Transcriptomics at maize embryo/endosperm interfaces identifies a transcriptionally distinct endosperm subdomain adjacent to the embryo scutellum

About this paper

Seeds are complex biological systems in which three genetically distinct compartments are nested one within another: the embryo (the future plant), the endosperm (a nutrient-storing tissue surrounding the embryo), and the maternal seed coat. Coordinated growth of these three compartments is essential for seed development, and requires constant nutrient transfer and molecular communication across their shared interfaces. However, the molecular identity of the cells forming these interfaces had remained largely unexplored.

In this study, we took advantage of the large size of the maize (Zea mays) kernel to dissect and profile by RNA-seq the transcriptomes of six kernel (sub)compartments at 13 days after pollination, with a focus on the embryo/endosperm interfaces. Our analysis reveals that the endosperm cells in direct contact with the embryo scutellum carry a transcriptional signature distinct from the rest of the endosperm, defining a novel subdomain we named the endosperm adjacent to scutellum (EAS). The EAS transcriptome is markedly enriched in genes encoding membrane transporters — including members of the SWEET and UMAMIT families involved in sugar and amino acid transport — suggesting an active role in nutrient partitioning toward the growing embryo. The EAS appears around 9 days after pollination, coinciding with scutellum emergence, and persists until approximately 20 days after pollination, dynamically tracking the embryo growth front. Cell death signatures and crushed cell walls at the innermost EAS layers, together with impaired EAS marker gene expression in embryo-deficient mutants, indicate that the EAS is a developmentally regulated, embryo-dependent interface zone.

About the data and scripts

The resources provided on this page supported the functional annotation and GO enrichment analyses described in the paper. The maize proteome annotation was generated de novo from the Zea mays B73 genome (AGPv4 / Gramene release 51) by combining BLAST searches against Swiss-Prot (complete dataset) and TrEMBL (plant subset) with InterProScan domain predictions (Pfam, Panther, ProSite). GO term associations were propagated using a local relational database (MySQL snapshot, February 2018). Enrichment analyses were performed using a custom pipeline built on a PostgreSQL annotation database, with hypergeometric testing. During analysis, the local GO DAG was represented using Graphviz. These resources are made available to facilitate reuse and reanalysis, particularly for researchers working on maize seed biology or seeking to apply similar GO annotation workflows to other plant transcriptome data sets. All RNA-seq data are deposited at GEO under accession GSE110060, and expression profiles can be explored interactively via the eFP Browser.