Skip Navigation


Human Molecular Genetics Advance Access originally published online on October 26, 2006
Human Molecular Genetics 2006 15(23):3436-3445; doi:10.1093/hmg/ddl421
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
15/23/3436    most recent
ddl421v2
ddl421v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Sirerol-Piquer, M. S.
Right arrow Articles by Pérez-Tur, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sirerol-Piquer, M. S.
Right arrow Articles by Pérez-Tur, J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

The epilepsy gene LGI1 encodes a secreted glycoprotein that binds to the cell surface

Maria Salomé Sirerol-Piquer1,{dagger}, Ana Ayerdi-Izquierdo1,{dagger}, José Manuel Morante-Redolat1, Vicente Herranz-Pérez1, Kristy Favell2, Philip A. Barker2 and Jordi Pérez-Tur1,*

1 Unitat de Genètica Molecular, Departament de Genòmica i Proteòmica, Institut de Biomedicina de València-CSIC, València, Spain and 2 Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada

* To whom correspondence should be addressed at: Institut de Biomedicina de València-CSIC, Unitat de Genètica Molecular, C/Jaume Roig, 11, E46010 València, Spain. Tel: +34 963391755; Fax: +34 963393774; Email: jpereztur{at}ibv.csic.es

Received July 19, 2006; Accepted October 20, 2006


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Autosomal dominant lateral temporal epilepsy (ADTLE) is a partial epilepsy caused by mutations in LGI1, a multidomain protein of unknown function. To begin to understand the biological function of LGI1, we have determined its pattern of glycosylation, subcellular expression and capacity for secretion. LGI1 is expressed as two different isoforms in the brain, and we show that the long isoform is a secreted protein, whereas the short isoform is retained in an intracellular pool. ADLTE-related mutants of the long form are defective for secretion and are retained in the endoplasmic reticulum and Golgi complex. Finally, we show that normal secreted LGI1 specifically binds to the cell surface of differentiated PC12 cells. We propose that LGI1 is a secreted factor important for neuronal development and that ADTLE is a disease that results from the loss of regulation in the protein available either extracellular or intracellularly.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Autosomal dominant lateral temporal epilepsy (ADLTE, OMIM no. 600512 [OMIM] ) (also known by some authors as autosomal dominant epilepsy with auditory features) is a neurological disorder characterized by auditory auras and focal seizures affecting the lateral temporal lobe of the brain. The disease was mapped to a 3 cM locus on human chromosome 10 by linkage analysis (1,2). In 2002, Morante-Redolat et al. (3) and Kalachikov et al. (4) demonstrated that mutations in LGI1 (leucine-rich glioma-inactivated 1) cause ADLTE. Numerous mutations have been found throughout the protein coding regions of this gene, but their effect on protein function is unknown. (3–11).

LGI1 was originally localized to the breakpoint of a balanced translocation between chromosomes 10 and 19 (t{10;19}{q24;13}) present in the glioblastoma cell line T98G (12). Because LGI1 showed low expression levels in human high-grade glial tumor samples and glioblastoma cell lines compared with normal brain samples, it was proposed to act as a tumor suppressor (12). However, more recent studies have challenged this view and concluded that LGI1 loss of function is unlikely to play a role in glial tumor progression (13). The protein encoded by LGI1 consists of a signal peptide and three leucine-rich repeats flanked by two cysteine-rich regions in the N-terminal part of the protein. Additionally, a novel domain composed of seven tandem repeats of 50 amino acids, termed EPTP repeats, was found in the C-terminal part of the protein. These EPTP repeats likely form a seven-bladed beta-propeller structure (14). The EPTP repeat is also found in three paralogues of LGI1 (LGI2, LGI3 and LGI4) as well as in two otherwise unrelated genes (TNEP1 and VLGR1) and together, these constitute the EPTP superfamily (14,15).

The function of LGI1 remains unclear. Recent studies from Senechal et al. (16) have shown that LGI1 is a secreted protein, consistent with our earlier prediction (3), and some studies have suggested that LGI1 is involved in regulation of cell movement, growth and survival (17,18). Interestingly, LGI4 has also been shown to be a secreted protein that may play a role in peripheral nerve myelination (19), whereas Schulte et al. (20) have provided evidence that LGI1 is a membrane-associated protein that regulates the activity of the Kv1.1 voltage-dependent potassium channel subunit.

In this work, we confirm that LGI1 is a secreted glycoprotein even in neuronal-like cells and show that a truncated naturally occurring spliced isoform is retained within cells. We show that the secretion of the full-length isoform depends on the integrity of each of its domains and that disease-causing ADLTE mutations block LGI1 secretion. Moreover, we provide evidence for an LGI1 receptor on cell surfaces. Together, our data indicate that manifestation of ADLTE is related to defects in the LGI1 secretion.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
The LGI1 isoforms are glycoproteins with different cellular fates
Human LGI1-FlagLONG and LGI1-FlagSHORT, shown schematically in Figure 1A, were transiently transfected in HEK293T cells, the extracellular media were concentrated and the cell extracts were immunoprecipitated with anti-Flag monoclonal antibodies. LGI1-FlagLONG was clearly present in both lysates and media, whereas LGI1-FlagSHORT was only detected in cell lysates, suggesting that only LGI1-FlagLONG was capable of being secreted (Fig. 1B). Similar experiments were performed on T98G cells, a glioblastoma-derived cell line lacking LGI1 expression (Fig. 1B), and on COS7 and CHO cells. In all cases, LGI1-FlagLONG accumulated in media, whereas LGI1-FlagSHORT did not, indicating that the secretion of LGI1-FlagLONG does not show cell-type specificity. In some instances, a minor band is observed in cell lysates over-expressing the LGI1-FlagLONG or the LGI1-FlagSHORT isoforms. The identity of such band has not been further investigated; therefore, we cannot exclude the possibility that LGI1 is being post-transcriptionally processed at a position close to the C-terminus end of the protein which is common to both isoforms. Finally, we tested the ability of LGI1 to be secreted on differentiated PC12 cells, a neuronal-like cell type. As is shown in Figure 1B, NGF-differentiated PC12 cells are also capable of secreting LGI1-FlagLONG, but retaining LGI1-FlagSHORT as well as an ADLTE-causing mutation (LGI1-Flag758delC) (data not shown).


Figure 4211
View larger version (40K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. LGI1-FlagLONG is secreted from HEK293T cells. (A) Schematic representation of the LGI1-Flag constructs with the discrete domains are used. The white box represents the signal peptide, ovals represent the N-term and C-term Cys-rich domains flanking the LRR domains (light grey rectangles). Dark grey rectangles represent the EPTP repeats. The pentagon indicates the position of the Flag epitope. (B) Analysis of the secretion of LGI1-FlagLONG and LGI1-FlagSHORT in NGF-differentiated PC12 (left), T98G (middle) and HEK293T (right) cells. Cells were transfected with each construct and the lysates and the concentrated media were analysed by western blot using an anti-Flag antibody, as detailed in the Materials and Methods section. Note the presence of an unspecific band close to the size of the LGI1-FlagSHORT isoform in T98G cells. (C) HEK293 cells transfected with plasmid encoding LGI1-Flag or control vector were incubated with 35S-Trans-label (ICN) for 30 min and then chased with media containing an excess of unlabelled cysteine and methionine for 10, 30 and 60 min, as indicated. LGI1-Flag was immunoprecipitated using an anti-Flag antibody and analysed by SDS–PAGE/fluorography. LONG refers to the LGI1-FlagLONG construct, and SHORT refers to LGI1-FlagSHORT.

 
In addition, to verify that the secretion was cell-line independent, we also tested the influence that the different tags used had on this process and found that the secretion of LGI1 was not altered by attaching a larger tag such as GFP at its C-terminus (data not shown). To confirm that accumulation of LGI1 in media reflects active secretion from living cells, pulse-chase analyses were performed. Figure 1C shows that HEK293T cells transfected with LGI1-FlagLONG and, 24 h later, labelled with 35S-Cys/Met showed accumulation of radioactive LGI1 in media within 30 min. This time course suggests that LGI1 is actively secreted.

If LGI1 enters the secretory pathway, it is likely to undergo N-linked glycosylation. To test this, we performed PNGase F treatment of LGI1-FlagLONG and LGI1-FlagSHORT. Figure 2B shows that both isoforms undergo a substantial molecular weight shift when treated with PNGase F, suggesting that both become N-glycosylated in the endoplasmic reticulum. We introduced point mutations at three predicted N-glycosylation sites in the full-length isoform (LGI1-FlagLONG-N192Q/N277Q/N422Q) and found that the protein produced no longer underwent a molecular weight shift when treated with PNGase F, indicating that some or all of these sites are required for N-glycosylation of the protein. We also examined truncated forms of LGI1-Flag in this assay and found that a truncated protein lacking the LRR domain was sensitive to the PNGase F treatment, whereas a corresponding construct in which the glycosylation sites had been mutated to glutamic acid (LGI1-Flag{Delta}LRR-N277Q/N422Q) was not (Fig. 2C). If glycosylation sites in this construct were mutated individually (i.e. LGI1-Flag{Delta}LRR-N277Q and LGI1-Flag{Delta}LRR-N422Q), the resulting proteins were sensitive to the PNGase F treatment. We also examined a construct in which the EPTP domain was deleted and, using a similar approach, found that a single potential N-linked glycosylation residue in the resulting product (LGI1-Flag{Delta}EPTP-N192Q) conferred sensitivity to the PNGase F treatment. Together, these data indicate that N192Q, N277Q and N422Q are sites of N-linked glycosylation in LGI1.


Figure 4212
View larger version (30K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 2. LGI1-Flag is N-glycosylated at N192, N277 and N422. (A) Constructs used in these experiments showing the position of the predicted N-glycosylation sites; see Figure 1 for details on the domains of the protein. (B) Both LGI1-FlagLONG (LONG) and LGI1-FlagSHORT (SHORT) were transfected in HEK293T cells and extracts treated in the absence (–) or presence (+) of PNGFase and subjected to SDS–PAGE followed by immunodetection with an anti-Flag antibody. (C) Analysis of the sites of glycosylation in the constructs. Each of the LGI1-Flag constructs shown with the N-glycosylation sites mutated was subjected to treatment without (–) or with (+) PNGFase. Whereas the LGI1-FlagLONG isoform is glycosylated, the equivalent construct with all three predicted glycosylation sites removed shows lack of glycosylation. Mutation of the N192, N277 or N422 residues clearly show a modification in the pattern of PNGFase-sensitive glycosylation (2x, construct with both predicted sites mutated simultaneously; 3x, construct with all three predicted N-glycosylation sites mutated). Note the presence of bands resulting from incomplete digestion, with PNGFase in some lanes. (D) Effect of abolishing the sites of N-glycosylation on the secretion of LGI1-FlagLONG. Removing any individual N-glycosylation site does not block the secretion of LGI1-FlagLONG; on the contrary, LGI1-FlagN192Q/1277Q/N422Q (first lane on the right) is not secreted (right panel) and gets retained within the cell (left panel).

 
We next assessed whether N-glycosylation of LGI1 was required for its secretion and evaluated the effect of the three potential N-glycosylation sites on the secretion of LGI1. For this, we compared the secretion of the normal protein to the triple mutant and to each of the single glycosylation mutants (LGI1-FlagN192Q, LGI1-FlagN277Q and LGI1-FlagN422Q). Figure 2D shows that the triple mutant (LGI1-FlagLONG-N192Q/N277Q/N422Q) was not secreted and that the secretion of the N192Q mutant was severely attenuated.

Role of LGI1 domains in the secretion of the protein
To determine whether the LRR and EPTP domains were required for secretion, two artificial constructs, LGI1-Flag{Delta}EPTP and LGI1-Flag{Delta}LRR, were transfected in HEK293T cells and their accumulation in media was assessed. Figure 3B shows that LGI1 lacking its LRR domain is efficiently secreted, although this is in contrast with the results of Senechal et al., and the constructs used in each work being not comparable. Whereas in this work our construct deleted the full LRR domain, including the flanking Cys-rich domains, in Senechal et al. the construct used in the same experiments maintained the Cys-rich domains; thus, it is possible that this construct behaves as a pathogenic mutation rather than as a wild-type protein. On the other hand, LGI1 lacking its EPTP domain is retained in the transfected HEK293T cells. We attempted to rescue LGI1 secretion in the latter mutant by stepwise addition of each of the EPTP domains from one to six, using the scheme described in (15). Only the intact wild-type protein that contains all seven EPTP domains was secreted from cells.


Figure 4213
View larger version (45K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 3. LGI1 domains required for protein secretion. (A) Representation of the constructs employed in this experiment; see legend of Figure 1 for details on the protein domains. (B) Analysis of the effect on the secretion of the two regions of the protein. As detailed in the Material and Methods section, HEK293T cells were transfected with the indicated constructs and, 48 h post-transfection, cell lysates and concentrated culture media were analysed by western blotting with an anti-Flag antibody. (C) The requirement for a complete EPTP-repeat region was tested by transfecting HEK293T cells with constructs containing increasing numbers of EPTP repeats and analysing both the cell extracts and concentrated culture media as before.

 
Secretion of LGI1, but not N-glycosylation, is blocked by ADLTE-causing mutations
We next assessed whether naturally occurring mutations in LGI1, which cause ADTLE, alter the secretion or glycosylation of the protein. HEK293T cells were transfected with normal LGI1-FlagLONG or with LGI1-Flag758delC, LGI1-FlagR474X, LGI1-FlagC46R, LGI1-FlagS145R, LGI1-FlagC200R, LGI1-FlagF318C and LGI1-FlagE383A. All of these constructs resulted in robust LGI1 expression but only LGI1-FlagLONG accumulated in media, indicating that the naturally occurring mutations in LGI1, which cause disease, do not exit the cell, regardless of whether they were truncating or missense mutations (Fig. 4B). We also examined the sensitivity of these proteins to the PNGase F treatment and found, despite these trafficking defects, that all the LGI1 mutants tested were N-glycosylated (Fig. 4C).


Figure 4214
View larger version (39K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 4. LGI1 mutations that cause ADTLE block secretion, but not the glycosylation, of LGI1-Flag. (A) Constructs employed in these experiments; see legend of Figure 1 for an explanation of the different domains of the protein. (B) HEK293T cells were transfected with each construct and cell lysates (left panel) and concentrated culture media (right panel) were subjected to SDS–PAGE followed by immunoblotting with an anti-Flag antibody. (C) PNGFase treatment of ADLTE-causing mutations. HEK293T cells were transfected with each construct and, 48 h post-transfection, cell extracts were subjected to treatment without (–) or with (+) PNGFase prior to western blotting.

 
Subcellular localization of LGI1
To determine the subcellular localization of LGI1-GFPLONG, LGI1-GFPSHORT and ADLTE-related mutants LGI1-GFP758delC, LGI1-GFPR474X, LGI1-GFPF318C and LGI1-GFPE383A, COS7 cells were transiently transfected with each construct and transfected cells were immunostained with an antibody recognizing calnexin, an endoplasmic reticulum resident protein, or labelled with the fluorescent Golgi marker Bodipy TR ceramide. Confocal analysis revealed that LGI1-GFPLONG is enriched in the Golgi apparatus, whereas LGI1-GFPSHORT and the mutants analyzed, excluding LGI1-GFPR474X, accumulate in the ER chaperone. LGI1-GFPR474X is present in both ER and Golgi (Fig. 5). Colocalization of LGI1 and the ER was also demonstrated by co-immunoprecipitating the wild-type and the mutant constructs with an antibody that recognizes calnexin (data not shown).


Figure 4215
View larger version (100K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 5. Subcellular localization of disease-causing mutations of LGI1. COS7 cells were transfected with each GFP-fused protein as indicated and treated with either an anti-calnexin antibody to identify the endoplasmic reticulum or the Bodipy probe to detect the Golgi apparatus. Confocal images for each construct are shown together with the merged image. Bar: 8 µm.

 
LGI1 binds to the surface of PC12 cells
Because LGI1 is secreted to the extracellular medium, we assessed whether LGI1 could specifically interact with cell surface proteins. For this, we prepared conditioned media containing either placental alkaline alone or placental alkaline phosphatase fused to LGI1LONG (AP-LGILONG) and compared these for their ability to bind NGF-differentiated PC12 cells as well as to the surface of non-differentiated PC12 and COS7 cells (data not shown). Figure 6 shows the highly specific, dose-dependent and saturable binding of AP-LGI LONG, with half-maximal binding observed at a concentration of 4 nM, suggesting that specific LGI1 receptor complex is present on these cells.


Figure 4216
View larger version (71K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 6. LGI1 binds to the cell surface of differentiated PC12 cells. AP-LGI1LONG was obtained from the concentrated culture media of transfected HEK293T and incubated with NGF-differentiated PC12 cells. After 1 h incubation, the reagents were removed and the cells washed and treated (A) to detect membrane-bound AP-LGI1LONG (left) or membrane-bound AP as a control (right) or subjected to a quantitative analysis, as detailed in the Materials and Methods section (B) Magnification of pictures: 200x.

 
Binding of AP-LGI1 to the cell surface results in reduction of activated ERK1/2
To determine the effect that AP-LGI1 binding had on the cells, we assessed the activation state of ERK1/2, as it has been shown that overexpression of LGI1 caused a decrease in their activated levels in T98G cells (21). Figure 7 shows that incubating NGF-differentiated PC12 cells with AP-LGI1LONG results in a significant reduction in the level of activated ERK1/2 when compared with AP-treated cells.


Figure 4217
View larger version (31K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 7. Binding of LGI1 to the surface of NGF-differentiated PC12 cells reduces the activation of ERK1/2. AP-LGI1LONG was incubated with NGF-differentiated PC12 cells, as shown in Figure 6. After the times indicated, cells were harvested and analysed to determine the level of phospho-ERK1/2 and total ERK1/2. (A) Western blot showing the reduction in phospho-ERK1/2 at 5 and 15 min after AP-LGI1LONG treatment when compared with the AP treatment. (B) Quantification of the relative levels of phospho-ERK1/2 over total ERK1/2 expressed as percentage over the levels at t=0. *P=0.02 using a Student's t-test. Values are given as mean±SEM for three independent experiments.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
We demonstrate here that LGI1 is a glycoprotein secreted to the extracellular media in several in vitro models including a neuronal-like cell type. The shorter and much less prevalent spliced isoform is retained within the cell. This is true even though the short isoform is expressed at higher levels in our in vitro system. In some instances, usage of alternative splice sites within coding exons has been shown to produce isoforms with different, even opposite, functions (22,23); thus, the existence of two LGI1 isoforms with different cellular locations suggests that these proteins may have different functions.

LGI1 is the gene responsible for ADLTE, and distinct mutations have been described in several families. Disease-causing LGI1 mutations have been identified throughout the LGI1 open reading frame (3,4,6–11,24), and in this study, we have tested the hypothesis that disease-causing mutations alter LGI1 secretion. We show here that each of the disease-causing forms of LGI1 are defective in secretion from transfected cells. We did not detect changes in the glycosylation of the mutant proteins, but instead demonstrate that the mutant proteins are largely retained in the endoplasmic reticulum. The only exception was LGI1-FlagR474X, which was present in both in the endoplasmic reticulum and the Golgi apparatus. This mutation produces a truncated protein lacking the C-terminal 83 amino acids, and it is possible that it retains structural elements that allow it to partially mimic the normal protein and reach the Golgi apparatus. Using artificial mutations, we also show that the LRR domain is not required for secretion, but that an intact EPTP domain with all seven tandem repeats is required for the protein to exit the cell, and find that N-linked glycosylation of LGI1 is necessary for its secretion, but individual mutation of any of the three putative N-linked residues does not block its exit from the cell. We conclude that the disease-related missense mutations are exerting their effect by altering the tertiary structure of the protein, causing their retention and degradation through protein quality control mechanisms.

The secretion of the full-length isoform together with the retention of the spliced isoform also raises interesting questions regarding the function of LGI1. Kunapuli et al. (21) showed that when T98G cells were forced to re-express the long isoform, the malignancy of this glioma cell line was severely reduced, and this correlated with reduced ERK1/2 phosphorylation and reduced metalloproteinase production. Given that, we have demonstrated that LGI1 binds in a dose-dependent manner to differentiated PC12 cells and that this binding has a functional effect in the cells. It is possible that forced overexpression of LGI1 in T98G cells may have activated a ligand-receptor autocrine loop in these cells, which altered cellular signalling pathways.

Schulte et al. (20) have recently shown that LGI1 is part of a Kv1 channel complex and provided evidence that LGI1 acts to prevent the inhibitory effect of the Kvß1 channel subunit. They report that mutant forms of LGI1 assemble into Kv1 channel complexes yet are incapable of altering the Kvß1 channel inhibitor activity. They propose that the C-terminal domain of LGI1 lost in truncated mutants of the protein is required for the Kvß1 channel antagonizing activity. We show here that a major consequence of LGI1 mutation is ER retention and a failure to enter the Golgi and the secretory pathway. Therefore, an equally plausible alternative explanation for the findings of Schulte et al. (20) is that the mutant LGI1 assembles into Kv1 complexes in the ER and the resulting assembly remains trapped in this compartment. In this scenario, only the Kv1 complexes lacking LGI1 would actually reach the cell surface.

Taken together, our data suggest a hypothesis for the molecular pathology of ADLTE based on the retention of LGI1 in intracellular compartments. This retention could have a gain of function with cell autonomous effects, such as on Kv1 channels, or could result in a lost of function from non-autonomous effects that include the loss of a critical cell communication between cells that secrete LGI1 and cells that respond to it.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
Constructs
Several constructs were prepared for this work as fusion proteins. Proteins fused to GFP were cloned into pEGFP-N1 (Clontech), proteins fused to AP were cloned into pc3.1AP6 and fusions to the Flag epitope were cloned into pSalo-Flag, a modified version of pEGFP-N1 with the GFP protein removed and substituted by a Flag epitope. For simplicity, each construct is named by placing the name of the tag showing its position with respect to the cloned LGI1.

Both LGI1 isoforms, the full-length (long, amino acids 1–557) and the splicing variant (short, amino acids 1–259), were cloned from a human cDNA library (Invitrogen). For this study, we have selected mutations producing truncated proteins [c.758delC and p.R474X (3)] and some missense mutations affecting different domains of the protein: the N-terminal cysteine-rich domain [p.C46R (24)], the third leucine-rich repeat [p.S145R (10)], the C-terminal cysteine-rich domain [p.C200R (8)] and the third and fourth EPTP repeats [p.F318C (6) and p.E383A (4), respectively]. In addition to these, we have also introduced the mutations affecting the three potential N-glycosylation sites predicted by NetNGlyc (http://www.cbs.dtu.dk/services/NetNGlyc/), p.N192Q, p.N277Q and p.N422Q. Finally, we have created two deletion constructs: one lacks the EPTP-repeat region ({Delta}EPTP, lacking amino acids 224–556) and one lacks the LRR region ({Delta}LRR) and contains the signal peptide (amino acids 1–34) fused to the EPTP-repeat region (amino acids 224–556). In addition to those, we have produced several chimeric forms lacking part or all of a particular domain, as shown in Figures 1–3. For the serial deletions of the EPTP repeats, different primers were used in order to obtain seven constructs containing from 0 to 6 EPTP repeats tagged with Flag: 0EPTPs (ending at amino acid 223), 1EPTP (ending at 267), 2EPTPs (ending at 314), 3EPTPs (ending at 364), 4EPTPs (ending at 415), 5EPTPs (ending at 462) and 6EPTPs (ending at 506). To obtain the AP-LGI1 vector, the cDNA encoding amino acids 35–557 was ligated into the pc3.1AP6 vector in frame with signal sequence and placental AP coding region. All mutagenesis were done with the Quick Change Mutagenesis kit (Stratagene), and all constructs generated by PCR were fully sequenced to ensure integrity of the cloned ORFs. All primers used as well as PCR conditions are available from the authors on request.

Cell culture, transient transfection and media collection
CHO, COS7 and HEK293T cells were grown in DMEM (Cambrex), supplemented with 10% (v/v) heat inactivated foetal bovine serum (FBS, Gibco), 2 mM glutamine (Gibco) and 2% (v/v) penicillin/streptomycin (10 000 U/ml penicillin and 10 mg/ml streptomycin; Gibco) on 10 cm plates and maintained at 37°C and 5% CO2. T98G cells, a glioblastoma-derived cell line that does not express LGI1 (12), were grown in DMEM (Cambrex), 10% (v/v) FBS, 2 mM glutamine, 1% non-essential amino acids (Gibco) and 1% sodium pyruvate (Gibco). PC12 cells were grown on poly-D-lysine coated plates in EMEM supplemented with 2 mM glutamine, 2% (v/v) penicillin/streptomycin and 0.1% BSA. When required, PC12 cells were differentiated with 50 ng/ml NGF in complete medium for 48 h.

T98G, COS7 and CHO cells were transiently transfected with liposomes (Fugene, Roche), and HEK293T were transfected by the calcium phosphate method. PC12 cells were transfected with lipofectamine and differentiated with NGF 24 h post-transfection. For media production, HEK293T cells were grown to ~60–70% confluence and were transiently transfected with 5 µg DNA. After 16–20 h of transfection, the media were replaced with a reduced serum media (Optimem I, Gibco) and incubated for an additional 24 h. Then, media were collected, supplemented with a protein inhibitor cocktail (Complete, Roche), filtrated through a 0.2 µm filter and concentrated 10–15 times to 300 µl using an Amicon Ultra15 device (MWCO: 30 000 kDa; Millipore) and analysed by SDS–PAGE.

For binding experiments, after 2 days on serum-free DMEM, secreted AP-LGI1 and AP were first concentrated as before and then the amount of protein was estimated according to (25,26), measuring the AP activity using p-nitrophenyl phosphate as substrate (Sigma). For the binding experiments, the concentrated media were diluted with PBS 1x medium to 5–40 nM.

SDS–PAGE and western blotting
Equivalent volumes of concentrated media and immunoprecipitates were loaded onto 10% SDS–PAGE gels to resolve proteins. Following transfer, nitrocellulose membranes (Amersham) were blocked at room temperature for 1 h with non-fat dry milk powder in Tris–HCl, 20 mM, pH 7.6, NaCl 140 mM and Tween-20 0.1% (v/v) (TBST) to reduce non-specific binding. The Flag epitope was detected by incubation with either mouse anti-Flag antibody (1:500, in blocking solution; Sigma) or rabbit anti-Flag antibody (1:1000, in 3% BSA; Sigma) at 4°C overnight. The membranes were washed in TBST prior to incubation with the HRP-conjugated anti-mouse IgG antibody or HRP-conjugated anti-rabbit IgG antibody (1:5000, in blocking solution, Jackson). The blots were further washed in TBST and subsequently incubated in ECL or ECL plus (Amersham).

Immunoprecipitation and pulse-chase analyses
HEK293T cells were lysed and scraped in 1 ml NP-40-containing lysis buffer (10 mM Tris–HCl, pH 8, 150 mM NaCl, 1% NP-40, 10% glycerol, 1% Triton X-100 and protease inhibitors) for 30 min at 4°C. Cell lysates were clarified by centrifugation at 13 000 r.p.m. at 4°C for 10 min and then the supernatants were incubated with protein G Sepharose beads (Amersham) conjugated with M2 mouse anti-Flag antibody (Sigma) at 4°C for 2 h. The beads were washed three times with lysis buffer and then resuspended in 60 µl 2x Laemmli sample buffer for SDS–PAGE. For pulse-chase analyses, cells were transfected with LGI1 or parental vector and 24 h later, pre-incubated in cysteine- and methionine-free media for 30 min, followed by incubation for 30 min in labelling media consisting of 50 µCi/ml 35S-Trans-label (ICN) and 3 mg/l unlabelled cysteine and methionine, respectively, for 30 min. Cells were then switched to normal media (30 mg/l cysteine and 30 mg/l methionine) for 10, 30 or 60 min. After lysis, immunoprecipitations were performed using monoclonal anti-Flag antibodies. Immunoprecipitates were separated by SDS–PAGE, gels were impregnated with Enhance (NEB) and dried, using the protocol provided by the manufacturer.

PNGase F digestion
Cell lysates and culture medium obtained as described from transiently transfected HEK293T cells were immunoprecipitated as mentioned earlier and resuspended and denatured in 25 µl of denaturing buffer at 100°C for 10 min and cooled to room temperature. The denatured samples were subjected to digestion with 4 mU of PNGase F (New England Biolabs) at 37°C for 90 min. The samples were then analysed by SDS–PAGE.

Subcellular localization of wild-type and mutant LGI1
Twenty-four hours after transfection with the LGI1-GFP constructs using FuGene (Roche), COS7 cells grown on glass cover slips were fixed with 4% (w/v) paraformaldehyde in PBS at 37°C for 30 min. Slides were blocked at room temperature for 2 h with 5% BSA in PBS and then immunostained with the ER marker anti-calnexin (1:500, Calbiochem) at 4°C overnight. Slides were washed and incubated with Texas Red-conjugated anti-rabbit antibody (1:5000, Molecular Probes) at room temperature for 2 h. Alternatively, COS7 cells were incubated with the Golgi marker Bodipy TR Ceramide (Molecular Probes) at a 3 µM concentration in culture media for 30 min at 37°C. Images were collected on a Leica TCS SL Spectral Confocal Microscope coupled to an inverted microscope (Leica DMIRE2; Leica Microsystems) at room temperature with a 63x immersion objective (HCX PLAPO CS, 63x/1.40–0.60. Images were captured on a computer system by the Leica Confocal Software v2.5 (Build 1347, Leica Microsystems).

Cell-binding assays of AP fusion proteins
AP-tagged fusion proteins were produced by transient transfection of HEK293T cells with either AP-LGI1 or empty vector producing secreted AP. PC-12 cells, 5x105 cells/well, were plated in 6-well plates coated with poly-D-lysine and differentiated with NGF (50 ng/ml) for 48 h. Afterwards, plates were washed twice with PBS 1x and then cells were incubated with dilutions of either AP or AP-LGI1 for 90 min at 37°C. Cultures were then washed three times with ice-cold Hanks balanced salt solution containing 20 mM sodium HEPES, 0.1% sodium azide and 0.5 mg/ml BSA. Plates, where the binding was quantified, were processed with lysis buffer (20 mM Tris–HCl, pH 8.0, 0.1% Triton X-100 and protease inhibitors). Endogenous AP activity was heat inactivated at 65° for 15 min and bound AP activity was assessed using p-nitro phenyl phosphate as substrate measuring absorbance at 405 nm. In order to normalize the results, total protein present in the lysates was measured by DC protein assay (Bio-Rad). Plates, where pictures were taken, were incubated at 65° for 90 min to heat-inactivate the endogenous AP activity, and the AP-LGI1 binding was revealed in the presence of nitro blue tetrazolium and BCIP (5-bromo-4-chloro-3-indolyl phosphate) (Sigma).

Determination of ERK1/2 levels
Conditioned media from HEK293T cells expressing either AP-LGI1LONG or the AP alone were incubated with NGF-differentiated PC12 cells for the times shown in Figure 7 at 37°C, normalizing for the AP activity at 15 nM. As the process of differentiation of the PC12 cells by NGF requires activation of the ERK1/2 pathway, and the incubation with conditioned media could affect the state of activation of ERK1/2 by itself, we pre-treated all cultures with an equivalent amount of conditioned media coming from mock-transfected HEK293T for 2 h prior to AP-LGI1LONG or AP treatment. After treatment, the cells were washed with PBS and harvested in RIPA buffer (1.5 M NaCl, 1% NP-40, 0.5% deoxycholate, 1% SDS, 50 mM Tris–HCl, pH 8.0) with a cocktail of protease inhibitors (Complete, Roche) supplemented with 1 mM sodium orthovanadate. Cell lysates were incubated on ice for 30 min prior to clarification at 4°C and 16 000g for 15 min. An aliquot of 15 µg of protein from each condition was assayed by western blotting using an antibody that recognizes phosphorylated ERK1/2. After stripping of the blot, it was incubated with a phosphorylation-independent antibody that recognized total ERK1/2. In both cases, the western blotting was developed and an image was captured and the intensity of each band was measured with the help of the MultiGauge v2.1 software (Fujifilm). After this, we estimated the relative amount of phospho-ERK1/2 over total ERK1/2.


   ACKNOWLEDGEMENTS
 
The authors thank two anonymous reviewers for their helpful criticism of the work presented herein. The authors are indebted to Silvia Aparicio-Domingo, Raquel Rodríguez-de Pablos and Benito Alarcón for their technical support as well as to the Unidad de Genética y Medicina Molecular and Unidad de Biología Vascular of the Instituto de Biomedicina de Valencia-CSIC for technical assistance. J.P.T. is part of the Grupos de Excelencia of the Generalitat Valenciana (Grupos 03/015) and P.A.B. is a CIHR Scientist. This work was supported by grants from the Ministerio de Educación y Ciencia (SAF2002-00060 and SAF2005-00136) to J.P.T. and from the Canadian Institute of Heath Research (PPP147918) to P.A.B. S.S.P. is funded by a fellowship of the Generalitat Valenciana (CTBPRB/2002/35), J.M.M.R. is funded by an FPU and a Bancaixa fellowship. Support from the Ministerio de Educación y Ciencia (BES-2003-0243, to A.A.I.) and from the Ministerio de Sanidad y Consumo (BF03/00182, to V.H.P.) is also acknowledged. K.F. is funded by a Canadian NSERC award.

Conflict of Interest statement. The authors do not have any conflict of interest regarding the work reported in this manuscript.


   FOOTNOTES
 
{dagger} The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors. Back


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Ottman R., Risch N., Hauser W.A., Pedley T.A., Lee J.H., Barker-Cummings C., Lustenberger A., Nagle K.J., Lee K.S., Scheuer M.L., et al. (1995) Localization of a gene for partial epilepsy to chromosome 10q. Nat. Genet. 10:56–60.[CrossRef][ISI][Medline]

  2. Poza J.J., Saenz A., Martinez-Gil A., Cheron N., Cobo A.M., Urtasun M., Marti-Masso J.F., Grid D., Beckmann J.S., Prud'homme J.F., et al. (1999) Autosomal dominant lateral temporal epilepsy: clinical and genetic study of a large Basque pedigree linked to chromosome 10q. Ann. Neurol. 45:182–188.[CrossRef][ISI][Medline]

  3. Morante-Redolat J.M., Gorostidi-Pagola A., Piquer-Sirerol S., Saenz A., Poza J.J., Galan J., Gesk S., Sarafidou T., Mautner V.F., Binelli S., et al. (2002) Mutations in the LGI1 gene on 10q24 cause autosomal dominant lateral temporal epilepsy. Hum. Mol. Genet. 11:1119–1128.[Abstract/Free Full Text]

  4. Kalachikov S., Evgrafov O., Ross B., Winawer M., Barker-Cummings C., Martinelli Boneschi F., Choi C., Morozov P., Das K., Teplitskaya E., et al. (2002) Mutations in LGI1 cause autosomal-dominant partial epilepsy with auditory features. Nat. Genet. 30:335–341.[CrossRef][ISI][Medline]

  5. Gu W., Brodtkorb E., Steinlein O.K. (2002) LGI1 is mutated in familial temporal lobe epilepsy characterized by aphasic seizures. Ann. Neurol. 52:364–367.[CrossRef][ISI][Medline]

  6. Fertig E., Lincoln A., Martinuzzi A., Mattson R.H., Hisama F.M. (2003) Novel LGI1 mutation in a family with autosomal dominant partial epilepsy with auditory features. Neurology 60:1687–1690.[Abstract/Free Full Text]

  7. Kobayashi E., Santos N.F., Torres F.R., Secolin R., Sardinha L.A., Lopez-Cendes I., Cendes F. (2003) Magnetic resonance imaging abnormalities in familial temporal lobe epilepsy with auditory auras. Arch. Neurol. 60:1546–1551.[Abstract/Free Full Text]

  8. Michelucci R., Poza J.J., Sofia V., de Feo M.R., Binelli S., Bisulli F., Scudellaro E., Simionati B., Zimbello R., D'Orsi G., et al. (2003) Autosomal dominant lateral temporal epilepsy: clinical spectrum, new epitempin mutations, and genetic heterogeneity in seven European families. Epilepsia 44:1289–1297.[CrossRef][ISI][Medline]

  9. Berkovic S.F., Izzillo P., McMahon J.M., Harkin L.A., McIntosh A.M., Phillips H.A., Briellmann R.S., Wallace R.H., Mazarib A., Neufeld M.Y., et al. (2004) LGI1 mutations in temporal lobe epilepsies. Neurology 62:1115–1119.[Abstract/Free Full Text]

  10. Hedera P., Abou-Khalil B., Crunk A.E., Taylor K.A., Haines J.L., Sutcliffe J.S. (2004) Autosomal dominant lateral temporal epilepsy: two families with novel mutations in the LGI1 gene. Epilepsia 45:218–222.[CrossRef][ISI][Medline]

  11. Ottman R., Winawer M.R., Kalachikov S., Barker-Cummings C., Gilliam T.C., Pedley T.A., Hauser W.A. (2004) LGI1 mutations in autosomal dominant partial epilepsy with auditory features. Neurology 62:1120–1126.[Abstract/Free Full Text]

  12. Chernova O.B., Somerville R.P., Cowell J.K. (1998) A novel gene, LGI1, from 10q24 is rearranged and downregulated in malignant brain tumors. Oncogene 17:2873–2881.[CrossRef][ISI][Medline]

  13. Piepoli T., Jakupoglu C., Gu W., Lualdi E., Suarez-Merino B., Poliani P.L., Cattaneo M.G., Ortino B., Goplen D., Wang J., et al. (2006) Expression studies in gliomas and glial cells do not support a tumor suppressor role for LGI1. Neuro-oncol. 8:96–108.[Abstract/Free Full Text]

  14. Staub E., Perez-Tur J., Siebert R., Nobile C., Moschonas N.K., Deloukas P., Hinzmann B. (2002) The novel EPTP repeat defines a superfamily of proteins implicated in epileptic disorders. Trends Biochem. Sci. 27:441–444.[CrossRef][ISI][Medline]

  15. Scheel H., Tomiuk S., Hofmann K. (2002) A common protein interaction domain links two recently identified epilepsy genes. Hum. Mol. Genet. 11:1757–1762.[Abstract/Free Full Text]

  16. Senechal K.R., Thaller C., Noebels J.L. (2005) ADPEAF mutations reduce levels of secreted LGI1, a putative tumor suppressor protein linked to epilepsy. Hum. Mol. Genet. 14:1613–1620.[Abstract/Free Full Text]

  17. Kunapuli P., Chitta K.S., Cowell J.K. (2003) Suppression of the cell proliferation and invasion phenotypes in glioma cells by the LGI1 gene. Oncogene 22:3985–3991.[CrossRef][ISI][Medline]

  18. Gabellini N., Masola V., Quartesan S., Oselladore B., Nobile C., Michelucci R., Curtarello M., Parolin C., Palu G. (2006) Increased expression of LGI1 gene triggers growth inhibition and apoptosis of neuroblastoma cells. J. Cell. Physiol. 207:711–721.[CrossRef][ISI][Medline]

  19. Bermingham J.R., Shearin H., Pennington J., O'Moore J., Jaegle M., Driegen S., van Zon A., Darbas A., Ozkaynak E., Ryu E.J., et al. (2006) The claw paw mutation reveals a role for Lgi4 in peripheral nerve development. Nat. Neurosci. 9:76–84.[CrossRef][ISI][Medline]

  20. Schulte U., Thumfart J.O., Klocker N., Sailer C.A., Bildl W., Biniossek M., Dehn D., Deller T., Eble S., Abbass K., et al. (2006) The epilepsy-linked lgi1 protein assembles into presynaptic kv1 channels and inhibits inactivation by kvbeta1. Neuron 49:697–706.[CrossRef][ISI][Medline]

  21. Kunapuli P., Kasyapa C.S., Hawthorn L., Cowell J.K. (2004) LGI1, a putative tumor metastasis suppressor gene, controls in vitro invasiveness and expression of matrix metalloproteinases in glioma cells through the ERK1/2 pathway. J. Biol. Chem. 279:23151–23157.[Abstract/Free Full Text]

  22. Boise L.H., Gonzalez-Garcia M., Postema C.E., Ding L., Lindsten T., Turka L.A., Mao X., Nunez G., Thompson C.B. (1993) bcl-x, a bcl-2–related gene that functions as a dominant regulator of apoptotic cell death. Cell 74:597–608.[CrossRef][ISI][Medline]

  23. Roos K.L. and Simmons D.L. (2005) Cyclooxygenase variants: the role of alternative splicing. Biochem. Biophys. Res. Commun. 338:62–69.[CrossRef][ISI][Medline]

  24. Gu W., Wevers A., Schroder H., Grzeschik K.H., Derst C., Brodtkorb E., de Vos R., Steinlein O.K. (2002) The LGI1 gene involved in lateral temporal lobe epilepsy belongs to a new subfamily of leucine-rich repeat proteins. FEBS Lett. 519:71–76.[CrossRef][ISI][Medline]

  25. Flanagan J.G. and Cheng H.J. (2000) Alkaline phosphatase fusion proteins for molecular characterization and cloning of receptors and their ligands. Methods Enzymol. 327:198–210.[ISI][Medline]

  26. Flanagan J.G., Cheng H.J., Feldheim D.A., Hattori M., Lu Q., Vanderhaeghen P. (2000) Alkaline phosphatase fusions of ligands or receptors as in situ probes for staining of cells, tissues, and embryos. Methods Enzymol. 327:19–35.[ISI][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
15/23/3436    most recent
ddl421v2
ddl421v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Sirerol-Piquer, M. S.
Right arrow Articles by Pérez-Tur, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sirerol-Piquer, M. S.
Right arrow Articles by Pérez-Tur, J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?