Actin-interacting protein 1 enhances stalk mechanical strength and yield in maize

Dear Editor,

Stalk lodging has a significant impact on maize yield and mechanized harvesting. Stalk perimeter (SP) is closely associated with lodging resistance (Zhang et al., 2023). However, SP and its genetic basis in maize have not been comprehensively investigated. Sclerenchyma tissues (STs) play an important role in stalk thickening and mechanical strength (Zhao et al., 2025a). The development of STs is closely related to cytoskeletal organization, particularly actin dynamics (Bao et al., 2012; Xie et al., 2022). Actin-interacting protein 1 (AIP1) regulates cell wall synthesis by modulating actin dynamics (Ono, 2018). Research on AIP1 function in plants is limited. In Arabidopsis, AIP1 dysfunction affects the organization of the actin cytoskeleton in the pollen tube (Diao et al., 2020). However, the specific biological functions of AIP1 in maize remain largely unknown. In this study, we identified one AIP1 gene controlling SP and examined its regulatory effects on cell wall development and plant complex traits in maize.

We first investigated SP, stalk length (SL), and the ratio of SP to SL (RPL) at the basal third, fourth, and fifth internodes (SP3–5, SL3–5, and RPL3–5) of 486 maize lines from the Complete-diallel plus Unbalanced Breeding-derived Inter-Cross (CUBIC) population (Supplemental Figure 1; Liu et al., 2020). Our genome-wide association study identified a total of 20, 17, and 38 quantitative trait loci (QTLs) associated with SP, SL, and RPL, respectively (Supplemental Figure 2; Supplemental Table 1). A co-localized QTL, qSP3, on chromosome 3 (23.02–30.05 Mb) was found to simultaneously regulate SP3–5 (Figure 1A; Supplemental Table 2). Further, we narrowed down the QTL region to 695 kb between bin 8370 and bin 8372 by identity-by-descent-based QTL refinement (Figure 1B and Supplemental Figure 3). In this region, only four genes were expressed in the stalk, and the candidate gene AIP1 (GRMZM5G853361) was located closest to the most significant SNP (Supplemental Figure 4).

Figure 1.

AIP1 interacts with ACT6/7 to modulate stalk mechanical strength and yield in maize.

(A) Manhattan plot showing genome-wide QTLs for stalk perimeter at the third internode (SP3). LRT, likelihood ratio test; qSP3, the QTL for SP3.

(B) Refinement of qSP3 (SP3 QTL) on chromosome 3. Left, six major haplotypes (n ≥ 6) at qSP3 across six bin markers (bin). Right, pairwise comparison of SP3 between haplotype 1 (Hap1) and the other 5 haplotypes. This analysis delimited the QTL to a 695-kb region containing 9 genes. Genes shown in gray are not expressed in parental stalks, whereas genes in black and red are expressed; the gene shown in red is closest to the most significant SNP.

(C) Gene structure ofAIP1 showing a GG-to-AA substitution at SNP256, which causes the conversion of amino acid 50 of AIP1 from alanine (A) to threonine (T).

(D) Association of SNP256 genotypes (AA and GG) with agronomic and yield-related traits in the AMP. ED, ear diameter; ERN, ear row number; KNPE, kernel number per ear; KWPE, kernel weight per ear. n = 94 and 388 lines for the AA and GG genotypes, respectively.

(E) Comparison of stalk phenotypes between AIP1 knockout (AIP1_KO) mutants generated by CRISPR–Cas9 and the WT. RPR3, rind penetrometer resistance at the third internode. Scale bar, 5 cm. n = 19, 18, 17, 12, and 53 for AIP1_KO#1, AIP1_KO#2, AIP1_KO#3, AIP1_KO#4, and WT, respectively.

(F) Cob and yield comparison between AIP1_KO mutants and the WT. CD, cob diameter. Scale bar, 5 cm. The number of lines and significance levels are the same as in (E).

(G) Stalk cross-section phenotypes and anatomical structures of AIP1_KO mutants and the WT visualized by microscopy and micro–computed tomography. VB, vascular bundle; ST, sclerenchyma tissue. Scale bar, 200 μm. r, Pearson correlation coefficient. ∗∗∗P < 0.001; two-tailed Student’s t-test.

(H) Comparison of green immunofluorescence signals of actin fibers in stalks of AIP1_KO mutants and the WT. Scale bars, 10 μm.

(I) Fluorescence in situ hybridization of stalk cross-sections. The two images from left to right show AIP1 transcripts (green, left) and merged fluorescence signals of blue (nucleus) and green. Red arrows indicate ST. Scale bar, 50 μm.

(J) Interaction assays between AIP1(A50) and ACT6 (left), AIP1(A50) and ACT7 (middle), and AIP1(A50/T50) and ACT6/7 (right) in tobacco leaves using luciferase complementation.

(K) Interaction between AIP1(A50/T50) and ACT6/7 assessed by co-immunoprecipitation. Values below the anti-GFP bands indicate relative band density, calculated as the ratio of anti-GFP to anti-MYC signal intensity.

(L) Proposed model illustrating the regulatory effects of AIP1 on maize stalk development and yield.

Asterisks in (B) and (D)–(H) indicate significant differences between the indicated samples (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, two-tailed Student’s t-test).

To investigate key sequence variations in AIP1, we amplified a 2 kb region upstream of the transcription start site and the coding sequence of AIP1 in 24 parents of the CUBIC population. Although a few tightly linked presence–absence variations in the promoter region were identified, they had no significant effects on SP3 values (Supplemental Figure 5). However, the GG-to-AA substitution at SNP256 caused the conversion of amino acid 50 in AIP1 from alanine (A) to threonine (T) (Figure 1C). The AA genotype of SNP256 was associated with increased SP3, indicating that the AIP1^256AA allele is favorable for an SP3 increase (Supplemental Figure 6). We then assessed the effects of AIP1^256AA on complex agronomic traits in a maize association mapping population (AMP) (Yang et al., 2014). AIP1^256AA significantly increased SP3, ear diameter (ED), ear row number (ERN), kernel number per ear (KNPE), and kernel weight per ear (KWPE) (Figure 1D). The results indicate that AIP1^256AA can improve both stalk thickness and yield. The synergistic regulation of stalk diameter and yield has also been reported in rice (Yano et al., 2015; Zhao et al., 2025a; 2025b). Notably, AIP1^256AA is present in only 19.29% of the AMP population (mainly in temperate lines), highlighting its great potential for improving maize varieties (Supplemental Table 3).

To investigate the biological functions of AIP1, we generated four homozygous knockout mutants using CRISPR–Cas9 technology (Supplemental Figure 7). The SP3 of AIP1 knockout (AIP1_KO) plants was significantly smaller than that of the wild type (WT). The mechanical strength of the stalk, including rind penetrometer resistance (RPR) and stalk bending strength, was significantly reduced in AIP1_KO plants compared with the WT (Figure 1E). Notably, cob diameter of mutants was significantly smaller than that of the WT (Figure 1F). Yield-related traits, including ED, ERN, KNPE, and KWPE, were significantly lower in AIP1_KO plants than in the WT (Figure 1F and Supplemental Figure 7). In addition, AIP1_KO mutants exhibited significantly reduced cellulose, hemicellulose, and lignin contents compared with the WT (Supplemental Figure 8A). Microscopy and computed tomography (CT) observation of stalks revealed that mutants exhibited thinner STs, smaller ST area, and smaller number and area of vascular bundles compared with the WT (Figure 1G). Importantly, both ST thickness and area were significantly positively correlated with SP3 and RPR3 (Figure 1G and Supplemental Figure 8B). In addition, immunohistochemistry revealed that actin fibers were less abundant in mutants than in the WT (Figure 1H and Supplemental Figure 9). In summary, AIP1_KO mutants displayed significant reductions in actin fiber content, ST and stalk thickness, stalk mechanical strength, and grain yield.

Fluorescence in situ hybridization revealed that AIP1 is highly expressed in the vascular bundles and STs of maize stalks and young cobs (Figure 1I and Supplemental Figures 10A–10D). Additionally, AIP1-EGFP was localized to both the cell membrane and cytoplasm, and AIP1 was constitutively expressed across all 13 selected tissues from the seedling to mature stage and highly expressed in tissues with dense vascular bundles (Supplemental Figures 10E and 10F).

RNA sequencing analysis showed that the expression of only two actin genes, ACT6 (GRMZM2G053284 on chromosome 3) and ACT7 (GRMZM2G082484 on chromosome 8), was significantly lower in AIP1_KO mutants compared with the WT (Supplemental Figure 11). Yeast two-hybrid assays demonstrated that AIP1 interacted with ACT6 and ACT7 in vitro (Supplemental Figure 12), and these interactions were further confirmed in vivo using luciferase complementation and co-immunoprecipitation assays (Figures 1J and 1K). To determine whether the AIP1^256AA allele enhances the AIP1–ACT6/7 interaction in vivo, AIP1 was divided into AIP1(T50) (favorable) and AIP1(A50) (unfavorable) variants. The results showed that the interaction between AIP1(T50) and ACT6/7 was stronger than that between AIP1(A50) and ACT6/7 (Figures 1J and 1K). Furthermore, we found that AIP1 interacts with both ACT6 and ACT7 in three-dimensional space, and that AIP1(T50) binds ACT6/7 through hydrophobic interactions or hydrogen bonding (Supplemental Figure 13).

Based on these findings, we propose that the favorable AIP1^256AA allele enhances the interaction between AIP1(T50) and ACT6/7, potentially increasing actin fiber content and sclerenchyma tissue thickness. This, in turn, promotes stalk perimeter and cob enlargement, ultimately improving stalk mechanical strength and increasing ear yield (Figure 1L).

Data and code availability

RNA sequencing data from the third stalk internode of the CUBIC parents and 67 progenies, as well as from AIP1_KO and WT stalks, have been deposited in the NCBI database (https://www.ncbi.nlm.nih.gov/) under accession numbers PRJNA1112727, PRJNA1116629, and PRJNA1111703, respectively. These data have also been deposited in the Big Sub database of the China National Center for Bioinformation (https://www.cncb.ac.cn/) under accession numbers PRJCA044958 and PRJCA054825.

Funding

This work was supported by grants from the Beijing Municipal Natural Science Foundation (6244040), the Outstanding Youth Foundation of the Beijing Academy of Agriculture and Forestry Sciences (YXQN202401), the National Natural Science Foundation of China (32201815), the Beijing Postdoctoral Research Foundation (2024-zz-087), and the Beijing Scholars Program (BSP041).

Acknowledgments

We thank Prof. Jianbing Yan from Huazhong Agricultural University for providing the CUBIC population. No conflict of interest declared.

Author contributions

X.W., R.W., W.S., and J.Z. designed the project. R.Z., X.S., J.W., S.W., J.X., and Y. Zhang performed yield assessments and collected the phenotypic data. X.W., J.L., T.W., M.L., D.C., Y. Zhao, S.W., and Y.J. conducted the molecular biology experiments. X.W., C.L., and Z.L. conducted the statistical analyses. X.W. wrote the manuscript. R.W., W.S., and J.Z. revised the manuscript.

Published: January 7, 2026