Abstract
Here we report the molecular characterization of Out-cold (Ocd) mutants of Drosophila melanogaster, which produce a dominant, X-linked, cold-sensitive paralytic phenotype. From its initial 1.5-Mb cytological location within 13F1-16A2, P-element and SNP mapping reduced the Ocd critical region to <100 kb and to six candidate genes: hangover, CG9947, CG4420, eIF2a, Rbp2, and paralytic (para). Complementation testing with para null mutations strongly suggests Ocd and para are allelic, as does gene rescue of Ocd semilethality with a wild-type para transgene. Pesticide resistance and electrophysiological phenotypes of Ocd mutants support this conclusion. The para gene encodes a voltage-gated sodium channel. Sequencing the Ocd lines revealed mutations within highly conserved regions of the para coding sequence, in the transmembrane segment S6 of domain III (I1545M and T1551I), and in the linker between domains III and IV (G1571R), the location of the channel inactivation gate. The G1571R mutation is of particular interest as mutations of the orthologous residue (G1306) in the human skeletal muscle sodium channel gene SCN4A are associated with cases of periodic paralysis and myotonia, including the human cold-sensitive disorder paramyotonia congenita. The mechanisms by which sodium channel mutations cause cold sensitivity are not well understood. Therefore, in the absence of suitable vertebrate models, Ocd provides a system in which genetic, molecular, physiological, and behavioral tools can be exploited to determine mechanisms underlying sodium channel periodic paralyses.
THE molecular mechanisms underlying cold sensitivity in metazoans are not well understood, not least because relatively few studies have addressed this question. A few cold-sensitive mutations have been described in Drosophila melanogaster (Siddiqi and Benzer 1976; Sondergaard 1980; Lilly et al. 1994), but again the precise mechanisms by which these specific mutations lead to cold sensitivity have not been elucidated. The objective of this study was to identify the biological processes underlying cold sensitivity.
The dominant cold-sensitive paralytic Out-cold (Ocd) mutations of D. melanogaster were originally isolated in a screen for X-linked male-lethal mutations (Sondergaard 1975, 1979). Further analyses of the male-lethal stocks identified seven lines in which heterozygous mutant females exhibited a dominant cold-sensitive paralysis. Within several of these lines a few mutant males did eclose, but they were weak, uncoordinated, and usually died in the wet food medium before they had the opportunity to mate. However, rare surviving males allowed crosses between strains to be undertaken and the seven apparently independently derived mutations were shown to be allelic. The seven mutant alleles of Ocd were designated Ocd1 to Ocd7 (Sondergaard 1975, 1979).
The effect of temperature on the Ocd mutant phenotype has been investigated in considerable detail (Sondergaard 1975). At 4°, all Ocd males are paralyzed within 1 min. Upon exposure to cold shock, Ocd mutants display paralysis-associated phenotypes such as leg shaking and wing flutter. Heterozygous (Ocd/+) females and the more severely affected rare hemizygous (Ocd/Y) males display reversible cold-sensitive paralysis. The survival of hemizygous mutant males is temperature dependent, with Ocd2 males eclosing from stocks maintained at 25°, but not at 18°. Surviving Ocd males walk in a reeling manner and are usually unable to fly, although the mutant phenotype varies greatly between individual flies. With respect to cold sensitivity and male viability, there is considerable variation between the mutant alleles in the severity of the phenotype (Sondergaard 1975, 1979). However, whether the phenotypic variation in severity is a property of the allele itself, or associated with the acquisition of modifier mutations in laboratory stocks, is at this stage unclear.
Initial investigations into the molecular basis of the Ocd cold-sensitive paralysis involved the study of the reaction kinetics of the mitochondrial enzyme complex succinate cytochrome c reductase (SCCR), which reflects the combined enzyme activity of complexes II and III of the respiratory chain (Sondergaard et al. 1975). Upon temperature decrease from 25° to 18–20°, abrupt changes in SCCR activation energy were observed in all Ocd lines tested (Sondergaard 1975, 1976, 1979), which correlates with the restrictive temperature for paralysis. This change in activation energy was also observed in wild-type flies, but occurs at a much lower temperature, 8°, below which wild-type flies also eventually succumb to paralysis. Furthermore, 2D gel electrophoresis apparently identified an abnormal pattern of mitochondrial polypeptides in Ocd adults (Sondergaard 1986), further implicating the Ocd gene product in mitochondrial function. However, it is equally possible that the mitochondrial defect is a downstream effect of susceptibility to cold-induced paralysis, rather than a direct result of the Ocd mutations themselves.
Classical genetic mapping using deletions and duplications placed the Ocd mutations within 13F1-16A2, a broad region on the X chromosome encompassing several hundred genes (Sondergaard 1975). Two recessive temperature-sensitive paralytic mutants, paralyticts1 (parats1) and shibirets (shits), and the leg-shaking mutant Hyperkinetic (Hk), were crossed to Ocd1, but no evidence of allelism was revealed (Sondergaard 1975).
Although there may be only one X-linked gene that mutates to result in dominant cold-sensitive paralysis (Sondergaard 1979), mutations in other X-linked genes do lead to recessive cold sensitivity. For example comatose (comt), which encodes an N-ethylmaleimide-sensitive fusion protein and is important for neurotransmitter secretion (Ordway et al. 1994; Dellinger et al. 2000), shows recessive cold-sensitive paralysis (Siddiqi and Benzer 1976). However, comatose maps genetically to 11E8, too far a distance from Ocd for the two to be allelic. A second X-linked mutant, hypoactive D (hypoD), which exhibits recessive cold-sensitive paralysis at 15° and maps within the Ocd region at 13F1-14B2, is an allele of slow receptor potential (slrp) (Homyk and Pye 1989). However, slrp has not yet been cloned.
Several temperature-sensitive paralytic Drosophila mutants have been isolated with malfunctions in ion pumps or channels. This includes recessive mutations in the voltage-gated calcium channel gene cacophony (cac) (Kawasaki et al. 2000) and the voltage-gated sodium channel gene paralytic (para) (Loughney et al. 1989). A P-element insertion in the sodium pump α-subunit gene has also been shown to result in recessive cold-sensitive lethality (Feng et al. 1997). The para gene encodes the major functional voltage-gated sodium channel in Drosophila (Loughney et al. 1989) and two para mutants isolated in screens for smell insensitivity, parasmellblind1 and parasmellblind2, are associated with recessive cold-sensitive lethality (Lilly et al. 1994). The para gene lies within the Ocd critical region, and several previously characterized para mutations exhibit heat-sensitive paralysis (Suzuki et al. 1971; Loughney et al. 1989). However, while para is an attractive candidate host of the Ocd mutation, it was previously excluded on the basis of complementation tests (Sondergaard 1975). Nevertheless a cold-sensitive human disorder associated with mutations of a sodium channel gene manifests as the rare autosomal dominant muscle disease paramyotonia congenita (Ptacek et al. 1992).
D. melanogaster models of genetic disease have proven to be invaluable in elucidating the pathogenic mechanisms behind many human disorders (Fortini and Bonini 2000; Kornberg and Krasnow 2000; Zoghbi and Botas 2002; O'Kane 2003; Jacobs et al. 2004; Bier 2005; Chintapalli et al. 2007). Such models therefore have the potential to aid the development of therapeutic interventions. Here we report the molecular and phenotypic characterization of the Drosophila Out-cold (Ocd) mutations, and speculate on how they may be exploited as a tool with which to investigate human disorders associated with seizures, paralysis, and cold sensitivity.
MATERIALS AND METHODS
Drosophila stocks:
Fly stocks were maintained in plastic vials or bottles (for large-scale amplification) containing food medium (1% (w/v) bacto-agar, 1.5% (w/v) sucrose, 3% (w/v) glucose, 3.5% (w/v) active dried yeast, 1.5% (w/v) maize meal, 1% (w/v) wheat germ, 1% (w/v) soya flour, 3% (w/v) treacle, 0.5% (v/v) propionic acid, 0.1% (w/v) nipagin m in H2O). Stocks were reared at 25° unless otherwise stated, with ambient humidity on a 12/12 hr light/dark cycle. Basic techniques for the laboratory culture of Drosophila were as described by Ashburner (1989). In fly crosses, six females were crossed with six males of the required genotype in each vial. The parent flies were removed prior to the next generation hatching.
The wild-type control strains, Oregon-R and Canton-S, have been maintained as mass-bred stocks in the O'Dell laboratory in Glasgow for many years. The generation of the EMS-induced Ocd lines has been described previously (Sondergaard 1975, 1979). The recessive lethal strain paraDf(1)D34 contains an internal deletion within para that causes a null mutation (Broadie and Bate 1993). The hangoverAE10NT strain has a P-element insertion within the coding region of the first exon of hangover (hang), causing an abnormal response to environmental stressors such as ethanol and paraquat (Scholz et al. 2005) and was kindly provided by H. Scholz. The slrp4 strain has an EMS-induced mutation in the slow receptor potential gene (Homyk and Pye 1989).
For the gene rescue experiments, the para strain P{UAS-para13.5} was used to expresses para splice variant 13.5 under P{GAL4} control (Warmke et al. 1997). UAS-para13.5 was produced via standard methods from a para13.5-containing pGH19 construct kindly provided by J. Warmke. The P{GAL4} lines used to drive P{UAS-para13.5} were P{GAL4-Heat-shock protein 70.PB} (Jarman et al. 1993), P{GAL4-daughterless.G32} (Wodarz et al. 1995), P{GAL4-1407} (Sweeney et al. 1995), P{GAL4-Myocyte enhancing factor 2.R} (Ranganayakulu et al. 1996), and P{GAL4-eyeless} (Bonini et al. 1997). The P-element insertion lines used for red/white selection in mapping Ocd, P{SUPor-P}kat80KG02315, P{EPgy2}EY04615, P{EPgy2}EY3459, P{EPgy2}CG4239EY01983, and P{GT1}BG00710, were obtained from the Berkeley Drosophila Genome Project (BDGP) Gene Disruption Project (Bellen et al. 2004).
The Out-cold strains have been kept in isolation since their creation (L. Sondergaard, personal communication). While they retain their cold sensitivity, it is very likely, even in a balanced stock where the mutant Ocd allele is usually present only in a heterozygous state, that selection will act to modify and mollify the dominant mutant phenotype. To address this the strains were isogenized into two wild-type backgrounds, Canton-S and Oregon-R. To avoid problems with selection, the Ocd mutations were isogenized using chromosome replacement. Note that the isogenized strains retain their original Ocd mutant X chromosome. Therefore ∼80% of the isogenized strain is from the wild-type Canton-S or Oregon-R line (chromosomes 2 and 3), and 20% (the X chromosome) is from the original Ocd strain.
SNP mapping:
Sequencing coding and noncoding regions of genes in the original Ocd critical region (13F1-16A2) identified eight SNPs distinguishing the Ocd1 and Oregon-R (ORR) strains (Table 1). The SNPs were identified during sequencing coding and noncoding regions of “mitochondrial” candidate genes within the original Ocd critical region (13F1-16A2). This identified SNPs in CG9240, CG8288 (mRpL3), CG8931, CG3525 (eas), and CG3560. To fill in appropriate gaps for fine recombination SNP mapping, SNPs were subsequently identified in disco-r, para, and rudimentary. Flies carrying a recombinant Ocd/ORR chromosome were generated, and in these the eight SNPs were genotyped using the ABI Prism SNaPshot primer extension kit (Applied Biosystems). The genotypes and phenotypes of recombinant flies allow an estimation of the relative position of Ocd1 to specific SNPs.
P-element mapping:
Using red/white selection (Zhai et al. 2003), the Ocd mutation was mapped relative to five molecularly defined P-element insertions obtained from the BDGP Gene Disruption Project (Bellen et al. 2004). Insertion sites were confirmed using inverse PCR, according to the BDGP protocol. The Ocd1 mutation was crossed into a w1118 background. Flies recombinant between Ocd1 and each P-element insertion were generated. The presence of each P-element was scored by the expression of the white gene. Presence of the Ocd1 mutation was scored by tests for cold sensitivity (at 4°). Genetic distances between Ocd and each P-element were calculated from the proportions of recombinant flies observed. Standard errors were calculated as the square root of PQ/n, where P is the percentage of recombinant genotypes, Q is the percentage of parental genotypes, and n is the sample size.
Gene rescue:
Flies carrying the construct P{UAS:para+} on either the second or third chromosome were crossed to Ocd1 flies using appropriate balancer chromosomes to generate Ocd1/FM7; P{UAS:para+}; + or Ocd1/FM7; + ; P{UAS:para+} females. In gene rescue crosses, these females were crossed to male flies that either were wild type (Canton-S or Oregon-R) or carried a P{GAL4} construct that directs wild-type para expression in a specific spatial and temporal pattern. The number of Ocd1 and FM7 males eclosing over 4 days from the first eclosion were counted, and the percentage of males carrying the Ocd1 mutant allele was calculated. Standard error of the proportion for each value was calculated as the square root of PQ/n, where P is the percentage of Ocd1 eclosing, Q is the percentage of FM7 males eclosing, and n is the sample size. 2 × 2 χ2-analysis was performed on the raw data to test for significant deviation from the pooled wild-type (Canton-S and Oregon-R) results.
DNA sequencing:
DNA sequencing reactions on purified PCR products were carried out by the Molecular Biology Support Unit at the University of Glasgow on a MegaBACE 1000 capillary sequencer (Amersham) using Big Dye chemistry (Applied Biosystems). Sequences were analyzed using ABI Prism EditView or 4peaks software.
Electrophysiology:
Electrophysiology experiments were performed using protocols described previously (Baines and Bate 1998; Mee et al. 2004). Voltage steps of 15 mV increments were applied to cells from a conditioning potential of −90 mV (steps ranged from −60 to +45 mV). To better resolve Na+ currents, an on-line leak subtraction protocol was used (P/4). Currents measured are the averages of five trials for each cell. Recordings were taken from aCC and RP2 neurons in first instar at three temperatures: 16°, 22°, and 28°. However, all larvae were raised at 25°. Student's t-tests were used to test for significant deviation from wild type.
DDT bioassay:
For DDT bioassays, 20 adult flies <72 hr posteclosion were placed in glass vials with interior surfaces evenly coated with varying concentrations of DDT (Sigma) dissolved in 200 μl acetone and allowed to air dry. The vials were sealed with cotton wool soaked in 5% sucrose solution. Mortality was scored after 24 hr with flies being unable to move being scored as dead. Dose-response curves were estimated from six different concentrations of DDT with three replicates per dose. Probit analysis was performed using the computer program POLO (Robertson et al. 1980).
Proteomics:
Protein was extracted from 12 adult male Drosophila aged between 1 and 2 days using protocols and statistical analyses described previously (Karp et al. 2005). Protein concentration was determined at the Cambridge Centre for Proteomics (CCP) using the BioRad DC protein assay (Bio-Rad). 2D DIGE was performed at the CCP as described previously (Swatton et al. 2004). The DeCyder version 4 Biological Variation Analysis (BVA) (GE Healthcare) software module was used to identify spots with increased or decreased expression between samples, on the basis of calculated standardized abundances. Standardized abundance was calculated by dividing spot volumes by the Cy2 internal standard for each spot. Statistical analysis was applied using standard ANOVA (Karp et al. 2005). Spots with P < 0.05 for random occurrence were considered to differ significantly between samples. Protein gels were fixed and stained using Colloidal Coomassie Brilliant Blue and spots of interest excised manually within a laminar flow cabinet. Mass spectrometry to identify proteins from 2D DIGE experiments was performed at the CCP. Protein spots within the gel were first reduced, carboxyamidomethylated, and then digested to peptides using trypsin on a MassPrepStation (Waters, Manchester, UK), before being applied to LC-MS/MS. For LC-MS/MS, the reverse phase liquid chromatographic separation of peptides was achieved with a PepMap C18 reverse phase, 75 μm i.d., 15-cm column (LC Packings, Amsterdam) on a capillary LC system (Waters) attached to QTof2 (Waters) mass spectrometer. The MS/MS fragmentation data achieved was used to search the National Center for Biotechnology Information (NCBI) database using the MASCOT search engine (http://www.matrixscience.com).Probability-based MASCOT scores were used to evaluate identifications. Only matches with P < 0.05 for random occurrence were considered significant (further explanation of MASCOT scores can be found at http://www.matrixscience.com).
RESULTS
Out cold is allelic to the voltage-gated sodium channel gene paralytic:
Genes that mapped within the Ocd critical region (13F1-16A2) and were considered to be directly or perhaps indirectly involved in mitochondrial function were sequenced. In the Ocd1 strain none of the five candidate genes, CG9240, mitochondrial ribosomal protein L3, CG8931, easily shocked, and CG3560, had coding mutations (Table 1). However, noncoding SNPs from each of the five genes were identified and, coupled with SNPs from three other genes, disco-r, paralytic (para), and rudimentary, were used to facilitate more precise mapping of Ocd. In addition, crosses performed between mutant alleles of Ocd and slrp (hypoD) did not provide any evidence that the two mutations are allelic (data not shown).
SNPs identified across the Ocd region
The Ocd (13F1-16A2) region of the X chromosome represents ∼2% of the fly genome, and ∼300 genes. Five P-element insertions (Figure 1) and eight SNPs (Table 1) were used to identify the position of the Ocd mutation more precisely. This identifies an insertion 5′ of the 5′-UTR of hangover (hang) and a SNP within intron 6 of para as the closest upstream and downstream flanking markers of Ocd (Figure 2). This mapping confined the Ocd critical region to only six candidate genes, hang, CG9947, CG4420, eIF2a, Rbp2, and para. Sequencing the coding regions of the first five of these genes in Ocd1 failed to find any mutations affecting their amino acid sequences.
The Ocd region showing relative positions of P-element insertions. P elements are represented by inverted triangles; the Ocd gene is represented by a circle. The map shows relative positions and the calculated genetic distances (cM) between each P-element insertion and Ocd.
Topographical representation of the Ocd region. The Ocd region encompasses six genes over the 100-kb 14C6–14E1 region. The para gene is ∼64 kb of the region. The P element defining the left-hand limit (BDGP line BG00710) is represented by an inverted triangle. The SNP defining the right-hand limit (AE003502.4 nt 123890 G/C; intron 6 of para) is indicated by a diamond. The arrow indicates the direction of transcription of para. The approximate position of the three Ocd point mutations is shown by a star.
Complementation tests are complicated by the fact that Ocd cold sensitivity is dominant. In addition, of the six candidate genes, mutant strains are only available for hangover and para. However, hemizygous Ocd/Y males and homozygous Ocd/Ocd females exhibit recessive semilethality and a recessive droopy wing phenotype. Therefore we would expect females heterozygous for a cold-sensitive Ocd allele and an Ocd null allele to have a similar semilethal and droopy wing phenotype. Irrespective of the temperature they are grown at, Ocd/hang females derived from Ocd/FM7 mothers and hang/Y fathers are cold sensitive (due to the Ocd dominant cold sensitivity). In addition they are viable and fertile and exhibit no other obvious mutant phenotypes, providing no evidence that Ocd and hang are allelic. However, the phenotype of Ocd/paraDf(1)D34 females depended on the temperature at which they were generated. When conceived and raised at 25° Ocd/paraDf(1)D34 females were cold sensitive and indistinguishable from their Ocd/+ sisters. However, when conceived and raised at 18° Ocd/paraDf(1)D34 females displayed phenotypes distinct from that of either heterozygous parent in that they were uncoordinated, holding their wings down in a drooped position, and were generally unfit, manifesting a phenotype that is remarkably similar to that of Ocd/Y mutant males (Figure 3, A and B). Given that paraDf(1)D34 is a null internal deletion of para, this strongly suggests Ocd and para are allelic.
Noncomplementation of Ocd and paraDf(1)D34. (A) If there was complementation between Ocd and para we would expect similar frequencies (50%) of Ocd/paraDf(1)D34 and Ocd/FM7 daughters. At 25° there is a significant deficit of Ocd7/paraDf(1)D34 daughters (χ2 = 12.16, P < 0.005). At 18° there is a significant deficit of all Ocd/paraDf(1)D34 daughters (Ocd1 χ2 = 6.52, P < 0.05; Ocd5 χ2 = 28.96, P < 0.005; Ocd7 χ2 = 26.28, P < 0.005). (B) All females with the Ocd7/paraDf(1)D34 genotype have the droopy-winged phenotype that is typical of Ocd7/Y mutant males.
Gene rescue:
Proof that the Ocd mutations are in the para gene can only be demonstrated by gene rescue, where expression of a wild-type para transgene in an Ocd mutant reverts the host fly to a wild-type phenotype. Again this is complicated by the fact that the primary Ocd phenotype, cold sensitivity, is dominant. However, Ocd1/FM7 mothers produce sons in the ratio of ∼9 FM7/Y to 1 Ocd1/Y rather than the expected 1:1, a phenotype described as semilethality. It is therefore theoretically possible to assay for any para-mediated rescue of Ocd1 male semilethality, by expressing the wild-type para transcript, which would be expected to restore the ratio of FM7/Y to Ocd1/Y males to ∼1:1. Gene rescue is further complicated by the size, splicing complexity, and RNA-editing of the para transcript. Given these complicating factors, we felt there was little prospect of achieving complete gene rescue, but nevertheless introduced para13.5 transgenes into Ocd1 stocks in the hope of detecting a significant shift from a 9:1 toward a 1:1 ratio.
For all four P{GAL4} drivers used to drive UAS-para13.5, χ2 tests reveal a significant increase (P < 0.001) in the number of Ocd1/Y males eclosing (Figure 4). The fact that para expression significantly rescues the viability of Ocd1/Y males strongly suggests that Ocd and para are allelic. Perhaps surprisingly, the observed rescue in the viability of Ocd1/Y males does not seem to discriminate between the different categories of P{GAL4} driver used. As para is normally expressed neuronally it is expected that P{1407:GAL4}-driven expression of para+ panneurally would result in rescue, as would ubiquitous P{da:GAL4}-driven expression of para+. However, it is difficult to explain why expression of para+ in the muscle by P{dmef2:GAL4} or in the eye discs by P{ey:GAL4} also results in significant rescue. One possible explanation is that the P{GAL4} drivers may not be as restricted in their expression pattern as previously described. Interestingly, relative to wild-type controls neuronal excitability, itself dependent on para expression, is increased in UAS-para13.5 stocks even in the absence of a P{GAL4} driver (R. Baines, unpublished data).
Rescue of Ocd1/Y semilethality by P{GAL4}-mediated expression of para+. Relative percentages of Ocd′/Y and FM7/Y males in the presence of named P{GAL4} drivers, P{UASG:para+} or both. χ2 tests reveal that the relative percentage of Ocd1/Y males eclosing is significantly higher when the para transgene is expressed (P < 0.001).
Ocd strains show missense mutations in highly conserved regions of para:
Sequencing of coding regions of the para gene revealed missense mutations within exon 28 in six of the Ocd strains, but no coding mutations were found elsewhere. The mutations uncovered were I1545M and G1571R, somewhat surprisingly occurring as a double mutation in Ocd1, Ocd3, Ocd5, Ocd6, and Ocd7, and T1551I in Ocd2 (Figure 5). None of these base-pair changes have been observed in any of the control lines tested, including the original progenitor Oregon-R stock. No mutation was found within the coding region of para in Ocd4. However, on closer investigation the Ocd4 strain exhibited male lethality without the expected dominant cold sensitivity and was not used further in this study.
Positions of Ocd mutations within Para. (A) Amino acid changes caused by Ocd mutations within specific domains of Para. The conserved IFMT box is also highlighted. (B) Positions of Ocd mutations on the Para sodium channel Ocd1 double missense mutation shown as dark circles; Ocd2 single missense mutation shown as dark square.
Amino acid sequence alignment of Para and the human voltage-gated sodium channels reveals striking similarity. Indeed, Para is 45% identical to the skeletal muscle sodium channel gene SCN4A, and the neural genes SCN1A and SCN2A, and 44% identical to the cardiac sodium channel gene SCN5A. The I1545M and T1551I Ocd mutations, caused by A-G and C-T transitions, respectively, both lie at the intracellular phase of segment S6 of domain III of the sodium channel. This is a highly conserved region and each of the three residues mutated in Ocd is conserved from Drosophila to man (Figure 5). The G1571R Ocd mutation, caused by a G-C transversion, resides within the cytoplasmic linker between domains III and IV, which contains the putative inactivation gate.
Ocd larval motor neurons have reduced voltage-gated sodium currents:
Previous analyses of human SCN4A derived from myotonic patients indicate that single amino acid substitutions change the activation threshold and/or rate of fast inactivation. A voltage clamp study was therefore undertaken to study the electrophysiological phenotype of Ocd1 (I1545M and G1571R) mutant larvae. Whole cell recordings were taken from wild-type and homozygous Ocd1/Ocd1 first instar larvae at three temperatures (16°, 22°, and 28°) to assess any temperature sensitivity. Recordings were taken from aCC or RP2 motor neurons, which do not differ in sodium conductance (INa). Peak transient INa was recorded at a range of voltages from −60 to 45 mV (in 15-mV steps). At each temperature, the sodium current-voltage relationships for INa were significantly reduced in Ocd1 larvae (Figure 6). This reduction is, however, more marked at 16° than at 28°, showing that INa is cold sensitive in these mutants (P < 0.05 at 22° and 28°, P < 0.001 at 16°). While temperature does not seem to have a great affect on wild-type currents, Ocd1 larvae display striking cold sensitivity. Analyses of current-voltage relationships, however, show no evidence for any change in activation voltage (Figure 6, C–E). This would suggest that the reduction in current flow is not due to changes in membrane voltage required for channel opening. The amino acid changes in Ocd1, particularly the G to R substitution in the III-IV linker (see Figure 5), are very close to the IFMT sequence that is a component of the fast inactivation particle that allows opened channels to close (i.e., inactivate) within a few milliseconds and to remain closed until membrane potential repolarizes. It is possible, therefore, that either one of these changes might reduce inward INa by increasing the speed of onset of fast inactivation. Further analysis will be needed to clarify whether this is indeed the case or whether current reduction is due to other mechanisms (such as reduced channel density).
INa is substantially reduced and shows cold sensitivity in Ocd1 homozygous mutant females. (A) Peak transient INa is significantly reduced in Ocd1 mutants at the three test temperatures used (16, 22, and 28°). (B) A typical voltage clamp recording of INa from an aCC motoneuron in a wild-type and Ocd1 mutant first instar larva (22°). (C–E) Current-voltage plots for INa in both wild-type and Ocd1 mutants at the three test temperatures. For all experiments n ≥ 5 ± SE. CS, Canton-S wild type.
Ocd mutants show altered resistance to DDT:
Because Ocd mutant lines encode sodium channels that differ at single-amino-acid sites, they may respond to specific neurotoxins in a novel fashion. Several neurotoxins, including DDT, target the S6 segment of domain III, the site of two of the three Ocd missense mutations identified, I1545M and T1551I. To test for resistance, Ocd flies were subjected to a standard DDT contact assay. Relative to the Ocd progenitor Oregon-R control, heterozygous Ocd2/+ (T1551I/+) females appear to show a mild DDT sensitivity (0.19 resistance compared to wild type), whereas Ocd5 (I1545M and G1571R) females display a remarkable 1000-fold increase in resistance (Figure 7 and Table 2). The I1545M mutation lies within domain III S6 and the G1571R substitution in the domain III-IV linker. This suggests that the I1545 and/or G1571 residues may be crucial novel targets for DDT.
DDT mortality assay in wild-type and Ocd mutants. Probit analysis of dose-response relationships for wild-type and Ocd mutants. Wild type (circles), Ocd2 (squares), Ocd5 (triangles): females (closed), males (open). Ocd2 males are not viable; Ocd2 females are heterozygous.
Resistance levels of wild-type and Ocd lines to DDT
Several mitochondrial proteins are differentially expressed in Ocd flies:
To determine, using state-of-the-art methods, whether Ocd mutants have a mitochondrial phenotype as reported previously, a comparative proteomics (2D DIGE) study was performed. In 2D DIGE, samples are labeled with spectrally distinct fluorescent dyes and run together on the same gel to overcome problems of reproducibility and intergel variability. The Ocd2 mutation (T1551I) was selected for analysis as this line yields the most viable males and also because these male flies display mutant phenotypes even at the permissive temperature of 25°. Protein analysis was performed on isogenized whole adult flies controlled for all aspects of environment. All flies were grown at 25° and protein was extracted in three biological replicates per line. Following gel electrophoresis and image analysis using DeCyder software, protein spots showing changes in abundance between wild-type and Ocd2 males were excised from the gel and identified using mass spectrometry. In total, >40 protein differences were identified. Of these, 22 were successfully identified, using mass spectrometry to search the NCBI databases for Drosophila matches. Those with a significant (P < 0.05) difference in abundance are shown in Table 3. Interestingly, of 9 different proteins identified, 5 are mitochondrially localized. Four proteins, fumarate hydratase, citrate synthase, aconitase, and isocitrate dehydrogenase, are all key enzymes of the tricarboxylic (TCA) cycle and are all upregulated in Ocd flies. Taken together this suggests an increase in flux capacity in the TCA cycle, which might support enhanced ATP synthesis. The fact that such a large proportion of the proteins identified in the 2D DIGE study are localized to the mitochondrion is consistent with the original proposal that Ocd may play some role in mitochondrial bioenergy metabolism. Alternatively, and in light of the fact that it is known Ocd encodes a voltage-gated sodium channel, it may be that the mitochondria alterations provide a route by which symptoms induced by the mutation may be secondarily alleviated.
Proteins identified as being differentially expressed in wild-type and Ocd flies
DISCUSSION
The Ocd mutations are alleles of para:
The mapping, complementation, gene rescue, and gene sequencing data provide compelling evidence that the Ocd mutations are alleles of the voltage-gated sodium channel gene paralytic (para). This is further supported by the observation that the Ocd mutations share phenotypes with previously identified para alleles, including resistance to DDT and reduced sodium ion current in motor neurons. It is interesting that the possibility of para being the host Ocd gene was previously excluded (Sondergaard 1975). The critical difference between previous complementation tests and those reported here is the use of a null para allele in a crossing scheme performed at the restrictive cold temperature of 18°.
The Ocd mutations are in conserved domains of Para:
Sequence analysis revealed two classes of missense mutations that underlie the Ocd phenotype, the single missense T1551I mutation unique to Ocd2, and the double missense I1545M and G1571R mutation which is curiously present in five Ocd alleles (Ocd1, Ocd3, Ocd5, Ocd6, and Ocd7). These lines were originally reported as being independently derived (Sondergaard 1979), but it is difficult to comprehend how five identical double missense mutations could arise independently, unless the variant was already present in the stock from which the mutants were isolated. Furthermore, while all five I1545M and G1571R stocks have similar dominant cold-sensitivity phenotypes, they do vary in terms of relative male viability and fertility. This strongly suggests that selection for different modifier genes has occurred within the different Ocd I1545M and G1571R stocks. All three missense mutations reside within conserved, functionally important parts of the sodium channel protein. Furthermore, analysis of the Para protein and closely related sequences using the SIFT amino acid substitution prediction program (Ng and Henikoff 2001) predicts that each of the Ocd mutations (I1545M, T1551I, and G1571R) will affect Para protein function.
The effect of Ocd mutations on sodium channel function:
The major functional sodium channel in Drosophila is encoded by para (Loughney et al. 1989). Although both para and DSC1, a putative voltage-gated sodium channel gene, are widely expressed in the CNS and PNS (Hong and Ganetzky 1994), mutations in para are sufficient to block nerve action potentials. For example paraDf(1)D34, an internal deletion within para, abolishes sodium current (Mee et al. 2004). Analyses of SCN4A derived from myotonic patients show that single-amino-acid substitutions alter the activation threshold and/or rate of fast inactivation (Cannon 2002; Wu et al. 2005). Ocd larval motor neurons show decreases in peak transient sodium currents, suggesting that the mutant channels are compromised in their ability to transport sodium ions. This might be due to a partial blockage of the sodium channel pore. In Ocd flies some sodium current is still present at 16°, indicating that action-potential firing is still occurring at this lower temperature, although it may be slowed or uncoordinated. Further characterization of the electrophysiological phenotype of Ocd utilizing heterologous channel expression is likely to establish how such mutations could lead to cold-sensitive sodium channel dysfunction.
Sodium channel function can also be dissected by exposure to neurotoxic insecticides. Indeed, a wide variety of neurotoxins, including DDT and pyrethroids, target Drosophila para sodium channels by binding to specific domains within the molecule (Baines and Bate 1998; Zlotkin 1999; Wang and Wang 2003; ffrench-Constant et al. 2004). Interestingly, DDT and pyrethroids target para sodium channel domain III S6 (Pauron et al. 1989), the site of the Ocd mutations. Several para mutants originally isolated on account of their temperature sensitivity have since been shown to confer insecticide resistance (Pittendrigh et al. 1997). This includes para74 (M1536I), which also resides in domain III S6 and is known to confer resistance to both DDT and pyrethroids, and paraDN7 (Pittendrigh et al. 1997; ffrench-Constant et al. 1998; Martin et al. 2000). Pyrethroids and DDT normally function to cause persistent activation of sodium channels (Soderlund and Bloomquist 1989). The fact that viable homozygous Ocd5 (I1545M and G1571R) females show a 1000-fold increase in resistance to DDT, and heterozygous Ocd2 (T1551I) females appear to show a mild DDT sensitivity, suggests that the I1545 and/or G1571 residues may be crucial targets for DDT.
Ocd and mitochondrial function:
Speculation that the Ocd mutation had a mitochondrial function was based on the observation of an unidentified aberrant protein in Ocd mitochondria (Sondergaard 1986) and abnormalities in the reaction kinetics of the mitochondrial respiratory chain SCCR activity in Ocd mutants (Sondergaard 1976, 1979). Interestingly, changes in the activation energy of mitochondrial enzymes in para mutants have also previously been observed (Sondergaard 1976). Using males tightly controlled for genetic background and inbreeding, 40 significant proteomic differences between wild-type and Ocd7 males were revealed. The 16 identified protein differences represent 10 proteins, 6 of which are mitochondrial. Interestingly 4 of these are critical components of the TCA cycle and the others are the α- and β-subunits of ATP synthase. All are expressed at high levels in Ocd7 mutant males. If the primary effect of Ocd on sodium channel function is to slow inactivation, then there will be an increased influx of sodium ions into the cell. The mitochondria may well respond to this increase in sodium ion concentration by increasing ATP synthesis, specifically by increasing levels of key TCA cycle enzymes and ATP synthase. This would facilitate subsequent energy-dependent clearance of excess intracellular sodium.
Ocd as a model of human sodium channel disorders:
Several human disorders associated with myotonia and periodic paralyses are caused by mutations in the skeletal muscle sodium channel SCN4A (reviewed by Caldwell and Schaller 1992; Cannon 1997, 2000; George 2005). These include hyperkalemic periodic paralysis (HYPP), paramyotonia congenita (PMC), potassium-aggravated myotonia (PAM), and hypokalemic periodic paralysis (HOKPP). These diseases are inherited in an autosomal dominant fashion and exhibit high penetrance. Symptoms are episodic and vary in severity both within and between affected individuals. SCN4A is only expressed at significant levels in skeletal muscle, so there are generally no cardiac or CNS symptoms in the patients. A plethora of SCN4A missense mutations have been described that result in the substitution of a highly conserved amino acid in the protein. In these mutant channels, permeation of sodium ions through the channel pore is apparently normal. The primary defect is instead an alteration in voltage-dependent gating, specifically the disruption of fast inactivation. This manifests itself as aberrant bursts of reopening of the channel and prolonged durations of the open state. Several identified mutations lie in the cytoplasmic loop, linking domains III and IV, which contains the inactivation gate. Other mutations lie at the cytoplasmic ends of S5 or S6, which may form part of the inner part of the ionic pore to which the inactivation gate binds, namely the docking site.
From a human genetic disease perspective the Ocd G1571R mutation is of particular interest. The mutated glycine residue is conserved from Drosophila to man and lies very close to the hydrophobic motif MFMT (residues 1576–1579), believed to directly interact with the docking site. Interestingly, mutations in the orthologous residue, G1306, in the human skeletal muscle sodium channel SCN4A, have already been found in cases of the human cold-sensitive muscle disorder paramyotonia congenita (G1306V) (McClatchey et al. 1992). It is remarkable that mutations at the same residue result in cold-sensitive phenotypes in both Drosophila and humans. Other SCN4A G1306 mutations have been uncovered in two families with potassium-associated myotonia, one of which is associated with a G1306A substitution (Ricker et al. 1994), while the second is caused by a G1306E substitution (Lerche et al. 1993). To date, no G1306R mutation in SCN4A has been reported. The effect of the three SCN4A G1306 mutations (G1306A, G1306E, and G1306V) on gating revealed both inactivation and activation defects (Mitrovic et al. 1995). The scale of channel defect correlates with the severity of symptoms observed in patients, with G1306A being the least affected and G1306E the most severe. G1306A has also been associated with succinylcholine-induced masseter muscle rigidity, a complication associated with anesthesia that can be fatal.
Cold-sensitive mutations in the human skeletal sodium channel genes, SCN4A, include R1448H and M1360V (Mohammadi et al. 2003) and T1313A (Bouhours et al. 2004). The molecular mechanisms underlying the temperature sensitivity of this muscle disorder are not well understood. Nevertheless, a decrease in temperature is known to slow down sodium channel kinetics and reduce the number of excitable sodium channels through hyperpolarizing shifts in slow inactivation (Ruff, 1999). Again the precise mechanism by which the R1448H, M1360V, and T1313A mutations induce cold sensitivity of sodium channels has not been elucidated (Yang et al. 1994; Mohammadi et al. 2003; Bouhours et al. 2004).
At present no mammalian model exists for human skeletal muscle (SCN4A) disorders. An appropriate genetic model would undoubtedly aid in establishing the mechanisms by which inactivation is disrupted in SCN4A-associated sodium channel disorders, and how this disruption may be exacerbated or mitigated by factors such as activity, serum potassium levels, and temperature. Once such interactions are better understood, it should also be possible to develop improved pharmacological treatments.
D. melanogaster models of genetic disease have proven to be invaluable in elucidating the pathogenic mechanisms behind many human disorders (Fortini and Bonini 2000; Kornberg and Krasnow 2000; Zoghbi and Botas 2002; O'Kane 2003; Jacobs et al. 2004; Bier 2005; Chintapalli et al. 2007). As it is likely that the pathogenic mechanism of sodium channel-mediated cold sensitivity is similar in flies and humans, further characterization of the Drosophila Ocd lines should clarify how equivalent human mutations cause disease phenotypes. The Ocd lines will also be invaluable in identifying specific chemicals and pharmaceutical products that may be able to alleviate symptoms of human sodium channel dysfunction.
Acknowledgments
For technical support with mass spectrometry, we thank Julie Howard and Svenja Hester of the Cambridge Centre for Proteomics. Helen A. Lindsay was supported by a Biotechnology and Biological Sciences Research Council Studentship.
Footnotes
↵1 Present address: Yorkshire Regional DNA Laboratory, St James's University Hospital, Leeds, UK, LS9 7TF.
Communicating editor: R. Anholt
- Received May 2, 2008.
- Accepted June 23, 2008.
- Copyright © 2008 by the Genetics Society of America
