PresentationCell Biology

Merlin/NF2 Functions Upstream of the Nuclear E3 Ubiquitin Ligase CRL4DCAF1 to Suppress Oncogenic Gene Expression

See allHide authors and affiliations

Science Signaling  30 Aug 2011:
Vol. 4, Issue 188, pp. pt6
DOI: 10.1126/scisignal.2002314
A presentation from the 50th Annual Meeting of the American Society for Cell Biology in Philadelphia, Pennsylvania, 11 to 15 December 2010.

Abstract

Integrin-mediated activation of PAK (p21-activated kinase) causes phosphorylation and inactivation of the FERM (4.1, ezrin, radixin, moesin) domain–containing protein Merlin, which is encoded by the NF2 (neurofibromatosis type 2) tumor suppressor gene. Conversely, cadherin engagement inactivates PAK, thus leading to accumulation of unphosphorylated Merlin. Current models imply that Merlin inhibits cell proliferation by inhibiting mitogenic signaling at or near the plasma membrane. We have recently shown that the unphosphorylated, growth-inhibiting form of Merlin accumulates in the nucleus and binds to the E3 ubiquitin ligase CRL4DCAF1 to suppress its activity. Depletion of DCAF1 blocks the hyperproliferation caused by inactivation of Merlin. Conversely, expression of a Merlin-insensitive DCAF1 mutant counteracts the antimitogenic effect of Merlin. Expression of Merlin or silencing of DCAF1 in Nf2-deficient cells induce an overlapping, tumor-suppressive program of gene expression. Mutations present in some tumors from NF2 patients disrupt Merlin’s ability to interact with or inhibit CRL4DCAF1. Lastly, depletion of DCAF1 inhibits the hyperproliferation of Schwannoma cells isolated from NF2 patients and suppresses the oncogenic potential of Merlin-deficient tumor cell lines. Current studies are aimed at identifying the substrates and mechanism of action of CRL4DCAF1 and examining its role in NF2-dependent tumorigenesis in mouse models. We propose that Merlin mediates contact inhibition and suppresses tumorigenesis by translocating to the nucleus to inhibit CRL4DCAF1.

Presentation Notes

Slide 1: Science Signaling logo

The slideshow and notes for this presentation are provided by Science Signaling (http://www.sciencesignaling.org).

Slide 2: Merlin/NF2 functions upstream of the nuclear E3 ubiquitin ligase CRL4DCAF1 to suppress oncogenic gene expression

Filippo Giancotti delivered this presentation at the 50th Annual Meeting of the American Society for Cell Biology in Philadelphia, Pennsylvania, 11 to 15 December 2010.

Slide 3: Adhesion signals converge on Merlin

Merlin, the protein product of the NF2 (neurofibromatosis type 2) tumor suppressor gene, mediates contact inhibition of proliferation and inhibits progression through the G1 phase of the cell cycle (14). Various extracellular cues converge to regulate Merlin’s phosphorylation status and thereby its activity. Simultaneous integrin and receptor tyrosine kinase signaling activates PAK (p21-activated kinase), which then phosphorylates Merlin on Ser518, thus disrupting the intramolecular association between Merlin’s FERM (4.1, ezrin, radixin, moesin) domain and its C terminus, maintaining it in an open and inactive conformation (5, 6). On the other hand, cadherin signaling initiated by cell-cell contact inhibits PAK, leading to an accumulation of the unphosphorylated, closed conformer of Merlin, which inhibits cell proliferation and mediates tumor suppression (4). In agreement with this model, genetic and structural analyses suggests that most, if not all, missense mutations in NF2 that have been identified in families or multiple unrelated individuals afflicted with neurofibromatosis type 2 disrupt the closed conformation of Merlin (79).

Merlin inhibits various mitogenic signals, including the recruitment of Rac to the plasma membrane [and, thereby, the activation of PAK (4, 10, 11)], the activation of mTORC1 (mammalian target of rapamycin complex 1) independently of the kinase Akt (1, 12), the Ras-ERK (extracellular signal-regulated kinase) pathway (13), PI3K (PI3 kinase)–Akt signaling (14), and FAK (focal adhesion kinase)–Src signaling (15). Several mechanisms have been proposed to explain the inhibitory effect of Merlin on Ras-ERK signaling, including inhibition of Ras, sequestration of EGFR (epidermal growth factor receptor) in an inactive conformation, inhibition of the export of EGFR and other RTKs (receptor tyrosine kinases) to the plasma membrane, and acceleration of EGFR endocytosis in Drosophila melanogaster (2, 4, 10, 13, 1517). Interestingly, experiments in Drosophila have revealed that Merlin cooperates with Expanded, another FERM domain protein, upstream of the Hippo pathway (18). Despite these major advances, the biochemical function of Merlin and therefore the mechanism through which it suppresses tumorigenesis has remained, until recently, elusive.

Slide 4: Wild-type Merlin interacts specifically with DDB1, DCAF1, and Cul4

To identify proteins interacting specifically with the growth-inhibitory conformer of Merlin, we performed tandem affinity purification (TAP) using N- or C-terminally Flag-hemagglutinin (HA)–tagged Merlin (FH-Merlin). The results indicated that Merlin associates specifically with proteins displaying apparent molecular weights of 169 and 127 kD. These proteins were recovered in similar amounts at an approximate ratio of 1:5 with respect to Merlin. The patient-derived Leu64→Pro (L64P) Merlin mutant, which lacks tumor-suppressor activity, did not associate with these proteins. These two proteins were identified by means of mass spectrometry as DCAF1 (DDB1- and CUL4-associated factor 1) and DDB1 (DNA damage-binding protein 1), which are components of the multimeric CRL4DCAF1 (Cul4-Roc1-DDB1-DCAF1) E3 ubiquitin ligase. In this E3 ligase, DDB1 functions as an adaptor that links the substrate recognition component, DCAF1, to the Cul4 and Roc1 subunits of the ligase. Nano LC-MS/MS (liquid chromatography, tandem mass spectrometry) on gel stacks further revealed that wild-type (WT) Merlin, but not the L64P mutant, also interacts specifically with the CRL4 component Cul4B at a lower stoichiometry.

Slide 5: Merlin interacts with CRL4DCAF1

Our TAP results suggest that the active form of Merlin, but not the inactive L64P mutant, interacts with the CRL4DCAF1 E3 ubiquitin ligase. The CRL4 ligases (Cullin Ring Ligases 4) are composed of Roc1 (which is the catalytic subunit of the ligase), Cul4 (which functions as a scaffold and regulates positioning of the catalytic subunit for ubiquitin conjugation), the DDB1 adaptor, and one member of a large family of DCAFs that serves as the substrate receptor. DCAF1 was originally identified as the cellular receptor for the human immunodeficiency virus Vpr protein and was designated VprBP (Vpr binding protein). This protein is one of many WD40 substrate receptors that combines with DDB1 to form CRL4 ligases (19). CRL4 ligases are implicated in chromatin remodeling, DNA replication, and the DNA damage response. CRL4DDB2 promotes ubiquitylation of histones and DNA repair components such as DDB2 and XPC (xeroderma pigmentosum complementation group C). CRL4CSA (Cul4-Roc1-DDB1-CSA) ubiquitylates the chromatin-remodeling enzyme CSB (Cockayne syndrome B protein). CRL4Det1-Cop1 (Cul4-Roc1-DDB1-Det1-Cop1) promotes ubiquitylation and proteasome-mediated degradation of the transcription factor c-Jun (19-21).

Slide 6: The closed, but not the open, form of Merlin interacts with DCAF1 and DDB1

We set out to examine whether transition between the open and closed conformations regulates Merlin’s binding to CRL4DCAF1. Merlin-Ser518→Ala (S518A), which cannot be phosphorylated at this residue and is therefore stabilized in the closed conformation, associated efficiently with both DCAF1 and DDB1 in 293T cells. Conversely, the phosphorylation site mimetic Merlin-Ser518→Asp (S518D) mutant exhibited dramatically reduced binding to DCAF1 and DDB1. As expected, the tumor-derived L64P mutant and the cytoskeletal adaptor Ezrin displayed essentially no binding in this experiment.

Slide 7: Merlin and DCAF1 colocalize in the nucleus

We next set out to identify the compartment in which Merlin interacts with CRL4DCAF1. Merlin-deficient Meso-33 mesothelioma cells were cotransfected with Myc-tagged DCAF1 (Myc-DCAF1) and HA-tagged Merlin (HA-Merlin). Immunofluorescent staining revealed that Myc-DCAF1 localizes predominantly to the nucleus. HA-Merlin localized not only underneath the plasma membrane but also was also strongly present within the nucleus. These results were confirmed biochemically by using subcellular fractionation and immunoblotting.

Slide 8: Endogenous Merlin combines with CRL4DCAF1 in the nucleus

We suspected that either poor permeabilization of the nuclear membrane or epitope masking may have hindered nuclear detection of Merlin in previous immunofluorescence studies performed by others. In order to assess the localization of endogenous Merlin, we used an enhanced fixation and permeabilization procedure that included treatment with the detergents Triton X-100 and sodium deoxycholate. Using this protocol, a monospecific antibody to an N-terminal fragment of Merlin revealed prominent nuclear staining in MCF-10A cells. Short hairpin RNA (shRNA)–mediated knockdown of Merlin confirmed specificity of the antibody, and similar nuclear staining was observed in other cell types, including HeLa, HEI286 Schwann cells, and LP9 mesothelial cells. Again, nuclear localization was confirmed by using biochemical fractionation. Immunoprecipitation of endogenous Merlin from either the nuclear soluble (NS) fraction or the crude cytosolic and membrane (CM) fraction of HeLa cells revealed that Merlin interacts with CRL4DCAF1 only in the soluble nuclear fraction, despite abundant quantities of these proteins in both fractions. These results suggest that Merlin’s association with CRL4DCAF1 in the nucleus is promoted by association with a cofactor, dissociation of an inhibitor, or a posttranslational modification that occurs in the nucleus.

Slide 9: The closed form of Merlin accumulates in the nucleus

Previous studies in our lab and others revealed that signals from neighboring cells or the extracellular matrix regulate cell growth through the activation of PAK and subsequent phosphorylation of Merlin, thus disrupting the intramolecular interaction between Merlin’s carboxy terminus and amino-terminal FERM domain, maintaining it in an open conformation that is permissive of growth signaling (5, 6). Whereas it seems that PAK is primarily responsible for phosphorylating Merlin on Ser518, in recent years it has become evident that other kinases, such as PKA, also phosphorylate this same residue and that phosphorylation of Merlin at additional sites may affect its function (22). Inhibition of PAK signaling or activation of a Merlin phosphatase such as MYPT-1/PP1d leads to an accumulation of dephosphorylated, closed Merlin, which translocates to the nucleus by an unknown mechanism and binds to the CRL4DCAF1 E3 ubiquitin ligase (23, 24). Biochemical fractionation experiments revealed that dephosphorylated Merlin accumulates predominantly in the nucleus, whereas the majority of phosphorylated Merlin remains in the nonnuclear fraction regardless of whether cells are taken from confluent or low-density cultures for lysis. Moreover, the S518A mutant, which is resistant to phosphorylation at this site, accumulated in the nucleus considerably more than wild-type Merlin or the S518D phosphomimetic mutant. Combined with data revealing that Merlin’s FERM domain is necessary and sufficient for nuclear localization and indeed displays greater nuclear accumulation than full-length Merlin, it seems that unmasking the C terminus of Merlin inhibits its nuclear accumulation and that intramolecular associations of the FERM domain with the C-terminal tail promotes nuclear accumulation.

Slide 10: Merlin binds through its FERM domain to DCAF1 and inhibits CRL4DCAF1 ligase activity

To further investigate the interaction between Merlin and CRL4DCAF1, we carried out deletion mapping and found that the FERM domain of Merlin binds directly to a 90–amino acid C-terminal segment of DCAF1 in vitro. Cycloheximide chase experiments revealed that Merlin is a long-lived protein, and those Merlin mutants that do not interact with DCAF1 (Merlin-L64P and Merlin-S518D) do not exhibit increased stability relative to WT Merlin. Moreover, knockdown of DCAF1 did not stabilize Merlin. These results suggested that Merlin’s association with DCAF1 does not promote Merlin degradation, and therefore Merlin is not a polyubiquitylation substrate of CRL4DCAF1. In vivo ubiquitylation assays revealed that expression of recombinant Merlin reduced the ubiquitin-conjugating activity of DCAF1, expression of a truncated DCAF1 mutant that lacks residues necessary for Merlin binding displayed increased ubiquitin-conjugating activity, and Merlin depletion increased this activity. Together, these experiments reveal that Merlin is a negative regulator of CRL4DCAF1.

Slide 11: DCAF1 is required for hyper­proliferation of Merlin-deficient tumor cells

To examine whether Merlin suppresses cell proliferation by inhibiting CRL4DCAF1, we silenced DCAF1 in NF2 mutant Meso-33 mesothelioma cells using a smart pool of small interfering RNAs (siRNAs). Because loss of Merlin promotes progression through G1, we examined whether loss of DCAF1 affects this phase of the cell cycle. Bromodeoxyuridine (BrdU) incorporation experiments revealed that depletion of DCAF1 inhibits the ability of Meso-33 cells to progress through G1 and enter S phase in response to mitogens. In contrast, silencing of DCAF1 exerted a moderate growth inhibitory effect in normal mesothelial Met-5A cells, suggesting that Merlin-deficient cells are more sensitive to inactivation of DCAF1 than their normal counterparts. The differential sensitivity of normal and NF2 mutant cells to depletion of DCAF1 is consistent with the hypothesis that DCAF1 is part of an oncogenic signaling pathway that is hyperactivated by loss of Merlin.

Slide 12: DCAF1 mediates exit from contact inhibition and proliferation after inactivation of Merlin

To examine whether CRL4DCAF1 is required for exit from contact inhibition and resumption of cell cycling after inactivation of Merlin, we silenced Merlin in confluent human umbilical vein endothelial cells (HUVECs). As anticipated, this manipulation induced a significant fraction of confluent HUVECs to enter S phase. Depletion of DCAF1 with either of two distinct siRNAs suppressed this ectopic event, suggesting that Merlin induces contact inhibition by inhibiting CRL4DCAF1.

To confirm that CRL4DCAF1 functions downstream of Merlin, we conducted an additional genetic epistasis experiment in HEI-286 Schwann cells, which are immortalized but not neoplastic. Silencing of Merlin caused these cells to undergo accelerated progression through G1 and entry into S phase in response to serum stimulation, which is in agreement with the observation that Merlin contributes to restraining G1 progression. Depletion of DCAF1 by means of siRNA reverted the hyperproliferation caused by siRNA-mediated Merlin depletion but did not affect the growth rate of control cells expressing endogenous Merlin, providing additional evidence that CRL4DCAF1 mediates the growth-inhibitory effect of Merlin.

Slide 13: Merlin induces a tumor suppressive program of gene expression through inhibition of CRL4DCAF1

CRL4 ligases can regulate gene expression by promoting ubiquitylation of histones and recruitment of enzymes involved in chromatin remodeling or histone methylation or by inducing proteasome-mediated degradation of transcription factors. To examine whether Merlin controls gene expression through inhibition of CRL4DCAF1, we compared the gene expression program activated by expression of Merlin or depletion of DCAF1 in Nf2−/− FC-1801 mouse Schwannoma cells that were derived from conditional Nf2 knockout mice. Unsupervised clustering of 1566 probe sets differentially expressed upon expression of WT but not L64P mutant Merlin or upon specific depletion of DCAF1 revealed largely overlapping changes in gene expression. In fact, expression of Merlin and depletion of DCAF1 induced a concordant up-regulation or down-regulation of 855 probe sets. In spite of this considerable functional overlap, expression of Merlin and knockdown of DCAF1 also induced specific effects on gene expression. These results indicate that Merlin regulates gene expression largely through inhibition of CRL4DCAF1 but also suggest that Merlin has CRL4DCAF1-independent functions. Pathway analysis using the Ingenuity modeling software revealed that expression of Merlin or inactivation of CRL4DCAF1 result in similar effects on gene expression, causing a concordant down-regulation of genes that promote cell-cycle progression and up-regulation of genes involved in growth arrest and apoptosis.

Slide 14: Validation of tumor-derived FERM domain missense mutations

Most NF2 missense mutations map to the region that encodes the FERM domain. To further evaluate the relevance of Merlin’s interaction with CRL4DCAF1 for tumor suppression, we examined several of these mutations. Leu46→Arg (L46R), Phe62→Ser (F62S), Leu64→Pro (L64P), Leu141→Pro (L141P), Ala211→Asp (A211D), and Glu270→Gly (E270G) mutations have been found in two or more unrelated NF2 patients or are known to correlate with disease in individual families and are therefore bona fide pathogenic mutations (8). In contrast, the Gly197→Cys (G197C) mutation, which was identified in a single patient affected by mild bilateral Schwannoma, may represent a polymorphism or a passenger mutation. Structural evidence suggests that the L46R, F62S, L64P, L141P, and A211D mutations disrupt the hydrophobic core of subdomain A or B, whereas the E270G mutation simply removes a surface charge from subdomain C. In contrast, the G197C mutation can be accommodated by a slight adjustment of the surface loop in which it resides without overt effects on the overall structure of subdomain B and does not introduce or remove a charge. To evaluate their growth-suppressive function, we introduced WT or point-mutant versions of Merlin into Meso-33 cells. As anticipated, WT Merlin and Merlin-S518A suppressed progression through G1 and entry into S phase in these cells, whereas Merlin-S518D inhibited proliferation to a much smaller degree. All the bona fide pathogenic mutants were found to be devoid of growth inhibitory activity in Meso-33 cells. In contrast, Merlin G197C suppressed proliferation as effectively as WT Merlin, suggesting that this is not a pathogenic mutation.

Slide 15: Tumor-derived mutant forms of Merlin do not combine with CRL4DCAF1 in vivo

To examine the ability of tumor-derived mutants of Merlin to interact with CRL4DCAF1 in cells, we subjected Cos7 cells expressing FH-Merlin or Merlin point mutants to immunoprecipitation with Flag M2 antibody. None of the bona fide tumor-derived mutant forms of Merlin interacted with CRL4DCAF1 in vivo as determined with Western blotting by using antibodies recognizing endogenous DCAF1 and DDB1, whereas the G197C mutant associated with the ligase as efficiently as did WT Merlin under the same conditions, providing evidence that Merlin needs to combine with CRL4DCAF1 to suppress tumorigenesis.

Slide 16: Some tumor-derived Merlin mutants are unable to accumulate in the nucleus

To gain insight into the mechanism by which each of these tumor-derived Merlin mutants is defective in binding CRL4DCAF1 in vivo, we set out to analyze the ability of these pathogenic mutants to localize to the nucleus and to bind DCAF1 in vitro. Subcellular fractionation and Western blotting revealed that several of these mutants, including L46R, L64P, L141P, and A211D, accumulated in the nucleus to a much smaller extent than did WT or G197C Merlin.

Slide 17: Other mutants display defective binding to DCAF1 in vitro

To determine whether these tumor-derived missense mutants are defective in binding to the ligase in vitro, we engineered each mutation into a glutathione S-transferase (GST) fusion protein comprising the isolated FERM domain of Merlin, which binds efficiently to in vitro–translated (IVT) DCAF1. Using these GST-fusion Merlin mutants, we performed a pull-down using IVT DCAF1 as bait. Robust detection of IVT DCAF1 was made possible by incorporating biotinylated residues into the protein during translation. A211D bound DCAF1 in vitro almost as efficiently as did WT or the G197C control, suggesting that this mutant is defective in interacting with the ligase because of its lack of accumulation in the nucleus. The remainder of the mutants showed decreased binding ability. In particular, L46R, L64P, and E270G exhibit almost no binding.

Slide 18: Tumor-derived mutants summary

Our data show that tumor-derived pathogenic Merlin mutants are unable to interact with the ligase because of defects in nuclear localization that result from aberrant translocation and, in some cases, reduced protein stability, or because they are defective in binding to DCAF1, or both. Subsequent experiments using Merlin truncation mutants revealed that progressive deletions from the C terminus actually increase nuclear localization as well as DCAF1 binding. However, these mutants are defective in inhibiting ubiquitin ligase activity, suggesting that Merlin’s C-terminal tail is necessary to suppress ligase activity. Taken together, the results of our mutational analyses provide strong genetic evidence that Merlin needs to translocate into the nucleus, bind to DCAF1, and suppress CRL4DCAF1 ubiquitin ligase activity in order to mediate growth inhibition and tumor suppression.

Slide 19: Inactivation of DCAF1 suppresses tumorigenicity of Merlin-mutant tumor cells

To test whether CRL4DCAF1 is required for tumorigenesis driven by mutations in NF2, we silenced DCAF1 in Meso-33 meso­thelioma and FC-1801 Schwannoma cells. Knockdown of DCAF1 inhibited the ability of both mesothelioma and Schwannoma cells to grow on soft agar, suggesting that inactivation of CRL4DCAF1 inhibits the malignant properties of Merlin-deficient tumor cells.

Meso-33 cells and other Merlin-deficient mesothelioma cell lines did not give rise to tumors upon subcutaneous or intrapleural injection in mice. In contrast, FC-1801 cells, which robustly produce DCAF1, gave rise to sizable tumors within 4 weeks of subcutaneous injection into nude mice. Silencing of DCAF1 with either of two distinct shRNAs inhibited tumor growth in vivo by approximately 75% at day 24. This effect was highly specific, because moderate overexpression of a shRNA-resistant form of DCAF1 restored the ability of FC-1801 cells expressing one of the DCAF1-targeting shRNAs to give rise to subcutaneous tumors. In fact, the tumors generated by these cells grew at a much faster pace than did those generated by the parental FC-1801 cells, suggesting that elevated abundance of DCAF1 can promote tumorigenesis.

To explore the relevance of our findings to the pathogenesis of NF2, we examined whether DCAF1 affects the ability of primary human Schwannoma cells derived from NF2 patients to proliferate in vitro. Silencing DCAF1 by means of shRNA suppressed the ability of these cells to progress through G1 and enter S phase in response to mitogens (DMEM + 10% FCS + 0.5 μM forskolin + 10 nM β1-heregulin177–244 + 0.5 mM 3-isobutyl-1-methylxanthine + 2.5 μg/ml insulin). In contrast, knocking down DCAF1 did not substantially affect the proliferation of control human Schwann cells. Together, these results indicate that CRL4DCAF1 is necessary for hyperproliferation driven by NF2 mutations, anchorage-independent growth, and tumorigenesis.

Slide 20: Merlin suppresses tumorigenesis through inhibition of CLR4DCAF1

Our current model for the role of Merlin in contact inhibition and tumor suppression is that multiple extracellular cues converge to regulate Merlin’s phosphorylation status, subsequently regulating its activity. Closed, active Merlin translocates to the nucleus through an unknown mechanism and binds to the Cul4-Roc1-DDB1-DCAF1 E3 ubiquitin ligase, thus inhibiting its activity. Merlin overexpression and DCAF1 depletion elicit largely overlapping gene expression programs. Moreover, Merlin’s inhibition of CRL4DCAF1 appears to be necessary for tumor suppression.

Slide 21: Acknowledgments

This work was made possible by the efforts of several past and present laboratory members as well as several other laboratories and core facilities at Memorial Sloan-Kettering Cancer Center that are acknowledged here. We are grateful to our collaborators for their help and review of this research.

Editor’s Note: This contribution is not intended to be equivalent to an original research paper. Note, in particular, that the text and associated slides have not been peer-reviewed.

References

View Abstract

Stay Connected to Science Signaling

Navigate This Article