Advanced identification of proteins in uncharacterized proteomes by pulsed in vivo stable isotope labeling-based mass spectrometry

Abstract

Despite progress in the characterization of their genomes, proteomes of several model organisms are often only poorly characterized. This problem is aggravated by the presence of large numbers of expressed sequence tag clones that lack homologues in other species, which makes it difficult to identify new proteins irrespective of whether such molecules are involved in species-specific biological processes. We have used a pulsed stable isotope labeling with amino acids in cell culture (SILAC)-based mass spectrometry method, which is based on the detection of paired peptides after [(13)C(6)]lysine incorporation into proteins in vivo, to greatly increase the confidence of protein identification in cross-species database searches. The method was applied to identify nearly 3000 proteins in regenerating tails of the urodele amphibian Notophthalmus viridescens, which possesses outstanding capabilities in the regeneration of complex tissues. We reason that pulsed in vivo SILAC represents a versatile tool to identify new proteins in species for which only limited sequence information exists.

Fig. 1.

Schematic outline of the experimental design used for the pulsed in vivo SILAC approach. a, newts were habituated to mouse liver diet for 28 days. Liver tissue from fully labeled SILAC mice was used to label newts over a period of 60 days. After 20 days, half of the newts were tail-tip amputated and allowed to regenerate; the remaining newts were left undamaged. Tissue from tail tips was isolated after 60 days and prepared for MS analysis. b, isotopic cluster pairs were identified for both damaged and undamaged tissue. The black circle in full MS spectrum defines the light peak; the red circle indicates the heavy Lys-6-labeled isotopic peak. The heavy peak is shifted by 6 Da. MS/MS spectra indicate a mass shift of 6 Da for all detected y ions, also marked in red. Masses for b3, y7, and y8 ions are displayed as examples. Peptide masses were used to perform a MASCOT search on several data bases. c, MASCOT searches for both time points were combined, and heavy light ratios were determined, resulting in several protein group lists. To identify the total number of unique protein groups, we combined all protein lists for both time points.

Fig. 2.

Tissue regeneration in newt results in accelerated [¹³C₆]lysine incorporation after tail amputation. The mean incorporation rate of [¹³C₆]lysine increased from 11.4% in undamaged tail tips to 46.7% in regenerating tails 40 days after amputation. The percentage of heavy-to-light peptide ratios is given on the y axis. The x axis displays the percentage incorporation rate of [¹³C₆]lysine peptide pairs. A shift in the frequency distribution was observed from 93.4% of all ratios within the mean ± S.D. interval in undamaged tail tissue after 60 days of feeding to 86.1% of all ratios within the mean ± S.D. interval in 40-day tail regenerates after 60 days of feeding.

Fig. 3.

Pulsed in vivo SILAC enables efficient identification of peptides in the newt proteome. Venn diagram showing the number of peptide pairs identified in both undamaged newt tail and 40-day tail regenerate after 60 days of feeding with [¹³C₆]lysine-labeled newt liver proteins (union middle) and peptide pairs identified in either undamaged (left), or regenerating tail (right). Human, mouse, and zebrafish IPI databases, as well as NCBI protein entries for Xenopus and Ambystoma, were used for cross-species database searches. N. viridescens NCBI protein entries were used to match the percentage of known newt proteins identified in labeled tail tissue.

Fig. 4.

Identification of previously unknown proteins with a dynamic labeling profile during tail regeneration. Peptide pairs of three selected contigs from undamaged tail and regenerating tail tissue 40 days after amputation (left panel) with no similarity to entries in public data bases with the corresponding peptide sequence in headline including charge, retention time and mascot score (M.S.). Left column displays peptide spectra from undamaged newt tail, and middle column displays spectra from 40-day regenerating tail tissue. Black circles indicate light isotopic peptide peaks, and red circles indicate heavy isotopic peptide peaks, shifting peak clusters by 6 Da in mass. Newt EST sequences with the corresponding sequence region matching to one unique peptide are highlighted in yellow, preceding codons for lysine are highlighted in red (right column, top). Additional identifying unique peptides are indicated in green. RT-PCR analysis from undamaged newt heart (H 0d), undamaged tail (T 0d), and 40-day regenerating tail (T 40d) for the corresponding EST sequences was used to verify expression on the mRNA level (right column bottom). The nucleotide sequences had been deposited in the GenBank database under GenBank Accession numbers GO934291, GO928959, and GO934397.

Fig. 5.

Workflow of peptide to contig assignment for newt ESTs. Individual sequence tags (EST, 5′ reads) from newt tissue were translated into three possible reading frames to generate an in silico peptide database (left) and aligned into contigs on the nucleotide level (right). In silico-translated peptides were identified via MASCOT search, and resulting peptide groups were compared with aligned contigs. More than 90% of peptide/nucleotide alignments were 1:1 assignments.