Rearrangement of mRNAs for costamere proteins during costamere development in cultured skeletal muscle from chicken

Distribution of mRNA transcripts and translation activity in skeletal myofibers

We examine the distribution of gene products in skeletal myofibers, which are highly differentiated multinucleated cells exhibiting a specific cellular architecture. In situ hybridization studies of adult rat myofibers with a single nucleus infected with influenza virus suggested that the viral mRNA products were distributed beneath the sarcolemma around the nucleus of origin. In situ hybridization studies with a poly-T oligonucleotide probe to detect endogenous mRNAs indicated their concentration around the nuclei and distribution beneath the sarcolemma in a cross-striated fashion at the A-I junctions (costamers). Labeling with bromouridine resulted in a similar distribution pattern. The ribosomal distribution pattern indicated concentration around the myonuclei but an intracellular component was also seen. Localization of the translating ribosomes by puromycylation revealed prominent spots perinuclearly and in the core regions of the myofibers. These spots flanked Golgi elements. Our results thus suggest that the total mRNA pool is heavily concentrated within the perinuclear and subsarcolemmal regions. However, the ribosomes and the translational activity did not follow this distribution pattern, so the mRNA transcripts were not restricted to a region beneath the sarcolemma. Furthermore, experiments utilizing green fluorescent protein showed the rapid movement of proteins within the endomembrane system, which thus facilitated proteins to reach their site of function irrespective of the site of synthesis.

Facioscapulohumeral muscular dystrophy (FSHD) region gene 1 (FRG1) is a dynamic nuclear and sarcomeric protein

Facioscapulohumeral muscular dystrophy (FSHD) region gene 1 (FRG1) is a candidate gene for mediating FSHD pathophysiology, however, very little is known about the endogenous FRG1 protein. This study uses immunocytochemistry (ICC) and histology to provide insight into FRG1's role in vertebrate muscle development and address its potential involvement in FSHD pathophysiology. In cell culture, primary myoblast/myotube cultures, and mouse and human muscle sections, FRG1 showed distinct nuclear and cytoplasmic localizations and nuclear shuttling assays indicated the subcellular pools of FRG1 are linked. During myoblast differentiation, FRG1's subcellular distribution changed dramatically with FRG1 eventually associating with the matured Z-discs. This Z-disc localization was confirmed using isolated mouse myofibers and found to be maintained in adult human skeletal muscle biopsies. Thus, FRG1 is not likely involved in the initial assembly and alignment of the Z-disc but may be involved in sarcomere maintenance or signaling. Further analysis of human tissue showed FRG1 is strongly expressed in arteries, veins, and capillaries, the other prominently affected tissue in FSHD. Overall, we show that in mammalian cells, FRG1 is a dynamic nuclear and cytoplasmic protein, however in muscle, FRG1 is also a developmentally regulated sarcomeric protein suggesting FRG1 may perform a muscle-specific function. Thus, FRG1 is the only FSHD candidate protein linked to the muscle contractile machinery and may address why the musculature and vasculature are specifically susceptible in FSHD.

Regulation of protein levels in subcellular domains through mRNA transport and localized translation

Localized protein synthesis is increasingly recognized as a means for polarized cells to modulate protein levels in subcellular regions and the distal reaches of their cytoplasm. The axonal and dendritic processes of neurons represent functional domains of cytoplasm that can be separated from their cell body by vast distances. This separation provides a biological setting where the cell uses locally synthesized proteins to both autonomously respond to stimuli and to retrogradely signal the cell body of events occurring is this distal environment. Other cell types undoubtedly take advantage of this localized mechanism, but these have not proven as amenable for isolation of functional subcellular domains. Consequently, neurons have provided an appealing experimental platform for study of mRNA transport and localized protein synthesis. Molecular biology approaches have shown both the population of mRNAs that can localize into axons and dendrites and an unexpectedly complex regulation of their transport into these processes. Several lines of evidence point to similar complexities and specificity for regulation of mRNA translation at subcellular sites. Proteomics studies are beginning to provide a comprehensive view of the protein constituents of subcellular domains in neurons and other cell types. However, these have currently fallen short of dissecting temporal regulation of new protein synthesis in subcellular sites and mechanisms used to ferry mRNAs to these sites.

Using a 3D virtual muscle model to link gene expression changes during myogenesis to protein spatial location in muscle

Background

Myogenesis is an ordered process whereby mononucleated muscle precursor cells (myoblasts) fuse into multinucleated myotubes that eventually differentiate into myofibres, involving substantial changes in gene expression and the organisation of structural components of the cells. To gain further insight into the orchestration of these structural changes we have overlaid the spatial organisation of the protein components of a muscle cell with their gene expression changes during differentiation using a new 3D visualisation tool: the Virtual Muscle 3D (VMus3D).

Results

Sets of generic striated muscle costamere, Z-disk and filament proteins were constructed from the literature and protein-interaction databases. Expression profiles of the genes encoding these proteins were obtained from mouse C2C12 cells undergoing myogenesis in vitro, as well as a mouse tissue survey dataset. Visualisation of the expression data in VMus3D yielded novel observations with significant relationships between the spatial location and the temporal expression profiles of the structural protein products of these genes. A muscle specificity index was calculated based on muscle expression relative to the median expression in all tissues and, as expected, genes with the highest muscle specificity were also expressed most dynamically during differentiation. Interestingly, most genes encoding costamere as well as some Z-disk proteins appeared to be broadly expressed across most tissues and showed little change in expression during muscle differentiation, in line with the broader cellular role described for some of these proteins.

Conclusion

By studying gene expression patterns from a structural perspective we have demonstrated that not all genes encoding proteins that are part of muscle specific structures are simply up-regulated during muscle cell differentiation. Indeed, a group of genes whose expression program appears to be minimally affected by the differentiation process, code for proteins participating in vital skeletal muscle structures. Expression alone is a poor metric of gene behaviour. Instead, the "connectivity model of muscle development" is proposed as a mechanism for muscle development: whereby the closer to the myofibril core of muscle cells, the greater the gene expression changes during muscle differentiation and the greater the muscle specificity.

SMN complex localizes to the sarcomeric Z-disc and is a proteolytic target of calpain

Spinal muscular atrophy (SMA) is a recessive neuromuscular disease caused by mutations in the human survival motor neuron 1 (SMN1) gene. The human SMN protein is part of a large macromolecular complex involved in the biogenesis of small ribonucleoproteins. Previously, we showed that SMN is a sarcomeric protein in flies and mice. In this report, we show that the entire mouse Smn complex localizes to the sarcomeric Z-disc. Smn colocalizes with alpha-actinin, a Z-disc marker protein, in both skeletal and cardiac myofibrils. Furthermore, this localization is both calcium- and calpain-dependent. Calpains are known to release proteins from various regions of the sarcomere as a part of the normal functioning of the muscle; however, this removal can be either direct or indirect. Using mammalian cell lysates, purified native SMN complexes, as well as recombinant SMN protein, we show that SMN is a direct target of calpain cleavage. Finally, myofibers from a mouse model of severe SMA, but not controls, display morphological defects that are consistent with a Z-disc deficiency. These results support the view that the SMN complex performs a muscle-specific function at the Z-discs.

Raver1 is an integral component of muscle contractile elements

Raver1, a ubiquitously expressed protein, was originally identified as a ligand for metavinculin, the muscle-specific isoform of the microfilament-associated protein vinculin. The protein resides primarily in the nucleus, where it colocalises and may interact with polypyrimidine-tract-binding protein, which is involved in alternative splicing processes. During skeletal muscle differentiation, raver1 translocates to the cytoplasm and eventually targets the Z-line of sarcomeres. Here, it colocalises with metavinculin, vinculin and alpha-actinin, all of which have biochemically been identified as raver1 ligands. To obtain more information about the potential role of raver1 in muscle structure and function, we have investigated its distribution and fine localisation in mouse striated and smooth muscle, by using three monoclonal antibodies that recognise epitopes in different regions of the raver1 protein. Our immunofluorescence and immunoelectron-microscopic results indicate that the cytoplasmic accumulation of raver1 is not confined to skeletal muscle but also occurs in heart and smooth muscle. Unlike vinculin and metavinculin, cytoplasmic raver1 is not restricted to costameres but additionally represents an integral part of the sarcomere. In isolated myofibrils and in ultrathin sections of skeletal muscle, raver1 has been found concentrated at the I-Z-I band. A minor fraction of raver1 is present in the nuclei of all three types of muscle. These data indicate that, during muscle differentiation, raver1 might link gene expression with structural functions of the contractile machinery of muscle.

The RNA-binding protein fragile X-related 1 regulates somite formation in Xenopus laevis

Fragile X-related 1 protein (FXR1P) is a member of a small family of RNA-binding proteins that includes the Fragile X mental retardation 1 protein (FMR1P) and the Fragile X-related 2 protein (FXR2P). These proteins are thought to transport mRNA and to control their translation. While FMR1P is highly expressed in neurons, substantial levels of FXR1P are found in striated muscles and heart, which are devoid of FMRP and FXR2P. However, little is known about the functions of FXR1P. We have isolated cDNAs for Xenopus Fxr1 and found that two specific splice variants are conserved in evolution. Knockdown of xFxr1p in Xenopus had highly muscle-specific effects, normal MyoD expression being disrupted, somitic myotomal cell rotation and segmentation being inhibited, and dermatome formation being abnormal. Consistent with the absence of the long muscle-specific xFxr1p isoform during early somite formation, these effects could be rescued by both the long and short mRNA variants. Microarray analyses showed that xFxr1p depletion affected the expression of 129 known genes of which 50% were implicated in muscle and nervous system formation. These studies shed significant new light on Fxr1p function(s).

Restricted distribution of mRNAs encoding a sarcoplasmic reticulum or transverse tubule protein in skeletal myofibers

Calsequestrin (CSQ) and dihydropyridine receptor (DHPR) are muscle cell proteins that are directed into the endoplasmic reticulum (ER) during translation. The former is subsequently found in the sarcoplasmic reticulum (SR) and the latter in the transverse tubule membrane. To elucidate the potential role of mRNA targeting within muscle cells, we have analyzed the localization of CSQ and DHPR proteins and mRNAs in primary cultured rat myotubes, in skeletal muscle cryosections, and in isolated flexor digitorum brevis muscle fibers. In the myotube stage of differentiation, the mRNAs distributed throughout the cell, mimicking the distribution of the endogenous ER marker proteins. In the adult skeletal myofibers, however, both CSQ and DHPRalpha1 transcripts located perinuclearly and in cross-striations flanking Z lines beneath the sarcolemma, a distribution pattern that sharply contrasted the interfibrillar distribution of typical ER proteins. Interestingly, all nuclei of the myofibers were transcriptionally active. In summary, the mRNAs encoding either a resident SR protein or a transverse tubule protein were located beneath the sarcolemma, implying that translocation of the respective proteins to the lumen of ER takes place at this location.

Characterization of Fxr1 in Danio rerio; a simple vertebrate model to study costamere development

The X-linked FMR1 gene, which is involved in the fragile X syndrome, forms a small gene family with its two autosomal homologs, FXR1 and FXR2. Mouse models for the FXR genes have been generated and proved to be valuable in elucidating the function of these genes, particularly in adult mice. Unfortunately, Fxr1 knockout mice die shortly after birth, necessitating an animal model that allows the study of the role of Fxr1p, the gene product of Fxr1, in early embryonic development. For gene function studies during early embryonic development the use of zebrafish as a model organism is highly advantageous. In this paper the suitability of the zebrafish as a model organism to study Fxr1p function during early development is explored. As a first step, we present here the initial characterization of Fxr1p in zebrafish. Fxr1p is present in all the cells from zebrafish embryos from the 2/4-cell stage onward; however, during late development a more tissue-specific distribution is found, with the highest expression in developing muscle. In adult zebrafish, Fxr1p is localized at the myoseptum and in costamere-like granules in skeletal muscle. In the testis, Fxr1p is localized in immature spermatogenic cells and in brain tissue Fxr1p displays a predominantly nuclear staining in neurons throughout the brain. Finally, the different tissue-specific isoforms of Fxr1p are characterized. Since the functional domains and the expression pattern of Fxr1p in zebrafish are comparable to those in higher vertebrates such as mouse and human, we conclude that the zebrafish is a highly suitable model for functional studies of Fxr1p.

Myocyte differentiation generates nuclear invaginations traversed by myofibrils associating with sarcomeric protein mRNAs

Certain types of cell both in vivo and in vitro contain invaginated or convoluted nuclei. However, the mechanisms and functional significance of the deformation of the nuclear shape remain enigmatic. Recent studies have suggested that three types of cytoskeleton, microfilaments, microtubules and intermediate filaments, are involved in the formation of nuclear invaginations, depending upon cell type or conditions. Here, we show that undifferentiated mouse C2C12 skeletal muscle myoblasts had smoothsurfaced spherical or ellipsoidal nuclei, whereas prominent nuclear grooves and invaginations were formed in multinucleated myotubes during terminal differentiation. Conversion of mouse fibroblasts to myocytes by the transfection of MyoD also resulted in the formation of nuclear invaginations after differentiation. C2C12 cells prevented from differentiation did not have nuclear invaginations, but biochemically differentiated cells without cell fusion exhibited nuclear invaginations. Thus, biochemical differentiation is sufficient for the nuclear deformation. Although vimentin markedly decreased both in the biochemically and in the terminally differentiated cells, exogenous expression of vimentin in myotubes did not rescue nuclei from the deformation. On the other hand, non-striated premyofibrils consisting of sarcomeric actinmyosin filament bundles and cross-striated myofibrils traversed the grooves and invaginations. Time-lapse microscopy showed that the preformed myofibrillar structures cut horizontally into the nuclei. Prevention of myofibril formation retarded the generation of nuclear invaginations. These results indicate that the myofibrillar structures are, at least in part, responsible for the formation of nuclear grooves and invaginations in these myocytes. mRNA of sarcomeric proteins including myosin heavy chain and alpha-actin were frequently associated with the myofibrillar structures running along the nuclear grooves and invaginations. Consequently, the grooves and invaginations might function in efficient sarcomeric protein mRNA transport from the nucleus along the traversing myofibrillar structures for active myofibril formation.

From the nucleus toward the cell periphery: a guided tour for mRNAs

RNA processing, directed transport along cytoskeletal tracks, and site-specific translation of mRNA at the cell periphery are considered discrete steps in the generation of microfilament-membrane adhesion complexes. A recently identified member of the heterogeneous nuclear ribonucleoprotein family, raver1, may couple these steps and contribute to the assembly and maintenance of these structures.

Fibre-type specific concentration of focal adhesion kinase at the sarcolemma: influence of fibre innervation and regeneration

In skeletal muscles, focal adhesion complexes (FACs) form part of the costamere, a sarcolemmal protein complex that enables lateral transfer of forces and ensures the stability of the sarcolemma. The present investigation tested whether localisation of a major assembly factor of FACs, focal adhesion kinase (FAK), to the sarcolemma parallels the known modulation of FACs by fibre type (innervation pattern) and fibre regeneration. Immunohistochemical experiments indicated that FAK is preferentially associated with the sarcolemma in a high proportion (>74 %) of the (slow-twitch) type I and(fast-twitch) type IIA fibres in normal rat soleus (N-SOL) muscle and of the type IIA fibres in extensor digitorum longus (N-EDL) muscle. In contrast, a low proportion (<15 %) of fast-twitch type IIB and type I fibres in N-EDL showed sarcolemmal FAK immunoreactivity. Cross-reinnervation of slow-twitch rat SOL muscle with the fast EDL nerve induced slow-to-fast fibre transformation and led to a significant reduction in sarcolemmal FAK immunoreactivity in type I and type IIA fibres. Transplantation of the fast EDL into the slow SOL bed with regeneration and reinnervation of the muscle by the SOL nerve (T-EDL) caused a significant increase in sarcolemmal FAK immunoreactivity in new type I and hybrid I/II fibres and a corresponding reduction in sarcolemmal FAK immunoreactivity in `normal' IIA and IIB fibres. Conversely, sarcolemmal FAK immunoreactivity in small IIB fibres of T-EDL muscle was increased. Correspondingly, the transplanted and regenerated SOL(reinnervated by the fast EDL nerve) maintained the percentage of FAK-positive sarcolemma in the (regenerated) type I and IIA fibres. Thus, the expression and association of FAK with the sarcolemma are regulated (i) by factors that determine the fibre type and (ii) during fibre regeneration. Our data suggest that the integrity of sarcolemmal FACs is dependent on the fibre type and that FAC turnover is increased during regeneration of muscle fibres.

Raver1, a dual compartment protein, is a ligand for PTB/hnRNPI and microfilament attachment proteins

By screening a yeast two-hybrid library with COOH-terminal fragments of vinculin/metavinculin as the bait, we identified a new protein termed raver1. Raver1 is an 80-kD multidomain protein and widely expressed but to varying amounts in different cell lines. In situ and in vitro, raver1 forms complexes with the microfilament-associated proteins vinculin, metavinculin, and alpha-actinin and colocalizes with vinculin/metavinculin and alpha-actinin at microfilament attachment sites, such as cell-cell and cell matrix contacts of epithelial cells and fibroblasts, respectively, and in costameres of skeletal muscle. The NH2-terminal part of raver1 contains three RNA recognition motifs with homology to members of the heterogeneous nuclear RNP (hnRNP) family. Raver1 colocalizes with polypyrimidine tract binding protein (PTB)/hnRNPI, a protein involved in RNA splicing of microfilament proteins, in the perinucleolar compartment and forms complexes with PTB/hnRNPI. Hence, raver1 is a dual compartment protein, which is consistent with the presence of nuclear location signal and nuclear export sequence motifs in its sequence. During muscle differentiation, raver1 migrates from the nucleus to the costamere. We propose that raver1 may coordinate RNA processing and targeting as required for microfilament anchoring in specific adhesion sites.

Perinuclear mRNA localisation by vimentin 3'-untranslated region requires a 100 nucleotide sequence and intermediate filaments

The role of the vimentin 3'-untranslated region (3'-UTR) in mRNA localisation was studied in cells transfected with a reporter sequence linked to subregions of the 3'-UTR. In situ hybridisation showed that nucleotides 37-137, including a previously identified protein-binding domain, were sufficient to localise transcripts to perinuclear cytoplasm. Transfection of two SW13 cell lines that do and do not express vimentin showed that perinuclear localisation due to either the vimentin or c-myc 3'-UTR requires intermediate filaments. The data suggest that both a specific protein-binding region of the vimentin 3'-UTR and intermediate filaments themselves are required to determine the site of vimentin synthesis.

Prosomes form sarcomere-like banding patterns in skeletal, cardiac, and smooth muscle cells

Prosomes (20S proteasomes) constitute the catalytic core of the 26S proteasomes, but were first observed as factors associated with unstranslated mRNA. Recently, their RNase activity was discovered together with the fact that their proteolytic function is dispensable in adapted human cells. By indirect immunofluorescence using monoclonal antibodies, we demonstrate as a general phenomenon, regular intercalation of specific types of prosomes into the sarcomeric structure of all types of striated muscle. Surprisingly, in cultured smooth muscle cells without sarcomeric organization, some prosomes also form regular striations in extended projections of cytoplasmic regions. The significance of their sarcomeric distribution is not understood as yet, but the pattern we observe is very similar to that shown by others for muscle-specific mRNAs, identified by in situ hybridization, and that of the cognate proteins. A role of prosomes in the cotranslational assembly of the myofibrillar proteins is suggested, since prosomes organize into pseudo-sarcomeric patterns prior to formation de novo of the actin-myosin arrangement.

The flightless I protein colocalizes with actin- and microtubule-based structures in motile Swiss 3T3 fibroblasts: evidence for the involvement of PI 3-kinase and Ras-related small GTPases

The flightless I protein contains an actin-binding domain with homology to the gelsolin family and is likely to be involved in actin cytoskeletal rearrangements. It has been suggested that this protein is involved in linking the cytoskeletal network with signal transduction pathways. We have developed antibodies directed toward the leucine rich repeat and gelsolin-like domains of the human and mouse homologues of flightless I that specifically recognize expressed and endogenous forms of the protein. We have also constructed a flightless I-enhanced green fluorescent fusion vector and used this to examine the localization of the expressed protein in Swiss 3T3 fibroblasts. The flightless I protein localizes predominantly to the nucleus and translocates to the cytoplasm following serum stimulation. In cells stimulated to migrate, the flightless I protein colocalizes with beta-tubulin- and actin-based structures. Members of the small GTPase family, also implicated in cytoskeletal control, were found to colocalize with flightless I in migrating Swiss 3T3 fibroblasts. LY294002, a specific inhibitor of PI 3-kinase, inhibits the translocation of flightless I to actin-based structures. Our results suggest that PI 3-kinase and the small GTPases, Ras, RhoA and Cdc42 may be part of a common functional pathway involved in Fliih-mediated cytoskeletal regulation. Functionally, we suggest that flightless I may act to prepare actin filaments or provide factors required for cytoskeletal rearrangements necessary for cell migration and/or adhesion.

Misdirected vimentin messenger RNA alters cell morphology and motility

Localized messenger RNAs were first observed as embryonic determinants that altered development when mislocalized. In recent years localized mRNAs have been found for several cytoskeletal proteins, including actin, vimentin and several microtubule associated proteins. We sought to determine whether redirecting mRNA for a cytoskeletal protein to an inappropriate address would alter cellular phenotypes. To do so we generated vimentin mRNAs with a myc epitope tag and the (beta)-actin 3′ untranslated region (3′ UTR) as a localization signal. When misdirected vimentin mRNAs are expressed in either fibroblasts or SW13 cells, cells develop numerous, extremely long processes; these cells also move more slowly to enter a wound of the monolayer. In situ hybridization revealed that the misdirected mRNA was often localized in the processes, in contrast to endogenous vimentin mRNA. The processes usually contained actin distal to the transgenic vimentin and microtubules proximal to it. SW13 cells lacking vimentin produced fewer and shorter processes, suggesting a dominant negative effect that involves recruitment of endogenous vimentin. Control experiments that transfected in constructs expressing tagged, correctly localized vimentin, or (beta)-galactosidase that localized through the (beta)-actin 3′ UTR, indicate that neither the shape nor the motility changes are solely due to the level of vimentin expression in the cell. This is direct evidence that the site of expression for at least one cytoskeletal mRNA alters the phenotype of the cell in which it is expressed. Messenger RNA localization is proving to be as essential for the normal maintenance of somatic cell phenotypes as embryonic determinants are for embryogenesis.

Specific types of prosomes distribute differentially between intermediate and actin filaments in epithelial, fibroblastic and muscle cells

First observed as components of non-translated mRNP complexes, prosomes harbour RNase and several proteinase activities; they are also the central constituent of the "Multicatalytic Proteinase (MCP) complexes" or "26S-proteasomes". In two recent publications (Arcangeletti et al., 1997b; De Conto et al., 1997) we have shown, by applying a new fixation technique, that these particles distribute differentially between the cytoskeletal networks of intermediate filament (IF) and actin types; previously they had been observed exclusively on the intermediate filaments. Here we further investigate the distribution of prosomes of several types, distinct by their subunit composition, between the IF of vimentin type and the actin network, as well as in the 3D space of the cell. It is shown that subtypes of prosomes occupy specific networks of the cytoskeleton, and that this pattern is specific for a given cell type. Confocal microscopy shows that prosome cytodistribution is not homogeneous in the 3D space: in the perinuclear area they colocalize most strongly with the IF, and more peripherally with the microfilament/stress fiber system; connections may exist between the two networks. Furthermore, new data indicate that the prosome-actin interaction may participate in the molecular structure of the stress fibers.

In the nucleus and cytoplasm of chicken erythroleukemic cells, prosomes containing the p23K subunit are found in centers of globin (pre-)mRNA processing and accumulation

Prosomes were originally identified as 20S particles associated with untranslated mRNA; they also constitute the core of the 26S proteasomes. The cellular distribution of three types of prosomes characterized by the presence of subunits with molecular masses of 23, 27, and 30 kDa was analyzed using an immunocytochemical approach on cultured chicken erythroblasts. The prosomes containing the p27K and p30K subunits were found in diffuse distribution in both nuclei and cytoplasm. In contrast, the prosomes containing the p23K subunit, although relatively rare in the nuclear space, were found concentrated in one or two large spots. Using in situ hybridization with an alpha(A)-globin gene-specific riboprobe we found that the p23K-type prosomes colocalize in the nucleus with centers of globin (pre-)mRNA processing, and of mRNA accumulation in the cytoplasm. This result suggests there is local coincidence of specific-type prosome function with processing and, possibly, transport of a particular kind of (pre-)mRNA.

Dynamic distribution and formation of a para-sarcomeric banding pattern of prosomes during myogenic differentiation of satellite cells in vitro

Myogenesis proceeds by fusion of proliferating myoblasts into myotubes under the control of various transcription factors. In adult skeletal muscle, myogenic stem cells are represented by the satellite cells which can be cultured and differentiate in vitro. This system was used to investigate the subcellular distribution of a particular type of prosomes at different steps of the myogenic process. Prosomes constitute the MCP core of the 26S proteasomes but were first observed as subcomplexes of the untranslated mRNPs; recently, their RNase activity was discovered. A monoclonal antibody raised against the p27K subunit showed that the p27K subunit-specific prosomes move transiently into the nucleus prior to the onset of myoblast fusion into myotubes; this represents possibly one of the first signs of myoblast switching into the differentiation pathway. Prior to fusion, the prosomes containing the p27K subunit return to the cytoplasm, where they align with the gradually formed lengthwise-running desmin-type intermediate filaments and the microfilaments, co-localizing finally with the actin bundles. The prosomes progressively form discontinuous punctate structures which eventually develop a pseudo-sarcomeric banding pattern. In myotubes just formed in vitro, the formation of this pattern seems to preceed that produced by the muscle-specific sarcomeric (alpha)-actin. Interestingly, this pattern of prosomes of myotubes in terminal in vitro differentiation was very similar to that of prosomes observed in vivo in foetal and adult muscle. These observations are discussed in relation to molecular myogenesis and prosome/proteasome function.

Mechanical loading regulates expression of talin and its mRNA, which are concentrated at myotendinous junctions

The hypothesis that mechanical loading regulates talin expression in developing and adult muscle was tested using in vitro and in vivo models. Talin was selected for study because it is a key structural link between the cytoskeleton and cell membrane. In the in vitro model, C2C12 myotubes were subjected to cyclic strains for 48 h. In the in vivo model, rat hindlimb muscles were unloaded for 10 days, then reloaded for 2 days. Cyclic loading of myotubes resulted in significant increases in the quantity of talin (68%) and its 190-kDa proteolytic fragment (70%), as well as talin mRNA (180%), relative to unloaded myotube cultures. Similarly, talin concentration and its mRNA increased by 68 and 136%, respectively, in soleus muscles reloaded for 2 days relative to ambulatory controls. Immunohistochemistry and in situ RT-PCR showed that talin and its mRNA are concentrated and colocalized at myotendinous junctions. Thus these findings indicate that increased mechanical loading promotes talin synthesis, which occurs principally at myotendinous junctions, according to talin mRNA distribution.

Nuclear domains in skeletal myotubes: the localization of transferrin receptor mRNA is independent of its half-life and restricted by binding to ribosomes

The retention of mRNAs near the nuclei that synthesize them may be an important feature of the organization of multinucleated skeletal myotubes. Here, we assess the possible role of two factors in this localization. First, we examine the role of mRNA half-life, by studying the distribution of the mRNA for the transferrin receptor (TfR), whose half-life can be manipulated in culture by changing the availability of iron. In situ hybridization of myotubes of the mouse muscle cell line C2 shows that TfR mRNA is concentrated in the core of the myotubes. Its distribution around the nuclei is often asymmetric and its concentration changes abruptly. Stable transcripts display the same asymmetric localization as unstable ones, suggesting that half-life does not determine subcellular localization of TfR mRNA. Differential effects of the protein synthesis inhibitors puromycin and cycloheximide suggest that the mRNA is retained in position by its association with ribosomes. We then examine the distribution of the rough endoplasmic reticulum (RER) and find it to be broader than the distribution of TfR mRNA. In contrast to TfR mRNA, the mRNA for a secreted immunoglobulin kappa light chain has a more uniform distribution. Taken together, the results suggest that TfR mRNA may associate with RER subdomains by specific targeting.

The 3' untranslated region plays a role in the targeting of metallothionein-I mRNA to the perinuclear cytoplasm and cytoskeletal-bound polysomes

The mechanism of localisation of metallothionein-I (MT-I) mRNA was studied in transfected cells by in situ hybridisation and cell fractionation. Hepatoma cells were transfected with the 5'-untranslated region and coding region of the beta-globin gene alone or linked to either the beta-globin 3'-untranslated region (3'-UTR) or the MT-I 3'-UTR. The wild-type beta-globin mRNA and the beta-globin mRNA lacking its native 3'-UTR were present in free and cytoskeletal-bound polysomes to a similar extent and showed no localisation. Chimaeric globin-metallothionein transcripts were significantly enriched in cytoskeletal-bound polysomes and were localised in the perinuclear cytoplasm. Chimaeric globin-metallothionein and wild-type globin transcripts were of similar stability. Chinese Hamster Ovary cells were transfected with constructs in which the MT-I 5'-untranslated region and coding sequences were linked to either the endogenous 3'-UTR or the glutathione peroxidase 3'-UTR. Wild-type MT-I transcripts were localised in the perinuclear cytoplasm but the chimaeric MT-I-glutathione peroxidase transcripts showed no distinct localisation. The results indicate that the 3'-UTR of MT-I mRNA contains a localisation signal which promotes both the association of the mRNA with the cytoskeleton and its perinuclear localisation.

Evidence for a localisation signal in the 3'-untranslated region from vimentin messenger RNA

There is increasing evidence that some mRNAs are localised in eukaryotic somatic cells, but it is unclear what proportion of mRNAs are localised and whether this sorting involves 3'-untranslated sequences. The presence of a localisation signal within the 3'-untranslated region of vimentin mRNA was investigated by studying mRNA distribution in fibroblasts transfected with beta-globin and hybrid globin-vimentin gene constructs. In cells transfected with constructs containing either a fragment of the rabbit beta-globin gene containing both coding sequences and 3'untranslated region or the beta-globin coding sequences alone in situ hybridisation showed that beta-globin mRNA was distributed throughout the cytoplasm without any evident localisation. In contrast, in cells transfected with globin coding sequences linked to the vimentin 3'-untranslated region there was a strong perinuclear localisation of the hybrid mRNA. The results show that loss of its endogenous 3'-untranslated region does not affect distribution of beta-globin mRNA whereas the vimentin 3'-untranslated region causes an altered localisation of beta-globin mRNA. We conclude that the vimentin 3'-untranslated region contains a localisation signal which can direct reporter sequences to the perinuclear cytoplasm.

Prosome cytodistribution relative to desmin and actin filaments in dividing C2.7 myoblasts and during myotube formation in vitro

Prosomes constitute the multicatalytic proteinase (MCP) core of the 26S proteasomes, but were first observed as subcomplexes of untranslated mRNP; this suggests that they play a putative role in the control of protein biosynthesis in addition to their catabolic enzymatic function. In previous investigations it was shown that some prosomes colocalize with the intermediate filaments (IF) of the cytoskeleton, of the cytokeratin type in epithelial cells, and of the vimentin type in fibroblasts. Studies on adult rat muscle carried out with prosome-specific monoclonal antibodies (p-mAbs) have shown, surprisingly, that specific types of prosomes predominantly occupy a particular zone in between the M and the Z lines of the sarcomeric structure. The data presented here show that the subunit composition of prosomes changes when the dividing C2.7 myoblasts fuse into myotubes. We show furthermore that, in dividing C2.7 myoblasts, prosomes colocalize with the desmin network as well as with that of actin, in a distribution that changes with the subunit pattern of the prosomes investigated by individual p-mAbs. Surprisingly, when myogenic fusion is induced, specific types of prosomes move first to the nuclei; later on, they reappear in the cytoplasm. There, superimposing initially onto the reorganizing desmin filaments that run from one pole of the prefusion myoblast to the other, prosomes gradually colocalize with the actin fibers in the fusing myotubes, finally forming a "pearl on a string" pattern. These results are discussed in relation to parallel observations of prosome distribution between the actin and IF networks not only in epithelial cells but also in fusing muscle satellite cells, which made it possible to monitor the complete buildup of the sarcomeric structure.

Effects of divalent cations on M-cadherin expression and distribution during primary rat myogenesis in vitro

In the process of myogenesis, cadherins are thought to be involved in the initial cell-cell recognition and possible initiation of myoblast fusion to form multinucleated myotubes. Of the cadherins, M-cadherin, but not N-cadherin, is down-regulated upon inhibition of myogenesis, suggesting that M-cadherin may be a key receptor involved in myogenesis. M-cadherin binds in a calcium-dependent manner, and depletion of divalent cations inhibits myoblast fusion. We analyzed the regulation of M-cadherin protein and mRNA levels in primary rat myogenic cultures in the presence and absence of divalent cations. In untreated cultures M-cadherin was localized to various myogenic cell-cell contacts. M-cadherin protein and mRNA levels showed a peak at day 2 after the initiation of growth. When divalent cations were removed from the cell culture medium, myoblast fusion was inhibited and immunocytochemical analysis revealed a failure of M-cadherin to localize to cell-cell contacts. Analysis of M-cadherin protein and mRNA in fusion-inhibited cultures still revealed a peak at day 2. However, by day 3, M-cadherin protein levels in the fusion-inhibited cultures were reduced in both the detergent-soluble and -insoluble fractions in comparison with the untreated cultures. Interestingly, beta-catenin, a protein associated with cadherins, was frequently observed at intercellular contacts in the fusion-inhibited cultures. We could also show that the intracellular levels of beta-catenin protein remained constant regardless of the presence or absence of divalent cations. In summary, the dynamic regulation of M-cadherin in muscle-fusion-related events is an indication of the importance of M-cadherin for myoblast fusion and myogenic differentiation.

A localisation signal in the 3′ untranslated region of c-myc mRNA targets c-myc mRNA and beta-globin reporter sequences to the perinuclear cytoplasm and cytoskeletal-bound polysomes

There is increasing evidence that in mammalian cells some mRNAs are localised to specific parts of the cytoplasm and a proportion of mRNAs and polyribosomes are associated with the cytoskeleton. It has been shown previously that c-myc mRNA is present in the perinuclear cytoplasm and associated with the cytoskeleton, and that this localisation is dependent upon the 3′ untranslated region of the mRNA. The present studies show that in transfected fibroblasts the c-myc 3′ untranslated region is able to localise beta-globin reporter sequences to the perinuclear cytoplasm. Studies with constructs containing deletions within the 3′ untranslated region identify the region between bases 194 and 280 as critical for localisation. Transfection of cells with constructs in which this region is linked to beta-globin sequences showed that it was sufficient to localise the chimaeric transcripts to the perinuclear cytoplasm and to cytoskeletal-bound polyribosomes. Transfection with constructs containing a mutated AUUUA sequence within the 194–280 base region showed that this conserved AUUUA is required for targeting of both c-myc mRNA and a chimaeric transcript of beta-globin transcripts linked to the c-myc 3′ untranslated region. The region between bases 194 and 280 did not induce instability of beta-globin transcripts and the AUUUA mutation had little effect upon mRNA stability. We propose that this 86 nt region of the 3′ untranslated region contains a localisation signal to target c-myc mRNA so that it is retained on cytoskeletal-bound polysomes in the perinuclear cytoplasm; a conserved AUUUA sequence appears to be a critical part of this signal.

Intracellular localization of the mRNAs of argininosuccinate synthetase and argininosuccinate lyase around liver mitochondria, visualized by high-resolution in situ reverse transcription-polymerase chain reaction

Argininosuccinate synthetase and argininosuccinate lyase, two cytoplasmic enzymes of the urea cycle, are released into the soluble phase in the absence of detergent when cells are disrupted. Yet previous biochemical studies, as well as immunocytochemistry at the electron microscope level, have shown that these enzymes are localized around mitochondria in situ. Such intracellular localization of soluble enzymes requires mechanisms to deliver the proteins to the appropriate sites, where they may then be anchored by specific protein-protein interactions. A method was developed to examine the intracellular distribution of the mRNA of argininosuccinate synthetase and argininosuccinate lyase in intact rat liver at the ultrastructural level by in situ reverse transcription and the polymerase chain reaction, using primers targeting regions of the coding sequences of the rat enzymes, digoxigenin-dUTP as the label, and anti-digoxigenin/1 nm [corrected] gold plus silver enhancement as the detection method. The tissue was fixed in 4% paraformaldehyde/0.1% glutaraldehyde and embedded in Lowicryl. Examination of the numbers and the location of the silver grains, coupled with morphometric analysis of the electron micrographs, permitted the calculation of the silver "enrichment ratio" for each type of cell structure. These ratios showed that the mRNAs for argininosuccinate synthetase and argininosuccinate lyase were located next to the cytoplasmic side of the mitochondrial membrane and in the nearby endoplasmic reticulum. Most of the silver grains that were observed in the endoplasmic reticulum were within 200 nm of the mitochondria; it was not possible, however, to determine if those grains were actually associated with the reticular membranes. These studies demonstrate that the mRNAs of these two soluble cytoplasmic proteins are localized to the same limited regions where the proteins are situated. Translation of the proteins, therefore, must occur at these specific sites. The targeting of argininosuccinate synthetase and argininosuccinate lyase mRNAs to the immediate vicinity of the mitochondria may be the first step of the mechanisms by which the spatial organization of these soluble proteins in situ is accomplished. The targeting of mRNAs for soluble cytoplasmic proteins of organized metabolic pathways has not been demonstrated previously. These studies also show that in situ reverse transcription and the polymerase chain reaction at the ultrastructural level, which has not been previously reported, can be used to detect specific mRNAs; it should be extremely valuable for the intracellular detection of low-abundance mRNAs.

Microfilament-membrane interactions in Xenopus myocytes

We used quick-freeze, deep-etch, rotary-replication transmission electron microscopy to determine at molecular resolution the organization of microfilaments at the cytoplasmic surface of the sarcolemma of Xenopus myocytes. We demonstrate that actin microfilaments interact with the sarcolemma in two distinct ways. In one, which resembled focal contacts in Xenopus fibroblasts [Samuelsson et al., 1993: J. Cell Biol. 122:485-496], bundles of microfilaments approached the sarcolemma at sites containing aggregates of membrane-associated particles. Immunogold cytochemistry showed that these particle aggregates contained vinculin, talin and beta 1-integrin. In the second, which covered most of the cytoplasmic surface of the sarcolemma, individual actin microfilaments formed an extensive, lattice-like array. Particle aggregates associated with this array of actin microfilaments also labeled with antibodies to vinculin, talin and beta 1-integrin. The unique, lattice-like association of actin microfilaments with the membrane in Xenopus myocytes suggests that the organization of actin filaments over most of the sarcolemma is distinct from focal contacts, mediating widespread associations of the actin cytoskeleton with the cytoplasmic membrane face.

The intracellular localization of messenger RNAs

As I hope this review has made clear, mRNA localization plays an important role in directing specific proteins to their correct position within a cell. Although the study of this process is still in its infancy, it is already apparent that there are several ways that mRNAs can be targeted to particular subcellular sites. However, the molecular mechanisms responsible for these different localization pathways are still largely obscure, and their elucidation must await the identification of the specific factors that mediate the interactions between the localized mRNAs and more general components such as the cytoskeleton. Most examples of localized mRNAs are likely to share several common features. First, the site of localization will be determined by the preexisting polarity of the cell, and this will most often depend on the organization of the cytoskeleton, either directly, in the case of active transport, or indirectly, when localization is mediated by localized anchoring sites or stability factors. Second, mRNA localization is likely to be tightly coupled to translational control. If it is important for a cell to synthesize a protein in a particular place, then the translation of the mRNA must be repressed until it is localized. Indeed, there are already several examples where the direct linkage between translational control and localization has been demonstrated, and these are discussed in the accompanying review by Curtis et al. (1995).