Functional study and pathogenicity classification of PRRT2 missense variants in PRRT2‐related disorders

Abstract Aims PRRT2 variants are associated with various paroxysmal disorders. To date, more than 90 PRRT2 variants have been reported in PRRT2‐related disorders. Lack of functional study in majority of missense variants makes their pathogenicity uncertain. We aim to evaluate the clinical significance of PRRT2 missense variants by performing in vitro experiments. Methods We systematically reviewed PRRT2‐related disorders and summarized reported PRRT2 missense variants. Protein expression and subcellular localization of mutant PRRT2 were investigated in mammal cells. American College of Medical Genetics and Genomics (ACMG) guidelines were used to analyze the pathogenicity of PRRT2 missense variants. Results A total of 29 PRRT2 missense variants were identified in PRRT2‐related disorders. Ten variants were observed to affect both subcellular localization and protein level, three variants only affect membrane localization, and two variants only affect protein level. According to ACMG guidelines, 15 variants were finally classified as “likely pathogenic”, three as “benign”, three as “likely benign”, and eight as “uncertain significance” variants. The likely pathogenic variants were concentrated in the C‐terminal of PRRT2. Conclusions The pathogenicity of eight uncertain significance variants needs further investigation. C‐terminal of PRRT2 is crucial for its physiological function.

a variable frequency ranging from one per month to hundreds per day. Incomplete penetrance is usually observed in PRRT2 variant carriers. 6,7 The potential mechanisms of PRRT2 variants in PRRT2-related disorders remain largely unclear. PRRT2 consists of four exons and encodes a 340-amino-acid protein with two predicted transmembrane (TM) domains in the C-terminal and one proline-rich domain in the N-terminal. Enriched in cerebral cortex, cerebellum, substantia nigra, and hippocampus, PRRT2 protein was found to involve in synaptic transmission by modulating soluble N-ethylmaleimide sensitive factor attachment protein receptor (SNARE) complex. 8 To date, more than 90 variants in PRRT2 have been included in human gene mutation database (HGMD) and most of them are nonsense or frameshift variants, causing the truncation of the protein. Among them, c.649dupC (p.R217Pfs*8) is the most frequent variant, accounting for 78.5% of mutation carriers. 9 Truncated variants leading to conspicuous reduced protein level have been reported in various in vitro studies. 1,3,8,10 Besides the truncated variants, about 29 missense variants of PRRT2 are documented in PRRT2-related disorders.
However, only three of them has been functionally studied. [11][12][13] Therefore, the clinical significance of these alleles of PRRT2 in paroxysmal disorders is difficult to evaluate. It is of great value to perform the functional studies to address the pathogenicity of PRRT2 missense variants.
In this study, we summarized the reported missense variants in PRRT2 and performed functional experiments to investigate the alternation of subcellular location and protein expression of PRRT2 missense variants. We further assigned the pathogenicity of the missense variants according to the guidelines of American College of Medical Genetics and Genomics (ACMG). 14

| Missense variants in PRRT2
To identify all the reported missense variants within PRRT2, we searched the HGMD (http://www.hgmd.org) that provided the most up-to-date version of PRRT2 mutation and the PubMed (https ://www. ncbi.nlm.nih.gov/pubmed) from November 2011 when PRRT2 was first reported as a disease-causing gene and up to December 2017.
After that, the pathogenicity of PRRT2 missense variants was preliminarily evaluated according to the ACMG guidelines.

| Cell culture and transient transfection
HEK293T cells and HeLa cells were cultured in DMEM (HyClone) supplemented with 10% fetal bovine serum (Gibco) in a 5% CO 2 incubator at 37°C. Cells were seeded in suitable wells the day before transfection. Transient transfection was performed using the Lipofectamine 3000 according to the manufacture's protocol (Invitrogen). Forty-eight hours of cultivation was needed for the protein expression after transfection.

| Western blot analysis
To get the protein lysate, cells were rinsed with phosphate buffer saline (PBS) and harvested in lysis buffer. After centrifuging, the supernatants were collected. Western blot was performed as previously described. 15 The GFP (1:5000) and β-actin (1:5000; Sigma-Aldrich) antibodies were used. The blots were semiquantified by gel densitometry using the Photoshop.

| Statistical analyses
All the experiments were repeated at least three times independently. For the Western blot, proteins were normalized to β-actin.

Differences in the mean values between wild-type or mutant PRRT2
were analyzed by one-way ANOVA using GraphPad Prism software. The decreased or abnormally localized mutant protein was considered functionally impaired.

| Overview of the missense variants in PRRT2
After a systematic review of PRRT2-related disorders, we found a total of 29 missense variants in the literature (Table 1). Of which, 26 were heterozygous, two (p.P279L and p.R311W) were homozygous, and one (p.G305R) was described in both homozygous and heterozygous condition. 16 The predominant phenotype of these missense variants was PKD, while BFIS and ICCA were also documented.
Two variants (p.P138A and p.D147H) had a minor allele frequency (MAF) ≥0.05 in a population database 17 and were predicted to be benign by functional software (Table 1). Another variant (p.P216L) was found to possess 5.2% of 115 controls in an Australian study, 18 although it was predicted to be deleterious (Table 1). Of note, these three variants were not co-segregated with the disease in the family pedigrees reported previously. 17 In addition, eight variants (p.R266W, p.S275F, p.A291V, p.G305R, p.A306T, p.S317N, p.G324E, and p.I327M) were reported to co-segregate with PKD or BFIS in multiple affected family members. [18][19][20][21][22][23][24] They were predicted to be deleterious by SIFT, Polyphen-2, and MutationTaster. All the eight variants but one (p.G305R) were absent in population database and our control. However, they were classified as uncertain significance variants for lack of sufficient evidence. Other 18 variants were also classified as uncertain significance variants combining the population MAF and prediction data.  (Figure 1).

| Missense variants decreased the protein level
The remaining 17 variants, including the benign ones, had undifferentiated expression of PRRT2 protein as wild-type (Figure 1). p.G324E) lost membrane targeting and were located in the cytoplasm (Figure 2A,B), indicating the alternation of subcellular localization of these missense variants. The remaining 16 variants were still retained in plasma membrane (Figure 2A,B). The red-fluorescent labeling plasma membrane was shown in the Figure S1.

| Classification of the missense variants in PRRT2
Decreased protein expression or alternation of subcellular localization was considered functionally impaired. We assigned the pathogenicity of the reported missense variants of PRRT2 with functional data. As a result (Table 1) 29 In our study, only the C-terminal amino acid change could result in mislocalization. Possibly, membrane orientation of PRRT2 is imposed by the C-terminal, especially the TM domain.

| D ISCUSS I ON
In conclusion, a total of PRRT2 missense variants reported is firstly assessed by the ACMG guidelines. It will be of great value for its instructive and meaningful role in clinical molecular diagnosis.
Missense variants could decrease the protein level and/or impair plasma membrane localization. C-terminal of PRRT2 is crucial for its physiological function. Functional study is vital for the classification and potential mechanisms associated with PRRT2 should be further explored.

ACK N OWLED G M ENTS
We sincerely thank the participants for their help and willingness to participate in this study. We also thank Novogene Company to share their whole exome sequencing data with us.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.