Pyrophosphate-fructose 6-phosphate 1-phosphotransferase (PFP1) regulates starch biosynthesis and seed development via heterotetramer formation in rice (Oryza sativa L.)

Abstract

Pyrophosphate-fructose 6-phosphate 1-phosphotransferase (PFP1) reversibly converts fructose 6-phosphate and pyrophosphate to fructose 1, 6-bisphosphate and orthophosphate during glycolysis, and has diverse functions in plants. However, mechanisms underlying the regulation of starch metabolism by PFP1 remain elusive. This study addressed the function of PFP1 in rice floury endosperm and defective grain filling. Compared with the wild type, pfp1-3 exhibited remarkably low grain weight and starch content, significantly increased protein and lipid content, and altered starch physicochemical properties and changes in embryo development. Map-based cloning revealed that pfp1-3 is a novel allele and encodes the regulatory β-subunit of PFP1 (PFP1β). Measurement of nicotinamide adenine dinucleotide (NAD+) showed that mutation of PFP1β markedly decreased its enzyme activity. PFP1β and three of four putative catalytic α-subunits of PFP1, PFP1α1, PFP1α2, and PFP1α4, interacted with each other to form a heterotetramer. Additionally, PFP1β, PFP1α1 and PFP1α2 also formed homodimers. Furthermore, transcriptome analysis revealed that mutation of PFP1β significantly altered expression of many essential enzymes in starch biosynthesis pathways. Concentrations of multiple lipid and glycolytic intermediates and trehalose metabolites were elevated in pfp1-3 endosperm, indicating that PFP1 modulates endosperm metabolism, potentially through reversible adjustments to metabolic fluxes. Taken together, these findings provide new insights into seed endosperm development and starch biosynthesis and will help in the breeding of rice cultivars with higher grain yield and quality.

Keywords: floury endosperm; map-based cloning; pyrophosphate-fructose 6-phosphate 1-phosphotransferase (PFP1); rice; starch synthesis.

Figure 1

Phenotypic characterization of pfp1‐3 mutant and wild‐type rice plants. (a) Whole plant phenotype of the pfp1‐3 mutant and the wild type. (b) Spikelets of the primary inflorescence branch at full maturation. (c) Grain‐filling process. (d) Morphology of dehusked seeds. (e) Horizontal sections of polished seed endosperms. The floury inner region of the endosperm is evident in the pfp1‐3 mutant. (f) Grain yield per plant in natural paddy field conditions. Data represent mean ± standard error (SE) of three biological replicates. Asterisks represent significant differences between pfp1‐3 mutant and wild‐type plants (**, P < 0.01; Student's t‐test). Scale bars, 10 cm in (a); 3 mm in (b, d); 1 mm in (e).

Figure 2

Abnormal endosperm development in pfp1‐3 seeds. (a–f) Scanning electron microscope analysis of transverse sections of wild‐type seeds (a–c) and pfp1‐3 mutant seeds (d–f). Starch granules in pfp1‐3 mutant seeds are irregularly shaped and loosely packed than those in wild‐type seeds. (g, h) Semi‐thin sections of wild‐type (g) and pfp1‐3 (h) centre endosperm at 15 days after flowering (DAF). Scale bars 1 mm in (a, d); 30 μm in (b, e); 10 μm in (c, f); 50 μm in (g, h).

Figure 3

Positional cloning of the pfp1‐3 mutation and complementation testing. (a) Fine mapping of the PFP1 locus. The PFP1 locus (red arrowhead) was mapped to a 94.5‐kb region between markers Q3 and P32 on chromosome 6 (Chr. 6), which contained 16 predicted open reading frames. Marker names and number of recombinants are shown above the map. (b) Schematic representation of the PFP1 gene structure showing the point mutation site of pfp1‐3. Mutant alleles of the PFP1 gene included a nucleotide substitution that generated an alternative splicing site, resulting in a 7‐bp insertion and premature stop codon (closed red rectangle). Black boxes indicate exons. White boxes indicate non‐coding regions. Lines represent introns. (c) PCR‐based confirmation of splice variants of PFP1 in the pfp1‐3 mutant (lane 1) and 16 wild‐type rice cultivars (lanes 2–17). (d) Co‐segregation analysis of the splicing error mutation with the pfp1‐3 mutant phenotype in the pfp1‐3 ×  ZH11 F₂ population based on PCR product size. P1, pfp1‐3 mutant; P2, ZH11; M, DNA marker. (e, f) Functional complementation of the PFP1 gene completely rescued normal grain appearance (e) and restored normal starch granule arrangement (f). Representative images of whole seeds (e) and scanning electron microscope images of transverse sections of seeds (f) of wild‐type, pfp1‐3 and pfp1‐3 complementation lines (L1 and L2). Scale bars represent 3 mm (e) or 10 μm (f).

Figure 4

Molecular and phenotypic characterization of wild‐type and pfp1‐3 mutant plants. (a) Expression of PFP1 in the root, stem and leaf tissues. (b) PFP1 enzyme activity assays. (c–f) Contents of protein (c), lipid (d), starch (e) and amylose (f) in mutant and wild‐type seeds. Values represent the mean ± SE of three biological replicates. Asterisks represent significant differences between the pfp1‐3 mutant and wild type (*, P < 0.05; **, P < 0.01; Student's t‐test).

Figure 5

Physicochemical properties of starch in the pfp1‐3 mutant. (a) X‐ray diffraction patterns of starch granules purified from mature endosperm of the pfp1‐3 mutant and its wild‐type cultivar, Hwacheong. (b) Size exclusion chromatography (SEC) weight chain length distributions (CLDs) of debranched starches. (c) Differential scanning calorimeter thermograms of pfp1‐3 mutant and wild‐type starches. (d) Differences in amylopectin chain length distributions between the wild type and pfp1‐3 mutant. (e, f) Starch digestion curves and LOST plots of endosperm starch in wild‐type seeds (e) and pfp1‐3 mutant seeds (f).

Figure 6

Analysis of seed germination and embryo development in the wild type and pfp1‐3 mutant. (a) Germination of mature seeds plated on half‐strength MS medium without sugar for 4 days. (b) Seed germination rate. Data represent mean ± SE of three biological replicates, with each replicate containing 30 seeds. Asterisks indicate significant differences between wild‐type and pfp1‐3 mutant seeds (**, P < 0.01; Student's t‐test). (c) Tetrazolium assay of wild‐type and pfp1‐3 mutant seeds. (d) Vertical sections of imbibed wild‐type and pfp1‐3 mutant embryos. The pfp1‐3 mutant seeds contained an abnormal embryo. Scale Bars 6 mm in (a); 3 mm in (c); 1 mm in (d).

Figure 7

Interactions PFP1β with PFP1α in planta. (a) BiFC assay showing interactions of PFP1β with three PFP1α proteins. The BiFC assay was performed in N. benthamiana leaves upon Agrobacterium tumefaciens‐mediated transient expression. Fluorescence was detected in epidermal cells of infiltrated tissues by confocal microscopy at 48 hpi. Scale Bars, 30 μm. (b) PFP1β and three PFP1α proteins interact in planta. Total proteins were extracted from N. benthamiana leaves expressing PFP1β–FLAG and PFP1α‐YFP. The immune complexes were pulled down by using anti‐FLAG or GFP agarose gel, and the coprecipitation of PFP1β or PFP1α was detected by Western blotting.

Figure 8

Effect of decreased PFP1 activity on metabolites and gene expression involved in starch biosynthesis. (a) Expression profiles of genes involved in starch biosynthesis. Total RNA was isolated from seeds at 3, 6 and 12 DAF, and subjected to reverse transcription using oligo (dT) primers. Blue and red colours indicate downregulated and upregulated expression levels, respectively, in comparison with expression levels in wild‐type rice plants. (b) Soluble sugars content of glycolytic intermediates and sugar signalling molecule was determined in mature seeds of rice by GC MS analyses. (c–g) The changes of five lipid classes were investigated from individual molecular species in mature rice seeds as revealed by LC‐SI‐MS. Galactolipids including DGDG and MGDG in (c–d). Phospholipids including PE, PG and PI in (e–g). Values represent mean ± SE of two biological replicates. Asterisks represent significant differences between the pfp1‐3 mutant and wild type (*, P < 0.05; **, P < 0.01; Student's t‐test).