Intellectual disability-associated UNC80 mutations reveal inter-subunit interaction and dendritic function of the NALCN channel complex

Abstract

The sodium-leak channel NALCN forms a subthreshold sodium conductance that controls the resting membrane potentials of neurons. The auxiliary subunits of the channel and their functions in mammals are largely unknown. In this study, we demonstrate that two large proteins UNC80 and UNC79 are subunits of the NALCN complex. UNC80 knockout mice are neonatal lethal. The C-terminus of UNC80 contains a domain that interacts with UNC79 and overcomes a soma-retention signal to achieve dendritic localization. UNC80 lacking this domain, as found in human patients, still supports whole-cell NALCN currents but lacks dendritic localization. Our results establish the subunit composition of the NALCN complex, uncover the inter-subunit interaction domains, reveal the functional significance of regulation of dendritic membrane potential by the sodium-leak channel complex, and provide evidence supporting that genetic variations found in individuals with intellectual disability are the causes for the phenotype observed in patients.

Fig. 1. Targeted disruption of UNC80 leads to apnea and neonatal lethality.

a The design of UNC80 knockout (KO) using the CRISPR technique to delete exon 3. Exon 3 sequence is in capital and the surrounding introns are in lower case. The 5′ and 3′ target sequences including the PAM motif (XGG) against which the two CRISPR sgRNAs targeted are underlined. Deleted sequences including exon 3 and part of the surrounding introns are shaded. Deletion of exon 3 (total of 157 nucleotides) leads to truncation after V47. The codon encoding R51 (CGA) corresponding to the residue mutated to a stop codon found in human patients are in red. PCR primers used for genotyping in b are in italic and boxed. b Genotyping PCR products from WT (+/+), heterozygote (+/−), and homozygous KO (−/−) pups using primers in a. c Total brain proteins from +/+ and UNC80 −/− were blotted with anti-UNC80 (upper), anti-NALCN (middle), or anti-UNC79 (lower) antibody. d Representative appearances of WT and KO P0 pup. For (b) and (c), three or more independent repeats were performed with similar results. For apnea phenotype in the KO, see Supplementary Movie 1. Source data are provided as a Source Data file.

Fig. 2. UNC80 is required for the TTX-resistant Na⁺ leak current and its regulation by extracellular Ca²⁺.

a, b Representative TTX-resistant Na⁺ leak current recorded in hippocampal neurons cultured from WT (+/+) (a) and UNC80 KO (−/−) (b) pups. Each trace represents 1 s of currents recorded at −70 mV in bath solutions with varying [Na⁺] (140 or 14 mM) and [Ca²⁺] (2 or 0.1 mM). The Na⁺-leak current (ΔI_LNa) is calculated as the difference of current sizes between those recorded under 140 mM [Na⁺] and 14 mM [Na⁺], as indicated by dashed lines. c–e Similar to a, but recorded from UNC80 KO neurons transfected with UNC80 (c), NALCN (d), or UNC79 (e) cDNA. f Averaged sizes of currents in a–e. I_LCa is the current activated by lowering [Ca²⁺] from 2 to 0.1 mM, as defined as the size difference of ΔI_LNa under the two [Ca²⁺] conditions). Data are presented as mean values ± SEM. Numbers of neurons are in parentheses. In (f), two sample t test of each group against the “+/+” group (n = 5) was performed: “−/−” (n = 8, p ≤ 0.001), “−/−; UNC80” (n = 6, p = 0.326), “−/−; NALCN” (n = 5, p ≤ 0.001), “−/−; UNC79” (n = 7, p ≤ 0.001). Asterisk “*” indicates p < 0.05. Source data are provided as a Source Data file.

Fig. 3. Haploinsufficient reduction in UNC80 is associated with severe intellectual disability.

a Sanger sequencing chromatograms confirming the whole-exome sequence findings of genetic variations in the proband leading to protein changes of Q340H, P341S (left), and E1295Q (right). Upper: from the proband; lower: control. Codons encoding Q340 (CAG), P341 (CCC) and E1295 (GAA) are underlined. b Protein sequence alignments showing conservation in the Q340, P341 (left) and E1295 (right) regions. Accession numbers of the sequences used are NM_032504 (human), NP_780719 (mouse), XP_015144792 (chicken), XP_009300567 (zebrafish), and XP_011676014 (sea urchin). c Western blots with proteins prepared from non-transfected HEK293T cells and those transfected with mouse UNC80 wild type, Q341H;P342S (corresponding to human Q340H;P341S) or E1296Q (corresponding to human E1295Q) mutants, blotted with anti-UNC80 antibody (upper) or anti-actin (lower) for loading control. More than three independent repeats were performed with similar results. d TTX-resistant Na⁺ leak current recorded from cultured UNC80 KO hippocampal neurons transfected with wild-type UNC80, E1296Q, Q341H;P342Q, or mixture of the two mutants, as indicated. Representative currents are in the left four subpanels and averaged current amplitudes are summarized in the right. Recordings were from −70 mV and were done with bath solutions with varying [Na⁺] (140 mM or 14 mM) and [Ca²⁺] (2 mM (2 Ca) or 0.1 mM (0.1 Ca)) (see Fig. 2 legend for details). Data are presented as mean values ± SEM. Numbers of neurons recorded are in parentheses. In the right bar graph, two sample t test of each group against the “−/−; WT UNC80” group (n = 17) was performed and the p values are as follows. ΔI_LNa (2 Ca): “−/−; E1296Q” (n = 17, p ≤ 0.001), “−/−; E1296Q + Q341H;P342S” (n = 10, p ≤ 0.001), “−/−; Q341H;P342S” (n = 9, p ≤ 0.001)). ΔI_LNa (0.1 Ca): “−/−; E1296Q” (p ≤ 0.001), “−/−; E1296Q + Q341H;P342S” (p ≤ 0.001), “−/−; Q341H;P342S” (p ≤ 0.001). ΔI_LCa (0.1 Ca–2 Ca): “−/−; E1296Q” (p ≤ 0.001), “−/−; E1296Q + Q341H;P342S” (p ≤ 0.001), “−/−; Q341H;P342S” (p ≤ 0.001). Asterisk “*” indicates p < 0.05. Source data are provided as a Source Data file.

Fig. 4. UNC80 and UNC79 are auxiliary subunits of the NALCN channel complex.

a Protein fractionation demonstrating that UNC80 and UNC79 co-segregate with NALCN to the membrane fraction. Total proteins (T) from adult brains were centrifuged at 200,000 × g and separated into the cytosolic (supernatant, S) and microsomal membrane (pellet, P) fractions, as illustrated in the schematic diagram (upper). Each fraction was blotted with anti-NALCN, anti-UNC80, anti-UNC79, or anti-actin antibody. b A knock-in mouse line with NALCN tagged with GFP, HA, and His tags (NALCN-GFP-HA-His mice). Upper, schematic design. Lower, total brain proteins (100 μg) prepared from the triple-tagged mice and wild-type (non-tagged) mice were immunoblotted with anti-His antibody. c TTX-resistant Na⁺ leak currents recorded from neurons cultured from the KI pups (n = 5). Left, representative current traces. Right, averaged current amplitudes. Recordings were from −70 mV and were done with bath solutions with varying [Na⁺] (140 mM or 14 mM) and [Ca²⁺] (2 mM (2 Ca) or 0.1 mM (0.1 Ca)) (see Fig. 2 legend for details). Numbers of neurons recorded are in parentheses. d Protein depletion demonstrating that all UNC80 and UNC79 proteins are associated with NALCN. Total brain protein lysates were prepared from the NALCN-GFP-HA-His mice. NALCN was depleted by incubating with Ni column (binding to 6-His) followed by further immune depletion with anti-GFP antibody. Lysates before (lane 1) and after (lane 2: with Ni-beads, lane 3: with α-GFP agarose) depletion were blotted with anti-NALCN, anti-UNC79, or anti-UNC80 antibody. Anti-actin was used as a control. For (a, b, d), three or more independent repeats were performed with similar results. Data are presented as mean values ± SEM. Source data are provided as a Source Data file.

Fig. 5. UNC80’s N-terminal half interacts with NALCN.

a Co-immunoprecipitation assays to locate fragments on UNC80 required for its association with NALCN. Upper panels, schematic presentation of mouse UNC80 truncation mutants used in the studies. Lower panels, HEK293T cells were co-transfected with NALCN and GFP-tagged full-length or truncated UNC80 containing residues as indicated. Cell lysates were immunoprecipitated (IP) with anti-GFP followed by immunoblotting with anti-GFP (lower) or anti-NALCN (top). GFP was used in “ctrl”. It migrated at ~20 kDa (outside the molecular ranges shown) and is not visible in the blots. More than three independent repeats were performed with similar results. b I_NALCN from cells transfected with NALCN and wild-type or truncated UNC80 mutants as indicated. Recordings were done using a ramp protocol from −100 to +100 mV in 1 s (holding voltage V_h = 0 mV). To ensure that the current was not from nonspecific leak, bath cations were replaced with large non-permeant ion NMDG after each recording (see “Methods” for details). Bar graphs show averaged I_NALCN amplitude (at −100 mV). Two sample t test of each group against the “full-length” group (n = 11) was performed and the p values are: “Mock” (n = 5, p ≤ 0.001), “1–2387” (n = 9, p ≤ 0.001), “1–2554” (n = 5, p = 0.019), “1–2885” (n = 6, p = 0.818), “1–3000” (n = 5, p = 0.933). Asterisk “*” indicates p < 0.05. Data are presented as mean values ± SEM. Numbers of cells are in parentheses. Source data are provided as a Source Data file.

Fig. 6. UNC80’s C-terminus truncated in human patients is not required for whole-cell I_NALCN but is essential for survival.

a, b I_NALCN (a done as in Fig. 5b, averaged amplitudes at −100 mV given in the bar graph) or Na⁺-leak current (ΔI_LNa) (b) from HEK293T cells (a) or cultured UNC80 KO neurons (b recorded with 2 mM Ca²⁺ (2 Ca) or 0.1 mM Ca²⁺ (0.1 Ca) in the bath) transfected with full-length or mutant UNC80 truncated at L2654 (aa1–2653). c Surface biotinylation assays. Left, representative western blots. Surface proteins in cells transfected with cDNA combinations as indicated were biotinylated, isolated using streptavidin-conjugated beads, and probed with α-NALCN. Total protein (whole-cell lysate, WCL) was probed with α-NALCN or α-GFP. In the lower panel, GFP alone in the control lane migrated at ~20 kDa (outside the molecular range shown) and is not visible in the blot. Right, quantification. Signal intensity of each band from cells co-transfected with NALCN and full-length or truncated UNC80 was normalized to that from cells co-transfected with NALCN and GFP. d Sanger sequencing of the knock-in mice (L2654*, lower) and WT (upper). In the L2654* mice, the leucine-encoding codon (CTA, L2654) is mutated to a stop codon (TAA,*). e Representative appearances of WT and homozygous L2654* mutant P0 pups. For apnea phenotypes, see Supplementary Movie 2. f ΔI_LNa recorded from neurons cultured from WT and L2654* pups. Data are presented as mean values ± SEM. In the bar graphs, numbers of recordings or repeats are in parentheses. Two sample t tests were performed. Asterisk “*” indicates p < 0.05. p values are as follows. a (against the “full-length” group): “Mock” (p ≤ 0.001), “1–2653” (p ≤ 0.001), b (between the group of “−/−; UNC80 full-length” (full-length UNC80-transfected) and the group of “−/−; (1–2653)” (truncation mutant-transfected)): ΔI_LNa (2 Ca) (p ≤ 0.001), ΔI_LNa (0.1 Ca) (p ≤ 0.001), ΔI_LCa (0.1 Ca–2 Ca) (p ≤ 0.001), c (against the group of “GFP” (GFP-transfected)): “full-length” (p = 0.025), “1–2653” (p = 0.049). Source data are provided as a Source Data file.

Fig. 7. UNC80’s C-terminal domain overcomes soma-retention for dendritic localization.

a C-terminally RFP (upper) or N-terminally GFP (lower)-tagged UNC80 was co-transfected with GFP (upper) or RFP (lower) into cultured wild-type hippocampal neurons. Scale bars: 50 μm. b–d Similar to (a) but transfected with UNC80 mutants truncated at L2654 (containing residues 1–2653) (b), lacking the last six residues (containing residues 1–3320) (c) or containing residues 2234–2758 only (d). Representative pictures from similar results of >10 are presented for each condition.

Fig. 8. UNC80’s C-terminus contains an UNC79-interacting domain.

a Association between UNC80 mutant truncated at L2654 (containing residues 1–2653) and NALCN. Cell lysates from HEK293T cells co-transfected with NALCN and GFP (as control), GFP-tagged full-length UNC80 or the truncation mutant were immunoprecipitated (IP) with anti-GFP and blotted with anti-NALCN. b Lack of association between UNC80 mutant truncated at L2654 and UNC79. Cell lysates from HEK293T cells co-transfected with UNC79 and GFP (as control), GFP-tagged full-length UNC80, or truncated UNC80 containing only residues 1–2653 or 1–1266 were immunoprecipitated (IP) with anti-GFP and blotted with anti-UNC79 (upper). Lower 3 panels: whole lysates were blotted with anti-UNC79, anti-GFP or anti-actin (for loading control). c Mapping UNC80’s UNC79-interacting domain. Cell lysates from HEK293T cells co-transfected with UNC79 and GFP-tagged full-length UNC80, or truncated UNC80 containing residues as indicated were immunoprecipitated (IP) with anti-GFP and blotted with anti-UNC79. d Disruption of UNC79-NALCN association in mouse brains lacking the C-terminal part of UNC80. Brain proteins were prepared from the NALCN-GFP-HA mice and those that also carry the homozygous L2654* mutation in the NALCN-GFP-HA-His background. Upper panel: NALCN was pulled down with anti-GFP antibody and the complex was probed with anti-UNC79. Lower 3 panels: whole lysates were blotted with anti-UNC79, anti-NALCN or anti-actin (lower 3 panels). e Summary of UNC80’s functional domains: NALCN-interacting domain (residues 301–1700), soma-retention domain (2387–2657), soma-retention relieve domain (2758-3000), and UNC79-interacting domain (2758–2947). f, g Reduced UNC80–UNC79 interaction strength associated with intellectual disability. f UNC79 was co-transfected with GFP, GFP-tagged wild-type UNC80 or GFP-tagged R2910Q UNC80 mutant in HEK293T cells. Immunoprecipitates (IP with anti-GFP, upper panel) and whole-cell lysates (lower two panels) were blotted (IB) with anti-UNC79 or anti-GFP as indicated. In the lower panel, GFP alone in the control lane migrated at ~20 kDa (outside the molecular range shown) and is not visible in the blot. g Protein levels normalized to that obtained with WT UNC80. Results were from four independent experiments. In (a–d, f), three or more independent repeats were performed with similar results. Data are presented as mean values ± SEM. Numbers of repeats are in parentheses. Source data are provided as a Source Data file.