Manipulation of the microRNA172–AP2L2 interaction provides precise control of wheat and triticale plant height

The REDUCED HEIGHT (RHT) dwarfing alleles Rht-B1b and Rht-D1b were essential in the ‘Green Revolution’. The RHT1 gene encodes a DELLA protein, which participates in the gibberellin (GA) growth-stimulating pathway (Peng et al., 1999), and truncations of this protein are responsible for the GA-insensitive semi-dwarf Rht1b alleles (Van De Velde et al., 2021). The growth-repressing effect of Rht1b alleles optimized plant height, reduced lodging and improved harvest index, but also reduced above-ground biomass and coleoptile length, limiting sowing depth and access to deeper soil moisture (Ellis et al., 2004). This has triggered the search for GA-sensitive dwarfing genes with fewer negative pleiotropic effects.

Plant height in grasses is regulated by a complex genetic network, which includes the conserved microRNA172 (miR172)–APETALA2-like (AP2L) module (Patil et al., 2019; Zhu and Helliwell, 2011). In wheat, miR172 expression is induced during the reproductive transition and regulates flowering time, plant height and both spike and floret development by repressing the expression of AP2L genes (Debernardi et al., 2017). Reduction of miR172 activity in the semi-dwarf tetraploid wheat variety ‘Kronos’ (Rht-B1b) using a transgenic target mimicry (MIM172) approach delayed reproductive transition a few days and generated shorter plants with more compact spikes (Debernardi et al., 2017).

Among the four AP2L genes targeted by miR172 in wheat, AP2L2 and AP2L5 regulate flowering transition, stem elongation and spike development (Debernardi et al., 2020). Point mutations in the miR172 target site of the AP2L genes reduce miR172 activity and generate resistant alleles designated hereafter as rAp2l. An rAp2l-A5 allele originated the domestication gene Q and the free-threshing wheats (Debernardi et al., 2017). Additional mutations in the miR172 target site of Q or in the homeolog AP2L-D5 result in plants with reduced height but, unfortunately, with associated spike defects (Greenwood et al., 2017; Zhao et al., 2018). In this study, we explore the effects of chemically induced alleles rAp2l-A2 from tetraploid and rAp2l-B2 from hexaploid wheat (Figure S1a) as well as multiple new CRISPR-induced alleles. All materials and methods are described in the Materials and Methods in Appendix S1.

The rAp2l-A2 EMS-mutation in the semi-dwarf Kronos reduced stem length by 21%, whereas the introgression of the rAp2l-B2 allele into Kronos or Kronos-rAp2l-A2 backgrounds, reduced stem length by 43–45% (Figure S1a–c, Data S1). We next used CRISPR-Cas9 with a gRNA specifically targeting the miR172 target site of AP2L-B2, because AP2L-A2 has a polymorphism that disrupts the gRNA target (Figure 1a, Figure S2a). We generated multiple independent CRISPR T₀ events into Kronos (Rht-B1b) and a near-isogenic tall line (Rht-B1a) (Figure S2, Data S2). Most of the CRISPR mutations were small frameshift indels in the miR172 target site (Figure 1a, Figure S2a), located downstream of the conserved AP2 domains and close to the stop codon (Figure 1a). Both in-frame and frameshift indels resulted in semi-dominant dwarfing effects, suggesting that disruptions of the reading frame at the end of the gene have limited effects on AP2L2 activity. The dominance effect of the dwarfing rAp2l-B2 alleles was similar in the tall Rht-B1a plants (Figure S2b,c) and the semi-dwarf Rht-B1b backgrounds (Figure S2d, Data S2).

Independent T₂ edited lines homozygous for different mutations in both Rht-B1a (Figure 1a–c) and Rht-B1b backgrounds (Figure S3a,b) showed significant effects on plant height that varied depending on the mutations. The strongest rAp2L-B2 alleles in the Rht-B1a background reduced plant height to similar levels as Rht-B1b (Figure 1b,c), suggesting that they can be used to replace the Rht1b alleles.

The rAp2l-B2 plants showed a higher spikelet density (Figure 1d, Figure S3c) as a result of reductions in spike length and slight increases in spikelet number per spike (Data S4). In the Rht-B1a background, the edited lines headed 1.8–2.9 days later, which was comparable to the delay generated by Rht-B1b (Figure 1e). The delay in heading time associated with the rAp2l-B2 alleles was slightly stronger in the Rht-B1b sister lines (4.4 to 5.7 days delay, Figure S3d, Data S4). Finally, plants with and without the rAp2l-B2 mutations showed similar coleoptile and first-leaf lengths in both the Rht-B1a (Figure 1f,g, Data S4) and Rht-B1b backgrounds (Figure S3e,f, Data S4). In summary, these results indicate that the rAp2l-B2 alleles can be used to reduce plant height with limited pleiotropic effects on spike architecture or heading time, and with beneficial effects in coleoptile length relative to the Rht1b alleles.

The highly efficient CRISPR vector makes it possible to rapidly induce different rAp2l2 dwarfing alleles in elite backgrounds without time-consuming crosses. To demonstrate this strategy, we generated semi-dwarf mutants for the triticale cultivar ‘UC-Bopak’ (PVP 202100269). Triticale is an anthropogenic allohexaploid combining wheat and rye (AABBRR genomes), which delivers significantly high biomass and grain yield (Tamagno et al., 2022). However, the taller plant stature of many triticale cultivars combined with their larger and heavier spikes can result in increased lodging. We transformed UC-Bopak using the same gRNA targeting the miR172 binding site in both AP2L-B2 and AP2L-R2 homeologs (Figure 1h). Under greenhouse conditions, we observed significant reductions in plant height in edited lines (Figure 1h,i, Figure S4a,b), which were larger in the lines with mutations in both genomes (Data S5). Plant height was correlated with the predicted effect of the mutations on miR172 binding energy both in lines with mutations in AP2L-R2 (R = −0.94) and in those with mutations in both AP2L-B2 and AP2L-R2 (R = −0.73, Data S5). A combined statistical analysis of the triticale and wheat results (Figure 1a–c and Figure S3a,b) showed that this correlation was highly significant (P = 0.0014, Data S5). By selecting different combinations of rAp2l2 mutations, we were able to fine-tune triticale plant height (Figure 1i, Figure S4b) without affecting coleoptile and first-leaf length or heading time (Figure 1j, Figure S4c,e). The edited plants showed more compact spikes but with the same number of spikelets (Figure S4f, Data S5).

Finally, we evaluated lines with 1-bp deletions in the miR172 target site of AP2L-B2 (B) or AP2L-R2 (R) under field conditions in two consecutive years. In 2023, we used headrows (Figure S5) and in 2024 small yield plots as experimental units (Figure 1k–n). The plants with the 1-bp deletions were 17–18 cm shorter the first year (Figure S5a,b) and 12–14 cm shorter the second year (Figure 1l), indicating some interaction with the environment. The spikes of the edited lines were more compact than the wildtype (Figure S5c), but this was not associated with significant differences in grain yield (Figure 1n, Figure S5d). In the second year, the plots of the wildtype variety suffered significantly more lodging than the edited lines (P < 0.0001, Figure 1k,m). Although the differences in grain yield were not significant (Figure 1n, Figure S5d), the edited lines showed a combined 9.5% increase in grain yield in the second year (P = 0.0528, Data S3), which was likely associated with their superior resistance to lodging.

In summary, we demonstrate that different induced mutations in the miR172 target site of AP2L2 genes can be used to precisely modulate wheat and triticale plant height. Breeders can use this technology to evaluate multiple plant heights in their top lines without lengthy backcrossing programs. Moreover, the rAp2l2 alleles did not reduce coleoptile and first-leaf length, suggesting that they can be a valuable replacement of the gibberellin-insensitive Rht1b alleles.

CZ performed experiments and contributed to data analyses and manuscript writing. JH contributed to project conceptualization, laboratory and field experiments, data analyses and manuscript writing. MP performed experiments. DMT conducted transformation experiments. JD contributed to project conceptualization, funding acquisition, CZ's supervision, data analyses and manuscript writing. JMD contributed to project conceptualization, performed and designed experiments, analysed data, supervised MP and wrote the original paper draft. JMD and JD generated the final version of the article.

The authors declare no conflict of interests.