Abstract

   β‐catenin signaling is abnormally activated in cancer. Here, this work
   screens the mevalonate metabolic pathway enzyme PMVK to stabilize
   β‐catenin signaling using a human genome‐wide library. On the one hand,
   PMVK‐produced MVA‐5PP competitively binds to CKIα to prevent β‐catenin
   Ser45 phosphorylation and degradation. On the other hand, PMVK
   functions as a protein kinase to directly phosphorylate β‐catenin
   Ser184 to increase its protein nuclear localization. This synergistic
   effect of PMVK and MVA‐5PP together promotes β‐catenin signaling. In
   addition, PMVK deletion impairs mouse embryonic development and causes
   embryonic lethal. PMVK deficiency in liver tissue alleviates
   DEN/CCl[4]‐induced hepatocarcinogenesis. Finally, the small molecule
   inhibitor of PMVK, PMVKi5, is developed and PMVKi5 inhibits
   carcinogenesis of liver and colorectal tissues. These findings reveal a
   non‐canonical function of a key metabolic enzyme PMVK and a novel link
   between the mevalonate pathway and β‐catenin signaling in
   carcinogenesis providing a new target for clinical cancer therapy.

   Keywords: β‐catenin, casein kinase Iα , hepatocellular carcinoma,
   mevalonate 5‐diphosphate, phosphomevalonate kinase
     __________________________________________________________________

   The mevalonate pathway kinase PMVK stabilizes β‐catenin via the
   PMVK‐generated intermediate MVA‐5PP. MVA‐5PP blocks β‐catenin
   phosphorylation at Ser45 by inhibiting the CKIα kinase. This stabilizes
   β‐catenin by preventing its cytoplasmic sequestration and proteolysis.
   Concurrently, PMVK also phosphorylates β‐catenin at Ser184, which
   facilitates its nuclear translocation. PMVK thus represents a novel
   therapeutic target in cancers that rely on aberrant β‐catenin
   signaling.

   graphic file with name ADVS-10-2204909-g006.jpg

1. Introduction

   Hepatocellular carcinoma (HCC) is the second leading cause of
   cancer‐related deaths worldwide.^[ [46]^1 ^] Although the current
   treatment of HCC mainly includes surgical resection, tumor ablation and
   chemotherapy, long‐term survival remains dismal.^[ [47]^2 ^] Metabolic
   reprogramming, considered one of the hallmarks of cancer,^[ [48]^3 ^]
   can involve changes in lipid metabolism, which supplies membrane
   components needed to support tumor cell proliferation. The targeting of
   lipid metabolism has thus been examined as a promising treatment
   option.^[ [49]^4 ^]

   The mevalonate anabolic pathway provides metabolites for multiple
   cellular processes in eukaryotes, archaea, and bacteria.^[ [50]^5 ^]
   Mevalonate, which is produced from acetoacetyl‐CoA, is converted to
   sterol isoprenoids such as cholesterol, which is an indispensable
   precursor of bile acids, lipoproteins, and steroid hormones. Mevalonate
   is also used in the synthesis of other hydrophobic molecules including
   nonsterol isoprenoids.^[ [51]^4 , [52]^5 ^] The role of these
   mevalonate pathways in tumorigenesis and progression has been
   extensively described.^[ [53]^5 , [54]^6 ^] In addition to the final
   metabolite cholesterol, mevalonate kinase promotes tumor growth by
   stabilizing mutant P53 protein levels through the metabolite mevalonate
   5‐phosphate (MVA‐5P).^[ [55]^7 ^] Geranylgeranyl diphosphate (GGPP) and
   farnesyl pyrophosphate (FPP) can be used as geranylgeranylation and
   farnesylation substrates in post‐translational modifications of
   proteins generically known as protein prenylation. Such
   post‐translational modifications are required for the proper
   sub‐cellular targeting and activation of many oncogenic proteins,
   including some members of the RAS family.^[ [56]^8 ^]

   Phosphomevalonate kinase (PMVK) catalyzes the cation‐dependent reaction
   of MVA‐5P with ATP to form mevalonate 5‐diphosphate (MVA‐5PP) and ADP
   and represents a key step in the mevalonate pathway for
   isoprenoid/sterol biosynthesis and protein prenylation. Intermediates
   of this network play important roles in the post‐translational
   modification of numerous proteins involved in inter‐ and intra‐cellular
   signaling. Although the mevalonate pathway has been implicated in
   various aspects of cancer development and progression, the role(s) of
   PMVK remains unknown.^[ [57]^9 ^]

   The highly conserved Wnt‐β‐catenin signaling pathway regulates
   embryonic development, hepatobiliary development and homeostasis, and
   the repair response to liver injury.^[ [58]^10 ^] Abnormal
   Wnt‐β‐catenin signaling is also closely related to the development of
   many cancers.^[ [59]^11 ^] In the absence of active WNT signaling,
   cytoplasmic β‐catenin is rapidly degraded by the “APC complex” that
   includes the adenomatous polyposis coli (APC) tumor suppressor protein,
   Axin, glycogen synthase kinase (GSK)‐3β, casein kinase Iα (CKIα) and
   the E3‐ubiquitin ligase β‐TrCP. β‐catenin's phosphorylation by GSK‐3
   and CKIα within the APC Complex facilitates β‐TrCP‐mediated β‐catenin
   ubiquitination and proteasomal degradation.^[ [60]^12 ^] Genetic
   alteration in the core components of Wnt/β‐catenin signaling, most
   notably β‐catenin itself, lead to WNT‐independent signaling, which can
   drive tumorigenesis.^[ [61]^12 , [62]^13 ^] Mutations of the CTNNB1
   gene, which encodes β‐catenin occur in most cases of HCC.^[ [63]^14 ^]
   Wnt‐β‐catenin activation can increase fatty acid utilization to promote
   tumor progression and mutant forms of β‐catenin can increase glutamine
   levels in tumor cells to drive anaplerotic oxidative metabolism.^[
   [64]^15 ^] The accumulation of cholesterol leads to reduced expression
   of squalene epoxidase which disrupts the GSK3β and p53 complexes,
   ultimately activating β‐catenin signaling.^[ [65]^16 ^] However, the
   cross‐talk mechanism between lipid metabolism and Wnt‐β‐catenin pathway
   are still poorly understood.

   We describe here a genome‐wide Crispr/Cas9 screen to identify genes
   that are responsible for imparting resistance to β‐catenin inhibitors
   in Huh7 cells. This identified PMVK as an enzyme that can stabilize
   β‐catenin protein. Mechanistically, PMVK‐generated MVA‐5PP inhibits
   CKIα‐mediated β‐catenin phosphorylation at Ser45. This stabilizes the
   protein by preventing its proteolytic degradation. Additionally, PMVK
   directly phosphorylates β‐catenin at an alternate site, Ser184 to
   facilitate its nuclear localization. Finally, we describe a novel small
   molecule inhibitor of PMVK that attenuates β‐catenin signaling and
   potentially represents a novel therapeutics. Collectively, our study
   reveals a novel, phosphorylation‐dependent mechanism by which the key
   lipid‐metabolizing enzyme PMVK regulates Wnt‐β‐catenin signaling and
   the ensuing oncogenicity of this otherwise highly regulated pathway.

2. Results

2.1. Genome‐Wide CRISPR Screen Identifies Regulators of β‐Catenin Signaling
in Huh7 Cells

   To circumvent the difficulty of targeting the β‐catenin signaling
   pathway itself in HCC, we instead used a human genome‐scale CRISPR‐Cas9
   knockout (GeCKO) v2 library to identify genes that support the
   pathway.^[ [66]^17 ^] We chose the CTNNB1 gene wild‐type (WT) HCC cell
   line Huh7 to avoid the abnormal activation of β‐catenin signaling
   caused by the gene mutations. Preliminary screening showed that the
   IC[50] of Huh7 cells for compound XAV‐939, a tankyrase1/2 inhibitor
   stabilizing of both Axin1 and Axin2, which indirectly promotes the
   degradation of β‐catenin ^[ [67]^18 ^] (Figure [68]S1A, Supporting
   Information). Huh7 cells were then infected with lentiviral stocks
   encoding the GeCKO v2 library, followed by selection with puromycin.
   The infected cells were treated with DMSO or XAV‐939 for 14 days, and
   then analyzed by high‐throughput sequencing for sgRNA enrichment of
   more than twofold (Figure [69]1A). Pathway enrichment showed that gRNAs
   targeted against the Wnt‐β‐catenin signaling pathway were significantly
   enriched as expected (Figure [70]S1B, Supporting Information).
   Interestingly, protein class analysis showed the highest enrichment to
   be for “metabolite interconversion enzyme,” implying that these were
   involved in regulating β‐catenin signaling (Figure [71]S1C, Supporting
   Information). Therefore, we selected 174 significantly down‐regulated
   enriched genes (potentially stabilizing β‐catenin levels, called
   negative gene) overlapping with metabolite interconversion enzyme,
   while the screened candidate genes should also be up‐regulated in HCC
   to fit the role of an oncogene (Figure [72]1B; Figure [73]S1D and Table
   [74]S1, Supporting Information). The results showed that PMVK, a
   metabolic enzyme of the mevalonate pathway, was significantly enriched
   (Figure [75]1C,D). However, because the relationship between PMVK and
   the β‐catenin signaling was unclear, PMVK was further investigated as a
   candidate gene.

Figure 1.

   Figure 1
   [76]Open in a new tab

   CRISPR‐Cas9 library screens for PMVK stabilizing β‐catenin protein
   levels in Huh7 cells. A) Workflow of GeCKO in Huh7 cells screens to
   identify potential resistance genes to β‐catenin inhibitors. B) Venn
   diagram showing 174 negative genes (significantly depleted sgRNAs), 264
   metabolite interconversion enzymes, and 5386 oncogenes upregulated in
   HCC. See Table [77]S1, Supporting Information, for details. C) Volcano
   map showing sgRNA library screening distribution. Blue dots indicate
   significant deletion of sgRNAs (negative selection). Orange dots
   indicate significant enrichment of sgRNAs (positive selection). Red
   dots show screening of positive gene representatives. FC, Fold change.
   D) Each PMVK sgRNA was analyzed in pairs in the DMSO and XAV‐939
   groups. E) Cell viability was determined by MTT after 2 µg shRNA#1 PMVK
   plasmid transiently transfected with Huh7, DMSO or XAV‐939 (30 µ m)
   treated for 72 h. Data are shown as mean ± SD. F) Representative IHC
   images of H&E, Ki67, PMVK, total β‐catenin and β‐catenin p‐Ser184 in
   wild‐type, PMVK^+/− and PMVK^−/− E12.5 embryos. Scale bar, 1 mm or 50
   µm. The IHC staining experiment was performed twice.

2.2. Knockout of PMVK Reduces β‐Catenin Protein Levels and Blocks Embryonic
Development in Mice

   Initial experiments showed that XAV‐939 has a stronger inhibitory
   effect on PMVK knockdown Huh7 cell lines (Figure [78]1E). To further
   determine the regulatory role of PMVK on β‐catenin, we constructed PMVK
   knockout mice and found that WT and PMVK^+/− mice reproduced normally,
   but that PMVK homozygous knockout was embryonic lethal (Figure
   [79]S1E,F, Supporting Information). H&E staining showed no obvious
   difference in embryonic morphology in PMVK^+/− mice compared to WT
   mice. In addition to PMVK β‐catenin protein levels were also
   significantly reduced in both PMVK^+/− and PMVK^−/− mice
   (Figure [80]1F). Given that Wnt is the major intracellular modality
   regulating β‐catenin signaling and its importance in embryonic
   development, we analyzed the role of PMVK in Wnt‐β‐catenin signaling
   and our results showed that knockdown of PMVK blocked Wnt3a‐activated
   β‐catenin signaling (Figure [81]S1H, Supporting Information). PMVK
   knockdown upregulated some β‐catenin‐dependent Wnt proteins, which may
   have triggered a negative feedback (Figure [82]S1I, Supporting
   Information). These results first suggested that PMVK is required for
   mouse embryonic development, and that its knockdown PMVK significantly
   reduces β‐catenin protein levels.

2.3. PMVK Knockdown Suppresses β‐Catenin Signaling Pathway and Tumor Growth

   To better study how PMVK regulates β‐catenin, we generated stable PMVK
   knockdowns in CTNNB1‐WT Huh7 cells and AXIN1‐mutant Hep3B cells. Each
   of these dramatically repressed β‐catenin protein expression and
   downstream signaling in a manner that could be rescued by WT PMVK
   (Figure [83]2A). The fact that PMVK knockdown did not affect β‐catenin
   mRNA levels indicates that PMVK regulates β‐catenin protein levels
   (Figure [84]2B). Consistent with this, PMVK knockdown significantly
   reduced β‐catenin protein half‐life while reducing its nuclear
   localization (Figure [85]2C–E) and the expression of known β‐catenin
   target genes (Figure [86]2F). Conversely, overexpression of PMVK
   retarded β‐catenin degradation while promoting the nuclear localization
   (Figure [87]S2A–C, Supporting Information). Consistent with this and
   the accompanying reduction in c‐Myc and Cyclin D1 levels, PMVK
   knockdown significantly inhibited cell proliferation in a manner that
   could be rescued by PMVK re‐expression (Figure [88]2G). Similar results
   were obtained in a mouse xenograft tumor model (Figure [89]2H,I and
   Figure [90]S2D, Supporting Information). Thus, PMVK appears to
   stabilize the β‐catenin protein, which is in turn necessary to maintain
   cell proliferation and tumor growth.

Figure 2.

   Figure 2
   [91]Open in a new tab

   Knockdown of PMVK reduces β‐catenin signaling and inhibits tumor
   growth. A) β‐catenin, c‐Myc, Cyclin D1, and PMVK protein expression in
   the indicated cancer cell lines. B) mRNA expression of β‐catenin in the
   indicated cancer cell lines. C) HEK293 cells were transfected with the
   indicated vectors. Cells were treated with CHX (50 µg mL^−1) for the
   indicated time and the expression of PMVK and β‐catenin were analyzed
   by western blotting. D) The intensity of β‐catenin expression for each
   time point in (C) was quantified by densitometry, with β‐actin as a
   normalizer. E) Distribution of PMVK and β‐catenin protein levels in the
   nucleus or cytoplasm in the indicated PMVK‐knockdown or control cancer
   cell lines. β‐actin as a cytoplasm normalizer and Histone H3 as a
   nucleus normalizer. F) qRT‐PCR analysis was performed to measure the
   mRNA levels of PMVK and β‐catenin target genes in Huh7 cells. G) Cell
   proliferation assays were performed in Huh7 (up) and Hep3B (down) cells
   stably expressing the indicated plasmids. rWT, rescued wild type PMVK.
   H) Subcutaneous xenograft experiments were performed in Huh7 (up) and
   Hep3B (down) cells stably expressing the indicated plasmids. n = 5 mice
   per group. I) Subcutaneous xenograft tumor weight for (H). Data are
   shown as mean ± SD.

   It has been shown that prenylation‐modified Rho‐GTPase can activate
   YAP/TAZ signaling, which contributes to β‐catenin activity.^[ [92]^19
   ^] To investigate whether PMVK stabilizes β‐catenin protein levels in
   relation to prenylation modification, we used shRNAs knockdown KRas
   protein. Results showed that knockdown of KRas did not affect β‐catenin
   levels (Figure [93]S2E, Supporting Information). To investigate whether
   PMVK‐mediated regulation of β‐catenin is related to cholesterol levels,
   we supplemented PMVK‐expressing HCC cells with cholesterol and showed
   that this only partially restored cell proliferation (Figure [94]S2F,
   Supporting Information). This suggests that cholesterol may be involved
   in the regulation of β‐catenin by PMVK but is not the main cause.
   Combined with the results of previous experiments in which knockdown of
   PMVK blocked Wnt3a‐activated β‐catenin (Figure [95]S1H, Supporting
   Information), we hypothesized that PMVK may directly interact with and
   stabilize β‐catenin.

2.4. PMVK‐Mediated Phosphorylation of β‐Catenin at Ser184 Enhances Its
Stability and Promotes Its Nuclear Translocation

   Co‐immunoprecipitation (co‐IP) experiments with the endogenous proteins
   indicated that PMVK interacts directly with β‐catenin (Figure [96]3A).
   Subsequent domain mapping demonstrated that this association requires
   PMVK's kinase domain and β‐catenin's N‐terminal domain (Figure
   [97]S3A,B, Supporting Information). Given that PMVK is a metabolic
   kinase,^[ [98]^21 ^] we asked whether it can directly phosphorylate
   β‐catenin by performing phosphorylation experiments in vitro and
   identifying phosphorylation sites by mass spectrometry (Figure [99]3B
   and Table [100]S2, Supporting Information). These studies showed that
   the evolutionarily conserved β‐catenin Ser184 residue was directly
   phosphorylated by PMVK and this was confirmed using a pSer184‐specific
   antibody (Figure [101]3C and Figure [102]S3C, Supporting Information).

Figure 3.

   Figure 3
   [103]Open in a new tab

   PMVK‐mediated β‐catenin Ser184 phosphorylation, protein stability and
   nuclear localization. A) Endogenous interaction between PMVK and
   β‐catenin in Huh7 and Hep3B cell lines. Anti‐PMVK Antibody (H‐9) from
   mouse as an IP antibody. Mouse lgG was used as a negative control. WCL,
   Whole cell lysate. B) Tandem mass spectrometry spectrum of β‐catenin
   pSer184. Detected productions are indicated in red (b ions) and blue (y
   ions). See Table [104]S2, Supporting Information, for details. C)
   Recombinant purified GST‐β‐catenin‐His and PMVK‐His proteins were
   subjected to an in vitro kinase reaction. β‐catenin phosphorylation was
   identified with β‐catenin p‐Ser184 antibody. Fast AP, Fast alkaline
   phosphatase. D) HEK293 cells were transfected with the indicated
   vectors. Cells were treated with CHX (50 µg mL^−1) for the indicated
   times and the levels of different β‐catenin mutants were analyzed by
   Western blotting. E) The expression intensity of different β‐catenin
   mutants at each time point in (D) was quantified by densitometry, with
   β‐actin as a normalizer. Data are shown as mean ± SD. F) HEK293 cells
   were transfected with the indicated plasmids, treated with MG132 for 6
   h and then lysed in RIPA buffer. Immuno‐precipitated HA‐β‐catenin
   proteins were subjected to immunoblot assay with Myc‐tag antibody. G)
   Distribution of different β‐catenin mutants protein levels in the
   nucleus or cytoplasm in Huh7 and Hep3B cells transfected with the
   indicated plasmids. β‐actin as a cytoplasm normalizer and Histone H3 as
   a nucleus normalizer. H) HEK293 cells were transfected with the
   indicated plasmids and then lysed in RIPA buffer. Immuno‐precipitation
   of different β‐catenin mutants proteins with anti‐HA magnetic beads and
   subjected to immunoblot assay.

   We then asked whether PMVK‐mediated β‐catenin phosphorylation was
   responsible for stabilizing β‐catenin. To this end, we generated
   [105]S184A (phosphorylation “dead”) and [106]S184D (phosphomimetic)
   phosphorylation site mutants and determined their half‐lives following
   cycloheximide (CHX) block. The [107]S184A mutant protein showed a
   significant decrease in its half‐life while [108]S184D’s half‐life was
   prolonged (Figure [109]3D,E). Consistently, [110]S184A showed increased
   levels of ubiquitination while [111]S184D showed the opposite
   (Figure [112]3F). Furthermore, β‐catenin [113]S184A showed markedly
   reduced nuclear localization (Figure [114]3G). In addition,
   immunofluorescence experiments also showed that [115]S184A was mainly
   localized in the cytoplasm while [116]S184D was significantly localized
   in the nucleus compared to β‐catenin WT. This further suggested that
   PMVK‐mediated phosphorylation of β‐catenin at Ser184 promotes its
   nuclear localization (Figure [117]S3E, Supporting Information). IHC
   staining of mouse embryos also showed that endogenous phosphorylated
   β‐catenin Ser184 preferentially localizes to the nucleus
   (Figure [118]1E).

   In adult hepatocytes, CKIα normally phosphorylates β‐catenin S45, which
   initiates GSK3β‐mediated β‐catenin phosphorylation as well as
   β‐TrCP‐mediated β‐catenin ubiquitination and proteasome‐mediated
   degradation.^[ [119]^20 ^] Therefore, we asked whether PMVK‐mediated
   β‐catenin Ser184 phosphorylation is regulated via changes in the above
   degradation complex proteins. The results showed that, whereas
   knockdown and over‐expression of PMVK respectively decreased and
   increased β‐catenin p‐Ser184 phosphorylation, neither affected the
   levels of CKIα, GSK3β, and β‐TrCP in HCC cells (Figure [120]S3D,
   Supporting Information). However, β‐catenin S184D showed significantly
   reduced association with CKIα, GSK3β, and β‐TrCP compared with S184A,
   whereas the phosphorylation of both Thr41 and Ser45 was significantly
   increased in S184A. Overexpression of WT β‐catenin or S184D, but not
   S184A, also activated β‐catenin downstream signaling (Figure [121]3H).
   Together the above results indicated that PMVK promotes β‐catenin
   protein stability and nuclear translocation by phosphorylating its
   Ser184 site, which in turn interferes with the de‐stabilizing
   phosphorylation events involving β‐catenin residues 41 and 45.

2.5. PMVK‐Generated MVA‐5PP Competitively Binds CKIα Causing Stabilization of
β‐Catenin

   β‐catenin's stability and turnover are regulated by the destruction
   complex.^[ [122]^21 ^] Our new findings show that, by directly
   phosphorylating β‐catenin, PMVK participates directly in this
   regulation. We therefore next asked whether the interaction between
   PMVK and β‐catenin occurs within the destruction complex or elsewhere.
   Co‐IP results showed that ectopic PMVK expression reduced the
   interaction between β‐catenin and several destruction
   complex‐associated proteins (Figure [123]4A). Interestingly, the
   over‐expression of a metabolically inactive mutant of PMVK,
   PMVK‐(17‐23),^[ [124]^22 ^] did not affect the interaction of β‐catenin
   with the destruction complex whereas PMVK WT did (Figure [125]S4A,
   Supporting Information). This result suggested that the PMVK‐mediated
   stabilization of β‐catenin protein levels relies upon downstream
   metabolites. Indeed, addition of the known PMVK metabolite MVA‐5PP to
   Huh7 and Hep3B cells increased their levels of both total and p‐S184
   β‐catenin and the down‐stream β‐catenin targets c‐Myc and Cyclin D1 in
   a concentration‐dependent manner (Figure [126]4B). Moreover, the
   addition of MVA‐5PP to cells lacking PMVK reduced β‐catenin's
   association with the destruction complex proteins (Figure [127]4C).
   Reverse in silico screening of the destruction complex's structure
   indicated a potential association between MVA‐5PP and CKIα (8.2 kcal
   mol^−1) (Figure [128]S4B, Supporting Information).

Figure 4.

   Figure 4
   [129]Open in a new tab

   PMVK‐produced MVA‐5PP competitively binds CKIα to stabilize β‐catenin.
   A) HEK293 cells were transfected with the indicated plasmids and then
   lysed in RIPA buffer. Immuno‐precipitated HA‐β‐catenin proteins were
   subjected to immunoblot assay. B) MVA‐5PP was added to Huh7 and Hep3B
   cells at the indicated final concentrations and subjected to immunoblot
   assay β‐catenin (total and p‐S184 form), c‐Myc and Cyclin D1 protein
   levels. C) HEK293 cells were transfected with the indicated plasmids,
   supplemented with MVA‐5PP (200 µ m) and then lysed in RIPA buffer.
   Immuno‐precipitation of HA‐β‐catenin proteins were subjected to
   immunoblot assay. D) The molecular docking of MVA‐5PP and CKIα is shown
   based on the crystal structure of CKIα (PDB code: 6GZD). Lys 46, Tyr
   64, and Asp 157 indicate the bound residues. E) Kd values were
   determined by MST for MVA‐5PP bound to purified human CKIα or CKIα
   mutant (K46A, Y64A and D157A) proteins. Data are shown as mean ± SD. F)
   Purified recombinant GST‐β‐catenin‐His protein and FLAG‐CKIα protein
   were subjected to in vitro phosphorylation assays at different
   concentrations of MVA‐5PP. Immunoblotting was performed using β‐catenin
   p‐Thr41/Ser45 antibody. G) Recombinant GST‐β‐catenin‐His and FLAG‐GSK3β
   proteins were subjected to in vitro phosphorylation assays at different
   concentrations of MVA‐5PP followed by immunoblotting using β‐catenin
   p‐Thr41/Ser45. H) Immunoblot assay of β‐catenin (total and p‐S184
   form), c‐Myc, Cyclin D1, and CKIα protein levels after MVA‐5PP (200 µ
   m) treatment in the indicated PMVK‐knockdown or control cancer cell
   lines.

   Diphosphomevalonate decarboxylase (MVD) catalyzes the fourth step of
   mevalonate biosynthesis involving the cation‐dependent reaction of
   MVA‐5PP with ATP to form isopentenyl diphosphate and ADP to release a
   carbon dioxide and a phosphate group, which binds to MVA‐5PP as a
   positive control (8.6 kcal mol^−1) (Figure [130]S4B, Supporting
   Information).^[ [131]^23 ^] To verify this result, we used Microscale
   thermophoresis (MST) to confirm binding of MVA‐5PP to the K46, Y64, and
   D157 sites of CKIα, with much less binding being seen with a mutant
   CKIα (K46A, Y64A, and D157A) (Figure [132]4D,E). In vitro
   phosphorylation assays also showed that MVA‐5PP inhibited, in a
   concentration‐dependent manner, the phosphorylation of β‐catenin S45 by
   CKIα but not β‐catenin T41 by GSK3β (Figure [133]4F,G). Finally,
   shRNA‐mediated depletion of MVD, which served an independent means of
   increasing MVA‐5PP levels, likewise increased β‐catenin protein. This
   suggested that the stabilization of β‐catenin was primarily caused by
   MVA‐5PP rather than its downstream metabolites (Figure [134]S4C,
   Supporting Information). In addition, the supplementation of cells with
   exogenous MVA‐5P increased the levels of total and p‐S184 β‐catenin as
   well as c‐Myc and Cyclin D1, indicating that PMVK's upstream
   metabolites also contribute to the activation of β‐catenin signaling,
   but that the primary effector is MVA‐5PP (Figure [135]S4D, Supporting
   Information).

   Cholesterol deficiency stimulates that transcription factor SREBP2,
   which in turn increases the expression of genes involved in the
   cholesterol biosynthesis pathway.^[ [136]^24 ^] To evaluate whether
   this can promote β‐catenin p‐Ser184 and β‐catenin signaling, we treated
   HCC cells with methyl‐β‐cyclodextrin (M‐β‐CD) which depletes
   cholesterol.^[ [137]^25 ^] Subsequent immunoblotting showed that M‐β‐CD
   treatment partially decreased the levels total and p‐S184 β‐catenin.
   This suggested that β‐catenin S184 phosphorylation is dependent on
   certain minimal levels of MVA‐5PP (Figure [138]S4E, Supporting
   Information). The data also suggest that the product of PMVK, namely
   MVA‐5PP, rather than PMVK itself, competitively binds to CKIα to
   prevent the phosphorylation‐dependent degradation of β‐catenin.

   Finally, in vitro supplementation with MVA‐5PP restored β‐catenin
   levels in PMVK knockdown cells but failed to rescue β‐catenin p‐Ser184
   or the downstream signaling defects. This may have been due to the lack
   of PMVK‐mediated β‐catenin Ser184 phosphorylation (Figure [139]4H).

2.6. PMVK Knockout in Hepatocytes Blocks HCC Generation in Mice

   To determine the role of PMVK in HCC development in vivo, PMVK
   hepatocyte‐specific conditional knockout (CKO) mice were constructed
   (Figure [140]S5A, Supporting Information). Immunoblotting confirmed
   that these mice also had significantly reduced levels of both total and
   p‐S184 β‐catenin (Figure [141]S5B, Supporting Information). Although
   the results of Ki67 staining showed that cell proliferation was
   suppressed in PMVK ^flfl Alb ^cre mice (Figure [142]S5C, Supporting
   Information), there were no significant differences in body or liver
   weights compared to WT mice at 8 weeks of age (Figure [143]S5D,E,
   Supporting Information). To evaluate PMVK's role on HCC development, we
   generated tumors in both groups of mice with DEN/CCl[4] (Figure
   [144]5A). CKO mice had both significantly reduced tumor numbers and
   smaller tumor volumes compared to control WT mice (Figure [145]5B–D).
   Serum TC and TG levels were also significantly reduced in the CKO group
   (Figure [146]5E,F) and Ki‐67 staining showed that tumor proliferation
   was significantly inhibited (Figure [147]S5F, Supporting Information).
   Adeno‐associated virus 8 (AAV8) restoration of PMVK expression promoted
   tumor growth and increased serum levels of TC and TG in CKO mice
   (Figure [148]5B–F). To further investigate the effects of PMVK and
   β‐catenin Ser184 phosphorylation on tumor growth, we delivered PMVK or
   β‐catenin (WT and S184A/D) via AAV8 in WT mice. Our results showed that
   overexpression of PMVK, β‐catenin WT or S184D significantly promoted
   HCC growth (Figure [149]5G–I). S184D also significantly reduced the
   survival rate (Figure [150]5J). Collectively, these data suggest that
   liver‐specific PMVK deletion significantly slows HCC development by
   reducing cell proliferation and cholesterol and triglyceride levels
   whereas PMVK‐mediated phosphorylation of β‐catenin Ser184 strongly
   promotes tumor growth.

Figure 5.

   Figure 5
   [151]Open in a new tab

   Liver‐specific deletion of PMVK inhibits HCC growth in mice. A)
   Schematic overview of DEN/CCl[4]‐induced HCC mice model. During this
   period, adeno‐associated virus containing PMVK were given via tail‐vein
   injection. B) Liver images extracted from the indicated mice. C) Tumor
   number of each liver from (B). D) Average tumor volume of each liver
   from (B). E) Serum TC levels in mice from (B). F) Serum TG levels in
   mice from (B). G) Representative IHC images of H&E and Ki67 in tumor
   tissues from (B). H) Adeno‐associated virus containing the indicated
   plasmids were delivered via tail‐vein injection. Liver images extracted
   from the indicated mice. I) Tumor number of each liver for (H). J)
   Average tumor volume of each liver for (H). K) Percent survival of mice
   for (H).

2.7. PMVK Is Overexpressed in HCC Patients and Correlates with Poor Prognosis

   In tumor samples from 10 HCC patients, PMVK, β‐catenin (total and
   p‐S184 form), c‐Myc and Cyclin D1 were significantly upregulated
   relative to matched normal liver (Figure [152]6A–F). In addition, PMVK
   positivity correlated with β‐catenin protein levels (Figure [153]6G),
   in contrast to other enzymes of the mevalonate pathway (Figure
   [154]S6A–D, Supporting Information). PMVK also positively correlated
   with β‐catenin p‐Ser184, which in turn correlated with the levels of
   downstream target gene proteins (Figure [155]6H–O). PMVK was also
   highly expressed in HCCs from the TCGA database (Figure [156]6P).
   Similar findings were shown in the GEO databases ([157]GSE45436,
   [158]GSE25097, [159]GSE102083, and [160]GSE14520). PMVK mRNA levels
   also positively correlated with the expression of β‐catenin target
   genes in the TCGA‐LIHC and [161]GSE14520 dataset (Figure [162]S6E–K,
   Supporting Information). The PMVK gene also tended to be highly
   amplified in HCC samples, thus explaining the basis for its
   over‐expression (Figure [163]6Q,R). Finally, PMVK upregulation
   correlated with poor HCC patient survival and advanced clinical stage
   (Figure [164]6S and Figure [165]S6L, Supporting Information). Together,
   these data identify PMVK as a potential therapeutic target in HCC.

Figure 6.

   Figure 6
   [166]Open in a new tab

   PMVK expression is increased in human HCC and correlates with poor
   survival. A) Western blot analysis for the indicated proteins in human
   HCCs. T, tumor; N, adjacent normal tissue. B–F) The intensities of
   PMVK, β‐catenin (total and p‐S184), c‐Myc and Cyclin D1 for (A) were
   quantified by densitometry, with β‐actin as a normalizer. Data are
   shown as mean ± SD. G–O) Correlation of different protein levels in HCC
   tissues from (A). Each point is an individual sample. P) PMVK mRNA
   levels in normal liver and HCCs from TCGA. Q) PMVK gene copy number in
   normal liver and tumor tissues from TCGA. R) Correlation of PMVK mRNA
   level with its DNA copy number in HCC tissues from TCGA dataset. Each
   point is an individual sample. S) Kaplan–Meier curves with univariate
   analysis of the survival of patients with HCC based on high versus low
   expression of PMVK. Data are shown as mean ± SD. The P values were
   determined by paired two‐sided Student's t‐test for (B–F) and unpaired
   two‐sided Student's t‐test for (P) and (Q). The correlation coefficient
   (r) and p values in (G–O and R) were determined using two‐tailed
   Pearson correlation analysis.

2.8. A Small Molecular Inhibitor of PMVK, Reduces HCC Growth In Vivo

   Using a virtual screen, we identified the putative PMVK inhibitor top 5
   (PMVKi5)
   (C[24]H[23]ClN[2]O[6],N‐[(E)‐[3‐[(4‐chloro‐3,5‐dimethylphenoxy)methyl]‐
   4‐methoxyphenyl]methylideneamino]‐3,4,5‐trihydroxybenzamide) from a
   library of 1.7 million ChemDiv small molecule compounds (Figure
   [167]7A). The molecular docking study based on the crystal structure
   PMVK showed that PMVKi5 contacts PMVK residues Lys17, Ser20, Gly21,
   Lys22, Asp23, Arg73, and Arg141 (Figure [168]S7A, Supporting
   Information). In support of this model, the results of MST showed
   PMVKi5 to have a higher affinity for WT PMVK than for mutants harboring
   any of the above‐discussed residues (Figure [169]7B). In vitro studies
   showed PMVKi5 to have IC[50]’s of 7.05 and 18.74 µ m in Huh7 and Hep3B
   cell lines, respectively (Figure [170]S7B, Supporting Information).
   Given that residues 17–23 of PMVK are necessary for ATP‐binding,^[
   [171]^22 ^] we speculated that PMVKi5 acts as an ATP mimetic. Indeed,
   an in vitro kinase assays showed that PMVKi5 significantly inhibited
   ATP consumption by PMVK and its phosphorylation of β‐catenin
   (Figure [172]7C,D). Finally treatment of Huh7 and Hep3B cells with 20 µ
   m PMVKi5 for 6 h significantly inhibited β‐catenin signaling without
   altering PMVK protein levels (Figure [173]7E).

Figure 7.

   Figure 7
   [174]Open in a new tab

   The PMVK inhibitor, PMVKi5, suppresses HCC growth in vitro and in vivo.
   A) Chemical structure formula of PMVKi5. B) Kd values were determined
   as the binding of PMVKi5 to purified human PMVK or PMVK mutant (K17A,
   S20A, G21A, K22A, and D23A) proteins using MST. C) In vitro kinase
   activity assays D) Reco9mnbinant GST‐β‐catenin‐His and PMVK‐His were
   used for in vitro kinase reactions, with or without the addition of 20
   µ m PMVKi5. β‐catenin phosphorylation was identified with β‐catenin
   p‐Ser184 antibody. E) Immunoblot for the indicated proteins in Huh7 and
   Hep3B cell lines following exposure to 20 µ m PMVKi5 for 6 h. F)
   Schematic overview of DEN/CCl4‐induced HCC mice model. For the
   prevention group, mice were treated between 16 and 20 weeks, with 30
   mg kg^−1 PMVKi5 i./p. every 2 days for a total of 14 doses for. In the
   treatment group, mice were treated from 23 to 27 weeks. After 28 weeks,
   livers were extracted. G) Liver images from the indicated mice. H)
   Tumor number of each liver for (G). I) Average tumor volume of each
   liver for (G). J) Immunoblot analysis for the indicated proteins in
   tumor tissues from (G). K) Representative IHC images of H&E and Ki67 in
   tumor tissues from (G). Scale bar, 50 µm.

   To further assess the role of PMVKi5 in tumorigenesis, we again
   utilized the DEN/CCl[4] HCC model in which mice were treated with
   PMVKi5 (30 mg kg^−1 intraperitoneally) every 2 days for a total of 14
   doses (Figure [175]7F). Both prevention and treatment group mice showed
   significant reduced tumor burdens compared to control tumor‐bearing
   mice (Figure [176]7G–I). PMVKi5 treatment also inhibited β‐catenin
   protein levels and downstream targets in mouse tumors (Figure [177]7J).
   Ki67 staining confirmed that PMVKi5 significantly inhibited tumor
   proliferation (Figure [178]7K).

   To further evaluate PMVKi5 efficacy in other cancers, we used the
   AOM/DSS induced colorectal cancer (CRC) mouse model (Figure [179]S7G,
   Supporting Information). Similarly, PMVKi5 also inhibited tumorigenesis
   in this model (Figure [180]S7H–K, Supporting Information).
   Collectively, these data confirm our previous results by showing that
   PMVKi5 blocks β‐catenin signaling by inhibiting PMVK.

   Importantly, PMVKi5 did not alter mouse body weight or serum ALS, ATL
   levels, indicating that PMVKi5 treatment does not cause liver injury
   (Figure [181]S7D–F,K, Supporting Information). PMVKi5 thus represents a
   promising lead compound that should be safe and effective for the
   prevention and treatment of HCC and/or CRC.

3. Discussion

   The WNT/β‐catenin signaling pathway is often activated in cancer due to
   the aberrant accumulation and nuclear localization of its effector
   β‐catenin.^[ [182]^10 ^] Although many β‐catenin inhibitors have been
   screened in the past decades, the eventual development of resistance
   has tended to stymie such efforts. There is thus an urgent need to
   screen for new targets that regulate β‐catenin indirectly.^[ [183]^26
   ^] This consideration was the primary motivation for our genome‐wide
   Crispr screen that identified PMVK as one such potential target.

   Of note, metabolic enzymes were significantly enriched in the
   above‐mentioned screen. Although metabolic reprogramming is one of the
   hallmarks of malignancy, its relationship to β‐catenin signaling has,
   until now, been unclear.^[ [184]^27 ^] Our Crispr/Cas9 screen led us to
   focus our attention on the mevalonate pathway enzyme PMVK as a novel
   target that regulates β‐catenin signaling. As a result, we have found
   that PMVK plays key roles in embryonic development and tumor growth,
   while PMVK‐mediated β‐catenin p‐Ser184 phosphorylation is also
   significantly upregulated in both key nodes (Figure [185]S5G,
   Supporting Information). The importance of this activity may be due to
   the uniqueness of the mevalonate pathway as the upstream pathway of
   cholesterol synthesis and the importance of β‐catenin signaling in
   embryonic development and malignancy.^[ [186]^28 ^]

   Previous studies have also shown that the mevalonate metabolic pathway
   contributes prominently to prominent tumor growth, including the
   ability of MVA‐5P to stabilize mutant p53.^[ [187]^7 ^] In addition,
   mevalonate pathway metabolites are also important for prenylation,
   geranylgeranylation, and farnesylation of several oncogenic proteins,
   such as Rho, Rac, Rab, and Ras.^[ [188]^29 ^] Although
   geranylgeranylation can activate the YAP/TAZ signaling pathway and thus
   promote the β‐catenin signaling pathway,^[ [189]^19 , [190]^30 ^] our
   results show that this activity is not involved in the PMVK‐mediated
   regulation of β‐catenin signaling (Figure [191]S2E, Supporting
   Information). In addition, knockdown of PMVK resulted in blocked
   synthesis of the mevalonate pathway, which also impaired cholesterol
   levels. Previous studies have shown that accumulated cholesterol
   stabilizes β‐catenin signaling.^[ [192]^31 ^] But in our study we
   showed that supplementation of additional cholesterol levels in HCC
   cells with knockdown PMVK only partially restored cell proliferation
   levels, suggesting that although cholesterol is an important factor for
   cell proliferation, it is not a major factor in the regulation of
   β‐catenin signaling by PMVK (Figure [193]S2F, Supporting Information).

   Interestingly, PMVK knockdown also impaired Wnt3a‐activated β‐catenin
   signaling, suggesting that the mechanism of PMVK involves the
   disruption complex or the nuclear localization of β‐catenin. qRT‐PCR
   showed that PMVK knockdown increased some Wnt family transcript levels,
   possibly due to a classical feedback mechanism triggered by the loss of
   β‐catenin signaling. This prompted us to explore whether PMVK interacts
   directly with β‐catenin. As a result, we determined that PMVK directly
   phosphorylates β‐catenin Ser184, thereby increasing β‐catenin protein
   stability, promoting its nuclear translocation and increasing target
   gene expression. Most HCCs contain mutations in CTNNB1, which stabilize
   β‐catenin by preventing its association with the destruction complex.
   However, we found that knockdown of PMVK still destabilized β‐catenin
   and impaired its downstream signaling in Hep3B (AXIN1 mutant) cells, in
   which the destruction complex is non‐functional. This suggested that,
   by phosphorylating the critical Ser184 residue, PMVK activates
   β‐catenin via a destruction complex‐independent mechanism. In the HCC
   mouse model, overexpression of β‐catenin p‐Ser184 strongly promoted
   tumor growth. Although stabilizing β‐catenin indirectly promotes its
   nuclear localization, the actual mechanism by which β‐catenin Ser184
   phosphorylation accomplishes this requires further clarification.
   Regardless of the actual means by which this is accomplished, β‐catenin
   Ser184 phosphorylation serves as a new and novel marker for β‐catenin
   pathway activation.

   PMVK phosphorylation of β‐catenin Ser184 is dependent on its classical
   metabolic enzyme function. Because PMVK mutants cannot affect the
   interaction of β‐catenin with the destruction complex, this implies
   that the downstream metabolites of PMVK may cooperate with mutations in
   the destruction complex or in β‐catenin itself to further stabilize the
   protein, promote its nuclear localization and enhance it's signaling.
   Interestingly, the PMVK metabolite, MVA‐5PP similarly inhibited the
   interaction of β‐catenin with the destruction complex with similar
   consequences. We found that the mechanistic underpinnings of this
   activity involves the binding of MVA‐5PP to CKIα. Genetic knockdown of
   MVD also increased β‐catenin protein level. This suggests that PMVK
   stabilization of β‐catenin is dependent on its metabolite MVA‐5PP
   rather than on the mevalonate pathway metabolite isopentenyl
   diphosphate or other intermediates. In the developing chicken liver,
   MVK and PMVK remain highly active for the first 5 days, while the
   activity of MVD is low.^[ [194]^32 ^] This suggests that increased
   MVA‐5PP may be elevated in a way that satisfies the strong dependency
   on β‐catenin signaling during early development.^[ [195]^33 ^]
   Collectively, we uncovered a non‐canonical function of MVA‐5PP by which
   it and the PMVK enzyme that generates it, play synergistic roles in
   regulating the β‐catenin signaling pathway.

   Analysis of clinical samples and data from TCGA indicated that PMVK is
   highly expressed in HCC, likely due to its genomic amplification. PMVK
   levels were also found to correlate closely with clinical stage and
   β‐catenin p‐Ser184 correlated with total β‐catenin and its downstream
   signals. This indicates that PMVK and β‐catenin p‐Ser184 protein levels
   are potentially useful as prognostic markers in HCC. Finally PMVK
   inhibitors represent a novel targeted therapy directed against a
   previously unsuspected target that drives HCC progression and may be
   useful in CRC as well. We also evaluated other inhibitors of the
   mevalonate pathway on PMVK‐regulated β‐catenin signaling. Simvastatin
   showed the same effect as PMVKi5 on β‐catenin, due to the fact that
   statins block the mevalonate pathway at its source (also blocking the
   synthesis of MVA‐5PP). Zoledronic acid instead promotes β‐catenin
   protein levels, the target of zoledronic acid is the inhibition of
   farnesyl pyrophosphate synthesis which may indirectly contribute to the
   accumulation of upstream metabolites, similar to knockdown of MVD
   (Figure [196]S7C, Supporting Information).

   In conclusion, we have demonstrated how the mevalonate pathway enzyme
   PMVK controls β‐catenin signaling. In doing so, we have revealed a new
   mechanism by which metabolic reprogramming controls HCC and CRC
   pathogenesis. PMVK inhibition, perhaps with PMVKi5 or its analogs thus
   represents a promising and rational therapeutic strategy.

4. Experimental Section

Vectors, Reagents, Antibodies, Cell Culture, and Stable Cell Lines

   Human PMVK shRNAs (Table [197]S3, Supporting Information) were
   expressed in the PLKO.1 vector. The open reading frames of human PMVK,
   β‐catenin, and CKIα were amplified and cloned into
   pHAGE‐CMV‐MCS‐PGK‐3×FLAG and pCMV‐HA vectors. Mutations in the PMVK,
   β‐catenin, and CKIα cDNA sequences were generated by overlap extension
   PCR. Primers for constructs used in this study are listed in Table
   [198]S4, Supporting Information, and were synthesized by GeneCreate
   Biological Engineering Co. Ltd (Wuhan, China). Transfection and the
   establishment of stable cell lines were performed as previously
   described.^[ [199]^34 ^]

   5‐diphosphomevalonate was purchased from Sigma‐Aldrich. More reagent
   information in this study is listed in Table [200]S5, Supporting
   Information. Anti‐HA magnetic beads and anti‐FLAG affinity agarose
   beads were purchased from Selleck (Houston, USA). Antibodies used in
   this study are listed in Table [201]S6, Supporting Information.

   Human HEK293 cells were obtained from the American Type Culture
   Collection in 2009. Human liver tumor‐derived cell lines Huh7 and Hep3B
   were obtained from the China Center for Type Culture Collection and the
   Cell Bank of the Type Culture Collection of The Chinese Academy
   Sciences and cultured as previously described.^[ [202]^34 ^] All cells
   were regularly authenticated by short tandem repeat analyses, were
   tested for the absence of Mycoplasma and were used within five passages
   after thawing.

Animal Experiments

   All animal studies were approved by the Animal Care Committee of Wuhan
   University (No. WQ20210130). 4‐week‐old female BALB/c nude mice were
   purchased from Gempharmatech (Jiangsu, China). PMVK ^−/− mice were
   generated, bred and raised at Cyagen Biosciences with CRISPR‐Cas9
   technology to delete exon E1‐E5 of the PMVK gene. PMVK ^fl/fl mice
   (Cyagen Biosciences) have loxP sites flanking exons 2 and 5 of the PMVK
   gene. Alb‐Cre mice were kindly provided by Prof. Yong Liu (Wuhan
   University). To generate PMVK ^fl/fl Alb ^cre mice, PMVK ^fl/fl mice
   were crossed with Alb‐Cre mice. All mouse genotype identification
   primer sequences used are shown in Table [203]S7, Supporting
   Information.

   For xenograft experiments, tumor growth was monitored every 3 days for
   a total period of 30 days. Tumor volumes were calculated by the
   equation V (mm^3) = a × b × (a + b)/2, where a is the length and b is
   the width.

   For induction of HCCs, C57BL/6J, PMVK ^fl/fl and PMVK ^fl/fl Alb ^cre
   male mice were intraperitoneally administrated 25 mg kg^−1 of
   N‐Nitrosodiethylamine (DEN) (Sigma) at 2 weeks after birth. From 4 to
   15 weeks, mice were intraperitoneally injected with 10% CCl[4] (5
   mL kg^−1) weekly. After 28 weeks, livers were examined. For induction
   of CRCs, C57BL/6J mice were intraperitoneally administrated AOM (12.5
   mg kg^−1) at 8 weeks after birth. From 9, 12, and 15 weeks, 2.5% DSS
   was added to the drinking water of mice for 1 week, after 25 weeks, the
   colons were examined.

   Littermate mice were randomly divided into control and various
   treatment groups including: control, PMVKi5 prevention, and PMVKi5
   treatment. All cohorts were maintained as blinded cohorts following
   these treatments.

Human Genome‐Wide CRISPR Screening

   Amplification of the sgRNA library followed the description of Feng
   Zhang et al.^[ [204]^17 , [205]^35 ^] Twelve T225 culture flasks were
   prepared for library virus generation according to the HEK293 culture
   conditions described above. Each flask was transfected with 20 µg of
   library plasmid (10 µg each of libraries A and B), 10 µg of pMD2.G and
   15 µg of psPAX2 were transfected using 100 µL of Lipofectamine 2000 and
   200 µL of Plus Reagent. The medium was replaced after 24 h and the cell
   supernatants were collected at 48 and 72 h of transfection and passed
   through 0.45 µm filters. The packed lentiviral library was then used to
   infect at least 5 × 10^7 Huh7 cells in the presence of 1 µg mL^−1
   polybrene, ensuring a multiplicity of infection (MOI) of 0.3 so that
   most cells receive only one stably integrated RNA guide. 24 h after
   infection, puromycin was added to select for stably transduced cells.
   Allow the cell count to expand to >8 × 10^7, ensuring that the initial
   sieve cell count is a library coverage multiple >300×. Each group of 4
   × 10^7 cells were treated with vector and XAV393 (30 µ m),
   respectively, and cells were passaged every 2 days for a total of 14
   days of drug treatment. All cells were collected and genomic DNA was
   extracted, enriched for fragments containing sgRNA sequences using
   amplification primers (forward primer, ATGGACTATCATATGCTTACCGTA;
   reverse primer, AGTACACGACATCACTTTCCC) and sequenced on an Illumina
   NovaSeq 6000 following a second round of PCR reactions with the
   addition of splice primers and indexT sequences used to identify
   different samples. The number of sgRNAs of XAV‐939 and vehicle was
   identified using the MAGeCK algorithm and analyzed using DESEq2 to
   obtain significantly enriched or depleted sgRNAs.

Immunohistochemistry and Immunofluorescence

   Liver sections or mouse embryos were fixed in 10% formalin and embedded
   in paraffin according to standard protocols. Liver sections or mouse
   embryo were sectioned and stained with H&E, Ki67, PMVK, β‐catenin, and
   β‐catenin p‐Ser184. Images were obtained on a Leica Aperio VERSA 8
   microscope.

   Immunofluorescence analysis was performed as previously described.^[
   [206]^34 ^] Briefly, HEK293 cells were transfected with the indicated
   vectors. 48 h later, they were fixed in 4% paraformaldehyde for 15 min,
   followed by 0.2% Triton X‐100 permeabilization for 5 min and 3% BSA
   blocking for 1 h. After incubation with primary antibody against FLAG
   (1:200) overnight in 4 °C, the cells were washed well with PBST and
   incubated with ABflo 488‐conjugated goat anti‐rabbit (1:500) for 1 h at
   room temperature. After additional extensive washing, the cells were
   labeled with DAPI (5 µg mL^−1) for 5 min at room temperature.
   Fluorescence imaging was performed with a LSM800 confocal laser
   scanning microscope.

Co‐IP and Immunoblotting

   Transfected cells were lysed in 1 mL lysis buffer [50 mm Tris (pH 7.4),
   150 mm NaCl, 1% NP‐40, 0.5% Sodium deoxycholate, 0.1% SDS, 5 mm EDTA,
   10 mg mL^−1 aprotinin, 10 mg mL^−1 leupeptin, and 1 mm PMSF]. Sepharose
   beads were washed three times with 1 mL lysis buffer containing 0.1%
   NP‐40. Co‐IP and immunoblot analysis were performed as described.^[
   [207]^36 ^]

qRT‐PCR

   Total RNA was isolated using Trizol followed by DNase treatment.
   Reverse transcription was performed with a cDNA Synthesis Kit and qPCR
   were performed using SYBR Green master mix using standard protocols.
   All qPCR primer sequences used are shown in Table [208]S8, Supporting
   Information. β‐actin was used as an internal control.

Protein Half‐Life Assay

   After achieving ≈70% confluence, cells in 12‐well plates were
   transfected with the indicated plasmids using Lipofectamine 2000. 36 h
   later, the cells were treated with cycloheximide (CHX, 50 µg mL^−1) for
   the indicated times before collection for immunoblot analysis.

MTT Assay

   For MTT assays, cells were seeded into 96‐well plates (10^3 cells per
   well) and cultured in DMEM supplemented with 10% FBS. After 3–7 days of
   incubation, MTT was added and cells were incubated for 4 h at 37 °C. At
   the indicated time, this was replaced with 200 µL of DMSO. Absorbance
   was then read at 570 nm using an ELX800 absorbance microplate reader.

In Vitro Phosphorylation Assay

   Recombinant GST‐β‐catenin‐His and PMVK‐His were prepared and purified
   from Rossita. 1 µg GST‐β‐catenin‐His protein was incubated with 1 µg
   PMVK‐His in a 50 µL kinase reaction buffer (20 mm Tris‐HCl (pH 7.5), 10
   mm MgCl[2], 50 mm KCl, 30 µ m ATP) at 37 °C for 0.5 h. Reactions were
   stopped by the addition of SDS sample buffer, and samples were then
   heated for 5 min at 95 °C. Phosphorylation of β‐catenin Ser184 was
   identified by mass spectrometry and immunoblot with β‐catenin p‐Ser184
   antibody that was generated by ABclonal by immunizing rabbits with a
   S184‐monophosphorylated peptide (KKEA(S‐pho)RHA) conjugated with
   keyhole limpet hemocyanin.

In Vivo Ubiquitination Assay

   36 h after transfection, cells were treated with 20 µ m MG132 for 6 h.
   They were then lysed in RIPA lysis buffer as described above. The
   samples were boiled for 10 min in SDS sample buffer and analyzed by
   immunoblot with the indicated antibodies.

Microscale Thermophoresis

   A Monolith NT.115 instrument (NanoTemper Technologies, Germany) was
   used to measure the Kd of binding of MAV‐5PP to CKIα‐GFP protein and
   PMVKi5 to PMVK‐GFP protein. All samples to be tested were first
   solubilized in RIPA buffer. The CKIα‐GFP or PMVK‐GFP was added to each
   dilution and incubated at room temperature for 5 min. The samples were
   loaded into silica capillaries (NanoTemper Technologies, Germany). The
   measurements were performed at 25 °C using 60% LED power and medium MST
   power. Data were analyzed with Nano‐Temper Analysis software, v.2.3.

Serum Cholesterol, Triglyceride, ALT, and AST

   Serum cholesterol, triglyceride, ALT, and AST were measured using
   colorimetric analysis kits (Elabscience, Wuhan). All experimental steps
   were performed according to the manufacturer's instructions.

Sources of Human HCC Samples

   Human HCC samples were used as previously described.^[ [209]^34 ^] All
   procedures that involved human HCC samples collection were approved by
   the ethics committee of Wuhan University (No.2022030) and the Union
   Hospital in Wuhan, China and conformed to the ethical guidelines of the
   1975 Declaration of Helsinki. Informed written consent of all
   participants was signed by all individuals or their families and
   obtained at the Union Hospital in Wuhan, China. The diagnoses of all
   samples were confirmed by histological review.

Docking and Molecular Modeling

   The PMVK‐based crystal structure (PDB: 3CH4) was used for
   computer‐based virtual docking of 1.7 million small molecules from the
   ChemDiv database. Hydrogenation and energy optimization of database
   molecules relied on Schrodinger's Ligprep program. Pre‐processing of
   protein structures used Schrodinger's Protein Preparation program.
   Next, the HTVS, SP, and XP docking procedures of Schrodinger's Glide
   module were used to cascade docking of the database.^[ [210]^37 ^]
   Finally, the top 5 candidate small molecules were selected for cellular
   experiments. ChemBioDraw Ultra 17.0 was used to draw the structure of
   MVA‐5PP, the protein structures of the destruction complexes were all
   from the PDB database (APC PDB, 3NMZ; Axin1 PDB, 1DK8; β‐catenin PDB,
   1QZ7; β‐TrCP PDB, 1P22; CKIα PDB, 6GZD; GSK3β PDB, 1I09; MVD PDB, 3D4J;
   PMVK PDB, 3CH4), and were docked using Discovery Studio.^[ [211]^38 ^]

Luminescent Kinase Assay

   A Luminescent Kinase Assay Kit was used to quantify enzymatic activity
   of PMVK by quantifying ATP consumption. The kinase reaction system
   consisted of 50 mm HEPES (pH = 8.0), 50 mm KCl, 3 mm MgCl[2], 200 µ m
   ATP, 40 µ m MVA‐5P, and 5 µ m PMVK, which was incubated at 37° for 5
   min. Fluorescence intensity was measured by a luminometer. All
   experimental steps follow the manufacturer's instructions.

Data Availability

   High‐throughput sequencing results for protein class analysis and
   pathway enrichment analysis were analyzed in the websites
   ([212]https://www.pantherdb.org). HCC datasets were downloaded from The
   Cancer Genome Atlas (TCGA) data portal
   ([213]https://www.tcgadata.nci.nih.gov). PMVK mRNA levels were analyzed
   from TCGA and NCBI GEO databases ([214]https://www.ncbi.nlm.nih.gov).
   Protein crystal structures were downloaded from PDB databases
   ([215]https://www.pdbus.org). Small molecules were downloaded from the
   ChemDiv database ([216]https://www.chemdiv.com).

Statistics and Reproducibility

   Data are presented as mean ± SD or mean ± SEM. Log‐rank tests were used
   for survival analyses. Student's t‐test was used for comparisons
   between two groups. One‐way ANOVA was used for multiple comparisons in
   patient populations. The correlation coefficient (r) and p values were
   obtained from Pearson correlation analysis, Pearson's r was calculated
   by GraphPad Prism 8. GraphPad Prism 8 was used for statistical
   calculations. p < 0.05 was considered to be statistically significant;
   *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n.s., not
   significant. The data analyses were performed using unbiased software
   programs or algorithms. All experiments were repeated with a minimum of
   three independent repeats unless otherwise noted. Each experiment was
   repeated independently with similar results.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   Z.C., F.W., and Y.L. designed the study; Z.C. performed most of the
   experiments; Y.T., M.L., and M.W. constructed plasmids; X.‐J.Z., Z.C.,
   and F.W. performed to analyze sequencing and clinical data; J.Z., C.T.,
   and X.‐Y.Z. performed some animal experiments. All authors discussed
   the results. Z.C., E.P., F.W., and Y.L. wrote the manuscript with
   comments from all authors.

Supporting information

   Supporting Information
   [217]Click here for additional data file.^ (1.9MB, pdf)

Acknowledgements