Determining DNA-Protein Binding Affinities and Specificities from Crude Lysates Using a Combined SILAC/TMT Labeling Strategy

Abstract

In recent years, quantitative mass spectrometry-based interaction proteomics technology has proven very useful in identifying specific DNA-protein interactions using single pull-downs from crude lysates. Here, we applied a SILAC/TMT-based higher-order multiplexing approach to develop an interaction proteomics workflow called Protein-nucleic acid Affinity and Specificity quantification by MAss spectrometry in Nuclear extracts or PASMAN. In PASMAN, DNA pull-downs using a concentration range of specific and control DNA baits are performed in SILAC-labeled nuclear extracts. MS¹-based quantification to determine specific DNA-protein interactions is then combined with sequential TMT-based quantification of fragmented SILAC peptides, allowing the generation of Hill-like curves and determination of apparent binding affinities. We benchmarked PASMAN using the SP/KLF motif and further applied it to gain insights into two CGCG-containing consensus DNA motifs. These motifs are recognized by two BEN domain-containing proteins, BANP and BEND3, which we find to interact with these motifs with distinct affinities. Finally, we profiled the BEND3 proximal proteome, revealing the NuRD complex as the major BEND3 proximal protein complex in vivo. In summary, PASMAN represents, to our knowledge, the first higher-order multiplexing-based interaction proteomics method that can be used to decipher specific DNA-protein interactions and their apparent affinities in various biological and pathological contexts.

Keywords: DNA pull-down; SILAC; TMT; binding affinity; binding specificity; higher-order multiplexing; protein−DNA interaction.

Figure 1

Schematic overview of the experimental setup to determine DNA–protein binding specificity and affinity in a single experiment. First, nuclear extracts from SILAC-labeled cells are prepared. DNA baits containing the motif of interest or a negative control motif are designed with 5′ biotinylation of the forward strand. A series with ten 3-fold dilutions of these DNA baits is prepared and immobilized in a 96-well filter plate. SILAC-labeled nuclear extracts are added to the DNA baits and incubated for 90 min. After extensive washes, bound proteins are digested, followed by 10-plex TMT labeling. Samples of a titration series, as well as samples of the forward or reverse experiment, are combined. Peptides are identified and quantified by mass spectrometry analysis. Obtained SILAC data are used to identify proteins that bind specifically to the motif of interest, whereas TMT data are used to generate Hill-like curves for K_d^App determination.

Figure 2

Benchmarking of PASMAN using the SP/KLF motif. (A) SILAC Log₂ ratios of the forward and reverse experiments are plotted against each other. Outlier statistics with IQR (interquartile range) cutoff of 1 was applied to call significant proteins. Significant binders for the SP/KLF motif are colored teal, and all other significant proteins are colored green. (B) Venn diagram showing the number of proteins for which a K_d^App could be calculated and how many proteins overlap when comparing binders of the SP/KLF motif with the control motif. (C) Hill-like curve for SP1 binding to the SP/KLF motif obtained by PASMAN. Data points represent the mean of two experiments, and standard errors are the standard error of the mean. (D) PAQMAN, followed by western blotting, to validate SP1 binding to the SP/KLF motif. (E) Hill-like curve for VEZF1 binding to SP/KLF and the exponential curve for VEZF1 binding to the negative control motif obtained by PASMAN. (F) Hill-like curve for PARP1 binding to SP/KLF and the negative control motif obtained by PASMAN. (G) PAQMAN, followed by western blotting, to validate PARP1 binding to the SP/KLF bait.

Figure 3

BANP and BEND3 interact with distinct motifs. (A) Standard DNA pull-down, followed by dimethyl labeling using the CGCG motif. The experiment was done in duplicate with a label swap of replicates. Significant proteins are labeled teal for the CGCG motif, and all other significant proteins are labeled green. (B) SILAC Log₂ ratios are plotted against each other for PASMAN profiling the CGCG motif. (C) Hill-like curve for BANP binding to the CGCG motif using TMT data obtained by PASMAN. (D) PAQMAN, followed by western blotting, to validate BANP binding to the CGCG motif. (E) Western blot analysis of BANP and BEND3 binding to the CGCG motif. (F) Hill-like curve for BEND3 binding to the BEND3 motif obtained from a standard PAQMAN experiment. (G) Western blot analysis of BANP and BEND3 binding to the BEND3 motif.

Figure 4

Identification of potential BEND3 interaction partners. (A) Design of the BEND3-V5-miniTurbo overexpression construct. (B) Western blot analysis of biotinylation activity of the BEND3-V5-miniTurbo overexpression construct in HeLa cells. (C) Mass spectrometry analysis of proteins captured by the biotin pull-down. Label-free quantitation of triplicates was used, and proteins were called significant with a p-value of <0.05 and a Log₂ fold-change of 5. Significant proteins are colored teal, significant proteins belonging to the NuRD complex are colored green, and the bait is colored magenta.