Selective Enrichment and Direct Analysis of Protein S-Palmitoylation Sites

Abstract

S-Fatty-acylation is the covalent attachment of long chain fatty acids, predominately palmitate (C16:0, S-palmitoylation), to cysteine (Cys) residues via a thioester linkage on proteins. This post-translational and reversible lipid modification regulates protein function and localization in eukaryotes and is important in mammalian physiology and human diseases. While chemical labeling methods have improved the detection and enrichment of S-fatty-acylated proteins, mapping sites of modification and characterizing the endogenously attached fatty acids are still challenging. Here, we describe the integration and optimization of fatty acid chemical reporter labeling with hydroxylamine-mediated enrichment of S-fatty-acylated proteins and direct tagging of modified Cys residues to selectively map lipid modification sites. This afforded improved enrichment and direct identification of many protein S-fatty-acylation sites compared to previously described methods. Notably, we directly identified the S-fatty-acylation sites of IFITM3, an important interferon-stimulated inhibitor of virus entry, and we further demonstrated that the highly conserved Cys residues are primarily modified by palmitic acid. The methods described here should facilitate the direct analysis of protein S-fatty-acylation sites and their endogenously attached fatty acids in diverse cell types and activation states important for mammalian physiology and diseases.

Keywords: S-palmitoylation; affinity enrichment; fatty acylation; mass spectrometry-based proteomics; posttranslational modification; site identification.

Figure 1.. Methods to enrich and identify S-fatty-acylated, alk-16 labeled and S-acylated proteins by proteomics.

(A) With the chemical reporter labeling, cells are treated with alk-16, an alkyne-tagged reporter of saturated fatty acids. Following cell lysis, modified proteins can be reacted with an azide-functionalized reagent via biorthogonal labeling and captured on Neutravidin™ beads. Lysate subjected to NH₂OH treatment or cells treated with DMSO are used as a negative control. Statistical comparison of samples enables the selective identification of candidate S-fatty-acylated, N/O-alk-16 labeled and alk-16 labeled proteins. (B) In the acyl-RAC method, following cell lysis, free Cys are capped with N-ethyl maleimide (NEM) in the presence of TCEP. Acylated Cys are hydrolyzed with hydroxylamine (NH₂OH) and reacted with the thiopropyl sepharose resin. Lysate not subjected to NH₂OH treatment is used as a negative control. Comparison of proteins enriched in the two different conditions as indicated by the two-headed arrows enabled the identification of candidate S-acylated proteins.

Figure 2.. S-fatty-acylated proteins in naïve and LPS/IFN-γ stimulated RAW264.7 macrophages.

(A-B) Volcano plot showing the results of the two-sample test (FDR = 0.05, s0 = 1) for the alk-16 chemical reporter strategy experiment (n= 4 replicates, LFQ quantification). (A) alk-16/DMSO and (B) alk-16/(alk-16+NH₂OH). Proteins reported to be S-fatty-acylated or GPI-anchored in Uniprot are shown in red and green, respectively. N-myristoylated proteins are shown in blue. Dually lipidated proteins (N-myristoylated and S-fatty-acylated) are shown in red. Putative S-fatty-acylated proteins, identified with high, medium or low confidence are shown in dark grey (Supporting Information, experimental procedures). All other proteins are shown in light grey. (D) Venn diagram showing the number of proteins identified with medium and high confidence in experiments A, B, E. (D) Table showing a subset of proteins identified in experiments shown in A, B, E. The four first columns indicate if proteins have been reported to be S-palmitoylated (palm), N-myristoylated (myr), GPI-anchored (GPI) or were found in previous proteomics experiments (prev prot). For the chemical reporter or acyl-RAC experiments, the means of log2 LFQ intensities and MS/MS counts, as well results of the two-sample tests (“+” = significant) are shown. “MS ratio” corresponds to the ratio of the mean MS/MS counts of the two samples being compared, and was used to further sort the data (“MS ratio” ≥5 was attributed to a high confidence hit (HC), 2.5≤”MS ratio”<5 to medium confidence and 1<“MS ratio”<2.5 to low confidence). (E) Volcano plot showing the results of the two-sample test (FDR = 0.01, s0 = 1, + NH₂OH versus – NH₂OH) for the acyl-RAC method (n= 4 replicates, LFQ quantification). Proteins shown in red are reported as S-fatty-acylated in Uniprot. Candidate S-fatty-acylated proteins, identified with high, medium or low confidence are shown in dark grey (Supporting Information, experimental procedures). All other proteins are shown in light grey. (F-G) Quantitative analysis of S-fatty-acylated proteome following LPS/IFN-γ stimulation. Volcano plot showing the statistical significance (FDR = 0.05, s0 = 1) of changes in enrichment levels between stimulated and naïve cells as a function of the t-test difference for (F) the chemical reporter method or (G) acyl-RAC.

Figure 3.. S-fatty-acylated proteins in HeLa cells.

(A) Volcano plot showing the statistical significance of enrichment level changes +/− NH₂OH as a function of the t-test difference for (A) the chemical reporter method (p =0.01, s0=1) or (B) acyl-RAC (p =0.001, s0=1). (C) Venn diagram showing the number of proteins identified with medium and high confidence in experiments A-B. (D) Table showing a subset of proteins identified in experiments shown in A and B. The four first columns indicate if proteins have been reported to be S-palmitoylated (palm). For the chemical reporter or acyl-RAC experiments, the means of log2 LFQ intensities and MS/MS counts, as well results of the two-sample tests (“+” = significant) are shown. “MS ratio” corresponds to the ratio of the mean MS/MS counts of the two samples being compared, and was used to further sort the data (“MS ratio” ≥5 was attributed to a high confidence hit (HC), 2.5≤”MS ratio”<5 to medium confidence and 1<“MS ratio”<2.5 to low confidence). (E-F) S-fatty-acylation of known and candidate S-fatty-acylation proteins were assessed by the chemical reporter (D) and acyl-RAC (E) methods combined with western blot analysis. (D) alk-16 chemical reporter. Beads = 5x Input or Supernatant (Sup), except for VCP and p62: Beads = 10x Input or Sup. E) Acyl-RAC. alk-16 chemical reporter. Beads = 2x Input or Sup except for β-tubulin where Beads = 4x Input or Sup.

Figure 4.. Direct identification of protein S-fatty-acylation sites.

(A). Protocol scheme. Following treatment with alk-16 and cell lysis, samples are subjected to CuAAC with Biotin azide and free Cys are capped with NEM. Samples are digested and enriched on Neutravidin beads. S-palmitoylation is selectively cleaved with NH₂OH and newly revealed free Cys are reacted with iodoacetamide. (B-C) MS/MS spectra of modified peptides. S-palmitoylation sites are labelled with carbamidomethyl (ca) and unmodified Cys are labeled with NEM (ne). (B) MS/MS spectra of SNAP23 was obtained following trypsin digest. (C) MS/MS spectra of caveolin-1 was obtained following chymotrypsin digest.

Figure 5.. Characterization of IFITM3 S-fatty-acylation sites and fatty acid composition.

(A) Schematic of IFITM3 and sites of S-fatty-acylation.^, (B) MS/MS spectra of IFITM3 C71-ca, C72-ca was obtained following chymotrypsin digest. (C) MS/MS spectra of IFITM3 modified peptide showing S-palmitoylation on C72. Ca = carbamidomethyl and pa = palmitic acid.