Fbxo4-mediated degradation of Fxr1 suppresses tumorigenesis in head and neck squamous cell carcinoma

The Fbxo4 tumour suppressor is a component of an Skp1-Cul1-F-box E3 ligase for which two substrates are known. Here we show purification of SCFFbxo4 complexes results in the identification of fragile X protein family (FMRP, Fxr1 and Fxr2) as binding partners. Biochemical and functional analyses reveal that Fxr1 is a direct substrate of SCFFbxo4. Consistent with a substrate relationship, Fxr1 is overexpressed in Fbxo4 knockout cells, tissues and in human cancer cells, harbouring inactivating Fbxo4 mutations. Critically, in head and neck squamous cell carcinoma, Fxr1 overexpression correlates with reduced Fbxo4 levels in the absence of mutations or loss of mRNA, suggesting the potential for feedback regulation. Direct analysis reveals that Fbxo4 translation is attenuated by Fxr1, indicating the existence of a feedback loop that contributes to Fxr1 overexpression and the loss of Fbxo4. Ultimately, the consequence of Fxr1 overexpression is the bypass of senescence and neoplastic progression.


Supplementary Figure 3. Prediction of the binding interface between Fbxo4 and
Fxr1 using bimolecular docking.
The Fbxo4 X-ray and Fxr1 homology model were analysed for the best possible dimer complex using ClusPro. (a) Superposition of the top 10 models. (b-k) Representative depiction of each mode l as individual dimer and in order from highest to lowest cluster size and weighted prediction score. In all images, Fbxo4 is depicted in black in the same orientation with Fxr1 in alternating colour for each model. Models 1 (b), 5 (f), and 9 (j) are highly similar in all dimensions, while models 1 (b), 4 (e), 5 (f), 7 (h), 9 (j) and 10 (k) use the same interface and only rotate about one axis. Model 1 (b) has the closest proximity of Fxr1 178-192 to Fbxo4 and this domain is circulated in turquoise oval.
Model 2 (c), 4 (e), 6 (g), and 9 (j) indicate the pose driven by interactions with the termini of Fxr1 and are potential artifacts of molecular docking process.

Transfection and infection
For transfection: the transient expression of genes was achieved using lipofectamine with plus reagents according to the instructions from the manufacturer. In addition, PolyJet™ in vitro DNA transfection reagent was also used for DNA delivery.
Virus production for mammalian cells in HEK293T cells: to overexpress proteins in mouse cells, pMX vectors with interested genes were packaged in ψ2 retroviral vector; while for human cells, pMX vectors with interested genes were packaged in Qψ retroviral vector. For lentiviral packaging , Fxr1 overexpression, Fxr1, Fbxo4, p21 and p27 knockdown, lentiviral backbone vectors were co-transfected with pMDLg/pRRE, CMV-VSVG, and RSV-Rev vectors. Generally, supernatant was collected 48 and 72 hours post-transfection. Infection was performed using 8 μg/ml polybrene if necessary, thereafter, cells were exposed to relative treatment.
Virus production in Sf9 cells: transfection was performed using the recombinant Bacimid with interested genes. Three to five days post-transfection, supernatant was collected for virus amplification (last for another 7 days). High titer viruses were utilized for Sf9 cell infection and protein purification.

MS analyses
Fbxo4 -/-MEFs w/o WT and ΔF Fbxo4 were exposed to medium containing 20 μM MG-132 for 6 hours. Cell pellets were lysed in Tween 20 lysis buffer (50 mM HEPES pH8.0, 150 mM NaCl, 2.5 mM EGTA, 1 mM EDTA, and 0.1% Tween 20 with protease and phosphatase inhibitors). After centrifugation, the supernatant was applied for immunoprecipitation with anti-FLAG® M2 affinity gel (A2220, Sigma-Aldrich) overnight. The following day, beads were washed and pulled-down proteins were eluted out using Flag peptides (F3290, Sigma-Aldrich). Eluted proteins were separated by SDS-PAGE gel and visualized with P ierce™ Silver Stain for Mass Spectrometry (24600, Thermo Fisher Scientific Inc). Excised bands were distained and sent out for LC-MS/MS in Taplin Biological Mass Spectrometry Facility at the Harvard Medical School. The data were analysed and proteins with at least two unique peptides were retained for further analysis.
For cBioPortal data mining, both Fbxo4 and Fxr1 genes were searched at the same time for the comparison of sequencing results as well as Copy Number Aberrations (CNAs) on the website: http://www.cbioportal.org/.

qRT-PCR
Upon treatment, total RNAs were extracted using RNeasy Mini Kit (Qiagen). One to two µg of total RNAs were reverse transcribed using SuperScript III Reverse Transcriptase according to the manufacturer's manual (Invitrogen). The cDNAs were diluted and run for quantitative analysis using predesigned and validated TaqMan probes with protocol indicated by Applied Biosystems. β-actin was used to normalize the RNA loading. Relative quantification was calculated using CFX Manager Software (Bio-Rad Laboratories, Inc.

CHX chase assay and quantification
Con shRNA and Fbxo4 shRNA infected 74B cells were treated with 40 μg/ml CHX for indicated time. Whole cell lysate was collected for Western blot analysis. Band quantification was performed using Quantity One (Bio-Rad Laboratories, Inc.). Signals were normalized to β-actin.

In vivo and in vitro ubiquitylation assay
For in vivo ubiquitylation assay, HEK293T cells were transiently transfected with myc-Fxr1 as well as WT or mutant Fbxo4 in the presence of HA-Ub or His-Ub w/o GSK3β. Twenty-four hours post-transfection, cells were collected, and boiled in 1% SDS at 95 °C for 10 min. Then, the reaction is quenched by 10% Triton X-100, finally, lysed in buffer (50 mM Tris pH7.5, 250 mM NaCl, 0.1% Triton X-100, 1 mM EDTA and 1 mM DTT) with protease and phosphatase inhibitors plus 5 mM NEM, and 20 μM MG132. After denaturing lysing, supernatant was incubated with anti-Myc agarose affinity gel or anti-His affinity resin. Thereafter, beads were washed and boiled; and SDS-PAGE gel was utilized to resolve the ubiquitylated proteins.
Purified proteins were added in Eppendorf tube w/o E1, E2 (UbcH5a), ATP, and ubiquitin for 60 min at 37°C. After reaction, proteins were resolved on 10% SDS-PAGE gel and detected by relative antibodies.

Proliferation assay and soft agar assay
HNSCC cells with Con and Fbxo4 shRNAs w/o Fxr1 knockdown were seeded in 24-well plate at the number of 2x10 3 . Six, 24, 72 and 120 hours after seeding, cells were washed, then fixed with 4% paraformaldehyde in PBS for 15 min. Cells were washed and stained with 0.1% crystal violet for 20 min. Thereafter, staining solution was sucked out and three times wash was performed using water. After final wash, the plates were airdried. Extraction was done using 10% acetic acid with shaking for 20 min. The absorbance was measured at 590 nM. Signals from cells with six-hour seeding were set up as Day 0.

Luciferase reporter assay
The 74B cells with Fxr1 knockdown and HEK293T cells with Fxr1 overexpression were used for luciferase assay. Different segments of human Fbxo4 3'-UTRs were systematically identified and cloned into a luciferase reporter vector system, pLightswitch-3'-UTR from Switchgear Genomics. Empty vector, or GAPDH (negative control) and p21 (positive control) were utilized as controls. Cells were incubated at 37°C prior to analysis. Twenty-four hours post-transfection, cells were lysed in LightSwitch Luciferase Assay kit following manufacturer's protocol. Luminescence signal was analysed via 20/20 n Luminometer (Turner Biosystems). The data were normalized to GAPDH 3'-UTR and 3'-UTR empty vector using LightSwitch normalization protocol (http://switchgeargenomics.com/sites/default/files/pdf/LightSwitch_3UTRnorm.pdf).