On the association of mRNA with the cytoskeleton in uninfected and adenovirus-infected human KB cells

Association of mRNA and eIF-2 alpha with the cytoskeleton in cells lacking vimentin

The human bladder carcinoma cell lines RT4 and T24 and the human breast adenocarcinoma cell line MCF-7 were found to be negative for vimentin when studied by means of immunofluorescence and immunoblotting. Northern blot analysis revealed that these cells lacked detectable levels of vimentin mRNA with the exception of T24, which contains trace amounts of vimentin mRNA compared to the RNA level in vimentin-containing HeLa cells. CAT assays performed on these cells showed that a hamster vimentin promoter is inactive in RT4 and MCF-7 cells. In the vimentin-lacking cells, the binding of polyribosomes, specific mRNAs, and translation factor eIF-2 alpha to the cytoskeletal fraction was examined. Our results indicate that the presence of a vimentin network is not crucial for the association of the translation machinery with the cytoskeleton. Furthermore, in these vimentin-negative cell lines the immunofluorescence staining pattern of eIF-2 alpha shows a fibro-granular structure that has no resemblance to the cytokeratin or actin cytoskeleton present in these cells.

Nontranslated cellular mRNAs are associated with the cytoskeletal framework in influenza virus or adenovirus infected cells

In an effort to understand the molecular mechanisms underlying the selective shutoff of host protein synthesis in influenza virus and adenovirus infected cells, we analyzed the subcellular location of representative cellular and viral mRNAs. Earlier work has shown that the majority of cellular mRNAs remain polysome associated after infection by either virus and that both the initiation and elongation steps of host protein synthesis were blocked in infected cells (M. G. Katze, D. DeCorato, and R. M. Krug, J. Virol., 60, 1027-1039, 1986). The present study was undertaken to test whether these cellular mRNAs were rendered nontranslatable during infection as a result of their dissociation from the cytoskeleton framework. HeLa cells were fractionated into subcellular components by first gently disrupting the cells with Triton X-100 yielding the soluble fraction (SOL); the cytoskeleton (CSK) fraction was obtained from the Triton insoluble material by the double detergent treatment of Tween-40 and sodium deoxycholate. In uninfected cells the majority of host mRNAs were associated with polysomes which were exclusively bound to the CSK as would be expected of actively translated mRNAs. The cellular mRNAs also remained almost totally associated with the cytoskeleton in adenovirus and influenza virus infected cells despite the fact that these mRNAs are not translated during infection. Indeed, the host mRNAs and the efficiently translated viral mRNAs were CSK associated to the same extent. In contrast to the adenovirus and influenza systems, significant amounts of cellular mRNAs were dissociated from the CSK and found in the SOL fraction of poliovirus infected cells as others have reported. In accordance with the biochemical analysis, morphological studies utilizing electron microscopy revealed that the cytoskeleton remained relatively intact during adenovirus and influenza infection but was substantially reorganized in poliovirus infected cells. We conclude that translational regulatory events are likely different in the poliovirus system and that cytoskeletal association of mRNAs may be required but is not sufficient for efficient mRNA translation during adenovirus or influenza virus infection.

Concurrent collapse of keratin filaments, aggregation of organelles, and inhibition of protein synthesis during the heat shock response in mammary epithelial cells

The sequence of heat shock-induced perturbations in protein synthesis and cytoskeletal organization was investigated in primary cultures of mouse mammary epithelial cells (MMEC). Exposure of the cells to 45 degrees C for 15 min caused a marked inhibition of protein synthesis through 2 h after heart. Resumption of protein synthesis began by 4 h, was complete by 8 h, and was accompanied by induction of four major heat shock proteins (HSPs) of 68, 70, 89, and 110 kD. Fluorescent cytochemistry studies indicated that heat shock elicited a reversible change in the organization of keratin filaments (KFs) and actin filaments but had a negligible effect on microtubules. Changes in the organization of KFs progressed gradually with maximal retraction and collapse into the perinuclear zone occurring at 1-2 h after heat followed by restoration to the fully extended state at 8 h. In contrast, actin filaments disappeared immediately after heat treatment and then rapidly returned within 30-60 min to their original appearance. The translocation of many organelles first into and then away from the juxtanuclear area along with the disruption and reformation of polyribosomes were concurrent with the sequential changes in distribution of KFs. The recovery of the arrangement of KFs coincided with but was independent of the resumption of protein synthesis and induction of HSPs. Thermotolerance could be induced in protein synthesis and KFs, but not in actin filaments, by a conditioning heat treatment. Neither protein synthesis nor induction of HSPs was necessary for the acquisition of thermotolerance in the KFs. The results are compatible with the possibility that protein synthesis may depend on the integrity of the KF network in MMEC. Heat shock thus can efficiently disarrange the KF system in a large population of epithelial cells, thereby facilitating studies on the functions of this cytoskeletal component.

Isolation of hamster brain polyribosomes-cytoskeleton complexes

We have developed a method for the isolation of a brain subcellular fraction enriched in both highly aggregated polyribosomes and cytoskeletal proteins. This method is based on gentle homogenization of brain tissue and low speed centrifugation. The mechanism of association of polyribosomes to cytoskeletal structures has been studied by in vitro treatment of this fraction with polyribosome-disaggregating agents. RNase and EDTA, while succeeding in completely disrupting them into monosomes or subunits, did not release them from cytoskeleton. Puromycin showed no noticeable effect.

Altered expression and distribution of the cytoskeletal-associated p35 protein in NIH 3T3 cells transformed with the Harvey sarcoma virus v-ras oncogene

1. Cytoskeletal events associated with retroviral oncogene (v-ras)-mediated transformation were studied in NIH 3T3 fibroblasts and their v-ras-transfected counterparts (3T3/H-1 cells). 2. Abnormal microfilament networks seen in 3T3/H-1 cells reflected significant decreases (approximately 90%) in two cytoskeletal-associated proteins (tropomyosin-1, p35). Neither actin content nor actin mRNA levels were altered, however, v-ras transfectants. 3. p35 mRNA activity in both NIH 3T3 and 3T3/H-1 cells was similar although differential compartmentalization of p35 to the detergent-resistant cytoskeletal fraction was evident only in normal fibroblasts. 4. Proper cytoskeletal organization may be a factor in the regulation of p35 mRNA translation in situ or influence the stability of p35 independent of translational rate.

Strategies for ultrastructural visualization of biotinated probes hybridized to messenger RNA in situ

A progressive development of the application of in situ methodology to ultrastructural procedures has resulted in the ability to detect individual molecules of mRNA with high probability. Beginning with whole-mount cells and then developing myotubes, both in culture and detergent extracted before fixation, we were able to progress to methods which allow detection of mRNA in tissue sections. Initial results confirm that the detection of mRNA in thin-sectioned tissue is very similar to observations on the extracted, cultured cells, and that the same methods of data analysis apply. Current work is devoted to the application of the methodology to other cellular structures, such as the nucleus, and to other tissue-probe systems, such as brain. Acknowledgements. The authors appreciate the skilled help from John McNeil and Shirwin Pockwinse in the laborious and time-consuming preparations of material and photography. FS was on sabbatical leave from the Department of Pathology at Southwestern Medical Center.

Association of ribosomes with in vitro assembled microtubules

Microtubules were purified from unfertilized eggs of the sea urchins Arbacia punctulata, Lytechinus pictus, Lytechinus variegatus, and Strongylocentrotus purpuratus. Numerous densely stained particles (24 x 26 nm) are associated with microtubules isolated from each of these sea urchins. The most striking aspect of this structure is an extended, slightly curved arm that appears to attach the particles to the microtubule. Morphologically similar particles are associated with microtubules of the isolated first cleavage mitotic apparatus. The particles are attached to the microtubules by ionic interactions and contain large amounts of extractable RNA. Based upon their size and density, RNA and protein composition, and sedimentation in sucrose gradients, the microtubule-associated particles are identified as ribosomes.

Modification of protein synthesis initiation factors and the shut-off of host protein synthesis in adenovirus-infected cells

A substantial body of data, largely derived from study of cell extracts, indicates that protein synthesis in adenovirus-infected cells requires VA RNAI at late times of infection to prevent the activation of a protein kinase known as DAI, and the consequent phosphorylation of the alpha-subunit of initiation factor eIF-2. To verify this conclusion, we have measured the steady-state levels of eIF-2 alpha phosphorylation in cells infected with wild-type virus (Ad2) and a mutant that produces no VA RNAI (Ad5dl331). Consistent with the proposed mechanism, the alpha-subunit was very highly phosphorylated (approximately 90%) at late times of infection with Ad5dl331. Surprisingly, eIF-2 alpha phosphorylation also increased (to approximately 30%) at late times of infection with Ad2, suggesting that VA RNA and DAI might be involved in the selective translation of viral mRNA and the shut-off of host cell protein synthesis during the late phase. In agreement with this model, host protein synthesis shut-off is defective in cells expressing low levels of DAI.

Effect of vasopressin on the regulation of protein synthesis initiation in liver cells

Vasopressin was found to be an effective inhibitor of protein labelling in isolated liver cells. Its effect shows the following distinct characteristics: (1) in contrast with alpha-adrenergic agonists, its effect is observable under a wide range of cellular Ca2+-loading conditions; (2) it is not influenced by the nutritional state of the animal. The lack of vasopressin effect on valine production, and its ability to decrease protein labelling from near-saturation concentrations of [3H]valine, indicate that the observed variations in protein labelling reflect actual changes in the rate of protein synthesis. The action of vasopressin is primarily exerted on the initiation step of protein synthesis and this effect is accompanied by a decreased activity of eukaryotic initiation factor 2. Activators of protein kinase C showed similar but not additive effects on protein synthesis, as did vasopressin. It seems plausible to conclude that protein kinase C activation may play an important regulatory role in hepatic protein synthesis as a transducer of hormonal and perhaps other type of signals.

Evidence that insulin increases the proportion of polysomes that are bound to the cytoskeleton in 3T3 fibroblasts

The association of polysome redistribution with changes in protein synthesis was investigated in insulin-stimulated fibroblasts. Free polysomes were released by Nonidet-P40 and 25 mM KCl, cytoskeletal-bound polysomes were retained at 25 mM KCl but released at 130 mM, while membrane-bound polysomes were released by deoxycholate. Insulin increased the proportion of polysomes which were retained at 25 mM KCl but had no effect when extraction was carried out at 130 mM KCl, suggesting that more polysomes were associated with the cytoskeleton. Insulin also reduced the amount of actin released from the detergent-insoluble cytoskeleton indicating that the hormone affects microfilament organization.

Relationship between heat-shock protein synthesis and thermotolerance in rainbow trout fibroblasts

The role of heat-shock protein synthesis in the development of thermotolerance by rainbow trout fibroblasts was examined. During the first 6 h after being shifted from 22 degrees C to 28 degrees C, cells of the rainbow trout fibroblast line, RTG-2, rapidly synthesized the major heat-shock proteins (hsps), hsps 87, 70 and 27, and developed tolerance to 32 degrees C. After 24 h at 28 degrees C hsp synthesis was drastically reduced but thermotolerance was maintained. If these thermotolerant cells were shifted to 32 degrees C, hsp synthesis continued at a very low level, but if they were subsequently returned to 22 degrees C, synthesis of hsps 70 and 27 was induced again. The addition of actinomycin D during the first 6 h at 28 degrees C prevented hsp synthesis and the development of thermotolerance. The presence of actinomycin D during the incubation of thermotolerant cultures at 32 degrees C blocked the reinitiation of hsps synthesis at 22 degrees C but had no effect on survival. Therefore, the hsps that accumulated at 28 degrees C were sufficient to allow cells to survive a subsequent thermal stress at 32 degrees C.

Embryonic corneal epithelial interaction with exogenous laminin and basal lamina is F-actin dependent

Between the third and sixth day of embryonic development, the avian corneal epithelium produces both a basal lamina and the primary corneal stroma composed of 20 orthogonally arranged layers of collagen fibrils. If the epithelium is removed by enzyme treatment from the basal lamina and stroma, the basal cell surface extends cell processes (blebs) which contain disorganized actin filaments and the epithelium decreases production of collagen. When placed on extracellular matrix or on Millipore filters in media containing soluble matrix molecules, the epithelium retracts the blebs, forms an organized basal actin cortical mat, and doubles its production of collagen. In the current investigation, we provide evidence for the hypothesis that organization of the RER by the actin cytoskeleton mediates this stimulation of collagen production. Laminin-treated epithelia and epithelia isolated with the basal lamina intact were treated with an actin-disrupting drug, cytochalasin D. Actin aggregates occur throughout the epithelium, the RER becomes disorganized, and the increase in collagen production expected to result from addition of laminin does not take place. Morphometrical analysis of the distribution of RER in the basal compartment of control and cytochalasin-treated epithelia shows that the decrease in collagen production is accompanied by displacement of the RER from the basal area of the cells, suggesting that attachment of RER to the intact actin cytoskeleton is essential to maintenance of normal RER organization and function. We also found that laminin-mediated bleb retraction requires intact actin microfilaments, whereas bleb extension does not, and that nocodazole does not inhibit bleb extension or retraction.

Cotranslational assembly of myosin heavy chain in developing cultured skeletal muscle

To examine how nascent myosin heavy chains associate with the cytoskeletons of developing muscle cells, we used pulse labeling, cell fractionation, and immunoprecipitation. More than 80% of nascent myosin heavy chains associate with the cytoskeleton. More than one-third of these nascent chains are not released by puromycin and/or RNase. The fraction of nascent heavy chains that resists release increases during development of muscle cells in culture. Treatment with cytochalasin D but not nocodazole decreases myosin heavy chain cotranslational assembly. These results indicate that (i) cotranslational assembly of myosin heavy chains is developmentally regulated, (ii) structures containing actin and not microtubules may mediate initial association of the heavy chains with the cytoskeleton, and (iii) the site of translation dictates where a significant fraction of the heavy chains will be inserted into the cytoskeleton.

Cytochalasin B selectively releases ovalbumin mRNA precursors but not the mature ovalbumin mRNA from hen oviduct nuclear matrix

Hen oviduct nuclear matrix-bound mature ovalbumin mRNA is released from the matrix in the presence of ATP, while the ovalbumin mRNA precursors remain bound to this structure. Detachment of the mature mRNA from the matrix by ATP as well as ATP-dependent efflux of mRNA from isolated nuclei were found to be inhibited by cytochalasin B. On the other hand, in the absence of ATP, cytochalasin B exclusively caused the release (and nucleocytoplasmic efflux) of the ovalbumin messenger precursors, but not of the mature mRNA. After cytochalasin B treatment, actin could be detected in the matrix supernatant. Phalloidin which stabilizes actin filaments did not cause RNA liberation in the absence of ATP, but inhibited the ATP-induced detachment of mature mRNA. RNA release was also achieved with a monoclonal antibody against actin but not with monoclonal antibodies against tubulin and intermediate filaments. These results suggest that actin-containing filaments are involved in the restriction of immature messengers to the cell nucleus.

Modulation of fibronectin gene expression in chondrocytes by viral transformation and substrate attachment

Chicken vertebral chondrocytes, which normally grow in suspension, synthesize large amounts of cartilage extracellular matrix proteins, but little fibronectin. We have analyzed the effects of both substrate attachment and transformation with a temperature-sensitive mutant of Rous sarcoma virus on fibronectin gene expression in these cells. Our experiments show that viral transformation increases fibronectin synthesis to a greater extent than substrate attachment. Furthermore, transformed chondrocytes have lost the ability to decrease fibronectin synthesis in response to suspension culture, suggesting that transformation alters the normal attachment-responsive control of fibronectin gene expression. Finally, infected substrate-attached chondrocytes shifted to the nonpermissive temperature for transformation use fibronectin RNA more efficiently in protein synthesis than cells grown under the other conditions, suggesting for the first time a role for translational control of fibronectin gene expression.

The function of proteins that interact with mRNA

Specific proteins are associated with mRNA in the cytoplasm of eukaryotic cells. The complement of associated proteins depends upon whether the mRNA is an integral component of the polysomal complex being translated, or, alternatively, whether it is part of the non-translated free mRNP fraction. By subjecting cells to ultraviolet irradiation in vivo to cross-link proteins to mRNA, mRNP proteins have been shown to be associated with specific regions of the mRNA molecule. Examination of mRNP complexes containing a unique mRNA has suggested that not all mRNA contain the same family of associated RNA binding proteins. The functions of mRNA associated proteins may include a role in providing stability for mRNA, and/or in modulating translation. With the recent demonstrations that both free and polysomal mRNPs are associated with the cytoskeletal framework, specific mRNP proteins may play a role in determining the subcellular localization of specific mRNPs.

Cytochalasin releases mRNA from the cytoskeletal framework and inhibits protein synthesis

Cytochalasin D was shown to be a reversible inhibitor of protein synthesis in HeLa cells. The inhibition was detectable at drug levels typically used to perturb cell structure and increased in a dose-dependent manner. The drug also released mRNA from the cytoskeletal framework in direct proportion to the inhibition of protein synthesis. The released mRNA was unaltered in its translatability as measured in vitro but was no longer translated in the cytochalasin-treated HeLa cells. The residual protein synthesis occurred on polyribosomes that were reduced in amount but displayed a normal sedimentation distribution. The results support the hypothesis that mRNA binding to the cytoskeletal framework is necessary although not sufficient for translation. Analysis of the cytoskeletal framework, which binds the polyribosomes, revealed no alterations in composition or amount of protein as a result of treatment with cytochalasin D. Electron microscopy with embedment-free sections shows the framework in great detail. The micrographs revealed the profound reorganization effected by the drug but did not indicate substantial disaggregation of the cytoskeletal elements.

Cytoskeletal association of muscle-specific mRNAs in differentiating L6 rat myoblasts

The importance of the cytoskeleton in protein synthesis was studied in differentiating L6 rat myoblasts. Soluble and cytoskeletal fractions obtained after gentle, non-ionic detergent lysis of myoblasts and myotubes were analysed for the presence of ribosomes and mRNPs. Polysomal mRNPs were predominantly associated with the cytoskeletal framework and free mRNPs were present in both soluble and cytoskeletal fractions. An examination of the distribution of specific mRNAs in the polysomal and free mRNP populations of both cytoplasmic fractions revealed differences in the pattern of their distribution. It is further demonstrated that in the L6 rat myoblast system, ribosomes and mRNA (or mRNP) are not associated with the microfilaments, unlike in other systems studied.

Translational control and the cytoskeleton in Physarum polycephalum

Translationally active plasmodia of the syncytial slime mold Physarum polycephalum develop into translationally dormant sclerotia during starvation. Although functional mRNA and ribosomes exist in sclerotia, protein synthesis is suppressed at the level of initiation. To test the possibility that alterations in the cytoskeleton may limit protein synthesis, we have examined the distribution of polysomes and actin mRNA in the cytoskeletal (CSK) and soluble (SOL) fractions of Triton X-100-extracted plasmodia and sclerotia. Most of the polysomes and actin mRNA were located in the CSK of plasmodia, while most of the ribosomes and actin mRNA were located in the SOL of sclerotia. The results suggest that ribosomes and mRNA shift from the CSK to the SOL as protein synthesis is suppressed during starvation. Plasmodia and sclerotia can be induced to accumulate excess polysomes by treatment with low levels of the elongation inhibitor cycloheximide. Treatment of plasmodia with cycloheximide caused excess polysomes to accumulate in the SOL, suggesting that the CSK contains a limited capacity for binding translational components and that the association of polysomes with the cytoskeleton is not required for protein synthesis. Treatment of sclerotia with cycloheximide, however, caused polysomes and actin mRNA to accumulate in the CSK, suggesting that the sclerotial cytoskeleton, although depleted in ribosomes and mRNA, is capable of binding translational components. It is concluded that alterations in the sclerotial cytoskeleton are not involved in translational control.

Association of messenger RNA with the cytoskeletal framework in rat L6 myogenic cells

The distribution of mRNA between the detergent-soluble and insoluble (cytoskeleton) fractions in rat L6 myoblast and myotube cells was examined. Approximately 85% of cytoplasmic mRNA in both myoblasts and myotubes was found associated with the cytoskeletal framework. The cytoskeleton-bound mRNA was present as polysomes. In contrast, the mRNA of the detergent-soluble fraction was not associated with ribosomes and was thus considered to be the repressed population. The association of mRNA with the cytoskeletal framework was not affected by treatments leading to dissociation of polysomes. Differential distribution of mRNA between the soluble and cytoskeleton-bound fractions was analyzed by in vitro translation. The mRNAs coding for polypeptides of molecular masses 40 kDa and 60 kDa were preferentially enriched in the soluble fraction. The nature of binding between mRNA and the cytoskeletal framework was examined following in vivo cross-linking of RNA and protein by irradiating muscle cells with ultraviolet light. It was observed that this treatment covalently linked RNA and the neighbouring protein moieties without any detectable damage to the cytoskeletal framework, as measured by the distribution of RNAs and proteins between the cytoskeleton and soluble fractions. Analysis of the polypeptide moieties cross-linked to the mRNA have shown that a large number of polypeptides of molecular masses between 15-220 kDa were associated with both cytoskeleton-bound and soluble mRNAs. The polypeptide moieties of these mRNA-protein complexes were not only similar in the cytoskeleton and soluble mRNA-protein complexes but also were similar between myoblasts and myotubes. However, one polypeptide of 165 kDa was preferentially associated with the cytoskeleton-bound mRNA-protein complexes. Interestingly this 165-kDa polypeptide was also preferentially enriched in the total proteins from the cytoskeleton fraction. This result suggests a possible role of the 165-kDa polypeptide in association between mRNA and the cytoskeletal framework. To examine the mechanism of interaction between mRNA and the cytoskeletal framework we have reported here a ghost monolayer transcription system from myotubes. This transcription system was able to synthesize rRNA and mRNA. The mRNA transcribed in vitro was preferentially associated with the cytoskeleton structure present in the ghost monolayer system.

Cytochalasin releases mRNA from the cytoskeletal framework and inhibits protein synthesis

Cytochalasin D was shown to be a reversible inhibitor of protein synthesis in HeLa cells. The inhibition was detectable at drug levels typically used to perturb cell structure and increased in a dose-dependent manner. The drug also released mRNA from the cytoskeletal framework in direct proportion to the inhibition of protein synthesis. The released mRNA was unaltered in its translatability as measured in vitro but was no longer translated in the cytochalasin-treated HeLa cells. The residual protein synthesis occurred on polyribosomes that were reduced in amount but displayed a normal sedimentation distribution. The results support the hypothesis that mRNA binding to the cytoskeletal framework is necessary although not sufficient for translation. Analysis of the cytoskeletal framework, which binds the polyribosomes, revealed no alterations in composition or amount of protein as a result of treatment with cytochalasin D. Electron microscopy with embedment-free sections shows the framework in great detail. The micrographs revealed the profound reorganization effected by the drug but did not indicate substantial disaggregation of the cytoskeletal elements.

Polyamine depletion results in impairment of polyribosome formation and protein synthesis before onset of DNA synthesis in mitogen-activated human lymphocytes

Lymphocytes, exposed to mitogens in culture, show enhanced protein and RNA synthesis before the onset of DNA synthesis. Inhibition by DL-alpha-difluoromethylornithine of polyamine synthesis in phytohaemagglutinin-activated human lymphocytes resulted in a suppression of protein synthesis, which was evident before the initiation of DNA synthesis. The mitogen-induced increase in the incorporation of [3H]thymidine into DNA was subsequently inhibited in parallel with the activity of thymidine kinase in the polyamine-depleted cells. Ultraviolet absorbance measurement of the ribosomes after sucrose gradient centrifugation revealed a suppression of polyribosome formation that coincided with the decrease in the rate of protein synthesis. The disturbance in the polysomal profiles did not appear to be due to a shortage of mRNA, since the synthesis of poly(A)-rich mRNA was reduced less than that of rRNA after inhibiting polyamine synthesis. Entry of both the pre-existing and newly synthesized ribosomal subunits into polysomal structures was found to be impaired. These results thus suggest an important role for polyamines in the initiation step of protein synthesis.

Association of rapidly-labelled RNAs with actin in nuclear matrix from mouse L5178Y cells

More than 90% of rapidly-labelled nuclear RNA was associated with a nuclear matrix prepared from mouse leukemia L5178Y cells. The binding was not affected with up to 4 M NaCl; however, these RNAs were released from the nuclear matrix by treatment with a low ionic strength buffer (5 mM Tris-HCl buffer, pH 7.5, containing 1 mM ATP, 1 mM dithiothreitol, 0.2 mM ethylenediaminetetraacetic acid (EDTA) and 0.4 mM calcium chloride), without destruction of the sphere of the nuclear matrix. Actin filaments in the nuclear matrix were depolymerized with this buffer accompanied with rapidly-labelled RNAs. When the depolymerization was inhibited by slight modifications of the low ionic strength buffer (replacement of ATP by the same concentration of GTP; replacement of calcium ion by the same concentration of magnesium ion; addition of 20 micrograms/ml of phalloidine, which is a specific inhibitor of actin depolymerization), the release of rapidly-labelled RNAs from the nuclear matrix was also inhibited. The complex containing rapidly-labelled RNAs and matrix proteins was solubilized by a sonication from the nuclear matrix, and subjected to cesium chloride equilibrium centrifugation. Rapidly-labelled RNAs were concentrated on the bottom of the gradient accompanied with a small number of proteins (68K, 60K, 43K and 40K). The 43K protein was identified as actin by immunoblotting. By RNase digestion before equilibrium centrifugation, actin in the bottom fractions disappeared. These results suggest that rapidly-labelled RNAs anchor on the actin filaments in the nuclear matrix.

Disruption of the three cytoskeletal networks in mammalian cells does not affect transcription, translation, or protein translocation changes induced by heat shock

Mammalian cells show a complex series of transcriptional and translational switching events in response to heat shock treatment which ultimately lead to the production and accumulation of a small number of proteins, the so-called heat shock (or stress) proteins. We investigated the heat shock response in both qualitative and quantitative ways in cells that were pretreated with drugs that specifically disrupt one or more of the three major cytoskeletal networks. (These drugs alone, cytochalasin E and colcemid, do not result in induction of the heat shock response.) Our results indicated that disruption of the actin microfilaments, the vimentin-containing intermediate filaments, or the microtubules in living cells does not hinder the ability of the cell to undergo an apparently normal heat shock response. Even when all three networks were simultaneously disrupted (resulting in a loose, baglike appearance of the cells), the cells still underwent a complete heat shock response as assayed by the appearance of the heat shock proteins. In addition, the major induced 72-kilodalton heat shock protein was efficiently translocated from the cytoplasm into its proper location in the nucleus and nucleolus irrespective of the condition of the three cytoskeletal elements.

Disruption of the three cytoskeletal networks in mammalian cells does not affect transcription, translation, or protein translocation changes induced by heat shock

Mammalian cells show a complex series of transcriptional and translational switching events in response to heat shock treatment which ultimately lead to the production and accumulation of a small number of proteins, the so-called heat shock (or stress) proteins. We investigated the heat shock response in both qualitative and quantitative ways in cells that were pretreated with drugs that specifically disrupt one or more of the three major cytoskeletal networks. (These drugs alone, cytochalasin E and colcemid, do not result in induction of the heat shock response.) Our results indicated that disruption of the actin microfilaments, the vimentin-containing intermediate filaments, or the microtubules in living cells does not hinder the ability of the cell to undergo an apparently normal heat shock response. Even when all three networks were simultaneously disrupted (resulting in a loose, baglike appearance of the cells), the cells still underwent a complete heat shock response as assayed by the appearance of the heat shock proteins. In addition, the major induced 72-kilodalton heat shock protein was efficiently translocated from the cytoplasm into its proper location in the nucleus and nucleolus irrespective of the condition of the three cytoskeletal elements.

Effect of viral infection on host protein synthesis and mRNA association with the cytoplasmic cytoskeletal structure

We studied the association of several eucaryotic viral and cellular mRNAs with cytoskeletal fractions derived from normal and virus-infected cells. We found that all mRNAs appear to associate with the cytoskeletal structure during protein synthesis, irrespective of their 5' and 3' terminal structures: e.g., poliovirus that lacks a 5' cap structure or reovirus and histone mRNAs that lack a 3' poly A tail associated with the cytoskeletal framework to the same extent as capped, polyadenylated actin mRNA. Cellular (actin) and viral (vesicular stomatitis virus and reovirus) mRNAs were released from the cytoskeletal framework and their translation was inhibited when cells were infected with poliovirus. In contrast, actin mRNA was not released from the cytoskeleton during vesicular stomatitis virus infection although actin synthesis was inhibited. In addition, several other conditions under which protein synthesis is inhibited did not result in the release of mRNAs from the cytoskeletal framework. We conclude that the association of mRNA with the cytoskeletal framework is required but is not sufficient for protein synthesis in eucaryotes. Furthermore, the shut-off of host protein synthesis during poliovirus infection and not vesicular stomatitis virus infection occurs by a unique mechanism that leads to the release of host mRNAs from the cytoskeleton.

Association of an aminoacyl-tRNA synthetase complex and of phenylalanyl-tRNA synthetase with the cytoskeletal framework fraction from mammalian cells

The intracellular distribution of several mammalian aminoacyl-tRNA synthetases was investigated by biochemical and immunocytological approaches. The fraction of amino-acyl-tRNA synthetases bound to the detergent-insoluble cytoskeletal framework obtained after extraction of NRK cells by 0.1% Triton X-100 was estimated, by activity measurements, to about 80% for phenylalanyl-tRNA synthetase and 40% for the high-molecular-weight (HMW) complex containing the seven aminoacyl-tRNA synthetases specific for glutamic acid, isoleucine, leucine, methionine, glutamine, lysine, and arginine. This association was shown to be salt-dependent. The subcellular localization of these enzymes was examined using an immunocytological approach. When cultured cells were fixed with paraformaldehyde and then permeabilized with Triton X-100, a fairly uniform cytoplasmic labelling was observed with antibodies directed to the aminoacyl-tRNA synthetase complex or to phenylalanyl-tRNA synthetase. By contrast, when cells were extracted with 0.1% Triton X-100 prior to fixation with paraformaldehyde, the staining patterns obtained with antibodies to aminoacyl-tRNA synthetases were very similar to that obtained with antibodies to rough endoplasmic reticulum, as assessed by single or double indirect immunofluorescence microscopy. These results suggest that free and bound forms of these aminoacyl-tRNA synthetases may coexist within the cell. In addition to cytoplasmic labelling, antibodies directed to phenylalanyl-tRNA synthetase stained the nucleus of rapidly growing cells. The possible significance of this finding is discussed.

Precursors of the nerve growth factor gamma subunit and renin bind to microtubules

Translation products of a reticulocyte lysate reaction, programmed with poly(A)-rich RNAs from the male mouse submaxillary gland, were subjected to affinity chromatography on a tubulin-Sepharose column. Analysis of the bound proteins in sodium dodecylsulfate/polyacrylamide gels revealed two polypeptides of Mr 27 000 and 45 000, that were shown to bind to tubulin in a specific manner. These polypeptides were absent from the translation products coded by poly(A)-rich RNAs from the female mouse. They were eluted from the tubulin-Sepharose resin under conditions similar to those employed for the dissociation of immune complexes. The Mr-27 000 and Mr-45 000 proteins were identified by immunoprecipitation with specific antisera as the precursors of the gamma subunit of the nerve growth factor (NGF) and renin respectively. These two precursors as well as a third, unidentified polypeptide of Mr 38 000, probably unrelated to the beta subunit of NGF, bound also to microtubules. The mature form of renin, purified from the submaximillary gland, also displayed an affinity for the microtubules. In contrast, the mature form of the gamma subunit of NGF did not bind to the microtubules. The possible involvement of the microtubules (tubulin) in the biosynthesis of these two secretory proteins is discussed.

Spatial distribution of messenger RNA in the cytoskeletal framework of ascidian eggs

Maternal poly(A)+RNA, histone mRNA, and actin mRNA exhibit unique spatial distributions in the different ooplasmic regions of ascidian eggs. These RNAs also appear to migrate with their respective ooplasms during the episode of extensive cytoplasmic rearrangement that occurs after fertilization, suggesting they are associated with a structural framework. The role of the cytoskeletal framework (CF) in determining the spatial distribution of maternal mRNA was tested by subjecting Triton X-100 extracted (Styela plicata) eggs and early embryos to in situ hybridization with poly(U) and cloned DNA probes. Grain counts indicated that substantial proportions of the egg poly(A)+RNA, histone mRNA, and actin mRNA were present in the CF and that there was no alteration in the extent of mRNA-CF interactions during the period between fertilization and the two-cell stage. Analysis of grain distributions indicated that poly(A)+RNA, histone mRNA, and actin mRNA were concentrated in the same regions of detergent-extracted eggs as they are in intact eggs. The proportions and spatial distribution of these RNAs in the CF were not affected when the actin cytoskeleton was destabilized by cytochalasin B or DNase I. The data suggest that maternal mRNA is associated with the CF, that this association is responsible for mRNA rearrangement during ooplasmic segregation, and that mRNA-CF interactions are not dependent on the integrity of the actin cytoskeleton.

Translational initiation factor and ribosome association with the cytoskeletal framework fraction from HeLa cells

The association of mRNA and ribosomes with the cytoskeleton of eucaryotic cells may be important for protein synthesis and its regulation. HeLa cells were gently lysed with detergent, and soluble and cytoskeletal framework subfractions were prepared by centrifugation. We analyzed these fractions for ribosomes and confirmed earlier findings that polysomes are preferentially associated with the cytoskeletal fraction. The levels of initiation factors eIF-2, eIF-3, eIF-4A, and eIF-4B were quantitated by immunoblotting; all are enriched in the cytoskeletal fraction relative to the soluble fraction. Heat shock, fluoride, pactamycin , and cytochalasin caused the release of both ribosomes and initiation factors into the soluble fraction. However, treatment of the cytoskeletal fraction with EDTA or low levels of ribonuclease resulted in polysome degradation but no release. Therefore initiation factor association with the cytoskeletal framework correlates with the presence of ribosomes, whereas ribosome association does not require intact mRNA.

Topography of the total protein population from cultured cells upon fractionation by chemical extractions

Chemical extractions are proposed as a major tool for a fractionation of cellular proteins. As a model system, proteins from cultured hamster lens cells have been divided by independent extractions into seven subcellular fractions, corresponding to water-soluble proteins and the proteins from membranes, microfilaments (and other deoxycholate-soluble proteins), intermediate filaments, microtubules, polysomes and nuclei respectively. The latter two fractions have been subfractionated yielding ribosomal proteins, the elongation and initiation factors of the protein-synthesis machinery, chromatin proteins and non-chromatin proteins. The protein compositions of the fractions have been analyzed by one-dimensional and two-dimensional gel electrophoresis. This resulted in an almost complete topography of the proteins detected on two-dimensional gels of total-cell lysates. Comparison of two-dimensional patterns of proteins from the total-cell lysate and proteins from hamster erythrocytes or from liver, muscle or brain tissue showed that the different cell types have only few proteins in common. Two proteins are common to all of these cell types, namely actin and a 68-kDa protein. The latter protein was, like actin, vimentin and the tubulin subunits, also present in most cell fractions. Evidence is presented that this protein is identical to a 68-kDa heat-shock protein.

Denaturation of RNA and DNA in situ induced by acridine orange

The products of interaction of acridine orange (AO) with single-stranded (ss) nucleic acids are precipitates which exhibit red luminescence. Titration of rRNA or thymus DNA with AO results in formation of such products suggesting that the dye, per se, denatures double-stranded (ds) sections of these biopolymers. This transition, measured as the increase of red luminescence, a concomitant decrease of green fluorescence, and followed by an increase of light scatter of the AO-nucleic acid complexes, is cooperative and at 0.15 N NaCl occurs at 4-20 and 10-50 microM range of AO concentration for rRNA and DNA, respectively. The changes in stainability of nucleic acids in situ, in permealized cells, occur at higher AO concentration. Thus, the transition of RNA in situ is biphasic and seen at 20-120 microM AO. In the presence of EDTA, however, the change is monophasic and shifted to the 10-30 microM range of AO concentration. The change in stainability of DNA also shows two phases: one at 30-60 microM and another at 70-120 microM of AO. Extraction of basic proteins with 0.08 N HCl shifts the transition of DNA to the 30-60 microM AO concentration and makes it monophasic. The observed differences in denaturability of RNA vs DNA explain the specificity of AO in differential staining of these bipolymers in histochemical reactions. In living cells the products of interaction of AO with nucleic acids are detected by electron microscopy. In the cytoplasm of interphase cells the formation of dense precipitates within ribosomes and polysomes, simultaneous with a specific retraction of ribosome-polysome complexes from the periphery of the cell to the nucleus is evident. The latter suggests higher order organization of these particles involving their association with each other or with the nucleus via polyanionic macromolecules which collapse upon binding with AO. The DNA in heterochromatin is more sensitive to AO-induced denaturation, as evidenced by the fact that the dense complexes are formed preferentially in the regions of condensed chromatin of the interphase nucleus, or in metaphase chromosomes.

Polyribosomes associated with microfilaments in cultured lens cells

Epithelial hamster lens cells, transformed by SV40 can be grown in suspension culture. Triton X-100 extraction of these cells grown under conditions when ribosome run off is blocked releases about 40% of the total amount of polyribosomes, designated as free- and loosely-bound polyribosomes. The Triton ghosts retain the remaining polysomal population which can be released by a combined treatment with deoxycholate and Nonidet P 40. Electron microscopic examination of the ghosts reveals microfilament-associated ribosome clusters next to a fraction of polysomes still attached to membranes. Preincubation of the cells with cytochalasin D prior to polyribosome isolation enables us to discriminate between these two latter polysome populations. The experiments indicate that about 25% of the polyribosomes are attached to microfilaments, while the remaining 35% are tightly bound to the membranes of the endoplasmic reticulum. When the different polyribosome classes were translated in a reticulocyte lysate, no significant differences could be observed in the patterns of the newly synthesized polypeptides. In all cases actin was one of the major products synthesized de novo.

Intracellular localization of adenovirus type 5 tumor antigens in productively infected cells

The intracellular localization of tumor antigens of human adenovirus type 5 (Ad5) during lytic infection of KB cells has been studied. The cells were pulse labeled with [35S]methionine early after infection and early proteins of 58,000 D (58K), 44K, 19K, 18.5K, and 14K detectable by immunoprecipitation with hamster antitumor serum were assayed for association with cytoplasm, nucleoplasm, chromatin, cytosol, cytoskeleton, and membranes. The 44,000 D (44K) tumor antigen encoded in early region 1A (E1A: 0-4.4%) was recovered in approximately equal amounts from cytoplasmic and nucleoplasmic fractions of pulse-labeled cells and within the cytoplasmic compartment was found in the cytosol as well as associated with the cytoskeleton. The E1B-58K (E1B: 4.5-11.2%) antigen was also found to be associated with the cytoplasmic and nucleoplasmic fractions in approximately equal amounts but unlike the E1A-44K showed no affinity for cytoskeletons. Pulse-chase and immunofluorescence experiments suggested the 58K antigen accumulated in the nucleus late in infection. The E1B-19K antigen was found almost exclusively associated with the membrane fraction of infected KB cells and was resolved in polyacrylamide gels into two related species of 18.5K and 19K. Immunofluorescence studies on the E1B 18.5-19K doublet suggested that within a population of infected HeLa cells a small minority seemed to be expressing copious amounts of stainable antigen. Cell fractionation and immunofluorescence studies showed that the E4-14K antigen was a nuclear protein and the only antigen in this study which showed a significant association with a nuclear subfraction composed almost entirely of histones. The implications of these findings for the roles of the Ad5 tumor antigens in lytic infection and transformation are discussed.

Synthesis and post-translational assembly of intermediate filaments in avian erythroid cells: vimentin assembly limits the rate of synemin assembly

The assembly of vimentin intermediate filaments and the high molecular weight filament crosslinking protein, synemin, was studied in erythroid cells from 10-day chicken embryos. Pulse labeling studies show that newly synthesized vimentin is present both in a Triton X-100-insoluble form and in a soluble form. The incorporation of labeled vimentin into the insoluble fraction increases linearly with time, while the soluble pool of labeled vimentin saturates quickly. In contrast, synemin accumulates rapidly in the Triton X-100-soluble fraction and begins to accumulate in the insoluble fraction only after a considerable lag of time. Pulse-chase studies reveal that the detergent-soluble pools of both vimentin and synemin contain precursors for their post-translational assembly into detergent-insoluble filaments and that the half-life of soluble synemin is about twice as long as that of soluble vimentin. Immunoprecipitation of solubilized filaments with synemin antiserum precipitates vimentin with synemin. On the other hand, soluble vimentin does not coimmunoprecipitate with soluble synemin. These results suggest that, in the assembly of vimentin and synemin into intermediate filaments, vimentin filament elongation generates synemin binding sites, and thus the rate of vimentin filament elongation limits the rate of synemin assembly.

Synthesis of spectrin in avian erythroid cells: association of nascent polypeptide chains with the cytoskeleton

The site of synthesis of spectrin was investigated in erythroid cells from 10-day chicken embryos. After various periods of [35S]methionine incorporation the cells were lysed in a Triton X-100 (TX-100)-containing buffer and were separated into a TX-100-soluble and -insoluble (cytoskeletal) fraction. Analysis of these two fractions by two-dimensional gel electrophoresis after a short pulse-labeling period reveals that alpha-spectrin nascent polypeptides are present predominantly in the TX-100-insoluble fraction. These polypeptides can be immunoprecipitated with alpha-spectrin antisera and the [35S]methionine incorporated into them during a short pulse can be chased into mature alpha-spectrin molecules. The alpha-spectrin nascent polypeptide chains are released quantitatively from the TX-100 cytoskeleton by treatment of lysed cells with puromycin, suggesting that they themselves are not associated with the cytoskeleton. A small fraction of the newly synthesized mature alpha-spectrin molecules is rapidly incorporated into the cytoskeleton, as shown by the fact that they are not released by the puromycin treatment; the rest are recovered in the soluble fraction. These results suggest that alpha-spectrin is synthesized in association with the cytoskeleton during chicken erythropoiesis and assembles onto the membrane-cytoskeleton posttranslationally.

Association of methionyl-tRNA synthetase with detergent-insoluble components of the rough endoplasmic reticulum

Using fluorescent antibody staining, we have established the association of methionyl-tRNA synthetase with the endoplasmic reticulum in PtK2 cells. After Triton X-100 extraction, 70% of the recovered aminoacyl-tRNA synthetase activity was found in the detergent-insoluble fraction. This fraction of the enzyme remained localized with insoluble endoplasmic reticulum antigens and with ribosomes, which were stained with acridine orange. By both fluorescence microscopy and electron microscopy the organization of the detergent-insoluble residue was found to depend on the composition of the extracting solution. After extraction with a microtubule-stabilizing buffer containing EGTA, Triton X-100, and polyethylene glycol (Osburn, M., and K. Weber, 1977, Cell, 12:561-571) the ribosomes were aggregated in large clusters with remnants of membranes. After extraction with a buffer containing Triton X-100, sucrose, and CaCl2 (Fulton, A. B., K. M. Wang, and S. Penman, 1980, Cell, 20:849-857), the ribosomes were in small clusters and there were few morphologically recognizable membranes. In both cases the methionyl-tRNA synthetase and some endoplasmic reticulum antigens retained approximately their normal distribution in the cell. Double fluorochrome staining showed no morphological association of methionyl-tRNA synthetase with the microtubule, actin, or cytokeratin fiber systems of PtK2 cells. These observations demonstrate that detergent-insoluble cellular components, sometimes referred to as "cytoskeletal" preparations, contain significant amounts of nonfilamentous material including ribosomes, and membrane residue. Caution is required in speculating about intermolecular associations in such a complex cell fraction.

A yellow crescent cytoskeletal domain in ascidian eggs and its role in early development

In this investigation, Triton X-100 extraction was utilized to examine the cytoskeleton of ascidian eggs and embryos. The cytoskeleton contained little carbohydrate or lipid and only about 20-25% of the total cellular protein and RNA. It was enriched in polypeptides of molecular weight (Mr)54, 48, and 43 x 10(3) Mr polypeptide was identified as actin based on its Mr, isoelectric point, and affinity for DNase I. Electron microscopy of the detergent-extracted eggs showed that they contained cytoskeletal domains corresponding to colored cytoplasmic regions of specific morphogenetic fate in the living egg. A yellow crescent cytoskeletal domain in the myoplasm was examined and shown to consist of a plasma membrane lamina (PML) and a deeper lattice of filaments which appeared to connect the yellow crescent pigment granules to the PML. The PML probably consists of integral membrane proteins stabilized by an underlying network of actin filaments since NBD-phallacidin stained this area of the egg cortex and the PML was extracted from the cytoskeleton by DNase I treatment. The yellow crescent cytoskeletal domain was found throughout the cortex of the unfertilized egg. During ooplasmic segregation it progressively receded into the vegetal hemisphere and was subsequently partitioned to the presumptive muscle and mesenchyme cells of the 32-cell embryo. It is suggested that contraction of the actin network in the yellow crescent cytoskeletal domain is the motive force for ooplasmic segregation. This structure may also serve as a framework for the positioning of morphogenetic determinants involved in muscle cell development.

The cytoskeletal framework of sea urchin eggs and embryos: developmental changes in the association of messenger RNA

Extraction of sea urchin eggs and embryos with Triton X-100 generated a cytoskeletal framework (CSK) composed of a cortical filamentous network and an internal system of filaments associated with ribosomes. The CSK contained only 10-20% of the cellular protein, RNA, and lipid. A specific subset of proteins was enriched in the CSK. Several lines of evidence suggest that mRNA is a component of the CSK of both eggs and embryos. First, the CSK contained poly(A) sequences which hybridized with [3H]poly(U). Second, the CSK contained polyribosomes. Finally, RNA extracted from the CSK showed translational activity in an in vitro system. The nonhistone messages present in the CSK were qualitatively similar to those solubilized by detergent, as determined by separation on polyacrylamide gels of the products of in vitro translation. In the unfertilized egg, most mRNA was present as nonpolyribosomal messenger ribonucleoprotein complexes which, along with monoribosomes, were efficiently extracted by Triton X-100. The converse was found in blastulae, as most of the mRNA was present as polyribosomes associated with the CSK, although monoribosomes were still efficiently extracted by detergent. These results indicate a correlation between the activation of protein synthesis in eggs and the association of polyribosomes with the CSK.

Many cytoskeletal proteins associate with the hela cytoskeleton during translation in vitro

Observations that cytoskeletal proteins assemble in vivo close to the time and site of synthesis have been confirmed and extended by an in vitro translation system. HeLa cytoskeletons prepared with Triton in a translation-extraction buffer without reticulocyte or wheat germ lysate efficiently incorporate 35S-methionine into polypeptides, and are stable during this translation. Cytoskeletal proteins translated in this way associate with the HeLa cytoskeleton independent of the concentration of soluble proteins. These associations are puromycin-resistant before the proteins are complete; the protein associations made in vitro show only minor differences from those made in vivo. The protein associations are not simply a consequence of protein solubility in the buffers used, as the associations require initiation in vivo. These results indicate that many cytoskeletal proteins associate with the cytoskeleton during translation.