Proteogenomics of Pristionchus pacificus reveals distinct proteome structure of nematode models

Abstract

Pristionchus pacificus is a nematode model organism whose genome has recently been sequenced. To refine the genome annotation we performed transcriptome and proteome analysis and gathered comprehensive experimental information on gene expression. Transcriptome analysis on a 454 Life Sciences (Roche) FLX platform generated >700,000 expressed sequence tags (ESTs) from two normalized EST libraries, whereas proteome analysis on an LTQ-Orbitrap mass spectrometer detected >27,000 nonredundant peptide sequences from more than 4000 proteins at sub-parts-per-million (ppm) mass accuracy and a false discovery rate of <1%. Retraining of the SNAP gene prediction algorithm using the gene expression data led to a decrease in the number of previously predicted protein-coding genes from 29,000 to 24,000 and refinement of numerous gene models. The P. pacificus proteome contains a high proportion of small proteins with no known homologs in other species ("pioneer" proteins). Some of these proteins appear to be products of highly homologous genes, pointing to their common origin. We show that >50% of all pioneer genes are transcribed under standard culture conditions and that pioneer proteins significantly contribute to a unimodal distribution of predicted protein sizes in P. pacificus, which has an unusually low median size of 240 amino acids (26.8 kDa). In contrast, the predicted proteome of Caenorhabditis elegans follows a distinct bimodal protein size distribution, with significant functional differences between small and large protein populations. Combined, these results provide the first catalog of the expressed genome of P. pacificus, refinement of its genome annotation, and the first comparison of related nematode models at the proteome level.

Figure 1.

Phylogeny of Pristionchus pacificus and proteogenomics workflow employed in this study. (A) Phylogenetic relationship of nematodes with sequenced genomes. The genome sizes (megabases) are written in brackets. The clades are depicted in the tree. (B) P. pacificus gene expression was assessed at the levels of transcription and translation and in three different developmental stages: dauer, J2, and “mixed stage” (containing all developmental stages, including eggs). In proteomics approach, several workflows for protein extraction and separation were used. ESTs detected with 454 pyrosequencing and peptide sequences detected with LTQ-Orbitrap mass spectrometry were used for genome reannotation.

Figure 2.

Overview of the proteomics results. (A) Application of complementary biochemical workflows for protein extraction and peptide separation led to enhanced proteome coverage. (sol) Soluble fraction; (pel) pellet. (B) Peptides were detected with a mean absolute mass deviation of 0.345 ppm. (C) Peptide sequences that mapped to the genome translation (“genomic peptides”) had a median size of 12 amino acids. (D) Distribution of posterior error probabilities (PEP) was markedly different in the genomic and reversed peptide sequences.

Figure 3.

Genome reannotation resulted in new gene predictions and new gene models. (A) Example of a new gene model. New gene model “Contig126-snap.64” contains the old model “Contig126-snap.71”. (B) Example of a new gene prediction. The gene model “Contig125-snap.27” appeared only after retraining of the SNAP prediction algorithm with gene expression data.

Figure 4.

Features of the P. pacificus predicted proteome. (A) Protein size distribution shows that the pioneer proteins are mainly responsible for the unusually low median protein size in P. pacificus. (B) BLAST results of the pioneer proteins against themselves show presence of highly homologous proteins that may have a common origin.

Figure 5.

Comparison and functional analysis of protein size distributions in nematode models. (A) Predicted protein sizes in P. pacificus and B. malayi have a unimodal distribution, whereas C. elegans and C. briggsae have distinct bimodal distributions. (B) Gene Ontology enrichment analysis for short and long proteins in C. elegans shows distinct functional differences between the two classes of proteins.