An integrative strategy to identify the entire protein coding potential of prokaryotic genomes by proteogenomics

Abstract

Accurate annotation of all protein-coding sequences (CDSs) is an essential prerequisite to fully exploit the rapidly growing repertoire of completely sequenced prokaryotic genomes.However, large discrepancies among the number of CDSs annotated by different resources, 5missed functional short open reading frames (sORFs), and overprediction of spurious ORFs represent serious limitations.Our strategy towards accurate and complete genome annotation consolidates CDSs from multiple reference annotation resources, ab initio gene prediction algorithms and in silico ORFs in an integrated proteogenomics database (iPtgxDB) that covers the entire protein-10 coding potential of a prokaryotic genome. By extending the PeptideClassifier concept of unambiguous peptides for prokaryotes, close to 95% of the identifiable peptides imply one distinct protein, largely simplifying downstream analysis. Searching a comprehensive Bartonella henselae proteomics dataset against such an iPtgxDB allowed us to unambiguously identify novel ORFs uniquely predicted by each resource, including 15 lipoproteins, differentially expressed and membrane-localized proteins, novel start sites and wrongly annotated pseudogenes. Most novelties were confirmed by targeted, parallel reaction monitoring mass spectrometry, including unique ORFs and variants identified in a re-sequenced laboratory strain that are not present in its reference genome. We demonstrate the general applicability of our strategy for genomes with varying GC content 20 and distinct taxonomic origin, and release iPtgxDBs for B. henselae, Bradyrhozibium diazoefficiens and Escherichia coli as well as the software to generate such proteogenomics search databases for any prokaryote.

A proteogenomic resource enabling integrated analysis ofListeriagenotype-proteotype-phenotype relationships

“…Comparative genomics identified both shared core gene clusters and gene clusters unique to EGD-e and ScottA (right upper panel). Moreover, an ab initio gene prediction based on Prodigal and an advanced in silico (six-frame translation) annotation (see Materials and Methods) were integrated with the RefSeq annotation to obtain a minimally redundant iPtgxDB [43] for EGD-e and for ScottA. DDAbased proteomics data were searched against the publicly available iPtgxDBs (https://iptgxdb.expasy.org) to obtain proteogenomic evidence for novel open reading frames (ORFs), novel start sites, and expressed pseudogenes (right lower panel).…”

“…Although a complete NCBI reference genome sequence existed for EGD-e, the NCBI reference genome sequence for ScottA was incomplete and consisted of five contigs [44]. Motivated by our recent finding of significant differences between an NCBI reference genome and the de novo assembly of the actual lab strain [43], and an earlier study on Pseudomonas aeruginosa PAO1 that had demonstrated substantial genomic fluidity between closely related strains [45], we sequenced and de novo assembled both genomes to create the best possible reference sequence for the two strains, an important aspect for the proteogenomics element. Next, a comparative genomics analysis was carried out to identify core and strain-specific genes, which, upon integration with protein abundance data, might provide clues to explain the different phenotypes of the strains.…”

“…After establishing a rapid and reproducible sample preparation workflow, we were able to quickly and quantitatively profile Listeria under various conditions and perturbations by DIA/SWATH. Lastly, in addition to the standard proteotype search against generic protein databases such as NCBI RefSeq and Uniprot, a search was carried out against a specialized, integrated proteogenomics search database (iPtgxDB; see Materials and Methods) [43], allowing us to identify protein expression evidence for as of yet unannotated small ORFs (smORFs), additional N-terminal start sites, and expressed pseudogenes (Fig 1).…”

“…This has two benefits: First, proteogenomics can be included in the initial genome annotation, thereby increasing its quality [43,48]. A public website (https://iptgxdb.expasy.org/) supports this for newly sequenced genomes [43], such as isolates from microbiomes or type strains from the Genomic Encyclopedia of Bacteria and Archaea (GEBA) initiative, which aims to expand the phylogenetic diversity of completely sequenced prokaryotic genomes [49]. Second, the identification of more comprehensive protein catalogs including functionally relevant smORFs, will better enable studies to model systems based on quantitative data of all functional elements.…”

“…The peptides from the two libraries were mapped against the protein-coding sequences of the other genome to identify strain-specific proteotypic peptides and to assess whether some of the strain-specific protein-coding genes were expressed at the protein level. This mapping was done using an in-house tool that extends the original version of PeptideClassifier [59] thereby enabling proteogenomics for prokaryotes [43]. Only 15.5% and 16.2% of peptides from EGD-e and ScottA, respectively, were strain-specific proteotypic, indicating that most peptides could be used to quantify protein abundance in both strains and potentially also in other Listeria strains.…”