Abstract
Objective: To describe a new FHM kindred, and to analyse the functional consequences of the disease-associated ATP1A2 p.G301R mutation in human cellular models grown at 37°C.
Patients and methods: Seven patients were clinically evaluated and gave informed consent for molecular analysis. Extra-pyramidal rigidity of the limbs was present in four subjects and in three of them tongue apraxia was also observed. ATP1A2 and CACNA1A were analysed by direct sequencing. Functional consequences of the mutation were investigated by cell viability assays, Western blots, and immunocytochemistry. Three-dimensional models of the human Na+/K+-ATPase α2 subunit were generated by homology modelling using SWISS-MODEL.
Findings: Analysis of ATP1A2 showed a heterozygous mutation, c.901G>A predicting the replacement of arginine for glycine at residue 301 (p.G301R). Functional analysis suggested that the mutation completely abolished Na+/K+-ATPase function.
Conclusions: The phenotypic spectrum of our FHM2 family includes some peculiar features. Functional data confirm that Na+/K+-ATPase haploinsufficiency caused by the ATP1A2 p.G301R mutation is responsible for FHM in the described family.
Introduction
Familial hemiplegic migraine (FHM) is a severe, autosomal dominantly transmitted subtype of migraine with aura associated with hemiparesis. FHM is genetically heterogeneous, with about 75% of the reported families linked to chromosome 19p13 (FHM1, OMIM 141500) and harbouring heterozygous mutations in CACNA1A, the gene coding for Cav2.1, the main α subunit of neuronal voltage-gated P/Q-type calcium channels (1). Approximately 20% of the reported families harbour mutations in the ATP1A2 gene (FHM2, OMIM 602481) on chromosome 1q21, encoding for the Na+/K+-ATPase α2 subunit; a FHM3 (OMIM 609614) form, associated with mutations in SCN1A on chromosome 2q24, a gene encoding for a neuronal voltage-gated sodium channel, has been only rarely described (2). Families unlinked to any of the reported loci as well as apparently sporadic cases have also been described. There is allelic heterogeneity in ATP1A2: mutations have been associated with FHM2, but also with benign familial infantile convulsions (BFIC) (3), alternating hemiplegia of childhood (AHC) (4) and additional features like seizures, prolonged coma and cerebellar signs as well as familial basilar migraine. Different mutations possibly causing different effects on the Na+/K+-ATPase pump activity might account for this phenotypic heterogeneity. Moreover, this highly variable clinical expression emphasizes the importance of functional studies to validate the described mutations as pathogenetically relevant.
In the present work, we describe a new FHM kindred harbouring a p.G301R mutation in ATP1A2 and we analysed the functional consequences of the disease-associated mutation identified in this family in human cellular models grown at 37°C.
Materials and methods
Patients
The FHM family was of Italian origin and included nine affected members in three generations (Figure 1). Seven patients (three males and four women) were clinically evaluated at least once by two of the authors (SL and MF) and two subjects (both women) were considered affected by clinical history. A diagnosis of migraine with or without aura, or FHM, was in agreement with the criteria set by the Headache Classification Subcommittee of the International Headache Society (5).
Pedigree of the family. Black circles and squares: clinically and genetically affected members. Grey circles: clinically affected members by history. The arrow indicates the proband.
The proband (II.5), a 58-year-old man, was admitted to the emergency room at the age of 38 years because of a sudden loss of consciousness and vomiting followed by coma lasting 3 days. Laboratory findings were normal and cerebrospinal fluid (CSF) analysis showed normal protein level and cell count. Electroencephalogram (EEG), brain computerized tomography (CT) or magnetic resonance imaging (MRI) of the brain were not performed. Migraine was not reported during this first episode. At the age of 45 years, he presented a new episode with altered mental status followed by aphasia and visual hallucinations that lasted 3 days. CSF analysis and brain CT and MRI, performed during and after the attack, were normal. Two years later he experienced a third episode with migraine and right hemiparesis, followed by nuchal rigidity and coma lasting a few days. Antiepileptic treatment with carbamazepine (CBZ) (800 mg/daily) was started. In the following years, the patient had additional, although less severe episodes, characterized by migraine and confusional state, but no hemiparesis. During one of these episodes, at the age of 53 years, CSF analysis showed a slight increase of protein level (58 mg/dl; upper normal value < 43 mg/dl), while brain MRI and EEG were again normal. At our evaluation, the interictal neurological examination showed only asymmetric deep tendon reflexes with a prevalence on the left side. Neuropsychological examination showed memory and executive function impairment. Moreover, brain MRI revealed some T2 hyperintense lesions in the frontal white matter, which were not enhanced after i.v. gadolinium injection and were considered as expression of a chronic cerebrovascular disease. The presence in the medical history of a single episode with hemiparesis and migraine did not fit the diagnostic criteria of the International Headache Society for FHM, but the family history allowed this hypothesis, and molecular analysis was therefore performed. After the diagnosis, CBZ treatment was stopped, and the patient started acethazolamide (500 mg/daily). In the last 3 years, the severity and frequency of migraine attacks clearly improved, and they were never associated with a confusional state. When the proband was seen for the first time, his parents had both died in their seventh and eighth decades, and he did not recall any complicated migraine episode in either parent. Medical records were uninformative.
Patient II.1, the proband’s sister, was a 70-year-old woman and referred her first episode of hemiplegic migraine at the age of 40 years. During the following years, she had several episodes of migraine with a frequency of about one per year and a duration from a few hours to 1 day. Migraine attacks were variably associated with visual aura, photophobia, aphasia, right hemiparesis and altered consciousness. When present, visual aura was characterized by scintillating scotomas and lasted about 15 minutes. At the last clinical evaluation (70 years), neurological examination revealed slight extrapyramidal rigidity at the lower limbs. The EEG was normal; brain MRI and CT showed calcification of basal ganglia.
Patient II.2, the proband’s sister, was a 67-year-old woman who reported her first clinically relevant episode at the age of 50 years. The episode was characterized by right hemiparesis and hypoesthesia without migraine. After 15 years, she experienced a second episode characterized by left hemiparesis and migraine. In the last 2 years, she suffered from diabetes and was treated with metformin. At the last clinical evaluation (67 years), neurological examination showed tongue apraxia (the patient was unable to push the tongue outside the mouth), and slight extrapyramidal rigidity at the upper and lower limbs. Her daughter (III.1), a 34-year-old-woman, was referred as affected by a frequent uncomplicated migraine, but she refused clinical evaluation, molecular analysis and treatment. The nephew (IV.1) of patient II.2 was a 14-year-old girl who, in the last year, reported several episodes of migraine with visual aura (scintillating scotomas lasting from 15 to 30 minutes), aphasia and alternating hemiparesis lasting a few days. During the last episode, she lost consciousness and became comatose. CSF analysis and brain MRI were normal. She recovered completely from coma in 2 days. The last interictal neurological examination and the EEG were normal.
Patient II.3, a 64-year-old man, referred his first episode of migraine at the age of 30 years with visual aura, aphasia, hemiparesis and confusional state, lasting 2–3 days. Since that time, similar episodes were reported with a frequency of two per year. The episodes were diagnosed as epilepsy, and the patient was treated with primidone until he came to our attention. At the last examination, neurological exam revealed tongue apraxia, as in his sister (II.2), slight extrapyramidal rigidity at the upper and lower limbs, and a finger tapping test showing reduced amplitude and velocity. The EEG was normal. His daughter (III.2) was referred as affected by similar episodes, but she was never seen by us.
Patient II.4, a 62-year-old man, experienced at the age of 10 years his first episode of migraine with visual aura (white spots for about 10 minutes), right hemiparesis and aphasia, which lasted a few hours. In the following years, similar episodes have occurred with a frequency of one or two per year. At the last neurological examination he showed a slight upward gaze limitation, tongue apraxia, mild rigidity at the upper limbs, and an impaired finger tapping test. He also referred a history of mood depression. Brain CT scan and EEG were normal. He has been taking acethazolamide for 2 years, and has been crisis-free since then. His daughter (III.3), a 30-year-old woman, suffered from migraine with visual aura (white spots for about 10 minutes), aphasia and alternating hemiparesis since she was 6 years old. The episodes had a frequency of three per year, each lasting a few hours. During the last episode, a few months before our evaluation, she was admitted to the emergency room in a comatose state. CSF analysis and brain MRI were normal. She recovered completely 3 days after hospital admission. At her last clinical evaluation, neurological exam was unremarkable, the EEG was normal, and she started to take acethazolamide (500 mg/day) with a progressive tapering of the antiepileptic drugs taken previously. Psychiatric consultancy revealed an obsessive-compulsive disorder. In the last 2 years she has been free from crisis.
Clinical data are summarized in Table 1
Genetic data
Clinical data
Roman numbers refer to generations and arabic numbers identify affected patients; EEG, electroencephalogram; MRI, magnetic resonance imaging; CT, computed tomography; +/−, presence/absence of symptoms or signs
Constructs and site-direct mutagenesis
For our experiments, we used the full-length cDNAs encoding for the ouabain-sensitive α2 (ATP1A2, also defined as α2ouaS) and β (ATP1B) subunits of the human Na+/K+-ATPase pump, each subcloned into the pcDNA3.1 expression vector (InVitrogen, Milan, Italy) (7). In order to differentiate Na+/K+-ATPase pump activity endogenously present in all vertebrate cells from that expressed heterologously, the exogenous Na+/K+-ATPase pump was rendered resistant to blockade by ouabain (IC50 in the millimolar range) (8,9) by incorporating into the ATP1A2 cDNA the 451A>G and the 483 A>G mutations (leading respectively to the p.Q116R and p.N127D amino acid substitutions), thus generating the α2ouaR construct. Furthermore, the nucleotide sequence encoding for the amino acids 408–419 of the human c-MYC (AEEQKLISEEDL) was introduced immediately after the translation start site of the cDNAs encoding for both α2ouaS and α2ouaR constructs, in order to generate cMYC- α2 fusion proteins. The 1005G>A mutation (leading to the p.G301R amino acid substitution found in the family herein investigated) was introduced in the α2ouaR cDNA by means of the sequence overlap extension (SOE) PCR technique; all constructs were verified by direct sequencing using the BigDye Terminator Cycle Sequencing Kit in an ABI PRISM 310 automated sequencer (Applied Biosystems, Foster City, CA, USA).
Transient transfections and cell viability assays
HeLa cells were grown in 12-well Petri dishes with RPMI 1640 (Gibco-Italia, Milan, Italy) supplemented with 10% fetal bovine serum (FBS) and penicillin (50 U/ml)/streptomycin (50 µg/ml) in a humidified atmosphere at 37°C with 5% CO2. Twenty-four hours after plating, HeLa cells were transiently transfected using Polyfect (Quiagen, Valencia, CA, USA) with either wild-type (α2ouaS or α2ouaR) or mutant (G301R-α2ouaR) cDNAs encoding for the Na+/K+-ATPase α2 subunits, together with that encoding for the β subunit. The maximal amount of total cDNA in the transfection mixture was 1–2 µg. The ratio between α2:β cDNAs was 1 : 1, except in experiments whose results are shown in Figure 2D. According to the experimental protocol, ouabain (1 µM) treatment was started 24 hours after transfection. Cell viability was evaluated with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; 0.5 mg/ml; 200,000 cells plated per well) reduction assay after 12, 24, 36 and 48 h from the beginning of the ouabain treatment (10). Formazan salts formation, indicative of cell vitality, was monitored spectrophotometrically (absorbance at 595 nm). Cell survival was expressed as percent of the value observed at the time of ouabain addition to the culture media (time 0). All experiments were carried out at least three times, with each data point performed in quadruplicate. In some experiments (Figure 3), a plasmid encoding for the enhanced green fluorescent protein (eGFP; Clontech, Palo Alto, CA, USA) was used as a transfection marker.
Functional analysis of wild-type and mutant ATP1A2 alleles. MTT survival assay of non-transfected HeLa cells (A) or of HeLa cells transfected with a plasmid encoding for the ouabain-sensitive (α2ouaS) or the ouabain-resistant (α2ouaR) Na+/K+-ATPase pump isoform, (plus the β subunit; ratio 1 : 1) (B). In both panels, as indicated, data obtained from cells exposed to ouabain (1 µM) are shown with filled symbols, whereas those from cells not exposed to the drug are shown with empty symbols. Time 0 reflects the time at which ouabain treatment was started (24 h post-transfection). In panel (C), HeLa cells were transfected with the indicated plasmids, and MTT viability assays were carried out at 12, 24 and 36 h after ouabain addition. (D) Effect of transfection with α2ouaR and G301R-α2ouaR cDNAs on HeLa cells viability after 24 h of ouabain exposure; the number in parentheses indicates the α2:β ratio utilized in each experimental group. In all panels, each data point is the mean ± SEM of at least three experiments each performed in quadruplicate. Asterisks indicate values significantly different (p < 0.05) vs. respective controls.
Analysis of the expression of the protein and of the mRNA for wild-type and mutant ATP1A2 alleles in transfected cells. (A) Confocal immunofluorescence analysis of HeLa cells transfected with wild-type α2ouaR cDNA, together with eGFP cDNA. As indicated, panels a, b, c and d describe images for c-Myc (red), eGFP (green), Hoechst-33258 (blue), and the merged image from the same microscopic field (× 40 objective). Confocal images shown in panels e, f and g (× 60 objective) are from HEK-293 cells transfected with wild-type α2ouaR cDNA showing immunofluorescent signals for c-Myc (red), chromomycin (green), and the merged image, respectively. (B) Confocal immunofluorescence analisys of HeLa cells transfected with G301R-α2ouaR cDNA, together with eGFP cDNA. As indicated, panels a, b, c and d describe the same microscopic field (× 40 objective) revealed for c-Myc (red), eGFP (green), Hoechst-33258 (blue), and the merged image, respectively. (C) Western blots on total cell lysates from non-transfected HEK-293 cells (Ctl), or from HEK-293 cells transfected with wild-type α2ouaR (wt) or G301R-α2ouaR (G/R) cDNAs. α2 subunits were detected using anti-Myc antibodies (top panel); the lower panel shows the expression, on the same filters, of the intracellular protein α-tubulin, used as an internal standard for equal protein loading. The arrows correspond to the expected molecular masses for the Na+/K+-ATPase α2 subunit and α-tubulin, indicated for clarity. (D) The mRNA encoding for the Na+/K+-ATPase α2 subunit from untransfected HEK-293 cells (Ctl) and from HEK-293 cells transfected with wild-type α2ouaR (wt) or G301R-α2ouaR (G/R) cDNAs was detected by RT-PCR, as indicated. The amplicon expected size is 292 bp for the Na+/K+-ATPase α2 mRNA, and 600 bp for the β-actin mRNA, as indicated by the arrows.
Confocal immunofluorescence analysis
HeLa cells were plated on poly-L-lysine-coated coverslips and transfected as aforementioned. Given the small size of HeLa cells, some morphological and biochemical experiments were also carried out using Human Embryonic Kidney-293 (HEK-293) cells (grown in Dulbecco’s minimal essential medium containing 10% FBS, 50 U/ml penicillin and 50 µg/ml streptomycin). Twenty-four hours after transfection, the cells were washed three times with PBS and incubated at room temperature (20–22°C) with freshly-made paraformaldehyde (4% w/v) for 10 min. Following paraformaldehyde treatment, the cells were washed for 5 min with glycine 0.1%, washed in PBS, and incubated for 1 hour with a mouse monoclonal anti-cMYC primary antibody (dilution 1 : 100; Sigma-Aldrich, Milan, Italy). The cells were then washed in PBS and incubated at room temperature for 1 h with an anti-mouse IgG secondary antibody (1 : 200; Jackson Laboratories, Bar Harbor, ME, USA) conjugated to CY3. Both primary and secondary antibodies were diluted in PBS containing 10% (v/v) FBS and 0.1% Triton X-100. In order to stain the cell nuclei, the markers chromomycin A3 (dilution 1 : 200; Sigma-Aldrich, Milan, Italy), or Hoechst-33258 (dilution 1 : 1000; Sigma-Aldrich, Milan, Italy) were used. Coverslips were placed onto untreated glass slides, allowed to air dry and then mounted in SlowfadeTM antifade (Invitrogen, Molecular Probes, San Giuliano Milanese, Italy) before coverslipping. Each image was acquired four times and the signal averaged to improve the signal-to-noise ratio. Coverslips were analysed using a Zeiss LSM 510 Meta argon/krypton laser scanning confocal microscope (Carl Zeiss, Jena, Germany), as described (11). The colour scheme used was red for CY3-labelled structures and green for the nuclear marker chromomycin.
Western blotting
Total lysates from non-transfected and transfected HEK-293 cells (24 h after transfection), were run on SDS-PAGE (6%; 80–120 µg of proteins per lane), transferred on PVDF membranes. Membranes were incubated overnight at 4°C with a mouse monoclonal anti-cMYC primary antibody (dilution 1 : 500; Abcam, Cambridge, UK), followed by an anti-mouse IgG secondary antibody (1 : 2000; Jackson Laboratories, Bar Harbor, ME, USA). Reactive bands were detected by chemiluminescence (ECL, Invitrogen, San Giuliano Milanese, Italy). An anti-α-tubulin antibody (1 : 8000; Sigma, Milan, Italy) was used to check for equal protein loading. Images were quantified on a GelDoc station (BioRad, Segrate, Italy) using the Uviscan software (Uvitech Ltd, Cambridge, UK).
RNA extraction and semiquantitative PCR
Total RNA isolated from transfected and untransfected HEK-293 cells using TRI-Reagent (Sigma-Aldrich, Milan, Italy) was treated with DNAse 1 U/µl for 15 min at room temperature and quantified by spectrophotometry (260 nm/280 nm ratio > 1.7). cDNA was synthesized by reverse transcription using 1 µg of isolated mRNA as template at 37°C for 2 h. cDNA amplification was carried out using AmpliTaq Gold 0.06 U/µl (Applied Biosystems, Monza, Italy). The following protocol was used for PCR amplification (25–35 cycles): denaturation at 95°C for 1 min, annealing at 60°C for 1 min, extension at 72°C for 1 min, with an intercycle interval at 72°C of 7 min. Optical density of the bands was quantified using the ImageJ software (available at http://rsb.info.nih.gov/ij/).
Homology modelling
Three-dimensional models of the human Na+/K+-ATPase α2 subunit were generated by homology modelling with known structures of the enzyme available in the Protein Data Bank (PDB) using SWISS-MODEL (12). In particular, the human α2 subunit was aligned with the shark α subunit Na+/K+-ATPase (PDB accession number 2ZXE, 86% of sequence identity), recently solved at 2.4 Å resolution (13); secondary structures and intra-chain polar interactions in wild-type and G301R mutant models were analysed using PyMOL (available at http://pymol.sourceforge.net/).
Statistics
Data are expressed as the mean ± SEM. Statistically significant differences between the data were evaluated with the Student’s t-test or with analysis of variance (ANOVA) followed by the Student–Newman–Keuls test (p < 0.05).
Results
Genetic data
Direct gene sequencing of CACNA1A and SCN1A in the proband did not disclose point mutations or heterozygous deletions/duplications. Analysis of ATP1A2 showed a heterozygous mutation, c.901G>A predicting the replacement of arginine for glycine at residue 301 (p.G301R). The same mutation has already been associated with FHM in an independent family (14). The mutation was also identified by direct gene sequencing in additional affected relatives, for a total of seven individuals in this family, but not in 150 healthy controls.
Functional analysis of wild-type and mutant ATP1A2 alleles
To assess the functional consequences of the G301R mutation in the α2 subunit of the Na+/K+-ATPase pump, cell survival assays using the MTT reduction assay in the presence and in the absence of the digitalis glycoside ouabain, a well-known blocker of the Na+/K+-ATPase pump, were carried out in HeLa cells. As shown in Figure 2A, the addition of ouabain (1 µM) to the culture medium of non-transfected HeLa cells caused a complete suppression of cell viability, leading to the death of the entire cell population within approximately 24 hours after drug addition. An identical loss of cell viability was also obtained when HeLa cells transiently transfected with the ouabain-sensitive form of the Na+/K+-ATPase pump (α2ouaS + β; 1 : 1 ratio) were exposed to ouabain (Figure 2B). By contrast, the heterologous expression of ouabain-resistant wild-type Na+/K+-ATPase (α2ouaR + β, 1 : 1 ratio) partially compensated for this loss of viability; in fact, at all time points (12, 24, 36 and 48 h) after ouabain (1 µM) addition, the cells expressing the α2ouaR constructs showed an enhanced survival when compared to those expressing the ouabain-sensitive wild-type Na+/K+-ATPase activity (α2ouaS + β; 1 : 1 ratio) (p < 0.05). Interestingly, in the absence of ouabain, the growth behaviour of HeLa cells transfected with the ouabain-sensitive (α2ouaS + β; 1 : 1 ratio) or with the ouabain-resistant (α2ouaR + β; 1 : 1 ratio) Na+/K+-ATPase pump, was identical. Altogether, these results suggest that HeLa cells survival assays effectively allowed discrimination between cells with functional and ouabain-resistant Na+/K+-ATPase activity, which were able to partially survive upon ouabain exposure, and cells in which the ouabain-sensitive Na+/K+-ATPase activity was entirely suppressed by the cardenolide, which were unable to survive. Noticeably, the extent of cell viability observed after transfection with the α2ouaR construct (about 20% at 24–48 h after ouabain exposure) is consistent with previously reported data (15).
To test the functional effect of the ATP1A2 allele carrying the G301R mutation found in the family herein investigated, cell survival assays were carried out in HeLa cells transfected with the G301R- α2ouaR construct, always together with the β subunit (1 : 1 ratio). As shown in Figure 2C, upon ouabain exposure, HeLa cells transfected with the G301R-α2ouaR plasmid displayed a growth behaviour identical to that of cells transfected with the ouabain-sensitive isoform of the pump (or of non-transfected cells); in fact, after 24 and 36 h from the beginning of the cardenolide treatment, the entire population of G301R-α2ouaR-transfected cells was not viable.
It should be noted that the experiments described in Figure 2C do not faithfully reproduce the heterozygous state of the family members carrying the G301R ATP1A2 allele. Therefore, in order to assess the possible consequences of the simultaneous expression of both α2ouaR and G301R-α2ouaR alleles, preliminary experiments were carried out to determine the optimal ratio of α2ouaR/β cDNA to be transfected such that the extent of cell survival was correlated to the amount of Na+/K+-ATPase pump activity expressed heterologously. As shown in Figure 2D, HeLa cells transfected with an α2ouaR/β cDNA ratio of 0.5 led to a 40 ± 2% cell survival after 24 h of ouabain treatment; by contrast, cell survival was significantly lower (27 ± 1%) when an α2ouaR/β cDNA ratio of 0.25 was used. For comparison, also in these experiments, transfection with an α2ouaS/β cDNA ratio of 0.5 completely failed to protect HeLa cells from ouabain toxicity. In order to mimic the heterozygous balance of the affected individuals, we transfected HeLa cells with α2ouaR and G301R-α2ouaR cDNAs, always in the presence of the β subunit-encoding plasmid (in a 0.25 : 0.25 : 1 cDNAs ratio). As shown in Figure 2D, these triple-transfected cells showed the same extent of cell survival observed in cells transfected with the same amount of α2ouaR alone (21 ± 1%; p > 0.05 vs. cells transfected with α2ouaR/β at a cDNA ratio of 0.25).
Immunocytochemical, Western blotting, and RT-PCR analysis of wild-type and mutant ATP1A2 alleles in transfected cells
Immunocytochemical experiments were carried out to investigate the expression and subcellular localization of the proteins encoded by the α2ouaR and G301R-α2ouaR ATP1A2 alleles upon HeLa (or HEK-293) cell transfection with their respective cDNAs. To this end, we used a monoclonal anti-cMYC primary antibody recognizing the cMYC tag in the fusion protein. As shown in Figure 3A (panels a, b, c and d), wild-type ATP1A2 subunits were clearly detected in transfected HeLa cells, as verified by the large extent of fluorescent signal overlap between eGFP (green) and c-MYC (red); in these experiments, cell nuclei were stained with Hoechst-33258 (blue). In HEK-293 cells, which, being larger, allowed a more precise cellular localization of the heterologously expressed protein, the α2 subunits appeared to be localized within the cytoplasmic compartment, but also around the cell contours, with a presumed distribution along the plasma membrane (Figure 3A, panels e, f and g); no evidence was found of nuclear localization of the protein, as revealed by the lack of co-localization of the anti c-MYC signal (red) and the fluorescent signal of chromomycin (green). By contrast, in G301R-α2ouaR-transfected cells (either in HeLa or in HEK-293 cells), we were unable to detect any fluorescence signal specifically corresponding to the α2 subunit carrying the G301R mutation (Figure 3B, panels a, b, c and d). Western blots using an anti-MYC antibody revealed a specific band of 112 kDa molecular mass in α2ouaR-transfected, but not in non-transfected HEK-293 cells; it is noteworthy that the 112 kDa band was also undetectable in lysates from G301R-α2ouaR-transfected cells (n = 3) (Figure 3C). Given these results, the potential effects of the 1005G>A substitution (responsible for the G301R mutation) on the stability of the RNA encoded by the mutant plasmid were investigated. Therefore, total mRNA was isolated from non-transfected HEK-293 cells and from cells transfected with the α2ouaR or the G301R-α2ouaR cDNAs; the isolated mRNA was then retro-transcribed, and the resulting cDNA used as template for a subsequent PCR amplification using a sense primer annealing with the fusion site between the nucleotide sequences of c-MYC and of the α2 subunit cDNA sequence. This strategy was chosen to allow amplification only of the RNA derived from transfected cDNAs, but not of the endogenous α2-encoding genomic DNA. The results obtained, shown in Figure 3D, revealed that a specific band, corresponding to an amplicon of 292 bp, could be amplified from both α2ouaR- and G301R-α2ouaR-transfected cells, but not from non-transfected cells. After 30 cycles of amplification, the intensity of this band did not differ between α2ouaR- and G301R-α2ouaR-transfected cells, since the ratios of the OD values for the ATP1A2 mRNA to that of actin were 0.26 ± 0.05 and 0.25 ± 0.07 for wild-type and mutant RNAs, respectively (p > 0.05; n = 3). Altogether, these results suggested that the G301R mutation is unlikely to prompt significant changes in mRNA stability, but caused a severe derangement of the normal maturation and intracellular trafficking of the Na+/K+-ATPase α2 subunit, thereby impeding its expression and, therefore, its functional activity.
Molecular modelling of the G301R mutation in the human α2 subunit of the Na+/K+-ATPase
To build a homology model of the Na+/K+-ATPase, the template of the shark α2 subunit (PDB accession number 2ZXE, 86% of sequence identity), recently solved at 2.4 Å resolution (13) was used. Using this model, the G301 residue is located toward the centre of the third transmembrane segment, in a region possibly located within the plane of the membrane (Figure 4A); replacement of the G301 residue with an R allowed the longer side chain to establish a novel interaction via an ionized hydrogen bond (1.87 Å distance) with the carbonyl groups of the protein backbone at the level of a leucine residue located at position 322 within the M4 transmembrane domain. Figure 4C shows the proximity of the G301 residue to the K+ binding sites (site I and site II) on the α2 subunit of the shark Na+/K+-ATPase (13).
Three-dimensional homology model of the Na+/K+-ATPase α2 subunit. (A) Modelling of a single human Na+/K+-ATPase α2 subunit, based on the crystal structure of the shark α2 subunit. (B) and (C) are enlarged images of the region corresponding to the M3 and M4 transmembrane segments, highlighting the G301 or the R301 residues, encoded by the wild-type and the mutant allele, respectively. The novel ionized hydrogen bond formed between the side chain of the R301 residue and the carbonyl of the protein backbone in TM4 is highlighted in yellow. (D) Model of the K+/Rb+ coordination sites contributed by the TM4, TM5 and TM6 domains, and relationship with the G301 residue. The two K+ ions are shown in yellow; a water molecule is depicted in purple.
Discussion
In the present work, a new Italian family affected with FHM2 and carrying a G301R mutation in the α2 subunit of the Na+/K+-ATPase has been described; moreover, by investigating the consequence of the mutation with biochemical, morphological and functional studies, the pathogenetic role of this mutation has been also proven. The same mutation has been described previously in another Italian family (14), which was unrelated to our family and was characterized by some clinical aspects overlapping with FHM1 (i.e. seizures and cerebellar dysfunction). In our family, none of the affected members had seizures, cerebellar dysfunction or high body temperature, either during the migraine or aura episodes or in the interictal periods. Moreover neuroradiological findings were normal during and after the attacks and did not show signs of permanent brain lesions such as cerebellar atrophy, which is frequently reported in FHM1 and was present in one member of the family described by Spadaro and colleagues (14).
However some other features, never described in other FHM families, were observed in the interictal condition. In particular, we revealed extrapyramidal rigidity of the limbs in four subjects (II.1, II.2, II.3 and II.4), who were the oldest of the family. In three of them, we also observed tongue apraxia (II.2, II.3, II.4). The attribution of these features to the FHM2 phenotype remains doubtful, but their high incidence among the family members is intriguing, even though we do not have a pathogenetic interpretation. Moreover, basal ganglia calcifications were present only in the CT scan of the brain in the oldest patient (II.1). Similar to the initial family harbouring the p.G301R mutation (14), migraine and hemiplegic aura appeared simultaneously in most patients of the kindred reported in the present work, although the hemiplegic aura preceded the first migraine episode by 15 years in one subject (II.2). The frequency of the hemiplegic aura was no different between the two families. A striking similarity between the two families was also the high incidence of repeated prolonged coma condition, an event described in several FHM2 families (16–19) but only in one FHM1 patient until now (20). This clinical aspect underlines the strict relationship between FHM2 and familial basilar migraine, both phenotypes associated with ATP1A2 mutations (21). Different effects on the Na+/K+-ATPase pump activity might account for the wide clinical heterogeneity and probably for the therapeutic effect of acethazolamide. In fact, this drug was ineffective in the previously described FHM families, but showed a clear efficacy in the three patients of our family. Although the effects of the ATP1A2 G301R mutation have been recently studied in Xenopus oocytes (22), in the present work the functional consequences of this mutation were assessed in human cellular models (HEK-293 or HeLa cells) grown at 37°C in order to achieve experimental conditions more closely reproducing the physiological context of operation of the Na+/K+-ATPase pump. Cell survival assays showed that the p.G301R mutation completely abolished the functionality of the Na+/K+-ATPase pump. Moreover, when the mutant plasmid was co-transfected together with the wild-type plasmid to reproduce the heterozygous state of the family members carrying the mutant ATP1A2 allele, we found no dominant-negative effect on wild-type Na+/K+-ATPase activity, suggesting that haploinsufficiency was the main pathogenetic mechanism responsible for the FHM phenotype in the described family. Western blotting experiments showed that the steady-state level of the mutant protein was below detection limits; in agreement with these biochemical results, we failed to detect any significant immunofluorescent signal from the mutant α2 subunit expressed in mammalian cells. It is worth noting that our results also revealed that the abundance of the mRNA encoded by the mutant plasmid appears comparable to that of the wild-type plasmid, arguing against a significant mutation-induced change in mRNA stability. At variance with our results, Western blot experiments performed in Xenopus oocytes grown at temperatures below 37°C revealed that the G301R mutation decreased (rather than abolished) the expression of the ATP1A2 protein (22); it seems possible to speculate that this quantitative difference might be due to a dramatic temperature-dependent decrease in protein stability. A similar hypothesis has been put forward to explain the molecular consequences of the FHM-associated P979L mutant in the ATP1A2 subunit (22,23). To provide further insight into the possible structural consequences of G301R mutation, as well as into the overall mechanism for the presumed effect of the mutation on ion translocation mediated by the Na+/K+-ATPase, we built a homology model of the human Na+/K+-ATPase α2 subunit using the crystal coordinates from the shark α2 subunit recently solved at 2.4 Å resolution (13). According to this model, replacement of the G301 residue with a longer and positively-charged R residue would allow the M3 transmembrane region to establish a novel hydrogen bond with the protein backbone at the level of the M4 transmembrane segment, thus decreasing the overall flexibility of the M3-M4 region. Note the G301 residue is rather close to the K+/Rb+ coordination sites mainly formed by negatively-charged residues contributed by different transmembrane regions (TM4, TM5 and TM6) in the occluded state of the enzyme (Figure 4D) (13,24). Thus, it seems plausible that the substitution of G301 with a positively-charged residue would reduce the net negative charge of this region and interfere with the transport of the K+ ions (25). In addition, the strong effect of the G301R mutation on the overall protein stability allows us to speculate that the mutation-induced destabilization of this region, in addition to ion coordination, might also impede the correct folding and exit from the endoplasmic reticulum of the mature protein (26). Altogether, the present results expand the phenotypic spectrum of FHM, and confirm that Na+/K+-ATPase haploinsufficiency caused by the G301R mutation in the human ATP1A2 gene is responsible for its autosomal-dominant transmission in the described family.