Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Outline of metabolism-focused CRISPR screen. b. Comparison of metabolic gene sgRNA abundance on day 24 between MLN4924-treated and DMSO-treated cells in CRISPR screen. c. Validation of MLN4924 sensitivity for select gene knockdowns via flow cytometry (FACS)-based cell competition experiment in U2OS cells. Data show n=3 biological replicates from independent experiments. d. Schematic of the de novo purine biosynthesis and purine salvage pathways in addition to the ADP-ribose hydrolase activity of NUDT5. Metabolites are italicized. e. Mass-spectrometry proteomics of NUDT5 3xFLAG immunoprecipitation from HEK293T cells. Statistical data are derived from n=3 technical replicates. f. Western blot of FLAG immunoprecipitation from endogenously edited NUDT5 3xFLAG HEK293T cells. Similar results were obtained in two independent experiments. g. Survey of NUDT5 post-translation modifications from the PhosphoSitePlus database filtered for n>5 observations. h. Western blot of FLAG immunoprecipitations from ΔNUDT5 HEK293T cells expressing wildtype NUDT5 3xFLAG and mutants. Similar results were obtained in two independent experiments. i. FACS-based growth competition experiment comparing ΔNUDT5 and rescue mutants to wildtype HEK293T cells treated with MLN4924. Data show n=3 biological replicates from a representative experiment. Similar results were obtained in two independent experiments.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: CRISPR, Comparison, Biomarker Discovery, Flow Cytometry, Activity Assay, Mass Spectrometry, Immunoprecipitation, Derivative Assay, Western Blot, Expressing

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Western blot of U2OS cells treated with indicated amounts of MLN4924 for 16 hours. Data are representative of two independent experiments. b. CellTiter Glo cell viability assay of U2OS cells treated with MLN4924 for 72 hours. Data are the mean and error bars are SEM of n=3 biological replicates from independent experiments. c. Western blot of sequential immunoprecipitation from HEK293T cells expressing PPAT Strep and NUDT5 3xFLAG d. Mass spectrometry of PPAT Strep -NUDT5 3xFLAG sequential immunoprecipitation compared to NUDT5 3xFLAG -only immunoprecipitation and normalized to NUDT5 bait protein abundance. Data show the mean of n=3 technical replicates for each condition.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Western Blot, Viability Assay, Immunoprecipitation, Expressing, Mass Spectrometry, Quantitative Proteomics

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Size exclusion chromatography (SEC) trace (left) and Coomassie-stained gel of PPAT-NUDT5-AMP complex (right) used for cryo-EM analysis. b. Motion-corrected and denoised micrograph of PPAT-NUDT5-AMP complex. c. Representative 2D-class averages for PPAT-NUDT5-AMP complex. Micrographs and 2D class averages from the 6-meTIMP and 6-benzylTIMP datasets were highly similar to that of the AMP-bound dataset. d, e. SEC and Coomassie-stained gels of PPAT-NUDT5 complexes formed in the presence of d. 6-meTIMP and e. 6-benzylTIMP. f. Cryo-EM data processing particle flow for the three nucleotide-bound PPAT-NUDT5 structures. All processing steps were performed in CryoSparc. g-i. Angular distribution of particles and FSC curves from refinements of PPAT-NUDT5-purine complexes with g. AMP, h. 6-meTIMP and i. 6-benzylTIMP.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Size-exclusion Chromatography, Staining, Cryo-EM Sample Prep

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Cryo-EM density of PPAT-NUDT5-AMP complex. b. Structural model of PPAT-NUDT5-AMP complex. The light orange box shows the location of the R70-Y74 loop interface, and the dark orange box indicates the positioning of the molecular glue interface from the same NUDT5 protomer. c. The R70-Y74 loop interface showing specific interactions between NUDT5 and PPAT including the phospho-Y74 site (right, bottom). d. PPAT-NUDT5-AMP molecular glue interface. Cryo-EM density for AMP is shown as a transparent surface. Specific interactions between protein components and the AMP ribose (top right) and purine ring (bottom right) are shown. e. Western blot of FLAG immunoprecipitations from ΔNUDT5 HEK293T cells re-expressing wildtype NUDT5 3xFLAG and C-terminal tail mutants. Similar results were obtained in two independent experiments.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Cryo-EM Sample Prep, Western Blot, Expressing

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: Example cryo-EM density from the PPAT-NUDT5-AMP map with fit molecular model highlighting: a. the R70-Y74 interface b. AMP-bound at the molecular glue interface/PRPP active site. c. the 4Fe-4S cluster bound by PPAT and d. the processed N-terminus of PPAT. e, f. Multiple sequence alignment (MSA) of e. NUDT5 R70-Y74 loop region and f. A NUDT5-binding region of PPAT. g, h. Alignment of PPAT protomers from the human PPAT-NUDT5-AMP complex and g. E. coli PPAT bound to a non-hydrolysable PRPP analog (PDB: 1ECC), h. E. coli PPAT bound to AMP (PDB: 1ECJ) and B. subtilis PPAT bound to GMP (PDB: 1AO0). Only the nucleotides are shown from the bacterial structures for clarity i. MSA of NUDT5 C-terminal tail region.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Cryo-EM Sample Prep, Sequencing, Binding Assay

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Biochemical activity of PPAT outlining effects of purines and NUDT5. b. PPAT activity assay measuring AMP-dependent inhibition in the presence and absence of wildtype NUDT5 and mutants conducted with 1 mM PRPP. Data points are the mean and error bars are SEM from n=3 independent experiments. c. Left – representative Western blot of FLAG immunoprecipitations from ΔPPAT HEK293T cells transiently expressing wildtype PPAT 3xFLAG or an AMP-binding deficient mutant in the presence and absence of adenine. Right – quantification of immunoprecipitated NUDT5 relative to PPAT 3xFLAG bait and normalized to the no adenine condition. Data show values from n=3 independent biological replicate experiments and error bars are SEM. d. [ 15 N-amide]-glutamine labeling of purines in the de novo biosynthesis pathway. e. Fractional enrichment of AMP (M+2) and GMP (M+3) isotopologs in [ 15 N-amide]-glutamine labeling experiments conducted without added hypoxanthine. Data show individual values of n=6 biological replicates from two independent experiments and error bars are SEM. f. Fractional enrichment of AMP (M+2) and GMP (M+3) isotopologs in [ 15 N-amide]-glutamine labeling experiments conducted in the presence of hypoxanthine. “EQ” denotes NUDT5 E112Q/E116Q. Data show individual values of n=8 (L217A/K218A only) or 9 biological replicates from three independent experiments and error bars are SEM. Statistical comparisons were performed using Welch’s two-tailed t-test with Bonferroni correction between wildtype and each mutant. *** denotes a Bonferroni adjusted p-value < 0.001 and “ns” is p-value > 0.05.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Activity Assay, Inhibition, Western Blot, Expressing, Binding Assay, Mutagenesis, Immunoprecipitation, Labeling, Two Tailed Test

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Western blot of engineered Expi293F cell line expressing PPAT V5-TwinStrep b. Coomassie-stained gel of purified PPAT and NUDT5 proteins. c. Time-course PPAT activity assay showing linearity of assays conducted with 0.25 and 1 mM PRPP. Data shown are the mean on n=2 two independent experiments. d. PPAT activity with NUDT5 in the absence of AMP conducted with 1 mM PRPP. NUDT5 mutants ‘WWAA’ is W26A/W48A and ‘LKAA’ is L217A/K218A. e-g. PPAT activity assay measuring inhibitory effects of e. AMP in the presence and absence of NUDT5 wildtype and indicated mutants (1 mM PRPP), f. AMP, GMP, and IMP in the presence and absence of wildtype NUDT5 (1 mM PRPP) and g. AICA-ribonucleotide in the presence and absence of NUDT5 wildtype and indicated mutants (0.25 mM PRPP). Data in enzyme activity assays in panels d-g show the individual values (d) or mean (e-g) and error bars are SEM from n=3 independent experiments. h. Western blot of in vitro PPAT Strep immunoprecipitation of recombinant NUDT5 in the presence and absence of AMP. The dilution of NUDT5 in the input is from 10 μM to 2.4 nM (4-fold serial dilution). Data shown are representative from two independent experiments.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Western Blot, Expressing, Staining, Purification, Activity Assay, In Vitro, Immunoprecipitation, Recombinant, Serial Dilution

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Left – Molecular structures of 6-meTIMP, 6-benzylthioinosine-5’-monophosphate (6-benzylTIMP), and 6-ethylthioinosine-5’-monophosphate (6-etTIMP). Right – PPAT activity assay measuring nucleotide-dependent inhibition in the presence and absence of NUDT5. Data points are the mean and error bars are SEM from n=3 independent experiments. b. Structural alignment of the region surrounding the molecular glue interface for the 6-benzylTIMP- and 6-meTIMP-bound PPAT-NUDT5 structures. Cryo-EM density for the 6-benzylTIMP nucleotide is shown as a transparent surface. c. Sharpened cryo-EM density of the I422-E436 loop in the 6-benzylTIMP and 6-meTIMP maps shown at the same contour. The dashed oval indicates a region of missing density in the 6-benzylTIMP map that is well-defined in the 6-meTIMP map. d. FACS-based growth competition experiment comparing ΔNUDT5 and endogenous NUDT5 L217A/K218A (LKAA) mutants to wildtype HEK293T treated with 6-ethylmercaptopurine riboside. Data show n=3 biological replicates from a representative experiment. Similar results were obtained in two independent experiments. e. Model of NUDT5-and purine-dependent molecular glue mechanism outlining effects on inhibition of de novo purine biosynthesis.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Activity Assay, Inhibition, Cryo-EM Sample Prep

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Simplified metabolism of 6-mercaptopurine (6-MP) and 6-thioguanine (6-TG). Asterisk denotes ability of 6-TGMP to be transformed into 6-meTGMP that may inhibit de novo purine synthesis. b. FACS-based growth competition comparing ΔNUDT5 and mutants to wildtype HEK293T cells treated with 6-TG. Data are individual values from n=3 biological replicates from a representative experiment. Similar results were obtained in two independent experiments. c. Chemical structures of adenosine-5’-monophosphate (AMP) and 6-methylthioinosine-5’-monophosphate (6-meTIMP). d. Left – alignment of molecular glue interface of AMP and 6-meTIMP showing cryo-EM density for the nucleotides. Right – rearrangement of PPAT interface residues in the 6-meTIMP structure (dark sidechains) compared to the AMP-bounds structure (light sidechains) e. Hydrophobic pocket of PPAT engaged by 6-meTIMP. f. 2D-ligand diagram of the 6-meTIMP molecular glue interface. g. PPAT activity assay measuring nucleotide-dependent inhibition in the presence of NUDT5 with 0.25 mM PRPP. Data points are the mean and error bars are SEM from n=3 independent experiments. h. Left – Western blot of endogenous NUDT5 3xFLAG immunoprecipitations following 16-hour treatment with methotrexate (2 µM), 6-MP (50 µM), and MTX + 6-MP. Right – Quantification of PPAT immunoprecipitation normalized to NUDT5 3xFLAG bait and compared to a DMSO-treated control condition. Data are individual values from n=3 independent biological replicate experiments and error bars are SEM. i. Fractional enrichment of AMP (M+2) and GMP (M+3) isotopologs in [ 15 N-amide]-glutamine labeling experiments conducted in the presence of 6-MP. Data points are individual values of n=6 biological replicates from two independent experiments and error bars are SEM. Statistical comparisons were performed using Welch’s two-tailed t-test with Bonferroni correction between wildtype and each mutant. *** denotes a Bonferroni adjusted p-value < 0.001 and ** is p-value < 0.01. j. FACS-based growth competition experiment comparing growth of ΔNUDT5 and endogenous L217A/K218A (LKAA) NUDT5 mutants to wildtype HEK293T treated with 6-MP and 6-TG. Data show n=3 biological replicates from a representative experiment. Similar results were obtained in two independent experiments.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Transformation Assay, Cryo-EM Sample Prep, Activity Assay, Inhibition, Western Blot, Immunoprecipitation, Control, Labeling, Two Tailed Test, Mutagenesis

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Example cryo-EM density of the PPAT-NUDT5 6-meTIMP molecular glue interface with model fit. b,c. PPAT activity assay measuring inhibitory effects of 6-meTIMP in the presence and absence of wildtype NUDT5 and indicated mutants and c. compared to AMP only. d. Left – Representative Western blot of immunoprecipitations from endogenous NUDT5 3xFLAG HEK293T cells treated with indicated drugs for 16 hours: methotrexate (MTX; 2 µM), lometrexol (LMX; 10 µM), 6-mercaptopurine (6-MP; 50 µM), MLN4924 (1 µM), brequinar (2 µM), and rapamycin (1 µM). Right – quantification of PPAT immunoprecipitation relative to NUDT5 3xFLAG bait and normalized to a DMSO-treated control condition. Data are individual values from n=3 biological replicates from independent experiments and error bars are SEM. e. Western blot of immunoprecipitations from endogenous NUDT5 3xFLAG HEK293T cells treated with MTX (2 µM) for the indicated amounts of time. Similar results were obtained in two independent experiments. f. Western blot of endogenous PPAT 3xFLAG immunoprecipitations following 16-hour treatment with MTX (2 µM), 6-MP (50 µM), and MTX + 6-MP g. Time-resolved microscopy (incucyte) growth assays of wildtype and mutant HEK293T cells treated with the indicated drugs. Data are the mean and error bars are SEM of n=6 biological replicates. h. Levels of intracellular 6-TIMP and 6-meTIMP metabolites following 16-hour treatment with 6-MP (20 µM). Data are individual values and error bars are SEM from n=3 biological replicates. i. PPAT activity assay measuring inhibitory effects of 6-meTGMP in the presence and absence of wildtype NUDT5 and indicated mutants. Activity data shown in panels b, c and i are the mean and error bars are SEM of n=3 independent experiments.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Cryo-EM Sample Prep, Activity Assay, Western Blot, Immunoprecipitation, Control, Microscopy, Mutagenesis

Journal: bioRxiv

Article Title: Metabolic glues as a means of purine sensing and chemotherapeutic response

doi: 10.64898/2026.05.05.723063

Figure Lengend Snippet: a. Example high-resolution cryo-EM density of the PPAT-NUDT5 6-benzylTIMP molecular glue interface with fit model. b. I422-E436 and A95-E101 loops of PPAT in the 6-benzylTIMP complex modeled into unsharpened cryo-EM density map and shown at a low contour threshold. c. PPAT activity assay measuring inhibitory effects of 6-etTIMP in the presence and absence of wildtype NUDT5 and indicated mutants. Data are the mean and error bars are SEM of n=3 independent experiments.

Article Snippet: For the PPAT V5-TwinStrep and NUDT5 3xFLAG sequential IP, supernatant was first bound to Streptactin XT 4Flow resin (IBA, 2-5010-010), washed 3x in lysis buffer, and eluted in lysis buffer containing 50 mM biotin (2-1016-005).

Techniques: Cryo-EM Sample Prep, Activity Assay