Not all neuroligin 3 and 4X missense variants lead to significant functional inactivation

Abstract Introduction Neuroligins are postsynaptic cell adhesion molecules that interact with neurexins to regulate the fine balance between excitation and inhibition of synapses. Recently, accumulating evidence, involving mutation analysis, cellular assays, and mouse models, has suggested that neuroligin (NLGN) mutations affect synapse maturation and function. Previously, four missense variations [p.G426S (NLGN3), p.G84R (NLGN4X), p.Q162K (NLGN4X), and p.A283T (NLGN4X)] in four different unrelated patients have been identified by PCR and direct sequencing. Methods In this study, we analyzed the functional effect of these missense variations by in vitro experiment via the stable HEK293 cells expressing wild‐type and mutant neuroligin. Results We found that the four mutations did not significantly impair the expression of neuroligin 3 and neuroligin 4X, and also did not measurably inhibit the neurexin 1–neuroligin interaction. These variants might play a modest role in the pathogenesis of autism or might simply be unreported infrequent polymorphisms. Conclusion Our data suggest that these four previously described neuroligin mutations are not primary risk factors for autism.

In our previous study, we identified one NLGN3 and three NLGN4X variants (p.G426S in NLGN3; p.G84R, p.Q162K, and p.A283T in NLGN4X) (Xu et al., 2014). All four were missense and located in the conserved extracellular noncatalytic acetylcholinesterase homology domain, which is essential for binding to neurexin and triggering synaptic activity. Prediction of the effect of these substitutions using MutationTaster (http://www.mutationtaster.org/), Poly-Phen-2 (http://genetics.bwh.harvard.edu/pph2/), and SIFT (http://sift.jcvi. org/) revealed that p.G84R and p.A283T were "probably damaging"; however, p.G426S and p.Q162K were likely to be "benign" (Xu et al., 2014). The four variants might impair the functional properties of neuroligin related to synaptic homeostasis, thereby increasing predisposition to autism. To test this hypothesis, here we analyzed the functional effect of these four missense variants.

| Plasmids and mutagenesis
Vectors expressing unmodified human neuroligin 3 (NM_018977.2) and neuroligin 4X (NM_020742.2) were subcloned into a green fluorescent protein (GFP)-tagged (C terminal, pEGFP-N1) vector, and verified by sequencing. Vectors expressing a soluble IgG-NX1β fusion protein (in a pCMV5 backbone) were provided by Professor Thomas C. Südhof (Zhang et al., 2009 The molecular weight for all of the GFP-NL3 fusion proteins is ≈110 kDa(including the wild-type protein, mutant protein, and the positive-control). The molecular weight for the GFP-NL4-WT and the three GFP-NL4-mutant fusion proteins is ≈120 kDa. The molecular weight for the GFP-NL4-R87W fusion proteins is ≈90 kDa. The molecular weight for the loading control beta-actin is ≈42 kDa.

| Immunofluorescence
Stable HEK293 cell lines were grown on coverglasses in a 24-well plate. After 24 hr, the culture medium was removed and the cells washed twice with 1× PBS. The cells were then fixed with 4% paraformaldehyde/4% sucrose for 15 min, permeabilized with 0.2% Triton X-100 in PBS for 10 min, and incubated with 5% bovine serum albumin in PBS for 30 min. The cells were then incubated with the indicated primary antibody (Sigma or Invitrogen) for 1 hr followed by Lakes, New Jersey) antibodies were used. Images were captured using a Leica TCS SP5 confocal microscope.

| Assay of cell surface biotinylation
Stable HEK293 cell lines were grown on coverglasses in 6-cm plate.
After 24 h, the culture medium was removed and the cells washed twice with 1× PBS. For surface biotinylations, the cells were washed twice with ice-cold PBS, pH 8.0 and incubated with 2 mg/ml sulfo-NHS-LC-biotin (Pierce) in PBS, pH 8.0 for 30 min on ice. in at 4°C) and the beads washed 3-5 times with 1× PBS (with protease inhibitors).
Then, 100 μl Quenching Solutio (with protease inhibitors) was added to terminate reaction. Subsequently, cells were washed with 1× PBS (with protease inhibitors). The proteins were extracted for 10 min at 4°C in Lysis Buffer (PBS containing 1% Triton X-100, 0.2% SDS, 5 mmol/L EDTA, 2 mmol/L DTT, and protease inhibitors), and the lysate saved in 1.5-ml centrifuge tubes. Insoluble material was removed by centrifugation (1,000g for 5 min), and the supernatant was quantified by Pierce BCA protein assay and retained for the biotin labeling. The 100 μl NeutrAvidin Agarose (Thermo Fisher Scientific) were washed three times with 1 ml 1× PBS (with protease inhibitors) by centrifugation (1,000g for 3 min). Then part of the supernatant (including 500 μg protein) was mixed with the NeutrAvidin Agarose and rotated at 4°C for at least 6 hr. The supernatant was removed by centrifugation (1,000g for 4 min at 4°C) and the beads washed 3-5 times in 1× PBS (with protease inhibitors) by centrifugation (1,000g for 4 min at 4°C). 1 mol/L HEPES-NaOH were added. Insoluble material was removed by centrifugation (2,500g for 20 min), and the supernatant was retained for the purification assay. Each binding assay used 10 μl IgG beads. The beads were washed three times with 1 ml 1× PBS (with protease inhibitors) by centrifugation (6,500g for 5 min). Then the supernatant was mixed with the beads and rotated at 4°C for at least 6 hr. The supernatant was removed by centrifugation (6,500g for 20 min at 4°C) and the beads washed 3-5 times in 1× PBS (with protease inhibitors). Then elution buffer (glycine, 100 mmol/L) was added and the tubes rotated at room temperature for 30 min.
The IgG beads were removed by centrifugation (6,500g for 5 min at 4°C). Neutralizing buffer (

F I G U R E 2
The two variants did not block the transport of neuroligin from ER to cell membrane. HEK293 cells expressing GFP-tagged wild-type neuroligin (NL3-WT and NL4-WT), mutant neuroligin (NL3-G426S and NL4-Q162K), and positive control (NL3-R451C and NL4-R87W) were immunostained with antibodies against the GFP tag (first column, green) and against the Golgi-marker GM130 (second column, red). The nucleus was labeled by DAPI (4′,6-diamidino-2-phenylindole) (third column, blue). Both the wild-type and mutant neuroligin are transported efficiently to the cell surface, whereas the positive control accumulates in the Golgi body, as confirmed by the overlap with the Golgi body-resident protein calnexin. The scale bar is 5 mm

| Neurexin 1-neuroligin interaction assay
The interaction between neurexin 1 and neuroligin was tested by

| Statistical analysis
In this study, all statistical analysis was performed using Prism 5 software (GraphPad). The two tailed Student's t-test was used for the comparison of the two groups. The two-way ANOVA and Dunnett's multiple comparison test was used for the comparison between the three groups or more than three groups. p values lower than .05 were considered to be significant.

| RESULTS
To study the functional effect of the four mutations, we analyzed the expression level, subcellular localization, degradation, and interaction with neurexin, using in vitro experiments in HEK293 cells stably expressing wild-type or mutant neuroligin 3 or neuroligin 4X.
Immunofluorescence showed that mutant neuroligin 3 and neuroligin 4X did not colocalized with calnexin and GM130. In contrast, the two positive control mutant proteins colocalized with calnexin but not colocalized with GM130, indicating that they were mostly retained in the ER. Cell surface biotinylation indicated that mutant neuroligin 3 and neuroligin 4X were abundant in the cell surface Because the four mutant proteins did not impair the folding of neuroligin 3 and neuroligin 4X and also did not inhibit their surface export, we measured the ability of these mutant proteins to bind to recombinant IgG-NX1β fusion protein. We observed that the mutant neuroligin 3 and neuroligin 4X showed a binding affinity for their cognate partner, β-neurexin, that was similar to that of wild-type neuroligin ( Figure 6). Therefore, neurexin-binding was not impaired by any of the four mutations. However, the two positive controls did not bind to recombinant IgG-NX1β fusion protein.

| DISCUSSION
Autism is a complex neurodevelopmental disorder in which mutations in neuroligin may play a role. In this study, we used in vitro F I G U R E 4 The four variants do not destabilize neuroligin 3 and neuroligin 4X. HEK293 cells expressing GFP-tagged wild-type neuroligin (NL3-WT and NL4-WT), mutant neuroligin (NL3-G426S, NL4-G84R, NL4-Q162K, NL4-A283T), and positive control (NL3-R451C and NL4-R87W) were treated with 20 μg/ml CHX for 0 to 12 hr (for NL3) or 0 to 24 hr (for NL4X). Inhibition of protein synthesis, and the accumulation of wild-type neuroligin, mutant neuroligin, or positive control were evaluated by Western blotting. Beta-actin was used as a loading control.
(a) Degradation rate assay of mutant neuroligin 3. (b) Degradation rate assay of mutant neuroligin 4X. (c, d) Quantification for the assay of mutant protein degradation rate. *** referred to p < .0001 assays in HEK293 cells to demonstrate that four identified neuroligin mutations did not affect protein expression or interaction with neurexin 1. We conclude that these four variants either play only a modest role in predisposition to autism, or are previously unreported infrequent polymorphisms rather than pathogenic mutations. We must bear in mind, however, that we cannot recapitulate the situation in a human patient: we were studying a single factor in isolation, ignoring other genetic, epigenetic, and environmental factors that may play a contributory role in disease causation.
In conclusion, our data indicate that our four previously described mutations in neuroligins 3 and 4X are not primary risk factors for autism.

ACKNOWLEDGMENTS
We thank Professor Thomas C. Südhof for presenting the vectors expressing soluble IgG-NX1β fusion protein (in pCMV5 backbone). We also thank patients and their families for agreeing to participate in the previous study and special professors whose participation made this project possible. We were appreciated for the help and advices of our colleagues. We will make their data, analytic methods, and study materials available to other researchers and we can supply these by E-mail or post. We did not preregister the research without an analysis plan in an independent, institutional registry. This study was supported by the National Natural Science Foundation of China (Grant No. 31301023).
F I G U R E 5 The four variants do not change the degradation pathway of neuroligin 3 and neuroligin 4X. HEK293 cells expressing GFPtagged wild-type neuroligin (NL3 and NL4X), mutant neuroligin (G426S, G84R, Q162K, and A283T), or positive control (R451C and R87W) were pretreated with the proteasome inhibitor MG132 (10 μmol/L) or the lysosome inhibitor chloroquine (CQ; 100 μmol/L) for 0 hr (untreated) or 10 hr. The accumulation of wild-type neuroligin, mutant neuroligin, and positive control was evaluated by Western blotting. Beta-actin was used as a loading control. (a) Degradation pathway assay of mutant neuroligin 3. (b, c) Degradation pathway assay of mutant neuroligin 4X. (d, e) Quantification for the assay of mutant protein degradation pathway. ** referred to p < .001 F I G U R E 6 The four variants do not impair the interaction of neuroligin 3 and neuroligin 4X with neurexin 1. The neurexin 1-neuroligin interaction was measured by IP. The soluble IgG-NX1β fusion protein could be immunoprecipitated by anti-IgG. Neuroligin could be detected by anti-GFP. Inputs: lysates of HEK293 cells expressing GFP-tagged wild-type neuroligin, mutant neuroligin, or positive control. IP: lysates after IP assay. (a) Assay for mutant neuroligin 3. (b) Assay for mutant neuroligin 4X