Towards knowledge-driven breeding

Manipulating three important genes in the CLAVATA–WUSCHEL signalling pathway quantitatively enhances grain-yield-related traits in maize.

With revolutionary genome-editing technologies and our ever-increasing understanding of biological traits, knowledge-driven breeding by design is becoming a reality. Through multiplex genome editing, researchers have recently made considerable achievements, including the coordinated targeting of multiple traits for better crop yield1, the establishment of synthetic apomixis for clonal reproduction from F1 hybrids2,3 and re-domestication or de novo domestication of wild species or orphan crops4,5,6,7. In this issue of Nature Plants, Liu et al. report that fine-tuning three important genes in the CLAVATA–WUSCHEL (CLV–WUS) signalling pathway could increase meristem size and quantitatively enhance yield-related traits in maize8.

Maize yield is a function of harvested kernel number and kernel weight. Kernel number per ear is determined by ear size, which depends to a great extent on the development of inflorescence meristems (IMs). It has been well established that IM activity is maintained by the CLV–WUS negative feedback circuit that coordinates stem cell proliferation with differentiation. This pathway was initially discovered in Arabidopsis, but appears to be conserved across plant species. The stem cell-promoting transcription factor WUS is expressed in the organizing centre of the meristem and activates expression of the secreted peptide CLV3, which is transduced by its receptors CLV1 and/or CLV2 and in turn represses WUS expression9 (Fig. 1a). Therefore, manipulating the CLV–WUS pathway could have great potential for crop yield improvement. This assumption was previously proven to be practical by Rodríguez-Leal et al. in tomato1, and now by Liu et al. in maize.

Fig. 1: Editing three CLV3 orthologues in the CLV–WUS pathway with two strategies quantitatively enhances maize yields.

a, The canonical CLV3–WUS negative feedback circuit. WUS is expressed in the organizing centre of the meristem and activates CLV3 expression, while CLV3 sends negative feedback signals to repress WUS expression through the CLV1/CLV2 receptors. b, Multiplex CRISPR–Cas9 promoter editing of ZmCLE7 and ZmFCP1. The promoter-edited weak alleles significantly decrease expression and confer enlarged, non-fasciated ears with significant increases in kernel row number and grain yield. c, Editing the coding regions of ZmCLE1E5 with CRISPR–Cas9. The edited null alleles exhibit normal ear development and significantly enhance grain-yield-related traits. Panel a adapted with permission from ref. 9, The Company of Biologists.

The authors demonstrate how prior knowledge and experimental data can be combined to improve traits of interest through genome editing. Specifically, three maize CLE genes in the CLV–WUS signalling pathway were selected for genome editing using CRISPR–Cas9, and distinct editing strategies were used. The previous data have shown that ZmCLE7 and ZmFCP1 are two major CLE genes controlling IM activity in maize10,11. Simply disrupting their functions causes meristem over-proliferation and produces fasciated ears that cannot be used in breeding. In this study, the authors edited the promoters of ZmCLE7 and ZmFCP1 to produce weak alleles that maintain normal ear architecture but exhibit a quantitative increase in kernel rows. During this process, the authors used the available chromatin state data (ATAC-seq and MNase-seq) to help predict which regions in the promoter of ZmCLE7 and ZmFCP1 could be effective for manipulating gene expression. Indeed, the edited alleles that have deletions in the accessible chromatin regions of promoters exhibited decreased ZmCLE7 and ZmFCP1 expression, which consequently led to an increase in kernel row number. Collectively, they obtained six promoter-edited alleles for ZmCLE7 and five for ZmFCP1, with large deletions or inversions in different regions of the promoters. Among these alleles, ones with decreased expression of ZmCLE7 or ZmFCP1 conferred enlarged, non-fasciated ears as well as increased kernel row number and grain yields, without undermining other important agronomic traits (Fig. 1b).

To identify new CLE genes that may further increase yield, the authors exploited the knowledge of partial compensation among paralogues and identified ZmCLE1E5, a new maize CLE gene that exhibited compensatory upregulation in Zmcle7 mutants. They edited the coding region of ZmCLE1E5 and obtained two null alleles. Interestingly, these Zmcle1e5 null alleles exhibited normal ear development and could quantitatively enhance grain-yield-related traits, exhibiting a similar effect to the weak promoter alleles of ZmCLE7 (Fig. 1c). These results address how biological knowledge could benefit the molecular breeding in agriculture.

Unlike in selfing species like rice, natural variations of traits in maize appear to be predominantly controlled by regulatory variants12,13. Moreover, in many cases the causative variants of quantitative trait loci isolated in maize reside in distant upstream or downstream regulatory regions of functional genes12,13. For example, a Hopscotch transposon insertion located ~60 kb upstream of teosinte branched1 (tb1) enhances tb1 expression to repress the branching, leading to the increased apical dominance in maize14. These distant cis-regulatory variants bring great challenges for precise editing. The work by Liu et al. demonstrates that proximal promoter engineering, as an alternative solution, could produce desirable effects on traits of interest. Given the predominance of regulatory variations for maize agronomic traits, editing promoters to generate diverse cis-regulatory alleles for selection will likely become a practical and routine strategy for maize breeding by design.

The work reported by Liu et al. shows that individually editing ZmCLE7, ZmFCP1 and ZmCLE1E5 could improve yield-related traits without causing obvious negative effects on other important agronomic traits. As ZmFCP1 and ZmCLE1E5 function redundantly with ZmCLE7, it would be very interesting to pyramid the new alleles of these genes and test whether they could act synergistically to maximize trait performances. Crop yield is a complex trait involving regulation of many different biological pathways. In the future, it would be also important to simultaneously target multiple pathways to co-ordinately improve various agronomic traits.

Notably, none of the three genes (ZmCLE7, ZmFCP1 and ZmCLE1E5) are linked to previously mapped yield-related quantitative trait loci. This suggests that either the three CLE genes lack natural variations or that the natural variations have not been utilized, which is unlike the CLE receptors FASCIATED EAR2 (FEA2) and FEA3 known to contribute to maize domestication or subsequent improvement10,15. The newly created alleles at ZmCLE7, ZmFCP1 and ZmCLE1E5 provide new and important resources for further optimizing maize yield traits. We are in an unprecedented era where genome editing opens up an extremely large variation space for crop improvement that natural and human selections were not able to touch during past evolution. With knowledge-driven breeding, crop traits could be precisely modified and new traits — or even new crops — could be created to meet increasing human demands and new agricultural modes.

Taken together, the study by Liu et al. shows that manipulating the CLV–WUS pathway with appropriate strategy can generate new alleles that balance meristem activity and modify traits quantitatively to enhance crop yields, taking an important step forward in knowledge-driven breeding by design.


  1. 1.

    Rodríguez-Leal, D., Lemmon, Z. H., Man, J., Bartlett, M. E. & Lippman, Z. B. Cell 171, 470–480 (2017).

    Article  Google Scholar 

  2. 2.

    Khanday, I., Skinner, D., Yang, B., Mercier, R. & Sundaresan, V. Nature 565, 91–95 (2019).

    CAS  Article  Google Scholar 

  3. 3.

    Wang, C. et al. Nat. Biotechnol. 37, 283–286 (2019).

    CAS  Article  Google Scholar 

  4. 4.

    Li, T. et al. Nat. Biotechnol. 36, 1160–1163 (2018).

    CAS  Article  Google Scholar 

  5. 5.

    Zsögön, A. et al. Nat. Biotechnol. 36, 1211–1216 (2018).

    Article  Google Scholar 

  6. 6.

    Lemmon, Z. H. et al. Nat. Plants 4, 766–770 (2018).

    CAS  Article  Google Scholar 

  7. 7.

    Ye, M. et al. Nat. Plants 4, 651–654 (2018).

    CAS  Article  Google Scholar 

  8. 8.

    Liu, L. et al. Nat. Plants (2021).

    Article  PubMed  Google Scholar 

  9. 9.

    Somssich, M., Je, B. I., Simon, R. & Jackson, D. Development 143, 3238–3248 (2016).

    CAS  Article  Google Scholar 

  10. 10.

    Je, B. I. et al. Nat. Genet. 48, 785–791 (2016).

    CAS  Article  Google Scholar 

  11. 11.

    Rodríguez-Leal, D. et al. Nat. Genet. 51, 786–792 (2019).

    Article  Google Scholar 

  12. 12.

    Chen, Q., Li, W., Tan, L. & Tian, F. Mol. Plants 14, 9–26 (2021).

    CAS  Article  Google Scholar 

  13. 13.

    Liang, Y., Liu, H.-J., Yan, J. & Tian, F. Annu. Rev. Plant Biol. 72, (2021).

  14. 14.

    Studer, A., Zhao, Q., Ross-Ibarra, J. & Doebley, J. Nat. Genet. 43, 1160–1163 (2011).

    CAS  Article  Google Scholar 

  15. 15.

    Bommert, P., Nagasawa, N. S. & Jackson, D. Nat. Genet. 45, 334–337 (2013).

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Feng Tian.

Ethics declarations

Competing interests

The authors declare no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chen, Q., Tian, F. Towards knowledge-driven breeding. Nat. Plants (2021).

Download citation


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing