Corrigendum: Elucidation of the Signatures of Proteasome-Catalysed Peptide Splicing

[This corrects the article DOI: 10.3389/fimmu.2020.563800.].

There were three minor errors in our original article.
1) The Materials and Methods section required additional detail about the search strategy implemented in Peaks and the discovery workflow to facilitate replication of the methodology. Two extra sentences have been added to expand on this, and a third sentence describing a step that was used in our previous HLA-I discovery workflow for spliced peptides, but that is not required in the interrogation of datasets derived from polypeptide digests, has been removed.
A correction has been made to Materials and Methods, Discovery workflow for identification of non-spliced and spliced peptides, paragraphs 1, 2 and 3: Paragraph 1: "PEAKS de novo assisted sequencing was implemented for the assignment of nonspliced peptides derived from individual polypeptide sequences following proteasomal digest of each of the 25 precursor substrates, and no PTMs were included in the de novo search space".
Paragraph 2: "Therefore, for DNUPs containing a total of 'n' leucine residues, all permutations (2 n ) of L/I variants were computed prior to in silico splicinge.g. for de novo sequence LTSLTLKE originating from polypeptide precursor LTSLVRRATLKENEQIPK, 2 3 combinations of the original de novo sequence would be computed (LTSLTLKE, LTSITLKE, ITSLTLKE, LTSLTIKE, LTSITIKE, ITSLTIKE, ITSITLKE, ITSITIKE) and each sequence input to the splicing algorithm".
Paragraph 3: "The (n-1) th fragment was first scanned for a contiguous match across the polypeptide, and when found, its corresponding splice partner fragment was scanned for a contiguous match within the remainder of the polypeptide sequence. Due to the lack of an applied false discovery rate (FDR) for identification of spliced peptides from de novo sequencing of LC-MS/MS spectra that were not matched to a pre-defined database, very short spliced peptide sequences (5-7 aa) were omitted from analysis, and only sequences with a length of 8 aa or greater were considered. Trans-spliced peptides were omitted from the analysis". 2) The sequence of precursor polypeptide PP13 was erroneously given in Table 1 as LQPQLIHLYYFDCFSESAIRNA, but this peptide in fact ended in 'K' and not 'NA'. The PP13 sequence has now been amended to LQPQLIHLYYFDCFSESAIRK; and overall amino acid frequencies within polypeptide precursors in Table 2 have been adjusted to reflect this minor change. The corrected Table 1 and Table 2 appear below.
3) In Supplementary Table 2, eight of the original undigested precursor polypeptide sequences were accidentally included in the non-spliced lists (PP7, PP8, PP16, PP17, PP19, PP20, PP21, PP25). These have now been removed. Four spliced peptides (EGCPMVVKF, ALIKPLPSV, FIRNLSFKCS, SFKCSEDDLKTVFAQFGAK) were also erroneously present in the non-spliced lists. Two of them have now been removed as they were 1-mer fusions (FIRNLSFKCS, SFKCSEDDLKTVFAQFGAK) which we did not consider in the manuscript, while the 2 cis-spliced peptides (EGCPMVVKF, ALIKPLPSV) have been added to the spliced peptide lists. The corrected Supplementary Material can be accessed via the link below.
Values in the sentence of the results text stating the total number and percentages of non-spliced and spliced peptides have been amended accordingly. The corrected sentence in the Results, The Relative Proportions of Unique Non-spliced and Cis-Spliced Peptides Generated by the Constitutive Proteasome Are Dependent on Precursor Peptide Length is now as follows: "Overall, we observed a total of 1,200 unique non-spliced (72.9%) and 446 cis-spliced (27.1%) peptides ( Figure 1A)".
The relevant panels of Figure 1 and Supplementary  Figures 1, 2 have also been amended, and figure legends have been revised as needed. The corrected Figure 1 and caption appear below.
"(A) Proportion of unique spliced and non-spliced peptides following a 2h in vitro digestion of 25 self-and HIV-1-derived polypeptides ( Table 1) by the constitutive proteasome. Proportions of spliced and non-spliced peptides within all unique peptides (n=1,646), unique peptides originating from only the 13 longest polypeptide substrates (n=1,331) and unique peptides originating from only the 12 shortest polypeptide precursors (n=315) are shown".
Given the large number of peptides observed overall, the changes made to the peptide lists in Supplementary Tables 1, 2 have not had any impact on the conclusions drawn from the data, and the revised figure panels are not discernibly different from the originals. Importantly, none of these 12 peptides were used for any of the analyses in the manuscript from Figure 2 onwards, as they all contained terminal amino acids of the polypeptide precursors.
In addition to these minor errors in the original version of our article, there was also an omission: the project accession numbers for the mass spectrometry datasets from the control undigested precursor substrates were not included in the Data Availability Statement. These have now been added.
The Data Availability Statement is now as follows: "Mass spectrometry proteomics datasets have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (https://www.ebi.ac.uk/pride) with project accession numbers PXD025893 (for undigested precursor substrates) and PXD021339 (for proteasomal digests of precursor substrates).