Identification of genes required for embryo development in Arabidopsis

Abstract

A long-term goal of Arabidopsis research is to define the minimal gene set needed to produce a viable plant with a normal phenotype under diverse conditions. This will require both forward and reverse genetics along with novel strategies to characterize multigene families and redundant biochemical pathways. Here we describe an initial dataset of 250 EMB genes required for normal embryo development in Arabidopsis. This represents the first large-scale dataset of essential genes in a flowering plant. When compared with 550 genes with other knockout phenotypes, EMB genes are enriched for basal cellular functions, deficient in transcription factors and signaling components, have fewer paralogs, and are more likely to have counterparts among essential genes of yeast (Saccharomyces cerevisiae) and worm (Caenorhabditis elegans). EMB genes also represent a valuable source of plant-specific proteins with unknown functions required for growth and development. Analyzing such unknowns is a central objective of genomics efforts worldwide. We focus here on 34 confirmed EMB genes with unknown functions, demonstrate that expression of these genes is not embryo-specific, validate a strategy for identifying interacting proteins through complementation with epitope-tagged proteins, and discuss the value of EMB genes in identifying novel proteins associated with important plant processes. Based on sequence comparison with essential genes in other model eukaryotes, we identify 244 candidate EMB genes without paralogs that represent promising targets for reverse genetics. These candidates should facilitate the recovery of additional genes required for seed development.

Figure 1.

List of 220 EMB genes of Arabidopsis. Gene identities confirmed through duplicate alleles or molecular complementation are highlighted in boldface. Missing from this list are 14 confidential or questionable identities from the community and 16 identities from the Syngenta collection that are undergoing further review.

Figure 2.

Functional diversity of indispensable genes of Arabidopsis. Numbers are percentages of two datasets: 250 genes with a mutant phenotype in the embryo and 550 genes with some other phenotype. Genes were manually assigned to functional classes based on current annotation and publications as described in “Materials and Methods.”

Figure 3.

Overlapping sets of essential and indispensable genes in model eukaryotes. BLASTP comparisons were performed using C. elegans genes with an RNAi phenotype (Kamath et al., 2003); 1,009 essential genes of S. cerevisiae (Giaever et al., 2002); and 800 indispensable genes of Arabidopsis. Numbers represent matches (e.g. 357 worm genes with an RNAi phenotype have at least 1 match among yeast essentials and 1,247 have no match in either the yeast or Arabidopsis datasets). Numbers in white represent genes with matches in all 3 datasets. Differences in reciprocal comparisons (e.g. 115 yeast genes with an Arabidopsis match; 88 Arabidopsis genes with a yeast match) reflect different levels of sequence redundancy and sizes of datasets.

Figure 4.

RT-PCR expression patterns for 30 EMB genes with confirmed identities but unknown cellular functions. Locus numbers for products loaded in lanes 1 to 30 are provided in Supplemental Table VI. Lane 32 contains an actin control (At3g18780). PCR products from genomic DNA (bottom section) differed in size from RT-PCR products.