Portal de Periódicos da CAPES

A Caenorhabditis elegans TGF-beta, DBL-1, controls the expression of LON-1, a PR-related protein, that regulates polyploidization and body length

2002; Springer Nature; Volume: 21; Issue: 5 Linguagem: Inglês

10.1093/emboj/21.5.1063

1460-2075

Koji Morita,

GDF15 and Related Biomarkers

Article1 March 2002free access A Caenorhabditis elegans TGF-β, DBL-1, controls the expression of LON-1, a PR-related protein, that regulates polyploidization and body length Kiyokazu Morita Kiyokazu Morita Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Present address: Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA Search for more papers by this author Anthony J. Flemming Anthony J. Flemming Department of Biological Sciences, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY UK Search for more papers by this author Yukiko Sugihara Yukiko Sugihara Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Search for more papers by this author Makoto Mochii Makoto Mochii Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Present address: Department of Life Science, Himeji Institute of Technology, Hyogo, 678-1297 Japan Search for more papers by this author Yo Suzuki Yo Suzuki Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA Search for more papers by this author Satoru Yoshida Satoru Yoshida Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Search for more papers by this author William B. Wood William B. Wood Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA Search for more papers by this author Yuji Kohara Yuji Kohara Genome Biology Laboratory, National Institute of Genetics, Mishima, 411-8540 Japan Search for more papers by this author Armand M. Leroi Armand M. Leroi Department of Biological Sciences, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY UK Search for more papers by this author Naoto Ueno Corresponding Author Naoto Ueno Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Search for more papers by this author Kiyokazu Morita Kiyokazu Morita Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Present address: Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA Search for more papers by this author Anthony J. Flemming Anthony J. Flemming Department of Biological Sciences, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY UK Search for more papers by this author Yukiko Sugihara Yukiko Sugihara Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Search for more papers by this author Makoto Mochii Makoto Mochii Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Present address: Department of Life Science, Himeji Institute of Technology, Hyogo, 678-1297 Japan Search for more papers by this author Yo Suzuki Yo Suzuki Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA Search for more papers by this author Satoru Yoshida Satoru Yoshida Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Search for more papers by this author William B. Wood William B. Wood Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA Search for more papers by this author Yuji Kohara Yuji Kohara Genome Biology Laboratory, National Institute of Genetics, Mishima, 411-8540 Japan Search for more papers by this author Armand M. Leroi Armand M. Leroi Department of Biological Sciences, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY UK Search for more papers by this author Naoto Ueno Corresponding Author Naoto Ueno Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan Search for more papers by this author Author Information Kiyokazu Morita1,2, Anthony J. Flemming3, Yukiko Sugihara1, Makoto Mochii1,4, Yo Suzuki5, Satoru Yoshida1, William B. Wood5, Yuji Kohara6, Armand M. Leroi3 and Naoto Ueno 1 1Department of Developmental Biology, National Institute for Basic Biology, and Department of Biomechanics, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan 2Present address: Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA 3Department of Biological Sciences, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY UK 4Present address: Department of Life Science, Himeji Institute of Technology, Hyogo, 678-1297 Japan 5Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, CO, 80309-0347 USA 6Genome Biology Laboratory, National Institute of Genetics, Mishima, 411-8540 Japan *Corresponding author. E-mail: [email protected] The EMBO Journal (2002)21:1063-1073https://doi.org/10.1093/emboj/21.5.1063 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Using cDNA-based array analysis combined with double-stranded RNA interference (dsRNAi), we have identified yk298h6 as a target gene of Caenorhabditis elegans TGF-β signaling. Worms overexpressing dbl-1, a TGF-β ligand, are 16% longer than wild type. Array analysis shows yk298h6 to be one of several genes suppressed in such worms. Disruption of yk298h6 function by dsRNAi also resulted in long worms, suggesting that it is a negative regulator of body length. yk298h6 was then mapped to, and shown to be identical to, lon-1, a known gene that affects body length. lon-1 encodes a 312 amino acid protein with a motif sequence that is conserved from plants to humans. Expression studies confirm that LON-1 is repressed by DBL-1, suggesting that LON-1 is a novel downstream component of the C.elegans TGF-β growth regulation pathway. Consistent with this, LON-1 is expressed mainly in the larval and adult hypodermis and has dose-dependent effects on body length associated with changes in hypodermal ploidy, but not hypodermal cell proliferation. Introduction Some of the most fundamental, but least understood, aspects of animal development are the mechanisms by which body size is determined (Conlon and Raff, 1999; Stern and Emlen, 1999; Day and Lawrence, 2000). Body size is also the most obvious way in which animal species differ from each other (Bonner, 1989). This is particularly true for nematodes, which vary in size between 3.0 × 10−1 and 8.0 × 104 mm, but are otherwise quite morphologically uniform (Flemming et al., 2000). Wild-type strains of the nematode Caenorhabditis elegans are 1.2 mm long. Mutations that affect one aspect of body size, length, have long been known in this worm (Brenner, 1974), causing adults to be either long (Lon) or small (Sma). The frequency and viability of these mutants indicate that the worm might be an especially good organism in which to study the molecular regulation of body size. Recent studies have shown that some C.elegans Sma mutants affect genes that encode components of a transforming growth factor-β (TGF-β) signaling pathway (Patterson and Padgett, 2000). These genes include: sma-6, a homologue of a vertebrate type I Ser/Thr kinase receptor (Krishna et al., 1999), daf-4, a type II Ser/Thr kinase receptor (Estevez et al., 1993) also used in dauer larva formation (Georgi et al., 1990; Ren et al., 1996) and sma-2, sma-3 and sma-4, cytoplasmic SMADs (Massague, 1996; Heldin et al., 1997), that translocate into the nucleus upon signal activation (Savage et al., 1996). Since DAF-4 can function as a type II receptor for BMP signaling in mammalian cells (Estevez et al., 1993) there appears to be substantial functional conservation between nematode and mammalian TGF-β signaling. More recently, DBL-1/CET-1, belonging to the TGF-β superfamily of proteins, has been identified as the ligand that triggers the sma signaling pathway (Morita et al., 1999; Suzuki et al., 1999). Loss-of-function mutations in dbl-1 also have a Sma phenotype, while worms that overexpress it are Lon. DBL-1 is presumed to be a secreted growth factor, but how it actually regulates body length remains largely unknown. One obvious way in which body length might be controlled by the TGF-β pathway is by regulation of cell proliferation. However, Sma and Lon mutants appear to have wild-type cell numbers (Morita et al., 1999; Suzuki et al., 1999; Flemming et al., 2000). This is not surprising, since ∼50% of the growth of C.elegans occurs during adulthood, when no cell proliferation takes place (Knight et al., 2001). This growth, and even part of larval growth, must then be due to increases in cell size (Flemming et al., 2000). Increases in cell size of at least some tissues such as the hypodermis and intestine are, in C.elegans, associated with increases in somatic ploidy via endoreduplication (Hedgecock and White, 1985). Recently, worms that lack TGF-β activity due to mutations in daf-4, sma-2 and dbl-1 have been shown to have reduced hypodermal ploidy relative to wild type (Flemming et al., 2000; Nyström et al., 2002). This observation also suggests a possible mechanism for body size evolution since the degree of somatic polyploidization appears to be, in part, correlated with evolved differences in body size among nematode species related to C.elegans (Flemming et al., 2000). Identification of further components of this pathway that affect endoreduplication is, therefore, of interest for an understanding not only of the regulation of body size, but also of nematode morphological diversity. Downstream target genes of dbl-1 signaling might be identified by mutational screens that enhance or suppress the Sma phenotype. However, this kind of genetic screen is often time-consuming and laborious. On the other hand, recent DNA macro/microarray technology (Galitski et al., 1999; Lockhart and Winzeler, 2000; Young, 2000) has permitted the simultaneous screening of thousands of expressed genes. It is even possible, in principle, to examine the expression of all 19 000 C.elegans genes in two different genetic backgrounds or growth conditions (Mochii et al., 1999; Reinke et al., 2000). Another powerful approach recently developed to address gene function is double-stranded RNA interference (dsRNAi), which permits the sequence-specific inactivation of genes (Fire et al., 1998; Bass, 2000; Zamore et al., 2000). dsRNAi has accelerated the high through-put analysis of gene function not only for C.elegans but also for other model organisms (Kennerdell and Carthew, 1998; Ngo et al., 1998; Sanchez Alvarado and Newmark, 1999; Chuang and Meyerowitz, 2000). Here we report large-scale analyses of gene expression profiles in different genetic backgrounds displaying Sma and Lon phenotypes. We show that one of the genes recovered from this screen has a Lon phenotype when its function is inhibited by dsRNAi. We show that loss-of-function mutations in a known regulator of body length, lon-1, interrupt this gene, and that lon-1 transcription is negatively regulated by dbl-1 signaling. lon-1 encodes an evolutionarily conserved protein that defines the ‘PR-1 related protein superfamily’ from yeast, plant, insect to human, whose biological functions are little known (Szyperski et al., 1998). Moreover, we provide evidence that DBL-1 regulation of hypodermal polyploidization is mediated by LON-1. Results Identification of DBL-1-regulated genes by differential hybridization analysis using cDNA-based macroarray As we reported previously, cDNA array of C.elegans mutants is an efficient way of screening signal-specific target genes (Mochii et al., 1999). To understand the molecular pathway regulating C.elegans body length, we used the same DNA macroarray and screened for genes regulated by dbl-1 signal. In this study, we compared the gene expression profile between dbl-1 null mutants displaying Sma phenotype, dbl-1(nk3) (Morita et al., 1999) and worms overexpressing dbl-1 (which contain a multi-copy of the dbl-1 genomic fragment) with Lon phenotype, ctIs40 (Suzuki et al., 1999). Out of ∼8000 genes arrayed on nylon membranes, 16 genes were found to be up-regulated and seven down-regulated. Gene identification and fold increases of hybridization intensity are summarized in Table I. Altered expression was confirmed for all identified genes by conventional northern blotting. Of the 23 genes identified by cDNA array signals as being regulated by DBL-1, we failed to confirm seven by northern blots. Table 1. List of genes identified and dsRNAi phenotype Clone name Gene Fold increase (HDF)a dsRNAi phenotype yk225d6 K04E7.1 2.3 Sma yk253f12 F25H8.5 2.2 no phenotype yk355c4 C17G1.6 3.4 molting defect yk412e5 Y38H6C.1 1.9 no phenotype yk608d12 no gene found 1.9 weak Sma yk174b12 Y57G11C.15 2.0 Let (embryonic lethal) yk479h1 C07E3.10 2.0 no phenotype yk532c7 ZK1290.8 2.0 no phenotype yk570a11 F27E11.3 2.5 no phenotype yk563h1 C44H4.3 1.9 no phenotype yk604a4 W04G3.8 2.0 Dpyb yk355e1 F44E2.4 2.8 no phenotype yk430h5 F23B2.11 3.0 no phenotype yk479e3 C44H4.2 2.0 no phenotype yk500g10 VC5.3 6.0 no phenotype yk44c1 K10C2.1 2.2 no phenotype yk257f11 T21C9.5 0.3 Uncc yk361c1 K10C2.3 0.5 no phenotype yk530e1 no gene found 0.4 no phenotype yk553c5 no gene found 0.5 no phenotype yk71c6 T21C9.2 0.5 no phenotype yk298h6 F48E8.1 0.5 Lon yk605h1 F55A4.2 0.5 no phenotype a The data are the average of four experiments. HDF, high density filter. b Dpy, Dumpy. c Unc, uncoordinated. yk298h6 encodes the lon-1 gene In order to examine the biological relevance of the identified genes to the determination of body length in relation to dbl-1 signaling, we carried out dsRNAi, a method developed recently by which particular genes can be rapidly inactivated (Fire et al., 1998). Among the genes regulated both up and down in our microarray screen, seven showed various dsRNAi phenotypes (Table I), some of which resembled the Sma and Lon phenotypes of known body length mutations. We focused on yk298h6 (F48E8.1), one of the genes suppressed by dbl-1 because its disruption by dsRNAi caused a typical Lon phenotype with a high frequency. Importantly, yk298h6 mapped close to the lon-1 locus (Figure 1A) by database analysis. Expecting that yk298h6 is the gene responsible for lon-1 mutation, we performed rescue experiments using a fragment of cosmid F48E8 harboring the entire yk298h6 sequence. The 8.0 kb genomic fragment was sufficient to rescue the lon-1 body length phenotype, indicating strongly that yk298h6 encodes the lon-1 product (Table II). This was also confirmed by identifying the mutational lesion in nine lon-1 alleles (Figure 1B). Sequence analysis of PCR-amplified lon-1 alleles identified missense and nonsense mutations that cause non-conservative amino acid substitutions in the following alleles: e1137, ct410, e185, ct411, n1130, e44 and sp1 (Figure 1B). Alleles sp3 and e43 each have a stop codon caused by a CG-to-TA transition that converts codon TGG (W83) to the opal terminator TGA and TGG (W150) into the amber terminator TAG, respectively. Based on these findings, we conclude that yk298h6 (F48E8.1 for gene) encodes lon-1. Figure 1.yk298h6 encodes the lon-1 gene. (A) The yk298h6 (F48E8.1) gene maps to the lon-1 locus. (B) Primary structure of LON-1 and point mutation site. LON-1 is a 312 amino acid protein. The N-terminal hydrophobic region is underlined. A putative N-glycosylation site is double underlined. The peptide sequence that generates the LON-1 antibody is boxed. Amino acid substitutions in lon-1 alleles are indicated below the sequence. e43 and sp3 each have a nonsense mutation (W150 to stop and W83 to stop, respectively). ct410, ct411, e44, e185, e1137, n1130 and sp1 have missense mutations. (C) The LON-1 Kyte–Doolittle hydrophobicity plot shows a hydrophobic stretch at the N-terminus of the protein (amino acids 4–18, see arrow). Download figure Download PowerPoint Table 2. Effect of lon-1 gene dosage on body length Genotype Body length (mm)a N N2 1.28 ± 0.03 85 lon-1(e185) 1.71 ± 0.01 93 lon-1(e43) 1.43 ± 0.04 87 lon-1(ct410) 1.60 ± 0.01 103 lon-1(sp3) 1.35 ± 0.05 19 lon-1(RNAi) 1.63 ± 0.06 35 lon-1(e43)/Dfb 1.49 ± 0.11 8 lon-1(sp3)/Df 1.29 ± 0.06 19 nkIs10c 1.00 ± 0.03 43 lon-1(e185);Exnk51[lon-1(+)] 1.24 ± 0.11 65 lon-1(e185);Exnk52[lon-1(+)] 1.19 ± 0.09 65 lon-1(e185);Exnk53[lon-1(+)] 0.89 ± 0.10 62 a Mean ± SE. N, measured animals. Body lengths of various mutants were measured using a micrometer at high magnification at 120 h post-hatching. b Deficiency line. c Lon-1-overexpressed Sma worm. LON-1 belongs to the PR-protein superfamily conserved among yeast, plant, insect and human, and may be a type II transmembrane protein The lon-1 gene encodes a 312 amino acid protein (Figure 1B). By a database search, we have found that genes encoding proteins homologous to lon-1 belong to the PR-protein superfamily and are present in other organisms including yeast, plant, insect and human (Szyperski et al., 1998) (see also Discussion). Although overall sequence similarity between them is not outstanding (>30%), 18 amino acid residues in the stretch of ∼100 amino acids compared are well aligned (Figure 2). Interestingly, almost all the amino acids identified as mutation sites in lon-1 alleles include those conserved between species, indicating that these amino acids are essential for LON-1 function (Figure 2, indicated by asterisk). There are no defined functional motifs within this family, but members of the family are predicted to be secreted or glycosylphosphatidylinositol (GPI)-anchored proteins. Two of the members of this family, pathogenesis related protein-1 (PR-1) in plant and glioma-pathogenesis related protein-1 (GliPR-1) in human, are secreted from cells (Szyperski et al., 1998). Yeast PR-1 like protein-3 (YPR-3) in yeast and cysteine-rich protease inhibitor are predicted to be GPI-anchored proteins (Hamada et al., 1998). However, the analysis of the LON-1 sequence shows that there is a hydrophobic stretch at the N-terminus of the LON-1 protein, indicating that LON-1 might be a transmembrane protein (Figure 1C). To characterize the LON-1 protein, we generated antibodies against a LON-1 peptide corresponding to 42–56 amino acid residues. This antibody (α-LON-1) detected ∼35 kDa of protein prepared from whole worm extract by immunoblotting (Figure 3A). The molecular mass of this protein is the same as that of the predicted LON-1 protein. Figure 2.Alignment of LON-1 protein with nine related proteins. The sequences are LON-1 (C.elegans), CG8483 (Drosophila), GliPR-1 (human) (Murphy et al., 1995), PR-1 (Triticum aestivum), SGP28 (human) (Kjeldsen et al., 1996), TPX-1 (testis-specific protein) (pig) (Kasahara et al., 1989), venom allergen 5 (Vespa mandarinia) (Lu et al., 1993), cystein-rich secretory protein (horse), cystein-rich protease inhibitor (mouse), YJH8 (yeast). Asterisks indicate the mutation sites identified in lon-1 alleles. Conserved amino acid residues (at least seven residues) are by bold type. Download figure Download PowerPoint Figure 3.LON-1 expression is negatively regulated by dbl-1 signaling. (A–C) Western blot analysis using α-LON-1 antibody. (A) Whole worm extracts of dbl-1(−), N2 wild-type, lon-1(++) and lon-1(−) alleles sp3, e185 and e43. The arrowhead indicates the 35 kDa LON-1 protein. Lower panel is the loading control staining with α-paramyosin antibody. (B) Cryostat sections of N2 worms were homogenized and extracted by Triton X-100 and fractionated by centrifugation (see also Materials and methods). LON-1 protein is detected in the membrane fraction but not in the soluble fraction. (C) The membrane fractions were treated with PI-PLC or alkaline wash (pH 11.5), or acid wash (pH 2.0). LON-1 protein remains associated with the membrane fraction. (D) Northern blot analysis. Poly(A)+ RNA (1.0 μg) prepared from mixed stages of dbl-1(−);dbl-1(nk3), N2, dbl-1(++);ctIs40 worms was blotted and hybridized with dbl-1, lon-1, CeIF (C.elegans initiation factor; for loading control) cDNA. The dbl-1 signal in dbl-1-overexpressing animals increased 3-fold compared with N2, and lon-1 in dbl-1-overexpressing animals decreased 5-fold compared with dbl-1 null animals. (E) lon-1::gfp expression is up- or down-regulated by dbl-1 dosage. The fluorescence indicating lon-1::gfp expression was increased with a dbl-1(−);dbl-1(nk3) background and decreased with a dbl-1(++);ctIs40 background. (F) Growth curve of N2 wild type, dbl-1(nk3), dbl-1(nk3);lon-1(e185) and lon-1(e185). The body length of mutants was measured at various times after hatching. Download figure Download PowerPoint To address the subcellular localization of LON-1 protein, we fractionated the extract of N2 worms. We detected the LON-1 protein in the membrane fraction, but not the soluble fraction, by western blot analysis (Figure 3B). These results indicate that LON-1 may be a membrane-integrated or -associated protein. Furthermore, the LON-1 protein remains associated with the membrane fraction after treatment with phosphatidylinositol-specific phospholipase C (PI-PLC) or an alkaline wash (Figure 3C). As the sequence following the signal sequence does not closely match with consensus cleavage sequence, we propose that the signal sequence is not cleaved and instead acts as an anchor region. We conclude that LON-1 is likely to be a type II integral membrane protein. Characterization of lon-1 alleles To analyze the relationship between the mutation site of lon-1 alleles and their phenotype, we compared the expression level of LON-1 protein and body length in mutants. lon-1(e185) is thought to be the strongest allele because they are the longest worms (Table II). By western blot analysis, LON-1 protein was found to be almost completely undetectable as a result of the point mutation (Figure 3A). Cysteine residues, including C185, are conserved in the PR-protein family and might form disulfide bonds, essential for the structural integrity of the protein. The missense mutation might cause misfolding of LON-1 by creating unpaired cysteine residues and may even prevent LON-1 from being exported from the endoplasmic reticulum, thereby resulting in a severe loss of LON-1 function. Compared with lon-1(e185), nonsense mutation alleles e43 and sp3 display mild phenotypes in terms of body length. LON-1 protein in e43 was weakly detected by western blot analysis (Figure 3A). The weak Lon phenotype of e43 and sp3 may be due to low levels of LON-1 protein produced by translational readthrough. Like other animals, C.elegans has tRNA[Ser]Sec that can insert selenocysteine at UGA codons (Lee et al., 1990). In prokaryotes, substitution of tryptophan at opal stops occurs due to third-position wobble in codon–anticodon recognition. A similar effect in daf-1 alleles is reported by Gunther et al. (2000). Next, we made e43/Df and sp3/Df animals. These animals have a longer phenotype than e43 and sp3 homozygotes. lon-1(e43) and some lon-1 alleles often produce stunted worms and show incomplete penetrance of the Lon phenotype. We cannot rule out that this is due to the secondary induced constriction behind the head that leads to auto-decapitation; nevertheless, dsRNAi of lon-1 seems to cause only the Lon phenotype (data not shown). We speculate that the lon-1 allele with the null or most severe phenotype is lon-1(e185). lon-1 expression is negatively regulated by dbl-1 signaling lon-1 was initially identified as being down-regulated in worms overexpressing dbl-1 in our cDNA macroarray (Table I). Northern blot analysis confirmed this, showing that lon-1 transcript levels seem to be regulated in a stepwise manner depending on the gene dosage of dbl-1 (Figure 3D). The change in expression level of lon-1 transcript between ctIs40;dbl-1-overexpressing worms and the dbl-1(−) mutation was estimated as ∼5-fold. As expected from the dbl-1-dependent repression profiled by cDNA array analysis and northern blotting, green fluorescent protein (GFP) fluorescence confirmed that the lon-1 promoter is negatively regulated by the dbl-1 signal. The fluorescence of lon-1::gfp appeared to be significantly down-regulated in worms overexpressing dbl-1;dbl-1 (++);ctIs40, compared with dbl-1(−) (Figure 3E). These results indicate that lon-1 expression is negatively regulated by dbl-1 signaling at the transcriptional level. To analyze the genetic relationship of dbl-1 and lon-1, we made double mutants of lon-1 alleles with dbl-1(−) and measured body length (Table IV). The lon-1(ct410);dbl-1 (nk3) double mutant was almost the same length as the lon-1(ct410) single mutant. This result indicates that the lon-1(ct410) mutation is completely epistatic to dbl-1. However, the double mutants of lon-1(e185) and lon-1 (e43) with dbl-1(nk3) had intermediate body lengths. The cause of this incomplete epistacy is unclear. Table 3. Expression of lon-1 cDNA in hypodermal, but not intestinal, cells rescued a lon-1 phenotype Genotype Transgene Expression sites Body length (mm)a N lon-1(e185) – – 1.71 ± 0.01 93 lon-1(e185) Pdpy-7 lon-1 hypodermal cells 1.12 ± 0.12 108 lon-1(e185) Pyk92e8 lon-1 intestinal cells 1.59 ± 0.01 60 lon-1(e185) Punc-54 lon-1 body wall muscle 1.60 ± 0.01 47 lon-1(e185) Pdbl-1 lon-1 neural cells 1.57 ± 0.01 37 lon-1(e185) Plon-1 lon-1 hypodermal and intestinal cells 1.30 ± 0.02 68 a Mean ± SE. Worms overexpressing LON-1 have a Sma phenotype Loss-of-function mutations in dbl-1 cause a Sma phenotype, while worms that overexpress dbl-1 are Lon (Morita et al., 1999; Suzuki et al., 1999). DBL-1 negatively regulates LON-1, and loss-of-function mutations in lon-1 give a Lon worm. If LON-1 is also a dose-dependent regulator of body length, overexpressing lon-1 should give a Sma worm. To address this possibility, we injected high concentrations of lon-1 DNA into lon-1 mutant or N2 worms (see Materials and methods). As predicted, higher dosages of lon-1 caused the Sma phenotype against both lon-1 and N2 backgrounds (Table II; data not shown). The body length of worms overexpressing lon-1;nkIs10 was close to sma mutants such as dbl-1, sma-2, sma-3 and sma-4 but, as discussed below, do not show the male tail phenotype (data not shown). To determine when mutants become longer than wild type, we measured the length of N2 wild-type, lon-1 (e185), dbl-1(nk3) and dbl-1(nk3);lon-1(e185) hermaphrodites at different times after hatching (Figure 3F). All mutants hatch as L1 larvae indistinguishable in size from wild-type larvae, but the Lon and Sma phenotypes are seen at late L1 stage, and these phenotypes are sustained through to adult stages. These results demonstrate that dbl-1 signaling is effective from all larval stages to adult. LON-1 is expressed in hypodermal and intestinal cells To examine the expression pattern of lon-1 in vivo at the cellular level, we stained the worms with LON-1 antibody (α-LON-1). Fluorescence was detected in almost all hypodermal and intestinal cells (Figure 4A). The signal of hypodermal cells is colocalized with the staining of MH27, which stains the adherens junction of the hypodermal cells (Figure 4A–C). Colocalization of LON-1 and MH27 antigen are also observed around hypodermal seam cells (Figure 4D–F). These expression patterns persisted through the subsequent larval stages to the adult stage. All signals were observed at the surface of the cell membrane. These observations are consistent with the result that LON-1 is a type II transmembrane protein. The signal disappeared completely in the presence of the antigen peptide (Figure 4G). In the lon-1(e43) nonsense mutation allele, LON-1 staining was substantially less (Figure 4F). Figure 4.LON-1 proteins are expressed in hypodermal and intestinal cells. All images are stained with purified anti-rabbit LON-1 peptide antibody (1:300 dilution) (green) and MH27 monoclonal antibody (1:100 dilution) (red). (A–C) N2 wild-type L1 stage animal. (A) Staining image of α-LON-1. The fluorescence is observed around the hypodermal cells and in the inner membrane of intestinal cells. (B) MH27. Signal is seen in the hypodermal adherens junction. (C) Merged image of (A) and (B). (D–F) Hypodermal region of N2 wild-type L3 stage animal. (D) α-LON-1, (E) MH27, (F) merged image of (D) and (E), the signals are observed around the seam cells. (