We are grateful to the genome project for sequencing cosmid C41G7 (chromosome I, +3.7 cM), which contains an ortholog of the SMN (Survival of Motor Neuron) gene. Mutations of the telomeric copy of the duplicate SMN gene leads in human to the second most frequent fatal autosomal recessive disease of childhood, spinal muscular atrophy or SMA (Lefebvre et al., Cell 80, 155-165, 1995). The illness is characterized by degeneration of motoneurons of the spinal cord. A tight correlation between clinical severity and SMN protein level was demonstrated in SMA patients and a strong protein expression was observed in the spinal motor neurons from control individuals (Lefebvre et al., Nature Genetics 16, 265-269, 1997). The function of the SMN gene product is unknown. The SMN protein appears to form a complex with spliceosomal snRNP proteins (Liu et al., Cell 90, 1013-1021, 1997) and to participate to spliceosomal snRNP biogenesis (Fischer et al., Cell 90, 1023-1029, 1997). We wish to characterize the nematode gene, its functions and regulations, in the hope that this knowledge may help device therapeutic approaches for the 1/6000 unfortunate homozygotes. No cDNA clone was available when we started the project a year ago. We isolated, cloned and sequenced the full length cDNA. It measures 0.7 kb, does not appear to be transpliced and exactly matches the genefinder prediction of the coding region (4 exons, 197 aa). We then generated transgenic strains expressing a translational fusion of GFP to the SMN genomic sequence, including the SMN promotor (0.53 kb of upstream sequence, touching the nearest upstream gene, which encodes a ncd-like kinesin). The SMN gene is fused to GFP-NLS (vector pPD95.70) 8 aa before the stop codon. Out of 44 Rol, only 2 transmitted. These transgenics fluoresce at very low levels in most tissues, with two dozen neurons (not motoneurons) a bit brighter than the other cells. In the course of these experiments, we observed that injection of the GFP fusion construct and the pRF4 Roller plasmid in the ratio of 50: 150 leads, 12 to 16 hours after the injection, to irreversible body muscle paralysis of the injected worms. The injectees look healthy, eat and lay between 60 and 140 eggs per hermaphrodite but move only the head. The eggs laid will either die as embryos (85%), or hatch as very short Lls, with a big head and an often knobbed tail (15%). We are intrigued by the paralysis which directly affects the injected animals and wonder if other cases of such "diffusible" effects were ever observed ? RNA interference experiments were also conducted. The injectees generated a few small round eggs, some dead embryos and a proportion of larvae that died as L2 or L3 and were either Unc (not paralysed, but uncoordinated like the acetylcholine receptor mutants) and/or strongly dumpy (like
dpy-17). But the most striking failure was in the differentiation of the germ line. Defects ranged from empty gonads (no or few germ line cells, but normal somatic gonad), to Glp-l like (only a few sperms), to small non growing ovocytes (vitellogenin accumulates in the intestine, a few tiny round eggs may be laid, none survives). A large number of animals also had misguided gonad(s), like Unc-5. Based on these phenotypic informations, we are selecting new alleles in standard mutagenesis screens, using mnDfl11 as uncovering deficiency and transgenics as wild type duplication. We are also looking at previously identified mutants in the area. In parallel, we started biochemical experiments. Rabbit antisera were raised against three peptides, and we are currently using them in immunoblots, and will try them soon in immunofluorescence.