Proteome and Secretome Analysis Reveals Differential Post-transcriptional Regulation of Toll-like Receptor Responses

Abstract

The innate immune system is the organism's first line of defense against pathogens. Pattern recognition receptors (PRRs) are responsible for sensing the presence of pathogen-associated molecules. The prototypic PRRs, the membrane-bound receptors of the Toll-like receptor (TLR) family, recognize pathogen-associated molecular patterns (PAMPs) and initiate an innate immune response through signaling pathways that depend on the adaptor molecules MyD88 and TRIF. Deciphering the differences in the complex signaling events that lead to pathogen recognition and initiation of the correct response remains challenging. Here we report the discovery of temporal changes in the protein signaling components involved in innate immunity. Using an integrated strategy combining unbiased proteomics, transcriptomics and macrophage stimulations with three different PAMPs, we identified differences in signaling between individual TLRs and revealed specifics of pathway regulation at the protein level.

Fig. 1.

Experimental design used to identify and quantify changes in the proteome and secretome during macrophage stimulation with TLR ligands. A, A schematic diagram of the canonical Toll-like receptor signaling through TLR2, TLR4 and TL7. B, 3-plex SILAC strategy with LC-MS/MS was used to study the effects of TLR stimulation. RAW264.7 cells were labeled with heavy (K8, R10), medium (K4, R6) and light (K0, R0), isotopes in culture and simulated with one of the three TLR ligands (LPS, P3C, or R848) or left unstimulated. The cell lysates were collected for the proteome study and conditioned media were collected for the secretome study. The proteins were extracted from each of the three samples, the samples were combined, the proteins were separated via SDS-PAGE and digested with trypsin as described in Experimental Procedures. Four biological replicates were used to perform independent experiments for each type of analysis.

Fig. 2.

Global comparison of the proteome and secretome data sets. A, Total numbers of proteins quantified in a minimum of two biological replicates for LPS, P3C, and R848. B, Principal Component Analysis with Perseus software. Squares represent the LPS treatment, circles - the P3C treatment, and triangles—the R848 treatment. The 6 h time point is shown in black and the 12 h time point—in red. C, Pairwise comparison between the six data sets. The triangle to the right of the diagonal with cells marked in red shows the absolute numbers of identified proteins overlapping between the two given data sets, and the triangle to the left of the diagonal with cells marked in green shows the percentage of proteins common between the two given data sets. The intensity of the color increases with the increase of the overlap.

Fig. 3.

Changes in the proteins common to all the data sets. The heatmap represents the hierarchical clustering of the common proteins in the secretome for all the timepoints for the cells treated with LPS. The color key above represents the changes (log2 scale), from dark blue representing the largest decrease, to red representing the largest increase. Cells colored gray represent missing data. Each row is a protein and each column is a sample. Samples are named based on data type, treatment type, time point and replicate as described in “Data analysis”.

Fig. 4.

Heatmap of proteins present in either the secretome or the proteome with fold change equal or larger than 2 induced by any TLR treatment. The color key is on the left and represents the changes (log2 scale, fold change equal or larger than 2 (in either direction), from dark blue representing the largest decrease, to red representing the largest increase. Cells colored gray represent missing data. Each row is a protein and each column is a sample. Samples are named based on data type, treatment type, and time point as described in Data analysis. Proteins (or rows) are sorted (decreasing order) based on average fold change across all treatments.

Fig. 5.

Correlation between the proteome, secretome and transcriptome. A, Pearson correlation values for the proteome and secretome data with microarray data from RAW264.7 cells treated with the same TLR ligands for one, two, or four hours. Log transformed (base 2) fold change data was used in call cases. B, The proportion of transcripts and proteins that exhibit more than 2-fold change in either direction at the protein level, but not at the transcript level (and vice versa) for each treatment.

Fig. 6.

Enrichment analysis using DAVID. Three most enriched processes are shown for (A) proteome and (B) secretome data sets. The color of the symbols indicates treatment type: the blue symbols represent data points from the LPS treatment, the red symbols represent data points from the P3C treatment, and the green symbols represent data points from the R848 treatment. The shape of the symbols indicates the cellular process: circles represent the immune response, diamonds represent the response to wounding, squares represent the DNA metabolic process, triangles represent chemotaxis and crosses represent translation.

Fig. 7.

The targeted proteomics results of the heat-killed pathogen challenge compared with the results obtained for the single ligand stimulations. Eight representative proteins are depicted. (A: Cd14, cluster of differentiation 14; B: Man2B1, lysosomal alpha-mannosidase; C: C3, complement factor 3; D: Ccl9, chemokine (C-C motif) ligand 9; E: Bax, Bcl-2-associated X protein; F: C1qb, complement factor C1qb; G: Lyz2, lysozyme C2; H: H2-K1, H-2 class I histocompatibility antigen, K-B alpha chain). The top graphs in each panel indicate fold changes upon 6h and 24h treatments with LPS (light gray), P3c (dark gray) and R848 (black). The bottom panels indicate fold changes upon 6h and 24h treatments with P. aeruginosa (white), S. aureus (dotted) and B. cenocepacia (hatched).