Identification of vimentin and novel vimentin-related proteins in Xenopus oocytes and early embryos

We have made antibodies against fusion proteins of Xenopus vimentin. We show for the first time the distribution of vimentin in larval stages, where it is found in cells of mesenchymal origin, and in radial glial cells. In sections of Xenopus oocytes and early embryos, immunocytochemistry reveals the presence of an extensive cytoplasmic network, distributed in an animal-vegetal gradient. Germ plasm stains particularly strongly. The form of the IF proteins in this network is unusual. In immunoblot experiments the anti-vimentin antibodies detect a number of distinct proteins. We have identified those that are the products of the two known vimentin genes, by injection of synthetic mRNA transcribed from cloned vimentin cDNAs into oocytes, followed by two-dimensional Western blotting. This has demonstrated unambiguously that one Xenopus vimentin, Vim1, is present in oocytes and early embryos. However, two other immunoreactive proteins detected in Triton extracts of oocytes and early embryos are not the products of Vim1, since depletion of vimentin mRNA by antisense oligonucleotide injection has no effect on the synthesis of these proteins. These results suggest that novel IF-like proteins are expressed in Xenopus oocytes and early embryos.     

Molecular cloning and sequence determination of four different cDNA species coding for alpha-subunits of G proteins from Xenopus laevis oocytes

A cDNA library prepared from Xenopus laevis oocytes in lambda gt10 was screened with a mixture of three oligonucleotide probes designed to detect sequences found in different mammalian genes coding for alpha-subunits of G-proteins. In addition to a clone coding for a G alpha o-type subunit previously reported [(1989) FEBS Lett. 244, 188-192] four additional clones have been found coding for different G alpha protein subunits. By comparison with mammalian alpha-subunits, these oocyte cDNAs correspond to two closely related G alpha s-1a, to a G alpha i-1 and to a G alpha i-3 species. The derived amino acid sequences showed that both G alpha s species contain 379 residues, corresponding to the short species without the serine residue and with a calculated Mr of 42720. The G alpha i-1 gene encodes a 354 amino acid protein with an Mr of 39,000 and the G alpha i-3 encodes an incomplete open reading frame of 345 residues, lacking the first 9 amino acid residues at the NH2 terminus. All these G alpha-subunits showed high identity with their respective mammalian counterparts (75-80%), indicating a great degree of conservation through the evolution and the important cellular regulatory function that they play.     

Characterization of metallothionein cDNAs induced by cadmium in the nematode Caenorhabditis elegans.

cDNAs of metallothioneins (MTs) in the nematode Caenorhabditis elegans were characterized. The MT-II clone encodes 62 amino acid residues and the predicted Mr is 6462. The MT-I clone contains an additional 12 residues at the C-terminal end, and the predicted Mr is 7959. There is a considerable similarity between MT-I and MT-II. Both of these proteins are cysteine-rich and, with a few exceptions, show a good alignment of cysteine residues. No obvious sequence relationship in the coding region was discernible between C. elegans MTs and mammalian MTs, aside from Cys-Cys, Cys-Xaa-Cys, and Cys-Xaa-Xaa-Xaa-Cys segments. However, 3'-untranslated region of cDNAs of C. elegans MT-I and -II have some consensus sequences found in mammalian MT cDNAs, suggesting that these regions may have some roles in the regulation of MT-gene expression.     

Role of the aromatic residues in the near-amino terminal motif of vimentin in intermediate filament assembly in vitro

Type III and IV intermediate filament (IF) proteins share a conserved sequence motif of -Tyr-Arg-Arg-X-Phe- at the near-amino termini. To characterize significance of the aromatic residues in the motif, we prepared vimentin mutants in which Tyr-10 and Phe-14 are substituted with Asn and Ser (Vim[Y10N], Vim[F14S] and Vim[Y10N, F14S]), and examined assembly properties in vitro by electron microscopy and viscosity measurements. At 2 s after initiation of assembly reaction at pH 7.2 and 150 mM NaCl, all the vimentin mutants formed so-called unit-length filaments (ULFs) that were slightly larger than ULFs of wild-type vimentin. In following filament elongation, Vim[Y10N, F14S] and Vim[Y10N] performed longitudinal annealing of ULFs very rapidly and formed IFs within only 2.5 and 5 min, respectively, while Vim[F14S] and wild-type vimentin gave IFs by 40-60 min. The IFs of Vim[Y10N, F14S] and Vim[Y10N], however, tended to intertwine each other and formed bundles in parts of the specimens. The intertwinements decreased as the salt concentration decreased, and optimal salt concentration for the two mutants to form normal IFs was 50 mM. These results suggest that the aromatic residues, especially Tyr-10, in the motif have a role in controlling intermolecular interactions involved in IF assembly in vitro and suppress undesirable filament intertwinements at physiological ionic strength.     

A whole-mount immunocytochemical analysis of the expression of the intermediate filament protein vimentin in Xenopus

We have developed a whole-mount immunocytochemical method for Xenopus and used it to map the expression of the intermediate filament protein vimentin during early embryogenesis. We used two monoclonal antibodies, 14h7 and RV202. Both label vimentin filaments in Xenopus A6 cells, RV202 reacts specifically with vimentin (Mr, 55 x 10(3] on Western blots of A6 cells and embryos. 14h7 reacts with vimentin and a second, insoluble polypeptide of 57 x 10(3) Mr found in A6 cells. The 57 x 10(3) Mr polypeptide appears to be an intermediate filament protein immunochemically related to vimentin. In the whole-mount embryo, we first found vimentin at the time of neural tube closure (stage 19) in cells located at the lateral margins of the neural tube. By stage 26, these cells, which are presumably radial glia, are present along the entire length of the neural tube and in the tail bud. Cells in the optic vesicles express vimentin by stage 24. Vimentin-expressing mesenchymal cells appear on the surface of the somites at stage 22/23; these cells appear first on anterior somites and on progressively more posterior somites as development continues. Beginning at stage 24, vimentin appears in mesenchymal cells located ventral to the somites and associated with the pronephric ducts; these ventral cells first appear below the anterior somites and later appear below more posterior somites. The dorsal fin mesenchyme expresses vimentin at stage 26. In the head, both mesodermally-derived and neural-crest-derived mesenchymal tissues express vimentin by stage 26. These include the mesenchyme of the branchial arches, the mandibular arch, the corneal epithelium, the eye, the meninges and mesenchyme surrounding the otic vesicle. By stage 33, vimentin-expressing mesenchymal cells are present in the pericardial cavity and line the vitelline veins. Vimentin expression appears to be a marker for the differentiation of a subset of central nervous system cells and of head and body mesenchyme in the early Xenopus embryo.     

Basic amino acid residue cluster within nuclear targeting sequence motif is essential for cytoplasmic plectin-vimentin network junctions

We have generated a series of plectin deletion and mutagenized cDNA constructs to dissect the functional sequences that mediate plectin's interaction with intermediate filament (IF) networks, and scored their ability to coalign or disrupt intermediate filaments when ectopically expressed in rat kangaroo PtK2 cells. We show that a stretch of approximately 50 amino acid residues within plectin's carboxy-terminal repeat 5 domain serves as a unique binding site for both vimentin and cytokeratin IF networks of PtK2 cells. Part of the IF-binding domain was found to constitute a functional nuclear localization signal (NLS) motif, as demonstrated by nuclear import of cytoplasmic proteins linked to this sequence. Site directed mutagenesis revealed a specific cluster of four basic amino acid residues (arg4277-arg4280) residing within the NLS sequence motif to be essential for IF binding. When mutant proteins corresponding to those expressed in PtK2 cells were expressed in bacteria and purified proteins subjected to a sensitive quantitative overlay binding assay using Eu3+-labeled vimentin, the relative binding capacities of mutant proteins measured were fully consistent with the mutant's phenotypes observed in living cells. Using recombinant proteins we also show by negative staining and rotary shadowing electron microscopy that in vitro assembled vimentin intermediate filaments become packed into dense aggregates upon incubation with plectin repeat 5 domain, in contrast to repeat 4 domain or a mutated repeat 5 domain.     

Overexpression of wild-type and dominant negative mutant vimentin subunits in developing Xenopus embryos.

Vimentin belongs to the diverse multigene family of intermediate filament proteins, each member of which is expressed in a tissue-specific and developmentally regulated pattern. The existence of vimentin filaments has been documented in oocytes, eggs, and early embryos of Xenopus laevis, but the role of these cytoskeletal components remains unknown. To investigate the functions of vimentin during early development in Xenopus, we induced the overexpression of wild-type and deletion mutant subunits in most of the cells of embryos by injecting synthetic RNA into fertilized eggs. Wild-type vimentin subunits, as well as subunits lacking most of the amino-terminal head piece, assembled into normal appearing filaments in vivo. Deletion mutants of the fourth alpha-helical rod domain were assembly incompetent and dominantly inhibited the polymerization of wild-type subunits when both types of subunit were co-expressed in cells. Expression of at least a tenfold excess of wild-type or mutant subunits within cells of embryos did not lead to any detectable morphological or developmental abnormalities, suggesting that the presence and proper regulation of vimentin expression is not essential during the initial stages of embryogenesis in Xenopus.     

Developmental expression of the protein product of Vg1, a localized maternal mRNA in the frog Xenopus laevis.

Vg1 is a maternal mRNA localized in the vegetal cortex of Xenopus laevis oocytes, that encodes a protein homologous to the mammalian growth factor TGF-beta. Using a polyclonal antibody to a T7-Vg1 fusion protein, we have identified the native protein. We find that a single protein of Mr 40 kd is immunoprecipitated following in vitro translation of oocyte poly(A)+ RNA, whilst two proteins of Mr 45 and 43.5 kd are immunoprecipitated from oocyte and embryo extracts. Synthesis of at least the 40 kd, in vitro, and 45 kd, in vivo, proteins is specifically inhibited following treatment of the respective systems with antisense Vg1 (but not histone H4) oligodeoxynucleotides. Tunicamycin treatment reveals the in vivo proteins to be glycosylated versions of a 40 kd protein, modified by the addition of either two or three N-linked oligosaccharide side chains. Both proteins are sensitive to digestion by the enzyme endoglycosidase-H, and are segregated within a membrane fraction from which they can be released by high pH treatment. Their synthesis is first detectable in stage IV oocytes and continues throughout early embryogenesis until the late gastrula. During embryogenesis the relative proportions of the two proteins change, the 45 kd protein being predominant in early embryogenesis and the 43.5 kd protein in late embryogenesis. Synthesis only occurs in the vegetal hemisphere at all stages; however, in the large oocyte diffusion of both proteins into the animal hemisphere occurs.     

Purification, characterization, and cDNA cloning of a novel metallothionein-like, cadmium-binding protein from Caenorhabditis elegans

Caenorhabditis elegans adapted for survival in high concentrations of Cd(II) express a heavy metal binding protein designated C. elegans metallothionein-like protein or MT-Ce. This protein was purified to homogeneity and characterized. MT-Ce binds 6 mol of Cd(II)/mol protein. The sequence of 39 amino-terminal residues in MT-Ce was determined. A radiolabeled 41-mer oligonucleotide, designed from the partial MT-Ce sequence, was used in conjunction with sucrose gradient centrifugation to obtain size-fractionated poly(A+) RNA enriched in MT-Ce sequences. Subsequently, cloned cDNAs, corresponding to MT-Ce mRNA sequences, were isolated from a lambda ZapII cDNA library prepared from the enriched template mRNA. cDNA and protein sequence analysis revealed that MT-Ce comprises 62 amino acid residues and has a predicted Mr of 6462. Seventeen of the 18 Cys residues in the nematode cadmium-binding protein are included in Cys-X-Cys and X-Cys-Cys-X motifs that are characteristic of mammalian metallothioneins (MTs). However, the resemblance of MT-Ce to mammalian MTs is superficial. The amino acid sequence of MT-Ce is unique, and neither its putative alpha and beta domains nor its Cys residues can be readily aligned with the corresponding regions of other eukaryotic MTs. This suggests that MT-Ce is an example of convergent evolution. The MT-Ce mRNA level in nematodes that were selected and grown with Cd(II) concentrations that are lethal for wild-type worms, was 55-fold higher than the level of MT-Ce mRNA in wild-type C. elegans. Comparison of the sequences of MT-Ce cDNAs revealed the occurrence of two types of MT-Ce mRNA. Each contains an identical coding region, but the cDNAs diverge markedly in their 5'-untranslated regions. This suggests the possibilities of regulation by alternative splicing and/or the presence of multiple MT-Ce genes encoding a single protein, but controlled by different regulatory elements.     

Xenopus laevis contains two nonallelic preproinsulin genes. cDNA cloning and evolutionary perspective

We undertook the cloning of preproinsulin cDNAs from the South African clawed toad, Xenopus laevis, in order to study the role of insulin during embryogenesis in this species. We found that X. laevis contains two different preproinsulin cDNAs, both of which code for peptides containing 106 amino acids of typical structure but which differ by eight amino acids: one in the signal peptide, two in the B-chain, four in the C-peptide, and one in the A-chain. Southern blot analysis indicates that the two preproinsulin cDNAs identified correspond to two different nonallelic genes which we believe arose through a recent gene duplication within the amphibian radiation possibly during the development of tetraploidy in this species. Both genes are expressed, since we have recently identified the two corresponding insulins in pancreatic extracts of adult toads (Shuldiner, A.R., Bennett, C., Robinson, E.A., and Roth, J. (1989) Endocrinology, in press). These cDNAs represent the first amphibian preproinsulin sequences to be elucidated.     

Sorting out IF networks: consequences of domain swapping on IF recognition and assembly

Vimentin and keratin are coexpressed in many cells, but they segregate into two distinct intermediate filament (IF) networks. To understand the molecular basis for the sorting out of these IF subunits, we genetically engineered cDNAs encoding hybrid IF proteins composed of part vimentin and part type I keratin. When these cDNAs were transiently expressed in cells containing vimentin, keratin, or both IFs, the hybrid IF proteins all recognized one or the other or both networks. The ability to distinguish networks was dependent upon which segments of IF proteins were present in each construct. Constructs containing sequences encoding either helix 1B or helix 2B seemed to be the most critical in conferring IF recognition. At least for type I keratins, recognition was exerted at the level of dimer formation with wild-type type II keratin, as demonstrated by anion exchange chromatography. Interestingly, despite the fact that swapping of helical domains was not as deleterious to IF structure/function as deletion of helical domains, keratin/vimentin hybrids still caused structural aberrations in one or more of the cytoplasmic IF network. Thus, sequence diversity among IF proteins seems to influence not only coiled-coil but also higher ordered associations leading to 10-nm filament formation and/or IF interactions with other cellular organelles/proteins.     

Vimentin expression in oocytes, eggs and early embryos of Xenopus laevis

Immunocytochemical studies using a monoclonal anti-porcine vimentin antibody reveal a well-organized pattern of staining in Xenopus laevis oocytes, eggs and early embryos. The positions of Xenopus vimentin and desmin in two-dimensional (2D) polyacrylamide gels were first established by immunoblotting of muscle Triton extracts with anti-intermediate filament antibodies (anti-IFA), which cross-react with all intermediate filament proteins (IFPs). The anti-porcine vimentin reacts with vimentin and desmin in muscle 2D immunoblots, but only reacts with one polypeptide in oocyte blots in the position predicted for vimentin (Mr 55 x 10(3), pI 5.6). Using an anti-sense probe derived from a Xenopus vimentin genomic clone in RNase protection assays, we show that expression of vimentin begins in previtellogenic oocytes. The level of expression remains constant throughout oogenesis and in unfertilized eggs. These data suggest that vimentin is expressed in oocytes and eggs. Most interestingly, the immunocytochemical results also show that vimentin is present in the germ plasma of oocytes, eggs and early embryos. It is therefore possible that vimentin has an important role in the formation or behaviour of early germ line cells.     

Comparison of mammalian, chicken and Xenopus brain-derived neurotrophic factor coding sequences

We have used the polymerase chain reaction (PCR) to amplify and clone coding sequences of the mature region of brain-derived neurotrophic factor (BDNF) from monkey, rat, chicken and Xenopus genomic DNA. Consistent with previous reports, the predicted amino acid sequences obtained in this manner from monkey and rat were identical to other mammalian BDNF sequences. The chicken and Xenopus BDNF sequences are also highly conserved, but contain 7 and 8 amino acid substitutions, respectively, compared to mammalian BDNF. Comparison of these sequences with the homologous NGF and NT3 coding regions provides further insight into amino acid residues that may be responsible for the different receptor specificities of these factors.