Proteomic analysis of the testa from developing soybean seeds

Soybean (Glycine max (L.) Merr. cv Jack) seed development was separated into nine defined stages (S1 to S9). Testa (seed coats) were removed from developing seeds at stages S2, 4, 6, 8, and 9, and subjected to shotgun proteomic profiling. For each stage "total proteins" were isolated from 150mg dry weight of seed coat using a phenol-based method, then reduced, alkylated, and digested with trypsin. The tryptic peptides were separated using a C18-reversed phase matrix, then analyzed using an LTQ Orbitrap Mass Spectrometer. Spectra were searched against the Phytozome G. max DB using the Sorcerer 2 IDA Sequest-based search algorithm. Identities were verified using Scaffold 3. A total of 306 (S2), 328 (S4), 273 (S6), 193 (S8), and 272 (S9) proteins were identified in three out of three biological replicates, and sorted into 11 functional groups: Primary Metabolism, Secondary Metabolism, Cellular Structure, Stress Responses, Nucleic Acid metabolism, Protein Synthesis, Protein Folding, Protein Targeting, Hormones and Signaling, Seed Storage Proteins, and Proteins of Unknown Function. In selected instances, individual seed coat proteins were quantified by spectral counting. The number of proteins involved in intermediary metabolism, flavonoid biosynthesis, protein folding and degradation are discussed as they relate to seed coat function. A database was constructed using the open reading frames of Phytozome G. max DB (updated August 8, 2009), plus a decoy DB generated by reversing the target sequences. Spectra were searched against the combined target-decoy DB using the Sorcerer 2 IDA Sequest-based search algorithm. Identities were verified using the Scaffold 3 program, and the false discovery rate (FDR) was estimated as (Ndecoy/Ntarget). For 186,364 peptide spectra examined, the FDR was between 5.18 and 5.26%. Samples were analyzed in biological triplicate; proteins were included in the results only if they were identified in all three biological replicates. Relative protein abundance was quantified by spectral counting ; data presented are the mean of three independent determinations. The coefficient of variation values calculated for each developmental stage are < 0.0516.