The Coordinated KNR6-AGAP-ARF1 Complex Modulates Vegetative and Reproductive Traits by Participating in Vesicle Trafficking in Maize

Abstract

The KERNEL NUMBER PER ROW6 (KNR6)-mediated phosphorylation of an adenosine diphosphate ribosylation factor (Arf) GTPase-activating protein (AGAP) forms a key regulatory module for the numbers of spikelets and kernels in the ear inflorescences of maize (Zea mays L.). However, the action mechanism of the KNR6-AGAP module remains poorly understood. Here, we characterized the AGAP-recruited complex and its roles in maize cellular physiology and agronomically important traits. AGAP and its two interacting Arf GTPase1 (ARF1) members preferentially localized to the Golgi apparatus. The loss-of-function AGAP mutant produced by CRISPR/Cas9 resulted in defective Golgi apparatus with thin and compact cisternae, together with delayed internalization and repressed vesicle agglomeration, leading to defective inflorescences and roots, and dwarfed plants with small leaves. The weak agap mutant was phenotypically similar to knr6, showing short ears with fewer kernels. AGAP interacted with KNR6, and a double mutant produced shorter inflorescence meristems and mature ears than the single agap and knr6 mutants. We hypothesized that the coordinated KNR6-AGAP-ARF1 complex modulates vegetative and reproductive traits by participating in vesicle trafficking in maize. Our findings provide a novel mechanistic insight into the regulation of inflorescence development, and ear length and kernel number, in maize.

Keywords: ADP-ribosylation factor (Arf) GTPase; Arf GTPase-activating protein; Golgi apparatus; inflorescence; kernel number; vesicle transport.

Figure 1

Loss of AGAP function significantly alters vegetative and reproductive traits. (A) A semi-dwarf plant generated from the gene knockout line (right, agap^cr1) compared with its non-transgenic sibling (left, AGAP^NT1). (B–D) Loss of AGAP function alters the architecture of inflorescences. In the gene knockout line agap^cr1 (right), ear growth was obviously inhibited (B,C), and the tassel was claw-like (D). (E,F) The gene knockout agap^cr1 plant (right) had shorter and compact internodes (E) and a reduced leaf size (F) compared with its non-transgenic sibling (left, AGAP^NT1). (G–J) Measured phenotypic characteristics of agronomically important traits: Plant height (G), ear height (H), leaf length (I) and leaf width (J) were significantly different between AGAP^NT1 and agap^cr1. Phenotypes were assessed at Wuhan, China, in spring 2020. The values in (G–J) are the means ± s.d.s, and the significance levels of differences were estimated using a one-way ANOVA. Scale bars = 10 cm in (A), 5 cm in (B,D–F), and 1 cm in (C); ** indicates a statistical difference at the p < 0.01 level, *** indicates a statistical difference at the p < 0.001 level.

Figure 2

AGAP genetically interacts with KNR6 to affect ear length and kernel number per row. (A,B) knr6^cr1 produced shorter mature ears (A) and ear inflorescence meristems (B) relative to its non-transgenic sibling (KNR6^NT1). (C–E) The ear inflorescence meristem length (D), mature ear length (E), and kernel number per row (F) of the KNR6 knock-out line (knr6^cr1) were significantly different from those of its non-transgenic sibling (KNR6^NT1). (F) Single-gene mutants (knr6^cr1/+ and +/agap^cr2) and the double-gene mutant (knr6^cr1/agap^cr2) showed smaller ears than wild type. (G–I) The double-gene mutation enhanced the phenotypic effects of the single-gene mutants. Phenotypic differences between mutants and wild type were revealed for kernel number per row (G), ear length (H), and kernel row number (I). Data are shown as the means ± s.d.s. The significances of the differences at p < 0.05 were determined using the Tukey HSD test. Scale bars = 2 cm in (A,F) and 200 μm in (B); ** indicates a statistical difference at the p < 0.01 level, *** indicates a statistical difference at the p < 0.001 level.

Figure 3

AGAP protein localized on the Golgi apparatus and alters structure of a partial Golgi apparatus. (A–E) Subcellular localization of AGAP and Golgi maker: (A) bright field, (B) mCherry for AGAP localization, (C) is the marker for Golgi, (D,E) shows the overlap of AGAP and Golgi maker. Scale bar = 10 μm. (F) Normal Golgi apparatus morphology in AGAP^NT cells. (G,H) Thin (G) and curved (H) Golgi apparatus were observed using a transmission electron microscope. (I) Proportions of different Golgi types in AGAP^NT and agap^cr. Scale bar = 100 nm; tg, thin Golgi apparatus; g, Golgi cisternae; t, trans-Golgi network; cg, curved Golgi apparatus.

Figure 4

The numbers and agglomeration of vesicles in AGAP^NT and agap^cr2 cells. (A,B) The FM4-64-labeled vesicles in AGAP^NT and agap^cr2 cells after 10 (i and vi, respectively), 30 (ii and vii, respectively), 60 (iii and viii, respectively), 90 (iv and ix, respectively), and 120 min (v and x, respectively) of FM4-64 staining. Arrowheads indicate FM4-64-labeled vesicles. (C) To test for FM4-64 treatment effect on the response variables with time, repeated measures ANOVA was performed and time as the within-subject factor (general linear model (GLM) in SPSS 16.0). (D,E) Vesicle agglomeration in AGAP^NT and the agap^cr2 cells. The fungal toxin Brefeldin A (BFA) bodies were revealed by BFA re-treatment in the FM4-64-labeled cells after 10 (i and vi, respectively), 30 (ii and vii, respectively), 60 (iii and viii, respectively), 90 (iv and ix, respectively), and 120 (v and x, respectively) min BFA treatments. (F) To test for FM4-64 and BFA treatment effect on the response variables with time, repeated measures ANOVA was performed the same as in FM4-64 treatment. Scale bars in (A,B,D,E) = 20 μm; *** indicates a statistical difference at the p < 0.001 level.

Figure 5

AGAP interacts with two ARF1 members on the Golgi apparatus. (A) In total, 16 putative maize ARF1s were grouped into 4 clades. (B) AGAP physically interacted with ARF1.1 and ARF1.2. The interactions were assessed using luciferase complementation image assays (B). (C) Subcellular localization of ARF1.1 and Golgi maker. (C-i) Bright field. (C-ii) mCherry for ARF1.1 localization. (C-iii) is the marker for Golgi. (C-iv) and (C-v) show the overlap of ARF1.1 and Golgi maker. Scale bar = 10 μm. (D) Subcellular localization of ARF1.2 and Golgi maker. (D-i) Bright field. (D-ii) mCherry for ARF1.2 localization. (D-iii) is the marker for Golgi. (D-iv) and (D-v) shows the overlap of ARF1.2 and Golgi maker. Scale bar = 10 μm.

Figure 6

A potential working model of the KNR6–AGAP–ARF1 complex in maize. Under normal conditions, KNR6 binds to and phosphorylates AGAP^NT, and the phosphorylated AGAP interacts with ARF1, leading to the release of vesicles from membranes. When treated with BFA, vesicles from exocytosis and ER-to-Golgi apparatus trafficking are disrupted, with the former gathering into TGN/EE and the latter forming BFA bodies. In agap^cr cells, endocytosis and Golgi apparatus-to-ER retrograde trafficking involving AGAP are deficient. The vesicle efflux from TGN/EE is greater during exocytosis than the protein influx into TGN/EE during endocytosis, leading to the smaller TGN/EE. The number of vesicles transported from the ER to Golgi apparatus is greater than the number of vesicles transported from the Golgi apparatus to ER, leading to larger Golgi cisternae. When treated with BFA, the BFA bodies that originate from endocytosis and Golgi apparatus-to-ER trafficking are inhibited; therefore, the agglomeration of BFA bodies is delayed compared with in AGAP^NT cells.