Rapid and direct control of target protein levels with VHL-recruiting dTAG molecules

Chemical biology strategies for directly perturbing protein homeostasis including the degradation tag (dTAG) system provide temporal advantages over genetic approaches and improved selectivity over small molecule inhibitors. We describe dTAGV-1, an exclusively selective VHL-recruiting dTAG molecule, to rapidly degrade FKBP12F36V-tagged proteins. dTAGV-1 overcomes a limitation of previously reported CRBN-recruiting dTAG molecules to degrade recalcitrant oncogenes, supports combination degrader studies and facilitates investigations of protein function in cells and mice.

M odulating protein abundance with small molecule degraders is a powerful approach for investigating functional consequences of rapid and direct protein loss, without alteration of corresponding mRNA levels. Degraders including heterobifunctional degraders (also known as PROteolysis-TArgeting Chimeras or PROTACs) and nonchimeric molecular glues, co-opt an E3 ubiquitin ligase to induce rapid and reversible proteasome-mediated degradation 1 . Achieving immediate target protein loss with degraders provides a crucial advantage over genetic knockout or knockdown approaches, which require a significant delay to achieve an impactful protein reduction 2 . However, degrader development is hindered by a reliance on target-specific chemical matter, which is unavailable for the majority of the proteome. To address this challenge, several strategies aimed at the direct control of cellular protein levels have been recently developed, including methods that use small molecules [3][4][5][6][7][8][9] , nanobodies 10 , or antibodies 11 .
We previously described a versatile approach known as the degradation tag (dTAG) system to rapidly deplete any tagged target protein in cells and in mice 6 . The dTAG system is a dual component platform requiring the expression of FKBP12 F36V inframe with a gene-of-interest and treatment with a heterobifunctional dTAG molecule (dTAG-13) 12 that engages FKBP12 F36V and cereblon (CRBN), an E3 ubiquitin ligase (Supplementary Fig. 1a, b). This interaction leads to exclusive degradation of the FKBP12 F36V -tagged protein. Studies degrading diverse targets including oncoproteins, transcription factors, chromatin regulators, and kinases illustrate the utility of the dTAG system for drug target validation and discovery 6,[12][13][14][15][16][17][18][19][20][21][22][23] . Despite this broad applicability, we observed context-specific and protein-specific differences in the effectiveness of dTAG-13 for inducing target protein degradation.
Here, we report the synthesis, characterization and utility of a second generation, in vivo-compatible dTAG molecule that recruits the von Hippel-Lindau (VHL) E3 ligase complex, dTAG V -1 (Fig. 1a, b). We demonstrate that dTAG V -1 degrades fusion proteins recalcitrant to CRBN-mediated degradation, exemplified by EWS/FLI, a driver of Ewing sarcoma. Collectively, this study describes an important extension to the dTAG platform, towards a universally applicable strategy for direct protein control.
We next evaluated the utility of dTAG V -1 for combinatorial degrader studies through co-treatment experiments with THAL-SNS-032, a previously reported CRBN-recruiting CDK9 degrader 27 . dTAG V -1 alone did not alter CDK9 protein levels (Fig. 1e,  lane 7), which was confirmed by mass spectrometry-based proteomics analysis (Supplementary Data 2). dTAG V -1 effectively combined with THAL-SNS-032 and co-treatment led to pronounced degradation of both LACZ-FKBP12 F36V and CDK9 at levels comparable to those with treatment of each degrader alone (Fig. 1e, lanes 8-12), avoiding potential substrate competition effects 28 . To confirm the utility of dTAG V -1 for target-specific validation studies, we evaluated KRAS G12V degradation, an oncogenic driver of PDAC, in PATU-8902 FKBP12 F36V -KRAS G12V ; KRAS −/− cells 15 . dTAG V -1 treatment led to rapid KRAS G12V degradation, which was not induced by dTAG V -1-NEG treatment ( Fig. 2a and Supplementary Fig. 2a). KRAS G12V degradation was rescued by pre-treatment with proteasomeinhibitor carfilzomib or Nedd8 activating enzyme inhibitor MLN4924, and by VHL knockout, consistent with the mechanism-of-action of a VHL-recruiting degrader ( Fig. 2b and Supplementary Fig. 2b). We also observed the expected collapse in cellular signaling and diminished cell proliferation in 2Dmonolayer and 3D-spheroid cultures upon FKBP12 F36V -KRAS G12V degradation with dTAG V -1 treatment, to levels comparable to CRBN-recruiting dTAG molecules ( Fig. 2c and Supplementary Fig. 2c-e).
To confirm the in vivo applicability of dTAG V -1, we characterized the pharmacokinetic (PK) and pharmacodynamic (PD) profile of dTAG V -1 in mice. dTAG V -1 demonstrated improved properties compared to dTAG-13, with a longer half-life (T 1/2 = 4.43, 2.41 h respectively) and greater exposure (AUC inf = 18,517, 6140 h ng mL −1 , respectively) by intraperitoneal administration at 10 mg kg −1 (Supplementary Table 1). To report on the PD profile of dTAG molecules, we employed MV4;11 luciferase-FKBP12 F36V (luc-FKBP12 F36V ) cells that allow noninvasive monitoring of bioluminescent signal upon dTAG molecule administration in mice 6 . Following tail vein injection of MV4;11 luc-FKBP12 F36V cells and establishment of leukemic burden, we performed daily bioluminescent measurements 4 h after vehicle, 35 mg kg −1 dTAG-13 or 35 mg kg −1 dTAG V -1 administration. Striking loss of bioluminescent signal was achieved 4 h after the first administration of dTAG V -1 ( Fig. 2d and Supplementary  Fig. 3). Consistent loss of bioluminescent signal was observed 4 h after each of the three dTAG-13 or dTAG V -1 administrations. Compared to dTAG-13, improved duration of degradation was also observed with dTAG V -1, with degradation evident 28 h after the final administration. These results support the use of dTAG V -1 as a potent and selective molecule to evaluate target-specific degradation in vitro and in vivo.
dTAG V -1 enables EWS/FLI degradation in Ewing sarcoma. There is long-standing interest in the oncology community in identifying targetable dependencies in Ewing sarcoma. Ewing sarcoma is driven by the translocation of EWSR1 and members of the ETS transcription factors, most commonly FLI, giving rise to a fusion transcription factor oncoprotein that activates an aberrant transcriptional program 29 . Currently, there are a shortage of model systems and direct-acting agents that allow modulation of EWS/FLI levels or activity with precise kinetic control. To evaluate the effects of EWS/FLI degradation in Ewing sarcoma, we selected EWS502 cells, which are highly dependent on EWS/FLI for proliferation ( Supplementary Fig. 4a). Through simultaneous expression of FKBP12 F36V -EWS/FLI and Cas9/ sgFLI, we developed FKBP12 F36V -EWS/FLI; EWS/FLI −/− cells ( Supplementary Fig. 4b). FKBP12 F36V -GFP was expressed in EWS502 cells as a control and was effectively degraded upon treatment with dTAG-13 and dTAG V -1, indicating the ability to co-opt both CRBN and VHL for targeted degradation in this cell line (Fig. 3a, b). The characteristic "hook effect" behavior 30 , which is commonly observed with heterobifunctional degraders 6 , was also evident at high doses (5000 nM) of dTAG molecule treatment due to saturation of FKBP12 F36V and E3 ubiquitin ligase binding sites. However, FKBP12 F36V -EWS/FLI was only susceptible to degradation upon treatment with dTAG V -1 in EWS502 cells, which was evident within 1 h of treatment, highlighting the necessity of comparative evaluation of both CRBN-recruiting and VHL-recruiting dTAG molecules (Fig. 3a, b and Supplementary Fig. 4c). To confirm FKBP12 F36V -EWS/FLI degradation and identify immediate downstream EWS/FLI targets, we performed acute quantitative mass spectrometry-based proteomics and observed pronounced loss of FKBP12 F36V -EWS/ FLI upon dTAG V -1 treatment ( Fig. 3c and Supplementary Data 1, 2). dTAG V -1 led to a decrease in downstream EWS/FLI targets such as NKX2-2, and Gene Set Enrichment Analysis (GSEA) of the differentially regulated target proteins (FDR < 0.05) identified known EWS/FLI and Ewing sarcoma signatures ( Fig. 3c-e). Potent antiproliferative effects were also observed upon EWS/FLI degradation at longer time-points, which were absent with dTAG V -1-NEG treatment and in FKBP12 F36V -GFP cell lines ( Fig. 3f and Supplementary Fig. 4d).
To evaluate the dependence of VHL-recruiting dTAG molecule activity on linker length and composition, we synthesized and evaluated dTAG-63, an analog of dTAG V -1. dTAG-63 is composed of a polyethylene glycol linker, (PEG) 3 , rather than an all carbon linker, between the FKBP12 F36V and E3 ubiquitin ligase binders ( Supplementary Fig. 5a), and is a matched molecular pair to our previously disclosed first-generation CRBN-recruiting compound, dTAG-7 6  in NKX2-2 protein levels and modest antiproliferative effects relative to dTAG V -1 ( Supplementary Fig. 5b, c). Similar trends in antiproliferation upon FKBP12 F36V -KRAS G12V degradation were observed upon treatment of PATU-8902 FKBP12 F36V -KRAS G12V ; KRAS −/− cells with dTAG-63 ( Supplementary  Fig. 5d). These results are consistent with our prior observation that dTAG-7 is less effective across multiple cell lineages than dTAG-13, a CRBN-recruiting dTAG molecule with an all carbon linker 6 .
Finally, while we observed pronounced antiproliferative activity upon degradation of EWS/FLI with dTAG V -1, our data indicates that combination approaches may be necessary to achieve complete loss of viability. Prior work indicates that BET bromodomain inhibitors and degraders may have applications in Ewing Sarcoma 31 . To investigate potential synergy between direct and indirect repression of the EWS/FLI transcriptional program, we evaluated degradation of EWS/FLI in combination with BET bromodomain degradation. We observed that dBET6, a CRBN-recruiting, pan-BET bromodomain degrader 32 , synergized strongly with VHLmediated EWS/FLI degradation ( Supplementary Fig. 6a-c).
Together, this data exemplifies the utility of VHL-recruiting dTAG molecules and provides model systems to evaluate the acute and prolonged consequences of EWS/FLI loss.

Discussion
We report dTAG V -1, a potent and exclusively selective VHLrecruiting degrader of FKBP12 F36V -tagged proteins. dTAG V -1 displays improved PK/PD properties and serves as an optimized tool for in vivo applications. Through evaluation of mutant KRAS degradation in PDAC models, we show that either CRBN or VHL can be co-opted to alleviate the aberrant signaling coordinated by this oncoprotein. By contrast, we observed contextual differences in the ability of these E3 ubiquitin ligase complexes to degrade EWS/FLI. This is consistent with our recent report demonstrating effective degradation of a core mediator subunit (MED14) with dTAG V -1 in HCT116 cells, a context in which CRBN-recruiting dTAG molecules were not effective 22 . We observed that rapid MED14 degradation abrogated lineage-specifying transcriptional circuits. Together, our studies provide support for use  of dTAG V -1 to overcome the current limitations of the dTAG system, enabling evaluation of the direct effects of fusion proteins recalcitrant to CRBN-recruiting dTAG molecules. Employing dTAG V -1 to study EWS/FLI, we demonstrate that VHL-mediated degradation of EWS/FLI rapidly alters downstream target protein expression and leads to pronounced growth defects in Ewing sarcoma cells, providing a powerful model system to investigate immediate consequences of EWS/FLI loss. This data supports that targeting EWS/FLI for degradation with directacting heterobifunctional degraders or molecular glues may be a tractable strategy and identifies potential combination strategies with BET bromodomain degraders. Together, the suite of dTAG molecules and paired controls provided in this study will facilitate evaluation of the functional consequences of precise posttranslational protein removal for an expanded target pool. The dTAG system enables rapid modulation of protein abundance and serves as a versatile strategy to determine whether targeted degradation is a promising drug development approach for a given target in vitro and in vivo.  Lentiviral dTAG plasmid construction. To generate pLEX_305-dTAG-GFP and pLEX_305-dTAG-EWS/FLI plasmids, gateway recombination cloning strategies (Invitrogen) were employed to clone GFP or EWS/FLI into pLEX_305-N-dTAG 6 . In brief, to first generate pDONR221-EWS/FLI, EWS/FLI was cloned into pDONR221 using BP clonase (Invitrogen) after PCR with the following primers containing BP overhangs: Forward-N-E/F-dTAG, 5′-ggggacaagtttgtacaaaaaagcaggcttcgcgtccacgga ttacagtacct-3′ and Reverse-N-E/F-dTAG, 5′-ggggaccactttgtacaagaaagctgggtcctagtagta gctgcctaagtgtgaaggc-3′. Second, pENTREGFP2 (Addgene #22450) and pDONR221-EWS/FLI were cloned into pLEX_305-N-dTAG using LR clonase (Invitrogen). pLEX_305-dTAG-KRAS G12V and pLEX_305-LACZ-dTAG plasmids were also employed in this study 6,15 .

Methods
Lentiviral CRISPR/Cas9 plasmid construction. pXPR007-sgGFP and pXPR007-sgKRAS were employed in this study 15 . To generate lentiCRISPR v2-Blast-sgFLI_Ex9, sgFLI_Ex9 (5′-GCCTCACGGCGTGCAGGAAG-3′) was cloned into lentiCRISPR v2-Blast vector (Addgene #83480) using BsmbI restriction sites. The PAM motif of sgFLI_Ex9 is present in an intron, enabling cutting of the endogenous locus only with cDNA rescue. FKBP12 WT and FKBP12 F36V dual luciferase assay. Dual luciferase assays were performed using 293FT FKBP12 WT -Nluc and FKBP12 F36V -Nluc cells 6 . In brief, cells were plated at 2000 cells per well in 20 µL of appropriate media in 384-well white culture plates (Corning), allowed to adhere overnight, and 100 nL of compounds were added using a Janus Workstation pin tool (PerkinElmer) for 24 h at 37°C. To evaluate Fluc signal, plates were brought to room temperature, 20 µL of Dual-Glo Reagent (Promega) was added for 10 min and luminescence was measured on an Envision 2104 plate reader (PerkinElmer). Subsequently, 20 µL of Dual-Glo Stop & Glo Reagent (Promega) was added for 10 min and luminescence was again measured to capture Nluc signal. DMSO-normalized ratios of Nluc/Fluc signal was analyzed and plotted using GraphPad PRISM v8 6 .
Analysis of cell viability. Cell viability was assayed in 2D-adherent or ultra-low adherent 3D-spheroids using CellTiter-Glo (Promega) 6,15 . Luminescence was measured on an Envision 2104 plate reader (PerkinElmer) and Fluostar Omega Reader (BMG Labtech) and data was analyzed using GraphPad PRISM v8. Synergy assessments were performed using CellTiter-Glo (Promega) with the following modifications to the protocol described in the ref. 34 . In brief, EWS502 cells were plated at 1000 cells per well in 50 µL of appropriate media in 384-well white culture plates (Corning) allowed to adhere overnight, and 100 nL of compounds were added using a Janus Workstation pin tool (PerkinElmer) for 72 h. Cell viability was measured by addition of 10 µL of CellTiter-glo (Promega), followed by incubation for 15 minutes at room temperature. Luminescence was measured on an Envision 2104 plate reader (PerkinElmer) and data was analyzed using GraphPad PRISM v8.
Quantitative proteomics: MS sample processing. Cell pellets from PATU-8902 and EWS502 cells were lysed using 8 M urea, 200 mM 4-(2-hydroxyethyl)-1piperazinepropanesulfonic acid (EPPS) at pH 8.5 with protease inhibitors. Samples were further homogenized and DNA was sheered via sonication using a probe sonicator (20 × 0.5 s pulses; level 3). Total protein was determined using a BCA assay and stored at −80°C until future use. A total of 100 µg of protein was aliquoted for each condition and TMT channel for further downstream processing. Protein extracts were reduced using 10 mM dithiothreitol (DTT) for 30 min at room temperature. Next, samples were alkylated with 20 mM iodoacetamide for 45 min in the dark at room temperature. To facilitate the removal of incompatible reagents, proteins were precipitated using chloroform methanol. Briefly, to 100 µL of each sample, 400 µL of methanol was added, followed by 100 µL of chloroform with thorough vortexing. Next, 300 µL of HPLC grade water was added and samples were vortexed thoroughly. Each sample was centrifuged at 14,000 × g for 5 min at room temperature. The upper aqueous layer was removed and the protein pellet was washed twice with methanol and centrifuged at 14,000 × g for 5 min at room temperature. Protein pellets were resolubilized in 200 mM EPPS buffer and digested overnight with Lys-C (1:100, enzyme:protein ratio) at room temperature. The next day, trypsin (1:100 ratio) was added and incubated at 37°C for an additional 6 h in a ThermoMixer set to 1000 RPM.
To each digested sample, 30% anhydrous acetonitrile was added and 100 µg of peptides were labeled using 200 µg of TMT reagent (TMT1-TMT11). Following labeling, a 5% hydroxylamine solution was used to quench excess TMT reagent. To equalize protein loading, a ratio check was performed by pooling 2 µg of each TMT-labeled sample. Samples were pooled and desalted using in-house packed C18 StageTips and analyzed by LC-MS/MS. Normalization factors were calculated from this label check, samples were mixed 1:1 across all TMT channels and desalted using a 100 mg Sep-Pak solid phase extraction cartridge. Eluted pooled peptides were further fractionated with basic-pH reverse-phase (bRP) HPLC using an Agilent 300 extend C18 column and collected into a 96 deep-well plate. Samples were consolidated into 24 fractions, and 12 nonadjacent fraction were desalted using StageTips prior to analyses using LC-MS/MS [35][36][37] .
Quantitative proteomics: MS data acquisition. All mass spectrometry data was acquired using an Orbitrap Fusion mass spectrometer in-line with a Proxeon NanoLC-1000 UHPLC system. Peptides were separated using an in-house 100 µm capillary column packed with 40 cm of Accucore 150 resin (2.6 µm, 150 Å) (Thermo Fisher Scientific) using a 180 min LC gradient per fraction. Eluted peptides were acquired using synchronous precursor selection (SPS-MS3) method for TMT quantification 38 . Briefly, MS1 spectra were acquired at 120 K resolving power with a maximum of 50 ms in the Orbitrap. MS2 spectra were acquired by selecting the top ten most abundant features via collisional induced dissociation (CID) in the ion trap using an automatic gain control (AGC) of 15 K, quadrupole isolation width of 0.7 m/z and a maximum ion time of 100 ms. For MS3 acquisition, a synchronous precursor selection of ten fragment ions was acquired with an AGC of 150 K for 150 ms and a normalized collision energy of 55.
Quantitative proteomics: MS data analysis. All acquired raw data were converted to mzXML using Raw File Reader (v3.0.77). Spectra were processed using Comet (2019.01.5) 39-41 , search results were filtered using the LDA function MASS Package in R, and data were processed using an in-house informatics pipeline [42][43][44] . Custom code was not developed in this study. Briefly, peptide spectral libraries were first filtered to a peptide false discovery rate (FDR) of less than 1% using linear discriminant analysis employing a target decoy strategy. Spectral searches were performed using a 2020 Uniprot Human database including canonical isoforms (96,788 total entries) fasta formatted database which included custom sequences for LACZ-FKBP12 F36V and FKBP12 F36V -EWS/FLI, common contaminants, reversed sequences (Uniprot Human, 2020) and the following parameters: 50 ppm precursor tolerance, fully tryptic peptides, fragment ion tolerance of 0.9 Da and a static modification of TMT (+229.163 Da) on lysine and peptide N-termini, carbamidomethylation of cysteine residues (+57.021 Da) were set as static modifications, while oxidation of methionine residues (+15.995 Da) was set as a variable modification. Resulting peptides were further filtered to obtain a 1% protein FDR and proteins were collapsed into groups. Reporter ion intensities were adjusted to correct for impurities during synthesis of different TMT reagents according to the manufacturer's specifications. For quantitation, a total sum signalto-noise of all report ions of 100 was required for analysis, isolation specificity at the MS2 and MS3 level >0.50 and no missing values (11/11 TMT channels). Protein quantitative values were normalized (column normalization) so that the sum of the signal for all protein in each channel was equal to account for sample loading. Mass spectrometry-based proteomics raw data files have been deposited to the ProteomeXchange Consortium via the PRIDE 45,46 partner repository (PXD018937). Gene Set Enrichment Analysis (GSEA) 47 was performed as indicated in the figure legends.
Animal studies: pharmacodynamic (PD) evaluation. All procedures for PD studies were compliant with ethical regulations for animal testing and research, and were approved by and performed in accordance with standards of the IACUC at Dana-Farber Cancer Institute. Housing conditions for mice at Dana-Farber Cancer Institute were as follows: lights on = 6 AM, lights off = 6 PM, average ambient temperature = 72°F, and average humidity = 35-55%. Evaluation of degradation of luciferase in vivo was performed using MV4;11 luc-FKBP12 F36V cells and bioluminescent measurements 6,48 . In brief, 250,000 viable MV4;11 luc-FKBP12 F36V cells were transplanted by tail-vein injection in 8-week-old immunocompromised female mice (NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ, NSG; Jackson Laboratory, #005557). Bioluminescent measurements were performed following IP injection of 75 mg kg −1 D-Luciferin (Promega). Mice were then anesthetized with 2−3% isoflurane, imaged on an IVIS Spectrum (Caliper Life Sciences), and total body bioluminescence was determined using a standardized region of interest. Two and seven days after tail-vein injection, bioluminescent measurements were performed to confirm engraftment 48 . Detectable bioluminescent signal in the cohort was observed at day 7 and was used as the baseline signal for each mouse. For treatments, compounds were formulated as described above, administered via IP injection and bioluminescent measurements were performed daily as described in Fig. 2d, beginning 8 days after tail-vein injection.
Statistical analysis. Information regarding center values, error bars, number of replicates or samples, number of independent experiments, and statistical analyses are described in the corresponding figure legends and data files. Experiments were not blinded nor randomized, and sample sizes were not predetermined using statistical analyses.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Mass spectrometry-based proteomics raw data files are provided in Supplementary Data 1 and have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD018937. Mass spectrometry-based proteomics processed data files underlying Figs. 1d and 3c and Supplementary Fig. 1f  Reagents are available upon request, following completion of a material transfer agreement, or at: http://graylab.dana-farber.org/probes.html. Source data are provided with this paper.