Proteome-Wide Characterization of Phosphorylation-Induced Conformational Changes in Breast Cancer

Abstract

Because of the close link between protein function and protein folding stability, knowledge about phosphorylation-induced protein folding stability changes can lead to a better understanding of the functional effects of protein phosphorylation. Here, the stability of proteins from rates of oxidation (SPROX) and limited proteolysis (LiP) techniques are used to compare the conformational properties of proteins in two MCF-7 cell lysates including one that was and one that was not dephosphorylated with alkaline phosphatase. A total of 168 and 251 protein hits were identified with dephosphorylation-induced stability changes using the SPROX and LiP techniques, respectively. Many protein hits are previously known to be differentially phosphorylated or differentially stabilized in different human breast cancer subtypes, suggesting that the phosphorylation-induced stability changes detected in this work are disease related. The SPROX hits were enriched in proteins with aminoacyl-tRNA ligase activity. These enriched protein hits included many aminoacyl-tRNA synthetases (aaRSs), which are known from previous studies to have their catalytic activity modulated by phosphorylation. The SPROX results revealed that the magnitudes of the destabilizing effects of dephoshporylation on the different aaRSs were directly correlated with their previously reported aminoacylation activity change upon dephosphorylation. This substantiates the close link between protein folding and function.

Keywords: MCF-7; SILAC; SPROX; aminoacyl-tRNA synthase; chemical denaturation; limited proteolysis; post-translational modification; protein folding; proteomics.

Figure 1

Schematic presentation of the SILAC-LiP protocol used in this work. (A) Experimental workflow for generation of the single and double digestion samples. (B) Data expected from single and double digestion samples.

Figure 2

Volcano plot showing the statistical significance of the H/L ratios for the peptides assayed in the SILAC-LiP study. Gray points represent the 2924 non-hit peptides (p-value > 0.05). Red points represent the 475 hit peptides (p-value < 0.05) that were identified in two or more biological replicates in both single and double digestion groups. Blue points represent the 91 hit peptides that were identified in only one biological replicate in at least one of the single or double digestion groups.

Figure 3

Schematic presentation of SILAC-SPROX protocol use in this work.

Figure 4

Representative SILAC-SPROX data. (a) Data obtained on the oxidized methionine-containing peptide VCM(ox)DFNIIR from isoleucine--tRNA ligase, which was destabilized upon dephosphorylation. (b) Data obtained on the methionine-containing peptide LTGMAFR from GAPDH, which was also destabilized upon dephosphroylation. (c) Data obtained on the oxidized methionine-containing peptide (ac)M(ox)EVTGDAGVPESGEIR from the DNA fragmentation factor subunit alpha, which was stabilized upon dephosphorylation. In all plots, the data point not included in each fitting is indicated with a red X. The solid black curves represent the best fit of the data to equation 1 (a) and (c) or 2 (b) in the Supporting Information. The blue and red dotted curves represent the predicted SPROX curves (i.e., peptide intensity versus [urea] data using arbitrary units for better visualization) for the heavy and light labeled peptides. The dashed vertical lines indicate the C_1/2,SPROX values determined from the SPROX curves.

Figure 5

Schematic representation of the three-dimensional structure of two selected protein hits identified in both the LiP and SPROX experiments. (a) Three-dimensional structure of PKM2 in complex with FBP, K⁺, Mg²⁺ and oxalate ion (PDB:3BJF). Y105 was shown in green. The regions to which selected peptide hits identified in SPROX mapped are shown in yellow, and both were stabilized in the dephosphorylated sample. (b) Three-dimensional structure of GAPDH in complex with NAD (PDB:1ZNQ). Y94, S98 and T99 were shown in green. The region to which selected the peptide hit identified in SPROX mapped is shown in yellow, and it was destabilized in the dephosphorylated sample. In both (a) and (b) the regions to which selected peptide hits identified in LiPs mapped are shown in red and blue, depending on whether the hit peptide was more or less (respectively) susceptible to proteolysis. Images were generated in Kinemage, Next Generation (KiNG).

Figure 6

Correlation between the dephosphorylation-induced changes in folding free energies (ΔΔG_f) measured for aaRSs hits identified in SILAC-SPROX experiment and the dephosphorylation-induced changes in activity of the aaRSs determined elshwere. The filled circles represent aaRSs with defined ΔΔG_f values and were included in the linear regression analysis. The open circles represent aaRSs that were not included in the linear regression analysis because either (1) the SILAC-SPROX data points were too few to calculate a ΔΔG_f value, which is the case for arginine (Arg) tRNA-synthase, or (2) the data point is an outlier, which is the case for histidine (His) tRNA-synthase data point that significantly reduced the correlation coefficient (R² = 0.2238). The ΔΔG_f and activity data used in the correlation are summarized in Table S-3 in the Supporting Information.