereafter referred to as MeCGTR1, was found to have transport capacity for linamarin, the most abundant CG in cassava (Jørgensen et al., 2017). The expression profile of MeCGTR1 is consistent with its role as a transporter (Figure S2) (Wilson et al., 2017). These findings provided a discrete target for gene editing to affect cassava CG transport. Leveraging the ability to make CRISPR‐Cas9‐mediated edits in cassava, we generated knockouts of MeCGTR1 (Figure 1b). Two unique events, each resulting in an early stop codon, were produced (Figure 1b).

Cyanide levels of wildtype (WT) and cgtr1 events were measured in roots, stems, top leaves and bottom leaves using a picrate assay (Gomez et al., 2023). Lower levels of CGs were detected in the top leaves and stems of cgtr1. No difference was found between genotypes for roots or bottom leaves (Figure 1c–g). Bottom leaves had the lowest overall CG levels, followed by roots, stems and top leaves in WT (Figure S3). The presence of CGs in stems is consistent with their reported phloematic transport. The very low detected levels of CGs in stems of knockout events therefore provide evidence for the function of MeCGTR1 as a systemic transporter of CGs (Figure 1f). CGs detected in stems of cgtr1 events may be a result of the activity of an alternative transporter. A putative paralog of MeCGTR1, Manes.17G021100, is a probable candidate.

cgtr1 events would hypothetically have lower root CGs if CGs are transported basipetally. However, no difference existed in cyanide levels between the roots of WT and cgtr1. Notably, there was a reduction of cyanide in the top leaves of cgtr1 (Figure 1d). The findings of the picrate assay suggest that cgtr1 is indeed a systemic transporter of CGs in cassava, and begins to suggest an acropetal mode of CG flow as an alternative to previously established evidence of exclusively basipetal movement. A root‐upwards mode of CG transport was further substantiated by a publicly available expression database captured from 3‐month‐old cassava plants (Wilson et al., 2017). Expression of CYP79D1 and CYP79D2, genes that encode CG biosynthesis enzymes, were found to be highest in fibrous roots by many fold relative to shoot tissues (Figure S2) (Wilson et al., 2017). We undertook a phloem girdling approach to provide further resolution of the directionality of CG movement. Incisions to the phloem to prevent movement of CGs were made, and measurements of CGs above and below the incision zone were subsequently taken (Jørgensen et al., 2005). Overall, lower levels of CGs were found in cgtr1 events, consistent with the stem picrate assay (Figure 1h). Comparisons of CG levels below and above the incision point were calculated. In WT plants, ratios were greater than one, indicating an acropetal direction of CG movement (Figure 1i). A higher ratio was observed in WT relative to cgtr1 events, consistent with MeCGTR1's function as a transporter (Figure 1i).

Efforts to improve the safety of cassava by editing cyanogenesis genes may be an effective approach in some contexts. Transporter editing as an alternative approach was considered in this work. Tissue‐specific metabolite levels have been successfully modulated by transporter engineering in other organisms (Nour‐Eldin and Halkier, 2013). Our work extends these findings to cassava and demonstrates the first in vivo validation of a systemic transporter in this crop. Leveraging the newly characterized function of MeCGTR1 as a systemic transporter of CGs, we sought to probe the extent to which de novo biosynthesis of CGs contributes to total CG content in cassava storage roots relative to transport from shoot apex tissues.

Our work suggests that the primary source of cassava storage root CGs is root biosynthesis. It is possible that greater production of root CGs is induced by mutations in MeCGTR1, but unlikely, considering the high root expression levels of biosynthesis genes in wild type plants. The reduction of top leaf CGs in edited events suggests a root‐to‐shoot method of CG movement. Lower CG levels in the top leaves of cgtr1, accumulation of CGs below the incision point of the phloem girdle and high expression of biosynthesis genes in fibrous roots, all indicate a root‐upward mode of CG movement. Multiple directions of CG movement are possible (Nour‐Eldin and Halkier, 2013). Shifts between basipetal and acropetal movements of CGs may be contingent upon developmental stage, environmental status or other conditions. It is also possible that a basipetal mode of transport may be confined to tissues vicinal to the shoot apex. It is known that CG levels in cassava storage roots increase during drought stress (Ernesto et al., 2002). Further investigation is required to determine the source of elevated CGs in these conditions. In the case that basipetal shuttling is the primary mode of enrichment in this condition, cgtr1 events could prove an effective strategy for limiting storage root CG content.

Author contributions

NGK developed and led the project. GAG and NGK designed guides and prepared plasmids. BKG transformed cassava and maintained in vitro plantlets in the lab of MJC. SAL genotyped transformed plants and phenotyped them alongside NGK with assistance from LL, AGC and JBL. NGK and SAL wrote the manuscript with feedback from JBL.

Supporting information

Data S1 Methods.

Figure S1 Yield and cyanogenic glucoside levels in Nigerian field trials.

Figure S2 Expression levels of MeCGTR1, CYP79D1 and CYP79D1 across cassava tissues.

Figure S3 Comparison of cyanide levels among tissues.