Proteomics‐enabled Hypothesis Generation in Eosinophils: the Cases of LOC100996504, ARHGEF18 and NHSL2

White blood cells (WBCs) can squeeze between capillary endothelial cells into tissues (extravasation). After activation with interleukin‐5 (IL5), the cytoskeleton of eosinophils (granulocytic WBCs) undergoes a dramatic conformational change, with ovoid cells becoming acorn‐shaped, and the bi‐lobed nucleus being propelled into the conical pole; the nucleopod. We have gathered an abundance of quantitative proteomic information for resting and IL5‐activated human eosinophils. Our attention was drawn to two proteins that were poorly annotated in the human protein database to which peptide spectral matches (PSMs) were made. Tryptic peptides from eosinophil lysates were labelled with isobaric tags. At the protein level, LOC100996504 abundance significantly decreased (p<0.01) over the 5‐min interval. Phosphopeptides were enriched using immobilized metal affinity chromatography prior to liquid chromatography tandem mass spectrometry. PhosphoRS was used to localize phosphorylation sites. Nance Horan Syndrome‐like protein 2 (NHSL2) was one of the most highly phosphorylated proteins (fold‐change>2, p<0.01 for 5 phosphoisoforms) after 5 min of activation. We hypothesized that these proteins play unique roles in shape change. Protein sequences predicted from publicly available RNA‐Seq data were used to further interrogate LOC100996504 and NHSL2. We found evidence that in WBC, LOC100996504 is transcribed as an extra N‐terminal exon of the ARHGEF18 Rho guanine nucleotide exchange factor. We dubbed the longer isoform LOCGEF and used polymerase chain reaction (PCR) and sequencing to verify the LOCGEF coding sequence (CDS). Immunoblotting of eosinophil extracts using a commercial anti‐ARHGEF18 antibody revealed that a ~175‐kDa protein (roughly the size expected for LOCGEF) was present rather than the canonical 114‐kDa ARHGEF18. Protein sequences were appended to the existing protein database, so that raw mass spectra files could be searched to see if any corresponded to our predicted CDS. Four peptides overlapping the exon‐exon junction of LOC100996504 and ARHGEF18 were discovered. Immunoblotting of eosinophil extracts suggested that several NHSL2 variants were present, none of which corresponded to the expected size for the 709‐residue protein predicted by the Uniprot‐annotated CDS. NHSL2's putative nucleotide sequence based on WBC RNA‐Seq data encodes two longer proteins of 1141 and 937 residues, arising from alternative splicing of 5′ exons. All mass spectra were re‐searched using the annotated protein database with added custom sequences and peptides corresponding to the custom sequences were identified. Remarkably, some of the discovered peptides overlapped our predicted exon‐exon junctions. Ongoing studies are testing our hypotheses about function. The examples of NHSL2 and LOCGEF highlight the importance of integrating proteomic and RNA‐Seq datasets to enable discovery and investigation of novel, cell type‐specific proteins and proteoforms. Support or Funding Information NIH R01 HL088394