The NF2 tumor suppressor merlin interacts with Ras and RasGAP, which may modulate Ras signaling

Inactivation of the tumor suppressor NF2/merlin underlies neurofibromatosis type 2 (NF2) and some sporadic tumors. Previous studies have established that merlin mediates contact inhibition of proliferation; however, the exact mechanisms remain obscure and multiple pathways have been implicated. We have previously reported that merlin inhibits Ras and Rac activity during contact inhibition, but how merlin regulates Ras activity has remained elusive. Here we demonstrate that merlin can directly interact with both Ras and p120RasGAP (also named RasGAP). While merlin does not increase the catalytic activity of RasGAP, the interactions with Ras and RasGAP may fine-tune Ras signaling. In vivo, loss of RasGAP in Schwann cells, unlike the loss of merlin, failed to promote tumorigenic growth in an orthotopic model. Therefore, modulation of Ras signaling through RasGAP likely contributes to, but is not sufficient to account for, merlin’s tumor suppressor activity. Our study provides new insight into the mechanisms of merlin-dependent Ras regulation and may have additional implications for merlin-dependent regulation of other small GTPases.


Fig. S3
In vivo interaction between merlin and KRas4B demonstrated by proximity biotinylation, related to Fig.  2. a Indicated constructs (with an N-terminal OLLAS tag) were transfected into 293 cells; ~24 h later, media was replenished with biotin added; ~48 h after transfection, cells were lysed for Streptavidin (SA) pulldown (PD), analyzed by Western blotting. Note that BioID2-KRas4B biotinylated merlin more efficiently than BioID2, even though BioID2 alone exhibited much higher non-specific biotinylation activity. b Overall biotinylation activity assessment for Fig. 2d

Fig. S6
RasGAP may prefer to bind merlin rather than ERMs. Western blot analysis of GST-RasGAP (aa 157-1047) pulldown of endogenous ERMs and merlin from 293 lysate. GST-fusion protein inputs were stained with Ponceau S. Ezrin was probed first with the 3C12 antibody (mouse monoclonal), with no specific signal detected in the pulldown. The membrane was treated with 30% H 2 O 2 to inactivate the bound anti-mouse IgG-HRP; then it was co-incubated with three rabbit monoclonal antibodies against radixin, moesin, and merlin, respectively (our prior experiments have showed that only ezrin and radixin partially overlapped on SDS-PAGE, whereas the others could be separated).
Moesin was non-detectable in the pulldown. Note that the signal from merlin was relatively stronger in the pulldown but weaker in the lysate, compared with those from radixin or moesin, suggesting more efficient binding to RasGAP by merlin, although the real expression levels of these proteins could not be inferred. # indicates a protein that strongly bound to glutathione sepharose under the N+T buffer condition, which was likely EF1 gamma, identified by mass spectrometry. Note that EF1 gamma contains a GST-like domain [13]; here it could serve as an additional control for sepharose inputs. N+T: 2% NOG + 1% Triton X-100.
The lentiviral vector pCDH-CMV.Bsd was generated by replacing the puromycin selection marker in pCDH-CMV-MCS-EF1-puro (System Biosciences) with blasticidin selection marker BSD, in two steps. First, the site upstream of the puromycin resistant gene was mutated into Avr II site. Second, the puromycin resistant gene was digested out with Avr II and Sal I; Bsd ORF was PCR amplified using a pcDNA6 vector (Invitrogen) as the template, digested and ligated into the two sites.
Lentiviral vectors with an SBP tag or without a tag were constructed as described in [9]. We later also constructed vectors with a long linker (LL) following the SBP-tag (SBPLL) or with an OLLAS tag. BioID2 cDNA [10] was amplified by PCR and cloned into the lentiviral vector with an EFS promoter and an OLLAS tag (pCDH-EFS-OLLAS).
Next, the BsrG I-Not I fragment (the MCS region) of pCDH-CMV-EGFP-C2.Bsd was replaced by a customized MCS.
NF2 iso2 tail with BamH I and Not I sites added was amplified by PCR and cloned into pGEX-4T-1. NF2 iso2 full-length ORF was re-assembled using iso1 ORF and a 3' fragment from iso2, and cloned into vectors for mammalian or bacterial expression.
pcDNA3/GAP was modified from h-RasGAP.dn3 by mutagenesis to delete the partial 5' and 3' UTR and to change the 3' EcoR I into BamH I. Subsequently, RasGAP ORF was sunbcloned into pGEX-4T-1, pEGFP-C2, pCerl-C2, and the lentiviral vector with an EF1a promoter using the 5' EcoR I site and an appropriate 3' site. All the deletion and point mutation mutants were generated by either mutagenesis or PCR cloning. More details are available upon request.
For co-transfection experiments, the plasmids were used at 1:1 ratio unless otherwise stated.

Lentivirus production and transduction
293T cells were transiently transfected with a lentiviral vector, the packaging and the envelope plasmids in the ratio of 2

Preparation of cell extracts for Western blotting analysis
Cells were lysed on plates with either 2× SDS sample buffer (100 mM Tris-Cl, pH 6.8, 4% SDS, 0.01% bromophenol blue, 20% glycerol) supplied with 30-50 mM DTT, or with lysis buffers for pulldown or co-immunoprecipitation (Co-IP). If lysed with a lysis buffer, a portion of clarified lysates were further mixed with equal or higher volume of 2× SDS sample buffer supplied with 30-50 mM DTT. Samples were boiled at 98°C for 8-10 min before being loaded into an SDS-PAGE gel.

Western blotting/Immunoblotting
Following SDS-PAGE, proteins were transferred onto nitrocellulose membranes by tank For sequential probing using antibodies from different species, before the second probing, signals from the first round of Western blotting were inactivated by incubation with 30% H2O2 at 37°C for 30 min [11].

Active Ras pulldown with cell lysates
The procedure was performed as per GST pulldown (see later), albeit with different lysis/wash buffers and GST-Raf1 RBD (either from Pierce Active Ras Pull-Down Kit or homemade) or GST-BRaf RBD (homemade) as the bait.

PDGF treatment prior to active Ras pulldown
MSC cells were seeded on poly-L-lysine-coated plates and cultured for ~30 hours; growth media were replaced with DMEM:F12-HAM (1:1) plus 1 μM forskolin, then cells were incubated overnight for growth factor starvation. For PDGF stimulation, media was replaced with the stimulation medium (the starvation medium plus 10 ng/ml PDGF-BB [Calbiochem]) equilibrated at room temperature, and cells incubated at 37°C for intended durations. Media were then removed and cells washed once with ice-cold PBS and lysed on ice; lysates were clarified by centrifugation at 15,000-20,000 g for 15 min at 4°C for active Ras pulldown.
Non-stimulated cells were processed likewise, without prior medium change before wash with PBS.

Proximity biotinylation by BioID2
Constructs were transfected into 293 cells cultured on 6-well plates with PEI at the ratio of 4:1 to DNAs. In Fig. S3a blotting. Western blotting with streptavidin-HRP was described in [12].
For RasGAP Co-IP, 2% n-Octyl-β-D-glucopyranoside (NOG; Carl Roth, Germany) was supplemented; for overexpressed merlin and RasGAP Co-IP, 2% NOG plus 1% Triton X-100 were supplemented in the lysis/binding buffer, 1% Triton X-100 plus 0.1% Sodium deoxycholate (DOC; Carl Roth, Germany) were supplemented in the washing buffer. 1× cOmplete protease inhibitor cocktail was included for all lysing steps, but omitted for washing steps. Briefly, cells were washed once with ice-cold PBS and lysed on ice; lysates were cleared by centrifugation at 15,000-20,000 g for 15 min at 4°C; for each IP, the supernatant was incubated with an antibody immobilized on Gamma-Bind G sepharose (GE Healthcare) for 0.5-2 hours with rotation at 4°C; the sepharose was spun down by centrifugation at 2,000 g for 30 s at 4°C and washed three or four times with ice-cold lysis buffer. IP were eluted with 2× SDS sample buffer by boiling, resolved on SDS-PAGE and immunoblotted with appropriate antibodies.

Small-scale purification of GST-tagged proteins from E. coli for pulldown
All the constructs were cloned in pGEX-4T-1. In brief, plasmids were transformed into E. coli BL21 (DE3) pLysS or Rosetta 2 (DE3); bacteria were grown in 5 ml of LB medium with 100 μg/ml ampicillin with shaking at 37°C, overnight. A 1-ml sample of each overnight culture was inoculated into 50 or 100 ml of 2× YT medium with 100 μg/ml ampicillin, and grown with shaking at 37°C for 6-10 hours until the E. coli density was close to saturation. The shaker temperature was lowered to 16°C and 0.2 mM IPTG was added to induce protein expression overnight. All subsequent steps were carried out at 4°C or on ice. Bacteria were pelleted by centrifugation for 3 min at 12,000 g, then either frozen at -80°C for later use or processed directly for lysis. For lysis, pellets were resuspended in ice-cold lysis buffer (PBS/1% Triton X-100/protease inhibitor cocktail, 1 ml for 50-ml culture) and lysed by sonication (30% amplitude, 10 s impulse, 10 s break, 1 min total impulse; Digital sonifier S-450D, Branson). In our later work, purified proteins (on beads) were stored in PBS/0.05% Triton X-100/50% glycerol/5mM DTT at -20°C.

Preparation of E. coli lysates for Western blotting or pulldown
SBP-merlin (isoforms and mutants) were expressed in E. coli; the preparation of lysates was same as described above for small-scale purification of GST-tagged proteins, except that the bacteria were resuspended in PBS/0.6% CHAPS/protease inhibitor cocktail before sonication. The clarified lysates were stored at -80°C for later use.

GST pulldown
GST-tagged proteins expressed in E. coli were immobilized on glutathione 4B sepharose (GE Healthcare). GST pulldown was performed in essentially the same way as Co-IP. For merlin:RasGAP interaction, 2% NOG plus 1% Triton X-100 were supplemented in the lysis/binding buffer; 1% Triton X-100 plus 0.1% DOC was supplemented in the wash buffer.
For merlin:Ras interaction, 0.025-0.05% DDM was supplemented in the binding/wash buffer. For experiments using CHAPS, 0.6-0.7% CHAPS was supplemented in the lysis/binding buffer.

Large scale purification of His-tagged proteins from E. coli
pET15TEV-HRas 1-166 was used to express N-terminal His-tagged HRas 1-166 protein; pET30a-NF2 was used to express N-terminal His-S-tagged full-length merlin. The procedure was similar to large scale purification of GST-tagged proteins, with the following modifications: for merlin purification, the lysis buffer was PBS with 1% Triton X-100, 1 mM DTT, 30 mM imidazole and the protease inhibitor cocktail; for HRas 1-166 purification, the lysis buffer was further supplemented with 2.5 mM MgCl2 and 100 μM GDP; protease inhibitor and GDP were omitted for wash; Ni-Sepharose 6 Fast Flow (GE Healthcare) was used for binding His-tagged proteins; for HRas 1-166, ProTEV Protease (Promega) was used to cleave the His-tag on beads (ProTEV Protease has a His tag and so is retained by Ni-Sepharose); for merlin, it was either eluted with PBS/500 mM imidazole and dialyzed to remove imidazole, or the His tag was cleaved on beads with thrombin as described for cleaving GST tag; for wash and storage of HRas 1-166, 2.5 or 2 mM MgCl2 was always included.

Purification of SBP-tagged proteins from E. coli
The procedure was similar to that described above for small-scale purification, with the following modifications: Streptavidin Agarose was used to bind SBP-tagged protein; the proteins were eluted with biotin elution buffer (25 mM Tris pH 7.5, 150 mM NaCl, 5 mM biotin, 2 mM DTT, 50% glycerol) at 4°C or -20°C, and the supernatants containing eluted proteins were transferred to new Protein LoBind tubes and stored at -20°C.

Far-western blotting
GST-RasGAP fragments were separated by SDS-PAGE and transferred onto a nitrocellulose membrane. The membrane was stored in TBS/T at 4°C for one day to allow certain spontaneous refolding of the proteins. The membrane was blocked with 5% milk in TBS/T for 1 hour, and incubated with purified His-S-merlin (2 μg/ml in PBS/0.01% DOC) for 1 hour at room temperature. Afterwards, the membrane was washed 3 times with PBS/0.01% DOC. Bound merlin was detected with NF2 (B-12) antibody, like a routine immunoblotting. The GST-fusion protein inputs were detected by Ponceau S staining.

In vitro Ras GAP activity assay by GST-Raf1 RBD pulldown
All proteins were purified from E. coli. Equal molar amount of RasGAP fragments (1 μg for FL RasGAP) were pre-incubated with or without 1 μg of merlin in 10 μl of GAP reaction buffer (50 mM Tris pH 7.5, 148 mM KCl, 2.5 mM MgCl2, 1 mM DTT), containing 2 μg of BSA in 1.5 ml-protein LoBind tubes (Eppendorf), for 1 hour at room temperature and transferred on ice; 1.6 μg of HRas 1-166 was preloaded with 1 mM GTP in 20 μl of buffer (50 mM Tris