|Home | About | Journals | Submit | Contact Us | Français|
Caenorhabditis elegans has a single lamin gene, designated lmn-1 (previously termed CeLam-1). Antibodies raised against the lmn-1 product (Ce-lamin) detected a 64-kDa nuclear envelope protein. Ce-lamin was detected in the nuclear periphery of all cells except sperm and was found in the nuclear interior in embryonic cells and in a fraction of adult cells. Reductions in the amount of Ce-lamin protein produce embryonic lethality. Although the majority of affected embryos survive to produce several hundred nuclei, defects can be detected as early as the first nuclear divisions. Abnormalities include rapid changes in nuclear morphology during interphase, loss of chromosomes, unequal separation of chromosomes into daughter nuclei, abnormal condensation of chromatin, an increase in DNA content, and abnormal distribution of nuclear pore complexes (NPCs). Under conditions of incomplete RNA interference, a fraction of embryos escaped embryonic arrest and continue to develop through larval life. These animals exhibit additional phenotypes including sterility and defective segregation of chromosomes in germ cells. Our observations show that lmn-1 is an essential gene in C. elegans, and that the nuclear lamins are involved in chromatin organization, cell cycle progression, chromosome segregation, and correct spacing of NPCs.
The nuclear lamina is a filamentous meshwork that is present between the inner nuclear membrane and the peripheral chromatin. The inner nuclear membrane and the nuclear lamina are involved in organizing nuclear structure and regulating nuclear events. These include the organization of the higher order structure of chromatin and regulation of nuclear assembly and disassembly. The nuclear lamina is a primary target for caspases in apoptosis (reviewed in Goldberg et al., 1999b ). Lamins are the major proteins of the nuclear lamina. They are classified as type-V intermediate filaments and are composed of an α-helical rod domain flanked by a short amino (head) and a long carboxy (tail) domains. The rod domain of lamins is 52-nm long and contains four α-helices, each composed of heptad repeats. Coiled-coil interactions and head-to-tail associations between lamin monomers form 10- to 200-nm thick lamin filaments (reviewed in Stuurman et al., 1998 )
In vivo, lamin filaments are closely associated with the chromatin fibers (Belmont et al., 1993 ). In vitro, lamins can bind interphase chromatin (Hoger et al., 1991 ; Yuan et al., 1991 ; Taniura et al., 1995 ; Ulitzur et al., 1997 ; Goldberg et al., 1999a ), mitotic chromosomes (Glass and Gerace, 1990 ; Glass et al., 1993 ), or specific DNA sequences (Shoeman and Traub, 1990 ; Luderus et al., 1992 ; Luderus et al., 1994 ; Baricheva et al., 1996 ; Zhao et al., 1996 ). The binding site of vertebrate lamins to chromatin is localized to specific sequences in the tail domain and can be displaced with the core histones H2A and H2B (Taniura et al., 1995 ; Goldberg et al., 1999a ).
The composition of the nuclear lamina varies in different cell types and is under developmental regulation in both vertebrate and Drosophila cells (Stuurman et al., 1998 ). Metazoan cells contain between one to seven lamin proteins. Mammalian lamin A, C, AΔ10, and C2 proteins result from alternative splicing of the lamin A gene. Separate genes encode lamin B1 and B2. Lamin B3 is a germ cell-specific splicing variant of lamin B2. Mutations in the human lamin A gene cause the autosomal-dominant form of Emery-Dreifuss muscular dystrophy (EDMD) (Bonne et al., 1999 ) and Dunningan-type familial partial lipodystrophy (Cao and Hegele, 2000 ). Homozygous mice with a knockout in the lamin A gene appear normal at birth but retarded in their growth and die at 4 to 8 wk. These lamin A-deficient mice exhibit a cardiac and skeletal myopathy, with symptoms similar to human EDMD patients, and lack fat cells (Sullivan et al., 1999 ).
Two lamin genes are known in Drosophila, coding for lamins Dm0 and C (Gruenbaum et al., 1988 ; Bossie and Sanders, 1993 ). Lamin Dm0 is an essential gene encoding a type-B lamin that is required for nuclear organization. Flies homozygous for mutations in the lamin Dm0 gene have aberrant nuclear structure and die after 9 to 14 h of development, as the maternal pool of lamin Dm0 is diluted (Harel et al., 1998 ). Hypomorphic mutation in the lamin Dm0 gene (<20% of lamin expression) causes reduced viability, defective nuclear envelopes, and accumulation of annulate lamellae (Lenz-Bohme et al., 1997 ).
A search of the nearly complete genome of C. elegans reveals a single lamin gene, termed lmn-1. The lmn-1 gene is located on chromosome I, is 2.7-kb long, and is composed of 6 exons and 5 short introns (Riemer et al., 1993 ). The lmn-1 gene encodes a putative 64-kDa type-B lamin protein, termed Ce-lamin.
C. elegans has several features for analysis of the biological roles of the nuclear lamina: (A) C. elegans has only one lamin gene, and (B) C. elegans gene expression can be readily manipulated during early embryogenesis. In this study we have characterized Ce-lamin in wild-type C. elegans and investigated the effects of partial or complete loss of Ce-lamin protein on cell and nuclear organization.
We produced a polypeptide that included coil 2 in the rod domain plus the tail domain of Ce-lamin (amino acid residues D-217 to F-550), by expressing a 1,087-bp EcoRI-BglII cDNA fragment in Escherichia coli JM109, using pRSET (Invitrogen, Leek, The Netherlands). The resulting His-tagged fusion protein was isolated by affinity purification on a nickel-NTA column (Qiagen, Hilden, Germany) in buffer containing 8 M urea and further purified by preparative gel electrophoresis. Rabbits were immunized with the purified Ce-lamin fusion peptide. The Ce-lamin-antibody was affinity purified by incubating the serum with the purified Ce-lamin peptide bound to CN-bromide-activated Sepharose (Pharmacia). Bound antibodies were eluted using a glycine buffer (pH 2.5), and immediately adjusted to neutral pH with NaOH. MAb414, which recognizes a subset of nucleoporins, was purchased from Babco (Richmond, California). T-9026, which recognizes alpha tubulin, was purchased from Sigma (St. Louis, MO). Cy3-conjugated goat-antimouse and goat-antirabbit antibodies were purchased from Jackson laboratories (West Grove, PA).
Immunostaining was performed essentially as described (Miller and Shakes, 1995 ). Mixed stages or adults C. elegans were placed on polylysine-treated slides, and 45-mm coverslips were placed above the nematodes. The slides were placed either in liquid N2 or on dry ice, and the coverslips were immediately removed. The nematodes were fixed for 4 min at −20°C in methanol and then incubated for 30 min at room temperature in PBST (phosphate buffer saline containing 0.1% Tween 20) with 3.7% formaldehyde. Animals were then washed once in PBST, incubated for 10 min at room temperature in PBST containing either 10% low-fat milk or 5% nonfat dry milk, washed once with PBST, and incubated overnight at 4°C with the primary antibody diluted in PBST (1:400 for Ce-lamin and 1:1000 for MAb414). Excess antibody was removed by washes in PBST: once for 1 min, once for 10 min, and twice for 30 min each. The nematodes were then incubated for 2 h at 22°C with the Cy3-conjugated goat-antirabbit (for anti-Ce-lamin) or Cy3-conjugated goat-antimouse (for MAb414), diluted in PBST. Double-label immunostaining for mAb414 and Ce-lamin was performed as follows: Animals were first stained with antibodies to Ce-lamin, followed by FITC-conjugated secondary antibody, and then washed in PBST (once for 1 min, once for 10 min, and twice for 30 min each). The animals were then incubated for 2 h at 22°C with mAb414, rewashed as above, and incubated for 2 h with Cy3-conjugated secondary antibodies. For both double- and single-label immunostaining, excess secondary antibody was then removed by washes in PBST: once for 1 min, once for 10 min, and twice for 30 min each. C. elegans were then incubated for 10 min in PBS containing 1 μg/ml 4′,6′-diamidino-2-phenylindole (DAPI), washed once with PBS, and mounted in glycerol containing 2% n-propyl gallate. C. elegans were viewed either with an Olympus IX70 microscope equipped with epifluorescence or a Bio-Rad MRC-1024 confocal scanhead coupled to a Zeiss Axiovert 135 M inverted microscope with a 40x/NA=1.3 oil immersion objective. Excitation light was provided by a 100-mW air-cooled argon ion laser run in the multi-line mode. Both 488-nm and 514-nm excitation were used, as described below. The emission filter in the Cy3 detection channel was a D580/32 interference filter (32-nm bandpass centered on 580 nm). In the GFP channel, a D522/35 interference filter (522-nm center wavelength, 35-nm bandwidth) was used with 488-nm excitation, and a D540/30 interference filter (540-nm center wavelength, 30-nm bandwidth) was used with 514-nm excitation. The confocal iris diameter was 2.5 to 3 mm, with the larger opening used for weaker signals. Vertical resolution was ~1 μm. If necessary, 2 to 4 images were averaged in order to reduce noise. Images of 512×512 pixels were acquired, using a hardware zoom of 1.0 (0.308 μm/pixel) or 1.8 (0.175 μm/pixel).
Fluorescence levels were quantified using a DAGE 300ET-RC CCD camera, with a series of neutral density filters used to confirm linearity of measurements in the range of the assay. Six embryos each were measured for wild-type and lmn-1(RNAi).
Three different lmn-1-GFP constructs were prepared (Figure (Figure1A).1A). The first construct, termed pDRNL1, contained a GFP-Ce-lamin fusion with GFP fused to the N-terminus of lmn-1. To prepare the first construct, a 6.5-kb HindII–SalI genomic fragment positioned between −7.95 kb and −1.45 kb upstream to the lmn-1 coding region, was cloned into the pPD95.77 vector. Precise structures of GFP vectors ppD95.77 and pPD102.33 are described on the Fire lab web site, www.ciwenb.edu. A 1.45-kb fragment, just 5′ to the lmn-1 coding region, was amplified by PCR with a novel NcoI cloning site introduced at its 3′ end, and inserted 3′ to the 6.5-kb genomic fragment. The 0.75-kb KpnI–SmaI fragment of the GFP gene was PCR amplified from pEGFP (Clontech) and inserted into the NcoI and SmaI sites. A 2.2-kb KpnI–SmaI fragment of the lmn-1 genomic region, containing the complete lamin coding region, was PCR amplified and inserted 3′ in frame to the GFP gene.
The second construct is a fusion of GFP to the C-terminus of Ce-lamin. To prepare this construct, termed pJKL380.4, long-range PCR (Expand long template PCR system, Boehringer Mannheim) was used to amplify from wild-type genomic DNA a 9.6-kb lamin genomic fragment, containing 4.7 kb of 5′ sequence, the entire coding sequence plus introns, and 2.8 kb of 3′ sequence. This 9.6-kb fragment was engineered such that unique NotI and SmaI sites were present at its ends and used for cloning into the pBluescript II SK+ vector. The resulting plasmid, in which the BamHI site in the vector backbone was destroyed, contained a single BamHI site 12 amino acids before the stop codon of lamin. This site was used to insert GFP (from pPD102.33) in frame into the lmn-1 gene.
A third construct, termed pDRCL2, contained 7.95 kb of 5′ promoter sequence of lmn-1, the first three exons, and two introns of lmn-1 fused in frame to the GFP gene.
To make transgenic lines, all three constructs were first linearized. pJKL380.4 construct was coinjected with linearized pMH86 marker plasmid into the temperature-sensitive dpy-20 (e1282) animals using the complex array injection method (Kelly et al., 1997 ). pDRNL1 and pDRCL2 constructs were each coinjected with the pRF4 plasmid containing the dominant rol(6) marker (Mello et al., 1991 ) into wild-type (N2) animals with no carrier DNA. Multiple transgenic lines were obtained for each construct, with GFP localized as expected in the nuclear envelopes in lines expressing pJKL380.4 and pDRNL1, and in the cytoplasm in lines expressing pDRCL2. One of the pDRNL1-derived transgenic lines, A12, was kept as an unintegrated line and used for further characterization.
For most of the transgenic lines, the earliest stage at which GFP expression could be detected was in embryos containing around 60 cells. Several of the pJKL380.4-derived transgenic lines initially showed GFP expression in the germ line also, including germ cells in the syncytial gonad, oocytes, and early embryos. The pJKL380.4 transgene from one of these lines was then integrated using the standard gamma-irradiation method (Mello and Fire, 1995 ). Three independent integrated lines were generated, and all were out crossed three times and mapped to the X chromosome. One such line, PD4810 (ccIs4810), was used for further studies. The germ line expression of lmn-1-GFP was lost after several generations even in the integrated lines. To circumvent this problem, the lmn-1-GFP transgene from PD4810 was crossed into the mut-7 (pk204) mutator background. mut-7 has been shown to desilence transgenes in the germ line to some extent (Tabara et al., 1999 ).
To assess the function of lamin in C. elegans development, RNAi was employed to deplete both the maternal and zygotic product of the lamin gene. Three different constructs were used for making double-stranded RNA in vitro (Figure (Figure1A).1A). yk63f8, a gift of Dr. Yuji Kohara (Japan), contained a 2.13-kb, nearly full-length, lmn-1 cDNA; pJKL253.1 contained a 659-bp PCR fragment from exon 5 (the largest exon) of the lmn-1 gene; pPRNA-1 contained a 995-bp BglII-EcoRI fragment of the lmn-1 gene (amino acids 217–550). Both pPRNA-1 and pJKL253.1 contained a pBluescript II SK+ vector backbone. Double-stranded RNA was made in vitro following the protocol described by Fire et al., (1998) . The dsRNA was injected at concentrations between 0.1 to 1.0 μg/μl into wild-type (N2) hermaphrodites. The injected animals were transferred to new plates every 6 to 12 h, and their progeny were scored for abnormalities during development.
One-cell embryos from wild-type or lmn-1 dsRNA injected hermaphrodites were mounted on agarose pads, and 4-D time-lapse recording was performed as described (Fire, 1994 ), with a complete series of 28 images every 1.2 microns in the Z axis taken at 30-s intervals. Complete image series were analyzed from three lmn-1(RNAi) embryos and several wild-type embryos. All nuclei from five additional embryos were analyzed by direct observation in real time with a similar phenotype. Confirming data was also obtained from a number of additional embryos from which shorter series of images were collected.
The C. elegans genome contains an open reading frame encoding a putative 64-kDa protein with homology to lamins. This protein has been termed Ce-lamin (Riemer et al., 1993 ). Analysis of the near-completed genome sequence data now available for C. elegans suggests that this is the only lamin homologue in this organism. Portions of the rod and tail domains of Ce-lamin (amino acids 217 to 550) were expressed in bacteria, purified to homogeneity by gel electrophoresis, and used to raise polyclonal antibodies in rabbits. Western blot analysis, using affinity-purified Ce-lamin antibodies, showed a single band of 64 kDa in SDS-PAGE of complete C. elegans extract from a mixed stage nematode population (Figure (Figure11B).
Immunostaining of C. elegans at different developmental stages with the Ce-lamin antibodies showed that Ce-lamin is predominantly localized to the nuclear periphery (Figure (Figure2).2). Additional confocal microscopy analysis revealed subregions of the nuclear envelope with a higher intensity staining (our unpublished results). Such subregions of intense staining are characteristic of lamin staining in other metazoans (Paddy et al., 1990 ) (Figure (Figure2).2). The localization of Ce-lamin to the nuclear periphery, its pattern in the nuclear periphery, and its sequence homology to other known type-B lamins, all suggest that lmn-1 is the unique type-B lamin ortholog in C. elegans.
Antibodies to Ce-lamin stained the nuclear envelope at all developmental stages and in essentially every region where the Ce-lamin antibodies penetrated, including embryos, all larval stages, and adults (Figure (Figure2a-d,g).2a-d,g). Nuclear envelope staining of Ce-lamin was also detected in all cells in the gonad (Figure (Figure2g),2g), with one exception: antibody staining was not detected in cells undergoing spermiogenesis that had condensed chromatin (Figure (Figure22e-f).
Transgenic lines expressing the GFP fused to Ce-lamin (Ce-lamin-GFP) were prepared in order to study the spatial and developmental distribution of Ce-lamin in living cells. With the exception of mature sperm, Ce-lamin-GFP fusion protein was localized to the nuclear periphery of all cells throughout the development of the nematode (Figure (Figure2h-i).2h-i). The Ce-lamin-GFP expression in these lines reproduced the antibody staining pattern of the native protein, including the presence of subregions in the nuclear envelope with higher fluorescence intensity (Figure (Figure3,3, bottom panel).
Intranuclear presence of Ce-lamin was also observed in most embryonic nuclei and in a fraction of nuclei in adult tissues, including the distal tip cell in the gonad (Figure (Figure2g).2g). In order to verify that the internal lamin staining was specific, embryonic and adult tissues were stained with both Ce-lamin antibodies and monoclonal antibody mAb414 (Davis and Blobel, 1987 ), which recognizes the nuclear pores in many eukaryotes, including C. elegans (Pitt et al., 2000 ). As expected, both MAb414 and Ce-lamin brightly stained the nuclear envelope. However, only Ce-lamin showed intranuclear location (top panels in Figure Figure3),3), while MAb414 gave a higher cytoplasmic background, as described (Pitt et al., 2000 ). The internal staining disappeared as well as the nuclear envelope staining in lmn-1(RNAi) embryos (Figure (Figure33).
To further verify the presence of Ce-lamin in the nuclear interior, the spatial distribution of the GFP-Ce-lamin fusion protein was analyzed by confocal microscopy in vivo in line A12, which expresses the lowest levels of Ce-lamin-GFP fusion (Figure (Figure1B).1B). Consistent with the antibody staining, a subset of cells showed GFP-Ce-lamin protein in the nuclear interior, in addition to the nuclear envelope (bottom panels in Figure Figure3).3). The intranuclear staining was not uniform and was always weaker than that seen on the nuclear envelope.
To understand the function of Ce-lamin in vivo, RNA interference experiments (RNAi) aimed at inhibiting lmn-1 expression, were performed by injecting lmn-1 dsRNA into the gonads of adult hermaphrodites. Equivalent results were obtained with three different dsRNA constructs (Figure (Figure1A),1A), injected at several different concentrations (0.1 to 1.0 μg/μl).
RNAi effects are most potent in a time window of 18 to 36 h after injection. All embryos laid during this time window arrested during development, indicating lamin is required for embryonic viability. We use the nomenclature the lmn-1 (RNAi) to describe the F1 progeny of hermaphrodites injected with dsRNA of lmn-1.
To monitor the extent of residual Ce-lamin expression in the lmn-1 (RNAi) embryos, embryos were stained at different stages with lamin antibodies. There was some variability between lmn-1 (RNAi) embryos and between regions in the same embryo (our unpublished results). However, quantitative fluorescence analysis carried out as described in Materials and Methods revealed a decrease of >60-fold in intensity of Ce-lamin staining in the lmn-1 (RNAi) embryos (Figure (Figure4a,b). 4a,b).
The intensity of the residual Ce-lamin staining in lmn-1(RNAi) embryos decreased as the embryos continued to develop; Ce-lamin was not detectable upon developmental arrest. A dramatic reduction in Ce-lamin staining in lmn-1(RNAi) embryo was also detected in the nuclear interior (Figure (Figure3,3, middle panels), proving further evidence for the presence of Ce-lamin inside the nucleus.
Although some lmn-1(RNAi) embryos arrested with <100 nuclei, DIC microscopy revealed that most lmn-1(RNAi) embryos were able to form several hundred nuclei before arrest (Figure (Figure4c,d).4c,d). All lmn-1(RNAi) embryos were abnormal and contained nuclei that varied in size (Figure (Figure4d).4d). Further incubation of the lmn-1 (RNAi) embryos caused them to degrade and to form regions in the embryos that were devoid of nuclei (our unpublished results). The reduced lmn-1 activity and embryonic arrest in the lmn-1 (RNAi) embryos indicated that lamins are an essential component of the nuclear envelope.
The phenotype of lmn-1(RNAi) embryos was examined by 4-D time-lapse microscopy on live embryos (Fire, 1994 ). The earliest observed defect was as early as the two-cell stage. While nuclei in the wild-type embryo had well defined and stable boundaries, the nuclei in lmn-1(RNAi) embryos appeared plastic, with rapid changes over time (see Figure Figure55 for an example of the AB nucleus in a two-cell embryo). Despite the gross defects in nuclear morphology, nuclear divisions still occurred in lmn-1(RNAi) embryos, with these embryos eventually producing several hundred cells.
The nuclear morphology of the lmn-1 (RNAi) embryos was further examined by staining with the DNA-specific dye DAPI. Whereas control embryos of mock-injected hermaphrodites or wild-type embryos had no apparent nuclear abnormalities, a fraction of nuclei in >95% of the lmn-1 (RNAi) embryos appeared abnormal, with defects as early as the first few nuclear divisions (Figure (Figure4e-i).4e-i). These phenotypes included inability to complete the cell cycle and distribution of unequal amounts of chromatin into the daughter nuclei. The most common phenotypes in the progression of the cell cycle were bridges of chromatin between nuclei (arrow in Figure Figure4e)4e) and chromatin that expanded abnormally (Figure (Figure4f).4f). Problems in the distribution of chromatin between daughter nuclei included large differences in the amounts of chromatin in the daughter nuclei (arrows in Figure Figure4g)4g) and chromosomes that were outside the daughter nuclei (arrows in Figure Figure4h,i).4h,i). These abnormalities were not due to gross defects in microtubule organization, since the pattern of staining of the lmn-1 (RNAi) embryos with tubulin antibodies was similar to that of wild-type embryos (our unpublished results).
Arrested embryos showed regions where the DNA was absent (our unpublished results), or regions with a large increase in the amount of DNA (black arrows in Figure Figure4d4d show ‘giant’ nuclei that are 8 times the size of wild-type gut nuclei in a comma stage embryo). In summary, the RNAi experiments revealed a broad requirement role for lmn-1 in cell cycle progression, and chromatin organization.
Reduction in levels of lamin Dm0 activity in Drosophila and elimination of lamin A gene in mouse have both shown to yield abnormalities in the pattern of NPCs (Lenz-Bohme et al., 1997 ; Harel et al., 1998 ; Sullivan et al., 1999 ). We therefore compared the spatial organization of the NPCs in lmn-1 (RNAi) embryos and normal C. elegans embryos by staining them with monoclonal antibody MAb414, which recognizes the FG repeats present in many nucleoporins (Davis and Blobel, 1987 ). While wild-type or mock-injected embryos showed a typical patchy nuclear rim staining, characteristic of nuclear pores in most species examined (Figure (Figure6a),6a), the NPCs in many (but not all) nuclei of the lmn-1 (RNAi) embryos showed an abnormal clustering pattern to one side of the nucleus (arrows in Figure Figure6b6b and both nuclei in Figure Figure6c;6c; a “normal” distribution of NPCs is shown by an arrowhead in Figure Figure6b).6b). Staining of the lmn-1(RNAi) embryos with both Ce-lamin antibodies and monoclonal antibody MAb414 revealed that NPCs clustering does not correlate with residual Ce-lamin; indeed, most NPC clusters were observed in nuclei where residual Ce-lamin could not be observed (Figure (Figure6d,e).6d,e). These results confirm the importance of lmn-1 in C. elegans for proper spacing of nuclear pore complexes.
We were able to examine germline development after conditions of Ce-lamin insufficiency by studying a few rare F1 animals that escaped the lethal effects of lmn-1 (RNAi). These animals resulted from eggs laid outside the most potent window of RNAi activity (Figure (Figure7).7). Significantly, most of these F1 adults were sterile or semisterile. Examination of the gonads of the sterile animals by DIC microscopy and DAPI staining revealed dramatic reductions in the number of germ cells (Figure (Figure7a-b).7a-b). In a fraction of such animals, small numbers of abnormal oocyte-like cells were observed (some with multiple nuclei and large vacuoles; Figure Figure7c,g);7c,g); nuclei with condensed chromatin (potentially spermatocytes or spermatocyte-like cells) were also observed (Figure (Figure7h).7h). Some of the F2 embryos from semisterile animals developed into fertile adults. Among these progeny, there was a high incidence of males (31 males out of 253 progeny or 12.25%, compared with 0.1% in wild-type animals; Hodgkin et al., 1979 ). These males were fertile. Their genotype was XO, since mating with dpy-11 unc-60 hermaphrodites produced wild-type progeny with 1:1 ratio of males and hermaphrodites.
In both vertebrates and Drosophila, the type-A lamins have a cell- and tissue-specific pattern of expression (reviewed in Stuurman et al., 1998 ), while type-B lamins are present in essentially all cells. The single C. elegans lamin protein, Ce-lamin, is similar to type-B lamins in both Drosophila and vertebrates (Goldberg et al., 1999b ). With the exception of mature sperm cells, Ce-lamin is detected in the nuclear envelope of essentially all cells during all developmental stages of C. elegans. C. elegans is a multicellular animal, which undergoes complicated differentiation processes to give rise to differentiated tissues and a nervous system. We therefore propose that Ce-lamin perform activities that in other animals are performed by both type-B and type-A lamins, and these activities maybe regulated by posttranslational modifications.
The only cells that do not contain detectable Ce-lamin are sperm cells with condensed chromatin. However, we cannot rule out the possibility that technical issues prevent expression of lamin-GFP and lamin antibody staining in sperm cells. Drosophila sperm cells have also been reported to lack a type-B lamin (lamin Dm0) and its associated protein, otefin (Ashery Padan et al., 1997 ). In contrast, vertebrate sperm cells contain both type-A and type-B lamins, which are probably involved in nuclear organization during meiosis (Furukawa and Hotta, 1993 ; Furukawa et al., 1994 ; Alsheimer et al., 1999 ).
In addition to the nuclear periphery, Ce-lamin was also detected in the nuclear interior. Internal Ce-lamin staining was not an immunofluorescence artifact, since (a) it was observed both in fixed animals stained with Ce-lamin antibodies and in GFP patterns from live animals that express low amounts of a GFP-Ce-lamin fusion protein, and (b) it was eliminated in lmn-1(RNAi) embryos. Most embryonic cells contained intranuclear lamin, while adult cells with internal nuclear lamin staining were found next to cells in which Ce-lamin could be detected only at the nuclear periphery. Internal type-A lamins have been reported in HeLa cells (Hozak et al., 1995 ) and in mammalian cells expressing lamin A-GFP fusion proteins (Broers et al., 1999 ). Type-A and type-B lamins are also seen as intranuclear foci in mammalian cultured cells (Bridger et al., 1993 ; Moir et al., 1994 ). These foci are thought to be intermediate stages of lamin processing, before their incorporation into the peripheral lamina (Goldman et al., 1992 ). Lamins also colocalize at sites of DNA replication, where they are required for the elongation stage of DNA replication (Spann et al., 1997 ). In addition, during apoptosis, lamins are distributed relatively uniformly in the interior of the nucleus (Rao et al., 1996b ). Thus, the roles of interior lamins will be interesting topic for a future study.
The RNAi experiments revealed that lmn-1 is an essential gene. The lethality caused by the lmn-1 dsRNA was specific since (A) embryos laid by mock-injected adults developed normally; (B) similar phenotypes were seen with three different RNAi constructs and different concentrations of lmn-1 dsRNA ranging between 0.1 to 1.0 μg/μl; and (C) the observed phenotypes were associated with a dramatic reduction in levels of Ce-lamin. The Ce-lamin staining became undetectable as the lmn-1 (RNAi) embryos continued to develop. In contrast to Ce-lamin levels, the level of nucleoporins was apparently unchanged after lmn-1 RNAi. It was not surprising that loss of Ce-lamin is lethal, since mutations in the Drosophila lamin Dm0 gene (Lenz-Bohme et al., 1997 ; Harel et al., 1998 ) and in mouse lamin A gene (Sullivan et al., 1999 ) also cause nuclear defects and lethality. The results of this study extend the results in Drosophila and mice in several ways. Unlike Drosophila or mice, the genome of the C. elegans contains only one lamin gene; the maternal contribution of this gene can be largely eliminated in a gene-specific manner by RNA-mediated interference (Fire et al., 1998 ). This allows the only example to date in which one can exam the effects of reducing all lamin on cell growth and development. In addition, due to large maternally contributed pools of Drosophila lamin Dm0 protein and RNA, the phenotypes did not appear until relatively late in embryogenesis, when lamin C is also being expressed. Likewise, the delayed phenotypes of mice lacking lamin A (4 to 8 wk after birth) were probably due to complementation by the lamin B1 and B2 genes.
We attribute the weak residual Ce-lamin signal in the lmn-1 (RNAi) embryos to the maternal contribution of stable proteins that cannot be eliminated completely by dsRNA. There is a large maternal pool of Ce-lamin in wild-type oocytes (our unpublished results), and lamins are very stable proteins (Harel et al., 1998 ). We consider it likely that the residual Ce-lamin probably allows nuclei to continue dividing in lmn-1 (RNAi) embryos until there is not enough lamin activity to support normal, or even abnormal, nuclear divisions. Consistent with this, residual Ce-lamin could not be detected in lmn-1 (RNAi) embryos with >100 to 200 nuclei (our unpublished results).
The nuclear lamina anchors the NPCs (reviewed in Stoffler et al., 1999 ). We found that loss of Ce-lamin caused NPCs to cluster. These results suggest that the association of the nuclear lamina with the NPCs is required for their correct spacing. The role of the nuclear lamina in spacing NPCs is probably conserved in evolution since similar phenotypes are seen in lines with reduced Drosophila lamin Dm0 and in mice lacking the lamin A gene (Lenz-Bohme et al., 1997 ; Harel et al., 1998 ; Sullivan et al., 1999 ).
A remarkable phenotype in the lmn-1 (RNAi) embryos is the rapid change in nuclear morphology in vivo. This change clearly demonstrates the role of nuclear lamina in determining the shape of the nucleus. However, the abnormality in nuclear shape did not interfere with the ability of most nuclei to undergo mitosis. In addition, the timing of nuclear divisions appeared similar in the lmn-1 (RNAi) and wild-type embryos (our unpublished results). We consider it likely that the instability of nuclear morphology in the lmn-1 (RNAi) embryos is a primary defect eventually leading to defects in cell cycle progression and chromatin organization (see below).
It was surprising to find that lamin activity may be required for the proper segregation of chromosomes. In the lmn-1 (RNAi) embryos, there were many examples of unequal segregation of chromatin to daughter nuclei, and of chromosomes that must have been lost from one of the daughter nuclei during the cell cycle. The difference in the amount of chromatin in daughter nuclei may be also the consequence of incomplete DNA replication, since lamins have been proposed to play an essential role in the elongation step of DNA replication (Moir et al., 2000 ). This difference can be also due to either nuclear fragmentation or fusion of chromatin during interphase. However, the high incidence of male offspring from the F1 animals that survived the initial RNAi indicates that Ce-lamin activity is required, either directly or indirectly, for proper segregation of chromosomes. Although Ce-lamin is present in a spindle envelope structure until early anaphase (Lee et al., 2000 ), its role in chromosome segregation is probably not through a change in microtubule distribution, since the spatial organization of microtubules in lmn-1 (RNAi) embryos is similar to wild-type.
A high fraction of the lmn-1 (RNAi) embryos contained some nuclei that were unable to complete the cell cycle. A requirement for lamins and lamin-associated proteins to complete the cell cycle in mammalian cells was suggested from a previous study, in which lamin antibodies injected into mammalian cells caused chromatin arrest in a telophase-like configuration (Benavente and Krohne, 1986 ). The roles of nuclear lamins in nuclear assembly are still a matter of debate (reviewed in Gant and Wilson, 1997 ). Lamins were suggested to play a role in assembling the nuclear envelope in Drosophila (Ulitzur et al., 1997 ), Xenopus (Dabauvalle et al., 1991 ) and mammalian cells (Burke and Gerace, 1986 ). The results presented here demonstrate that lamins, either directly or by affecting the distribution of their associated proteins, are required for proper cell cycle progression, which may reflect a defect in nuclear assembly in vivo.
Many nuclei in the arrested lmn-1 (RNAi) embryo and in gonads of the lmn-1 (RNAi) F1 adults had abnormally condensed chromatin. We propose that chromatin might condense in these cells due to insufficient attachment to the nuclear envelope. Given that many nuclear membrane proteins bind directly to a chromatin partner, our results suggest that (A) attachments between chromatin and nonlamin components of the nuclear membrane are insufficient to keep the chromatin decondensed; (B) the membrane proteins may not be fully functional in the absence of lamins; or (C) loss of the lamina may trigger the activity of apoptotic factors that cause chromatin condensation, which normally only occurs after caspase-induced lamin degradation. Indeed, in C. elegans the nuclear envelope is a primary docking site for CED-4 (Chen et al., 2000 ), and in mammalian cells lamin degradation is required for the dissociation of chromatin from the nuclear membranes during apoptosis (Rao et al., 1996a ). Further experiments will be aimed at examining apoptotic patterns in C. elegans lacking normal Ce-lamin levels.
This study was funded in part by grants from the USA-Israel Binational Science Foundation (BSF to Y.G.), the Israel Science Foundation (BRF to Y.G.), and the German-Israel Foundation (GIF #1–573-036.13 to Y.G. and K.W). Support to J.L. and A.F. was provided by the NIH grants F32HD08331 and GM37706.
We thank Margalit Eshel and Jamie Fleenor for technical help, Aryeh M Weiss and Naomi Melamed-Book for helping with the confocal microscope, Bill Kelly for C. elegans strains, and Mary Osborn and Kenneth Lee for critical reading of the manuscript. We are also in great debt to Kathy Wilson for her wonderful comments and inspiring discussions.