Multiplex Substrate Profiling by Mass Spectrometry for Kinases as a Method for Revealing Quantitative Substrate Motifs

Abstract

The more than 500 protein kinases comprising the human kinome catalyze hundreds of thousands of phosphorylation events to regulate a diversity of cellular functions; however, the extended substrate specificity is still unknown for many of these kinases. We report here a method for quantitatively describing kinase substrate specificity using an unbiased peptide library-based approach with direct measurement of phosphorylation by tandem liquid chromatography-tandem mass spectrometry (LC-MS/MS) peptide sequencing (multiplex substrate profiling by mass spectrometry, MSP-MS). This method can be deployed with as low as 10 nM enzyme to determine activity against S/T/Y-containing peptides; additionally, label-free quantitation is used to ascertain catalytic efficiency values for individual peptide substrates in the multiplex assay. Using this approach we developed quantitative motifs for a selection of kinases from each branch of the kinome, with and without known substrates, highlighting the applicability of the method. The sensitivity of this approach is evidenced by its ability to detect phosphorylation events from nanogram quantities of immunoprecipitated material, which allows for wider applicability of this method. To increase the information content of the quantitative kinase motifs, a sublibrary approach was used to expand the testable sequence space within a peptide library of approximately 100 members for CDK1, CDK7, and CDK9. Kinetic analysis of the HIV-1 Tat (transactivator of transcription)-positive transcription elongation factor b (P-TEFb) interaction allowed for localization of the P-TEFb phosphorylation site as well as characterization of the stimulatory effect of Tat on P-TEFb catalytic efficiency.

Figure 1

Depiction of the MSP-MS assay for kinases. (a) Recombinant or immunoprecipitated kinases are added to the peptide library. Aliquots are removed from the reaction at several time points, quenched, and then analyzed by LC–MS/MS. Phosphopeptides were identified and quantified at each time point, then used to develop specificity motifs and generate catalytic efficiency values for each peptide in the multiplex assay. (b–g) Validation of the use of MSP-MS for kinases using a variety of disparate, well-studied kinases and (h–k) characterization of less-studied kinases. All substrate signatures were developed using the phosphopeptides identified after 1200 min of kinase reaction with the peptide library. (a) GCK. (b) MAPK. (c) cSrc. (d) Casein kinase II. (e) Protein kinase A. (f) CAMKK2. (g) CSNK1γ1. (h) BMPR2. (i) ALK2. (j) TbERK8. (k) LegK1.

Figure 2

Delineation of cyclin-dependent kinase substrate specificities. (a) Development of a positional scanning synthetic combinatorial library for further substrate specificity determination after initial MSP-MS characterization. A scaffold peptide is chosen from the library and two positions are fixed, based on the two key determinants of substrate specificity, as identified in MSP-MS. Every other position is then varied, one at a time, to a mixture of a diversity of amino acids representing all distinct physicochemical groups. Kinetic PS-SCL assays allow for generation of a more detailed substrate specificity heat map. (b) Substrate profile from MSP-MS analysis of CDK9/cyclin T1. (c) Heat map generated using maximum percent conversion values from CDK9 PS-SCL experiment. Positive enrichment was visualized as blue, and zero was set to yellow. Residue indicated in red is highlighted for comparison between kinases. (d) Substrate profile from CDK1/cyclin B. (e) Heat map generated using maximum percent conversion values from CDK1 PS-SCL experiment. (f) Substrate profile from CDK7/cyclin H/MNAT1. (g) Heat map generated using maximum percent conversion values from CDK7 PS-SCL experiment. (h) Differential phosphorylation of the “P + 3 K” peptide, RVYLTSPKKPES, by CDK1, CDK9, and CDK7, highlighting each enzyme’s preference for basic residues at the P + 3 position. P + 3 K is a better substrate for CDK1 than for CDK9 (20-fold difference in k_cat/K_M) or for CDK7 (substrate is too poor to obtain k_cat/K_M value). (i) CDK1 phosphorylation of PS-SCL scaffold peptide RVYLTSPKAPES and P + 3 A9 → K9 peptide RVYLTSPKKPES. Alteration of the P + 3 residue makes the peptide a 2.5-fold better substrate.

Figure 3

MSP-MS analysis of recombinant and immunoprecipitated P-TEFb and the HIV-1 Tat–P-TEFb interaction. (a) Silver stain gel of FLAG-CDK9 immunoprecipitation from HEK293 cells stably expressing FLAG-CDK9. Lane contents: 1, ladder; 2, unbound to beads; 3, first wash; 4, FLAG eluent; 5, lysate. Indicated bands correspond to the correct molecular weight expected for FLAG-CDK9 and cyclin T1. In lane 4, 5 μL out of the 110 μL elution from the FLAG beads was loaded for SDS–PAGE and silver staining. This method has an approximately 50 fmol limit of detection, which for 124 kDa P-TEFb corresponds to approximately 5 ng as a lower limit of protein abundance. Total yield from this preparation was estimated as approximately 100 ng. An amount of 0.4 ng of active P-TEFb is required per time point (Figure S10). (b) Anti-cyclin T1 Western blot of FLAG-CDK9 immunoprecipitation from HEK293 cells stably expressing FLAG-CDK9 reveals that cyclin T1 is the primary coimmunoprecipitant of FLAG-CDK9. Lane contents: 1, ladder; 2, unbound to beads; 3, FLAG eluent. (c) Substrate profile from immunoprecipitated P-TEFb, which aligns with recombinant P-TEFb substrate specificity. (d) Schematic representing the super elongation complex (SEC) in which Tat associates with P-TEFb and the scaffolding protein in order to coordinate P-TEFb phosphorylation of the RNA Pol II CTD and the two negative elongation factors NELF and DSIF, which allows the stalled RNA Pol II to resume transcription of the integrated viral genome. Residues phosphorylated by P-TEFb (RNA Pol II CTD Ser5; NELF Ser181, Thr272, Ser281, and Ser353; DSIF Thr775 and Thr784) and the proline residues driving specificity are represented in the figure in bold. (e) Kinetic experiments including HIV-1 Tat reveal that some peptides have an enhanced k_cat/K_M in the presence of Tat while others are relatively unchanged. (f) Weighted specificity motif generated using the sequences of peptides that Tat increased the k_cat/K_M values for by more than 2-fold. Alignment of this motif with RNAP II CTD Ser5 and RNAP II CTD Ser2 as the phosphorylation site indicates that Tat selectively increases P-TEFb catalytic efficiency toward peptides that are similar to CTD Ser5.