Predicting small molecule binding pockets on diacylglycerol kinases using chemoproteomics and AlphaFold

Diacylglycerol kinases (DGKs) are metabolic kinases involved in regulating cellular levels of diacylglycerol and phosphatidic lipid messengers. The development of selective inhibitors for individual DGKs would benefit from discovery of protein pockets available for inhibitor binding in cellular environments. Here we utilized a sulfonyl-triazole probe (TH211) bearing a DGK fragment ligand for covalent binding to tyrosine and lysine sites on DGKs in cells that map to predicted small molecule binding pockets in AlphaFold structures. We apply this chemoproteomics-AlphaFold approach to evaluate probe binding of DGK chimera proteins engineered to exchange regulatory C1 domains between DGK subtypes (DGKα and DGKζ). Specifically, we discovered loss of TH211 binding to a predicted pocket in the catalytic domain when C1 domains on DGKα were exchanged that correlated with impaired biochemical activity as measured by a DAG phosphorylation assay. Collectively, we provide a family-wide assessment of accessible sites for covalent targeting that combined with AlphaFold revealed predicted small molecule binding pockets for guiding future inhibitor development of the DGK superfamily.


SUPPORTING FIGURES
. TH211 treatment blocks catalytic activity of recombinant DGKa and DGKz as determined by ADP-glo substrate assay. Production of active recombinant DGKa-or DGKζ was determined by enhanced activity in DGKa-or DGKζ-HEK293T-versus mock-transfected soluble proteomes as measured using an ADP-glo substrate assay. Specificity of activity to recombinant protein was confirmed by abolishment of the observed activity with heat denatured recombinant lysates (95° C for 5 min). Pretreatment with TH211 (100, 50, 25, 10 and 1 µM) resulted in concentration dependent blockade of recombinant DGKa (A) and DGKζ (B) activity. Data shown are mean +/-SEM for n = 2 independent biological replicates; ****p < 0.0001, ***p < 0.001, or not significant (ns) for samples compared with DMSO-treated recombinant DGK lysate group. Statistical significance was determined using a Dunnett multiple comparison following a one-way ANOVA test.   Figure S2. Expression of recombinant DGKs was comparable in SILAC light and heavy HEK293T cells. Recombinant human DGK proteins were co-expressed in SILAC HEK293T cells and used for chemical proteomic evaluation. Expression of recombinant DGKs was detected by western blots using anti-FLAG antibodies except for DGKz and DGKi, which were detected with anti-HA antibody. Equivalent protein loading was confirmed by anti-GAPDH. Recombinant protein expression was comparable between HEK293T cells cultured in light and heavy media. The catalytic domain (DAGKc and DAGKa region) is shaded in light green. Probe modified Lys and Tyr are shown in dark blue and red, respectively. Lys and Tyr residues confidently predicted by AlphaFold ("Confident" and "Very Confident" predictions, pLDDT > 70) but not modified by our probe are shown in gray. Lys and Tyr residues predicted less confidently ("Low" and "Very Low" predictions, pLDDT < 70) are not highlighted and were not included in the analysis. All data shown are representative of 3 experiments (n = 3 biologically independent experiments). Predicted structures were visualized using PyMOL (Version 2.6; https://pymol.org).

Reagents
TH211 was synthesized and used for live cell chemoproteomics as previously described 1 .

Western blot analysis
Western blot analysis of recombinant protein expression was performed as previously described 3 .

SILAC sample preparation for MS-based chemical proteomic assay
Light and heavy proteomes from TH211-treated cells were prepared for LC-MS chemical proteomic assay as previously described 1 .

LC-MS/MS chemical proteomics
The Results from Byonic and Skyline were combined and filtered in R to retain high confidence peptides as determined by the following criteria: Byonic score ≥ 500, a precursor mass error within 5 ppm; normalized SILAC ratio (SR) > 5, with both isotope dot-product (iDOTP) and ratio dotproduct (rDOTP) ≥ 0.8. These results were used for all analyses except the generation of Table S1, which was based on Byonic analysis only and the results were filtered using these criteria: Score ≥ 300, precursor mass error within 5 ppm, Delta ≥ 25, DeltaMod Score ≥ 20 and Log Prob ≥ 3.0.

Biochemical substrate assay of DGK chimeras
A liposomal substrate assay for measuring DAG kinase activity was performed as previously

ADP-glo substrate assay of TH211 inhibition
TH211 inhibitory activity against recombinant DGKa-or DGKz-HEK293T soluble proteomes was evaluated by ADP-glo following published methods 2 except TH211 pretreatment with lysate was performed in the presence of free ATP (55 µM) in reaction buffer.

APPENDIX
Human DGKa and DGKz chimera construct plasmids were custom synthesized by GenScript.