Sorbs1 and -2 Interact with CrkL and Are Required for Acetylcholine Receptor Cluster Formation

Abstract

Crk and CrkL are noncatalytic adaptor proteins necessary for the formation of neuromuscular synapses which function downstream of muscle-specific kinase (MuSK), a receptor tyrosine kinase expressed in skeletal muscle, and the MuSK binding protein Dok-7. How Crk/CrkL regulate neuromuscular endplate formation is not known. To better understand the roles of Crk/CrkL, we identified CrkL binding proteins using mass spectrometry and have identified Sorbs1 and Sorbs2 as two functionally redundant proteins that associate with the initiating MuSK/Dok-7/Crk/CrkL complex, regulate acetylcholine receptor (AChR) clustering in vitro, and are localized at synapses in vivo.

FIG 1

Crk/CrkL RNAi-treated myotubes display AChR clusters that are fewer and lack complexity. (A) Myotubes plated on laminin-111 were transfected with siRNAs directed against a nontargetable control transcript (NT), Crk, CrkL, or both Crk and CrkL. At 48 h posttransfection, myotubes were stained with fluorophore-conjugated bungarotoxin to label AChR aggregates and quantified by intensity threshold gating. In Crk RNAi-treated myotubes, AChR clusters were modestly reduced (∼19%) compared to those in NT-treated myotubes, Similarly, in myotubes treated with CrkL RNAi, AChR aggregates were reduced ∼33% compared to those in NT-treated myotubes. In contrast, myotubes treated with both Crk and CrkL RNAi displayed ∼88% fewer AChR aggregates than NT-treated myotubes (n = 3; means ± standard errors of the means [SEM] are shown; P < 0.05). (B) Myotubes were treated with siRNAs against an NT control or Crk/CrkL as described for panel A and then transfected 4 h later with in vitro-transcribed CrkL mRNA that was resistant to RNAi-mediated degradation. Myotubes transfected with both CrkL mRNA and Crk/CrkL RNAi partially rescued AChR aggregation (∼44% of NT-treated myotubes versus ∼15% for Crk/CrkL siRNA-treated myotubes) (n = 3; means ± SEM are shown; P < 0.05). (C) AChR aggregates that form in Crk/CrkL RNAi-treated myotubes are small and simplified. We scored AChR aggregates for the presence or absence of cortical actin puncta, markers of maturation, and found ∼27% of AChR aggregates from NT-treated treated myotubes harbored actin puncta and that the number was reduced to nearly zero in Crk/CrkL RNAi-treated myotubes (n = 3; means ± SEM are shown).

FIG 2

Exogenous Crk and CrkL are both enriched within AChR aggregates. In vitro-transcribed mRNAs encoding CrkL or both isoforms of the Crk gene, CrkI and CrkII, were fused to GFP and transfected into differentiated C2C12 myotubes. We analyzed soluble cellular lysates at 12 h posttransfection and confirmed the predicted sizes of our GFP fusion proteins by LiCor Odyssey infrared imaging using antibodies specific for CrkL or Crk. In parallel experiments, myotubes plated on laminin-111 were stained with fluorophore-conjugated bungarotoxin to label AChRs as well as Alexa Fluor 488-conjugated antibodies directed against GFP. Significant enrichment of CrkL-GFP, CrkI-GFP, and CrkII-GFP was observed within the area occupied by clustered AChRs and to a lesser degree the gaps within AChR clusters.

FIG 3

Identification of CrkL binding proteins from myotube lysates. (A) Cartoon representation of the TAP-tagged CrkL transcript that was tested. The transcript contains an N-terminal TAP tag, the CrkL coding sequence, an IRES-GFP, and an SV40 poly(A) sequence. (B) Myotubes transfected with serial dilutions of TAP-CrkL RNA and cell lysates were analyzed at 18 h posttransfection. CrkL RNA scales linearly over a 5-fold range and permits reconstitution of exogenous CrkL at levels similar to physiological CrkL protein levels. (C) TAP-CrkL proteins were purified from myotube lysates and subjected to SDS-PAGE and silver staining. The entire lane from either the sham or TAP-CrkL condition was cut into 18 individual bands and subjected to MS/MS analysis. Arrows indicate the size of the TAP-CrkL doublet band. The green bar indicates an area of the gel that was enriched with CrkL-specific interactors. Arrowheads indicate clear CrkL-specific proteins. (D) MS/MS summary of proteins identified under the two conditions. There were roughly 1,500 proteins in total, of which ∼400 proteins showed a >5-fold peptide enrichment specific to the TAP-CrkL condition. (E) Pie chart summary of CrkL-interacting proteins, summarizing proteins known, predicted, and not predicted to interact with CrkL. Of the predicted interacting proteins, many are predicted to interact with the SH3 domains of CrkL (79%).

FIG 4

Sorbs1 is enriched at AChR aggregates, and Sorbs1 RNAi blocks AChR clustering in vitro. (A) Treatment of myotubes with siRNA directed against Sorbs1 blocks AChR clustering in C2C12 cells. Montages containing sixteen fields at a magnification of 10× were analyzed with ImageJ software (NIH). (B) Sorbs1 siRNA significantly reduces Sorbs1 protein expression in myotubes. (C) Sorbs1 protein is highly enriched at sites where AChRs aggregate. (D) Agrin stimulates tyrosine phosphorylation of MuSK and Dok-7 at similar levels in myotubes treated with Sorbs1 siRNA.

FIG 5

Reduced AChR aggregation in a Sorbs1 knockout cell line. (A) Sorbs1 mutant C2C12 myotubes lack protein expression of all splice variants of Sorbs1 in myotubes. (B) Protein enrichment of Sorbs1 at AChR aggregates is specific to Sorbs1, as antibodies fail to label AChR aggregates in Sorbs1 mutant cells. (C) AChR aggregates are reduced 3-fold in Sorbs1 mutant cells. (D) Relative expression of Sorbs2 is upregulated 1.7-fold in Sorbs1 myotubes compared to wild-type control (n = 3, means ± SEM are shown). (E) A single muscle section was stained with bungarotoxin-Alexa Fluor 594 and an antibody to Sorbs1 or Sorbs2. Sorbs1/2 protein enrichment is detected at adult murine synapses in vivo.

FIG 6

Sorbs2 knockdown inhibits AChR clustering in muscle cells. (A) Myotubes were transfected with Sorbs1 or a nontargeting siRNA, and total RNA was isolated 48 h later. Total Sorbs2 transcript levels were assessed by quantitative PCR and found to be modestly elevated (∼25%) in myotubes treated with Sorbs1 siRNAs. (B) Myotubes were transfected with siRNAs directed against Sorbs1, Sorbs2, or a nontargeting control (siScr). Cell lysates were probed with antibodies directed against Sorbs2. Sorbs2 protein is significantly reduced by Sorbs2 RNAi treatment and modestly increased with Sorbs1 RNAi treatment. (C) Knockdown of Sorbs2 protein levels inhibits AChR clustering independent of Sorbs1.