Proc. Natl. Acad. Sci. USA Vol. 92, pp. 11534-11538, December 1995 Cell Biology

Regulation of microfilament organization and anchorage- independent growth by tropomyosin 1 JEFF BOYD*, JOHN I. RISINGER, ROGER W. WISEMAN, B. ALEX MERRICK, JAMES K. SELKIRK, AND J. CARL BARRETTt Laboratory of Molecular Carcinogenesis, Environmental Carcinogenesis Program, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, NC Communicated by Martin Rodbell, National Institute of Environmental Health Sciences, Research Triangle Park NC, July 12 1995 (received for review March 6, 1995)

ABSTRACT Variants of chemically immortalized Syrian To explore the molecular basis of anchorage-independent hamster embryo cells that had either retained (supB+) or lost growth and the potential relationship of the cytoskeleton to (supB-) the ability to suppress tumorigenicity when hybrid- this phenomenon, we utilized a series of Syrian hamster ized with a fibrosarcoma cell line were subcloned. Both supB embryo (SHE) cell lines in which loss of a tumor suppressor cell types are nontumorigenic; however, the supB- but not function correlates with the acquisition of conditional anchor- supB+ cells exhibit conditional anchorage-independent age-independent growth potential (21-25). We describe the growth. Alterations of actin microfilament organization were use of these closely related clonal variants to examine the observed in supB- but not supB+ cells that corresponded to potential relationships among microfilament organization, tro- a significant reduction of the actin-binding protein tropomy- pomyosin expression, and anchorage-independent growth.t osin 1 (TM-1) in supB- cells. To examine the possibility of a direct relationship between TM-1 expression and the supB- MATERIALS AND METHODS phenotype, supB+ cells were transfected with an expression vector containing the TM-1 cDNA in an antisense orientation. Cell Lines and Culture. The establishment, characterization, The antisense-induced reduction of TM-1 levels in supB+ and culture of the SHE cell lines used in this study have been clones caused a microfilament reorganization and conferred described in detail (23, 24). anchorage-independent growth potential that were indistin- Photomicroscopy. To visualize actin microfilaments, sub- guishable from those characteristic of supB- cells. These data confluent cells were fixed, permeabilized, and stained with the provide direct evidence that TM-1 regulates both microfila- F-actin-specific rhodamine-phalloidin conjugate as described ment organization and anchorage-independent growth and by the supplier (Molecular Probes). Photomicrographs were suggest that microfilament alterations are sufficient for an- taken with a Leitz Orthomat microscope equipped with an epifluorescence light source, a Leitz N2 filter unit, and an oil chorage-independent growth. immersion objective at 630X and with Kodak Tri-X film (ASA 400). Cytoskeletal alterations are frequently observed in tumor cells Two-Dimensional PAGE. For comparison of total proteins (1), but the significance and molecular basis of these changes among the various supB clones, metabolic labeling of total are unclear. The actin microfilament component of the cy- cellular proteins and quantitative two-dimensional PAGE toskeleton appears to regulate cell shape in many contexts were performed as described (26, 27). [14C]Methylated pro- (2-4), and growth control is one of several biological processes teins (14.3-200 kDa, Amersham) were used as molecular mass that appears tightly linked to cell shape in normal cells (5-8). markers on slab gels. The proliferating cell nuclear antigen One common change in neoplastically transformed cells is the protein was utilized as a quantitative control because of its loss of anchorage- or shape-dependent controls on growth (9). electrophoretic migration in the TM region and its apparently It has been hypothesized that the cytoskeletal perturbations invariant expression in these cells under the conditions de- observed in tumor cells, particularly of the microfilament scribed. The identity of TMs on two-dimensional gels was component, may mediate the uncoupling of cell shape and confirmed by two independent methods. High molecular growth control. weight TM isoforms were immunoprecipitated from radiola- The regulation of microfilament organization is complex, beled SHE cell extracts by using rabbit polyclonal antiserum and numerous actin-binding proteins are involved in this against smooth muscle TM from chicken gizzard (T651, Sigma) process (10). Synthesis of several members of the tropomyosin as described (14). The immunoprecipitated proteins were (TM) family of actin-binding proteins (11), particularly the electrophoresed concomitantly with experimental samples to isoforms of apparently higher molecular weight, is frequently confirm TM localization. In addition, second-dimension gels decreased in association with neoplastic transformation by containing unlabeled SHE cell proteins were subjected to various chemical and viral agents (12-16), as well as in human Western blotting procedures (28) with the mouse anti-chicken carcinoma cells (17). In addition, transfection of TM-1 into gizzard TM monoclonal antibody TM311 (T2780, Sigma). viral oncogene-transformed rodent cells suppresses the tumor- cDNA Probes and Cloning. cDNA probes used in the cloning igenic phenotype (18). These observations suggest that down- of homologous Syrian hamster cDNAs and in Northern blot regulation of TM synthesis is related to the microfilament analysis included human ,B-actin clone pHFO3A-3'UT-HF (29), aberrations observed in diverse types of tumor cells (19). human y-actin clone pHF-yA-3'UT-Fnu (29), rat TM-1 clone Consistent with this hypothesis is a report demonstrating that pSM10 (30), and rat TM-4 clone pREF-102 (31). All cDNA disruption of the single TM gene in the yeast Saccharomyces cerevisiae leads to disappearance of cytoplasmic actin cables Abbreviations: SHE, Syrian hamster embryo; TM, tropomyosin. and perturbations in cellular growth rate (20). *Present address: Department of Obstetrics and Gynecology, Univer- sity of Pennsylvania Medical Center, 415 Curie Boulevard, Philadel- phia, PA 19104. The publication costs of this article were defrayed in part by page charge ITo whom reprint requests should be addressed. payment. This article must therefore be hereby marked "advertisement" in *The sequence reported in this paper has been deposited in the accordance with 18 U.S.C. §1734 solely to indicate this fact. GenBank data base (accession no. U29167). 11534 Downloaded by guest on September 28, 2021 Cell Biology: Boyd et al. Proc. Natl. Acad. Sci. USA 92 (1995) 11535

FIG. 1. Fluorescence photomicrographs depicting actin microfilament organization in supB+ and supB- cell lines. (A) 10WsupB+; (B) 1OWsupB-; (C) DES4supB+; (D) DES4supB-. (Photomicrographs taken at a magnification of 630X.) inserts were radiolabeled with [a-32P]dCTP (3000 Ci/mmol; 1 method (35). Following selection in G418 (800 ,ug/ml), resis- Ci = 37 GBq; Amersham) to a specific activity of > 109 cpm/,ug tant cells were further analyzed as either mass cultures or two by using the Prime-It random priming kit (Stratagene). Total subclones each of sense and antisense transfectants. Anchor- cellular RNA was prepared from early passage, normal SHE age-independent growth potential in soft agar was assessed as cells by the guanidinium isothiocyanate/acid/phenol method described (24). All antisense- and sense-expression experi- (32). Poly(A)+ RNA was prepared from total RNA by two ments were carried out in the presence of 1 ,uM dexametha- cycles of selection on an oligo(dT)-cellulose column, and sone to elicit maximal expression of the cloned cDNA se- cDNA was synthesized using oligo(dT) primers, Moloney quence. murine leukemia virus reverse transcriptase, and a Stratagene ZAP-cDNA synthesis kit (Stratagene). The AZAP library was subcloned into pBluescript SK(-), and hamster cDNAs were RESULTS identified by hybridization with the rat TM-1 cDNA probe at low stringency. Candidate cDNA clones were sequenced in Cytoskeletal Organization. Significant differences were ob- both directions with Sequenase 2.0 (United States Biochem- served in actin microfilament organization between supB+ ical). clones and supB- clones (Fig. 1). Well-defined bundles of Northern Blotting. Total RNA was isolated from cultured microfilaments (stress fibers) generally overlapping and ex- cells as above, and Northern blotting was performed as tending entirely across the cells were observed in both described (33). Blots were rehybridized with a f-actin cDNA 10WsupB+ and DES4supB+ cell lines. Conversely, in the to control for RNA integrity and quantity. Quantitation of 1OWsupB- and DES4supB- cells, stress fibers were reduced in TM-1 mRNA levels in transfected clones was performed by number and integrity; many cells lacked them entirely, while in scanning densitometry of autoradiography films with a USB others they were shorter and thinner than in the supB+ cells. SciScan 5000 system (United States Biochemical). The observed differences in microfilament organization were TM-1 Antisense Expression Experiments. A partial hamster not the result of alterations in soluble vs. polymerized actin TM-1 cDNA, beginning 23 bp after the initiation codon and ratios, biosynthesis rates, protein turnover rates, or ratios of extending to the poly(A) terminus, was subcloned into the cytoplasmic ,3- vs. y-actin isoforms (data not shown). Immun- pMAMneo expression vector (Clontech) (34). DESsupB+ ofluorescent staining of the cells with antibodies specific for cells were transfected with vectors containing this TM-1 cDNA vimentin intermediate filaments and microtubules indicated in either the sense or antisense orientation, as determined by no discernible differences among the cell lines (data not sequence analysis by the calcium phosphate precipitation shown). Downloaded by guest on September 28, 2021 11536 Cell Biology: Boyd et al. Proc. Natl. Acad. Sci. USA 92 (1995) densitometric analyses of autoradiograms from metabolically radiolabeled proteins and photographic films of silver-stained proteins, and the results agreed closely (Table 1). Five TM isoforms were detected in all cell lines examined, but profound A~~~~~3 downregulation of the high molecular weight isoforms TM-1 and TM-2 was observed in both 1OW- and DES4supB- clones. 1 ' x The high molecular weight isoform TM-3 was also downregu- lated in 1OWsupB- cells, but both DES4 clones expressed *,. ...f~~~~46 similarly low levels of this isoform. No significant differences were observed between any supB+ and supB- clones in the levels of lower molecular weight isoforms TM-4 and TM-5. Hamster TM-1 cDNA Cloning and Expression Analysis. A 951-nt Syrian hamster TM-1 cDNA sequence, representing the entire coding region of 284 amino acids, the entire 3' untrans- supB+ supB- lated region, and 28 bases of the 5' untranslated region, was generated. The hamster cDNA exhibited 95% and 92% sim- ilarity to the published sequences of rat (30) and human (36) B + = q 0 TM-1, respectively. The predicted amino acid sequence of the eS hamster cDNA showed 99% similarity to both the rat and the ~ ~~~~A human predicted proteins. Under high-stringency conditions, the hamster TM-1 cDNA probe detected a single message of -0.9-kb by Northern blotting. The rat and hamster TM-1 probes also gave identical hybridization patterns on Southern blots containing hamster DNA. Northern analysis indicated that steady-state levels of TM-1 mRNA were similar in supB+ clones but were at least 95% lower in supB- clones (Fig. 2B). TM-1 Effects of TM-1 Antisense RNA. DES4supB+ cells were stably transfected with an expression vector containing the partial hamster TM-1 cDNA cloned in either the antisense or sense orientation. Four independent drug-resistant clones were then randomly selected: two antisense and two sense. As ,-actin shown in Fig. 3, perturbations of microfilament organization were observed in both supB+-transfected subclones expressing an antisense TM-1, while neither supB+ subclone expressing FIG. 2. TM-1 biosynthesis and expression in supB+ and supB- cell this cDNA in the sense orientation showed any apparent lines. (A) Autoradiograms of two-dimensional PAGE gels containing alteration in microfilament organization. On the basis of 35S-labeled total cellular proteins from 1OWsupB+ and -supB- cell previous results showing that the supB- but not supB+ cells are lines; the migration region of TM proteins in these cells is shown in capable of anchorage-independent growth in the presence of detail. TM isoforms, numbered according to apparent molecular certain growth factors (24), we then examined the effects of weights, are indicated with labeled, solid arrowheads and proliferating antisense TM-1 expression on anchorage-independent growth cell nuclear antigen (internal control) is indicated with an unlabeled, potential. Both supB+ subclones transfected with antisense open arrowhead. (B) Northern blot of TM-1 mRNA levels (Upper) and control TM-1 cDNA displayed marked anchorage-independent l3-actin loading (Lower). growth, forming large colonies in soft agar with a colony- Differential TM Biosynthesis. To address the hypothesis forming efficiency of 10-15%, while both sense TM-1 trans- that differences in microfilament organization between pairs fectants produced colonies with an efficiency of less than of supB+ and supB- clones were the result of variations in 0.01% (Fig. 4). actin-binding proteins, we performed a comparative analysis of Southern blotting confirmed the presence of multiple copies total cellular proteins by two-dimensional PAGE using wide- of exogenous TM-1 sequences in genomic DNA of sense and range pH ampholytes that allowed optimal visualization of the antisense transfectants, and Northern blotting revealed that greatest number of proteins. Under these conditions, repro- steady-state levels of TM-1 mRNA were significantly lower in ducible differences between supB+ and supB- clones were both DES4supB+ subclones expressing the antisense TM-1 observed in the region containing the TM family of proteins construct than in either subclone expressing the TM-1 cDNA (Fig. 2A). Quantitation of TM protein levels was performed by in the sense orientation. To confirm that the observed effects of antisense TM-1 expression were the result of decreased Table 1. Quantitative analysis of TM isoforms in SHE cell clones TM-1 protein production and, further, that antisense TM-1 Clone TM-1 TM-2 TM-3 TM-4 TM-5 PCNA* RNA did not affect the production of other proteins, com- parative two-dimensional PAGE experiments were performed [35S]Methionine incorporation, ppm with total proteins isolated from all transfected clones. Visual low and computerized densitometric analyses of two-dimensional supB+ 96 158 64 409 128 142 protein gels indicated no significant (>1.5:1), reproducible supB- 12 63 7 264 166 167 difference between sense and antisense clones for any detect- DES4 able protein on these gels other than one protein in the TM supB+ 81 61 10 211 132 181 region. The identification of this protein as TM-1 was con- supB- 22 20 8 224 114 195 firmed as described in Materials and Methods. Densitometric Silver stain, ppm quantitative analysis of replicate gels from both antisense low clones and both sense clones indicated that TM-1 protein levels supB+ 343 1073 486 6493 2756 2791 were 19% and 30% in antisense clones 1 and 2, respectively, of supB- <46 520 <46 5365 2931 2380 the average TM-1 protein level in sense clones 1 and 2 (Table *PCNA, proliferating cell nuclear antigen. 2). Downloaded by guest on September 28, 2021 Cell Biology: Boyd et al. Proc. Natl. Acad. Sci. USA 92 (1995) 11537

FIG. 3. Fluorescence photomicrographs depicting actin microfilament organization in DES4supB+ subclones stably expressing sense (A and B) or antisense (C and D) TM-1 RNA. (A) Sense clone 1, (B) sense clone 2, (C) antisense clone 1, and (D) antisense clone 2. (Photomicrographs taken at a magnification of 630X.) DISCUSSION from destruction or depolymerization of filamentous actin but rather from a more subtle redistribution of the microfilament These data indicate that the reorganization of microfilaments network and is independent of detectable differences in the may occur at a discrete, early stage in chemically induced cell expression of cytoplasmic 3- and y-actin isoforms. Previous transformation and is associated with loss of a tumor suppres- studies have provided indirect evidence that actin microfila- sor gene function. The alteration in microfilament organiza- ment organization may be related to tumor suppressor gene tion corresponds precisely with loss of negative controls on function (37-40). Although TM itself is also capable of tumor anchorage-independent growth by the supB- clones. This suppression in a transformed rodent cell model (18), it is likely change in microfilament organization does not appear to result that TM-1 is not a tumor suppressor gene per se but rather a downstream mediator of tumor suppressor gene function. This hypothesis is consistent with the well-characterized downregu- lation of TM-1 expression during cell transformation by a variety of oncogenic agents (19). The finding that TM-1 downregulation occurred concomi- tantly with microfilament reorganization led us to further investigate the possibility of a direct association between these phenomena. While several studies have documented the af- finity of TM for actin microfilaments in vitro (41-45), leading to the supposition that TM functions to assemble or stabilize Table 2. Quantitative analysis of TM-1 protein biosynthesis in transfectants Clone Density, arbitrary units DES4supB+ Sense clone 1 5131 ± 354 FIG. 4. Effect of sense or antisense TM-1 RNA expression on Sense clone 2 5075 ± 1023 anchorage-independent growth potential of transfected subclones DES4supB+ after 10 days of culture. (A) DES4supB+ sense transfectants, mass Antisense clone 1 982 ± 311 culture; (B) sense clone 1; (C) sense clone 2; (D) DES4supB+ antisense Antisense clone 2 1563 ± 416 transfectants, mass culture; (E) antisense clone 1; and (F) antisense clone 2. (x 11.) Data are mean ± SEM; n = 2. Downloaded by guest on September 28, 2021 11538 Cell Biology: Boyd et al. Proc. Natl. Acad. Sci. USA 92 (1995) these structures, the downregulation of TM may not necessar- 13. Matsumura, F., Lin, J. J.-C., Yamashiro-Matsumura, S., Thomas, ily lead to the rearrangement of microfilaments observed in G. P. & Topp, W. C. (1983) J. Biol. Chem. 258, 13954-13964. vivo. Our data are consistent with the hypothesis that the actin 14. Hendricks, M. & Weintraub, H. (1984) Mol. Cell. Biol. 4, alterations observed in supB- cells are the result of decreased 1823-1833. TM-1 levels of 15. Cooper, H. L., Feuerstein, N., Noda, M. & Bassin, R. H. (1985) expression. Thus, the lower high molecular Mol. Cell. Biol. 5, 972-983. weight TM isoform expression that have been associated with 16. Varma, M. & Leavitt, J. (1988) Mutat. Res. 199, 437-447. diverse types of transformed cells (14-19) may represent a 17. Bhattacharya, B., Gaddamanugu, L. P., Valverius, E. M., critical and early molecular event in the multistep process of Salomon, D. S. & Cooper, H. L. (1990) Cancer Res. 50, 2105- neoplastic transformation. 2112. The redistribution of actin microfilaments induced by anti- 18. Prasad, G. L., Fuldner, R. A. & Cooper, H. L. (1993) Proc. Natl. sense TM-1 effectively removed a negative constraint on Acad. Sci. USA 90, 7039-7043. anchorage-independent proliferation. Folkman and Green- 19. Matsumura, F. & Yamashiro-Matsumura, S. (1986) Cancer Rev. span (46) proposed that "geometric regulators of growth 6, 21-39. appear to act as gates; they give or withhold permission for 20. Liu, H. & Bretscher, A. (1989) Cell 57, 233-242. 21. Oshimura, M., Gilmer, T. M. & Barrett, J. C. (1985) Nature growth" under certain conditions. The normal actin cytoskel- (London) 316, 636-639. eton, regulated by high molecular weight TM isoforms, ap- 22. Thomassen, D. J., Gilmer, T. M., Annab, L. A. & Barrett, J. C. pears to represent one such "gate" in normal cells: allowing (1985) Cancer Res. 45, 726-732. growth when cells are attached and spread but serving as a 23. Koi, M. & Barrett, J. C. (1986) Proc. Natl. Acad. Sci. USA 83, growth restraint in unattached, rounded cells. Thus, TM-1 may 5992-5996. serve a tumor suppressor function, insofar as the acquisition of 24. Koi, M., Afshari, C. A., Annab, L. A. & Barrett, J. C. (1989) Proc. anchorage-independent growth potential may represent a crit- Natl. Acad. Sci. USA 86, 8773-8777. ical step in the neoplastic progression of some tumor cells. It 25. Afshari, C. A. & Barrett, J. C. (1993) Mol. Carcinog. 7, 249-256. is unclear how the phenomenon of tumor suppression by 3' 26. Garrels, J. I. (1979) J. Biol. Chem. 254, 7961-7977. untranslated sequences of TM may relate to the data 27. Garrels, J. I. (1989) J. Biol. Chem. 264, 5269-5282. (47, 48) 28. Merrick, B. A., He, C., Craig, W. A., Clark, G. C., Corsini, E., described in this report. It cannot be ruled out that some or all Rosenthal, G. J., Mansfield, B. K. & Selkirk, J. K. (1992) Appl. of the effect that we observed in antisense TM-1-transfected Theor. Electrophor. 2, 177-187. cells was the result of sequestration or destruction of the 29. Erba, H. P., Gunning, P. & Kedes, L. (1986) Nucleic Acids Res. endogenous 3' untranslated region of TM-1, as opposed to the 14, 5275-5294. presumed mechanism of downregulation of TM-1 protein. 30. Yamawaki-Kataoka, Y. & Helfman, D. M. (1985) J. Biol. Chem. Together, these data suggest an area for future studies on 260, 14440-14445. tumor suppression, especially in regard to the role of structural 31. Yamawaki-Kataoka, Y. & Helfman, D. M. (1987) J. Biol. Chem. proteins and their respective mRNA molecules. 262, 10791-10800. Other critical tasks include determination of precisely how 32. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162, 156- variations in the normal actin act to 159. cytoskeleton deregulate 33. Boyd, J. & Risinger, J. I. (1991) Mol. Carcinog. 4, 189-195. growth control. Changes in cytoskeletal organization would be 34. Lee, F., Mulligan, R., Berg, P. & Ringold, G. (1981) Nature expected to alter tensional integrity of the cytomatrix system, (London) 294, 228-232. including the nuclear matrix (4, 49). We have shown that 35. Chen, C. & Okayama, H. (1987) Mol. Cell. Biol. 7, 2745-2752. nuclear shape is significantly more irregular in supB- cells than 36. Prasad, G. L., Meissner, S., Sheer, D. G. & Cooper, H. L. (1991) in supB+ cells (50). This difference in nuclear shape could Biochem. Biophys. Res. Commun. 177, 1068-1075. result in altered patterns of chromatin interaction with the 37. Kopelovich, L., Conlon, S. & Pollack, R. (1977) Proc. Natl. Acad. nuclear matrix, leading to new patterns of gene transcription Sci. USA 74, 3019-3022. (51). It is likely that changes in the distribution of actin 38. Antecol, M. H., Darveau, A., Sonenberg, N. & Mukherjee, B. B. microfilaments would also affect the association of mRNA (1986) Cancer Res. 46, 1867-1873. an 39. Der, C. J., Ash, J. F. & Stanbridge, E. J. (1981) J. Cell Sci. 52, with the cytoskeleton, interaction that is well documented 151-166. (52, 53), possibly resulting in altered mRNA translation. Such 40. Bouck, N., Stoler, A. & Polverini, P. J. (1986) Cancer Res. 46, a limitation on the availability of all mRNA or perhaps specific 5101-5105. messages (54, 55) would exert a profound influence on gene 41. Lazarides, E. (1976) J. Cell Biol. 68, 202-219. expression and, hence, biological processes such as differen- 42. Matsumura, F., Yamashiro-Matsumura, S. & Lin, J. J.-C. (1983) tiation and growth control. J. Biol. Chem. 258, 6636-6644. 43. Bretscher, A. (1986) Nature (London) 321, 726-727. 1. Ben-Ze'ev, A. (1985) Biochim. Biophys. Acta 780, 197-212. 44. Fattoum, A., Hartwig, J. H. & Stossel, T. P. (1983) Biochemistry 2. Clarke, M. & Spudich, J. A. (1977) Annu. Rev. Biochem. 46, 22, 1187-1193. 797-822. 45. Ng, S.-Y., Erba, H., Latter, G., Kedes, L. & Leavitt, J. (1988) Mol. 3. Korn, E. D. (1978) Proc. Natl. Acad. Sci. USA 75, 588-599. Cell. Biol. 8, 1790-1794. 4. Ingber, D. E. & Folkman, J. (1989) in Cell Shape: Determinants, 46. Folkman, J. & Greenspan, H. P. (1975) Biochim. Biophys. Acta Regulation, and Regulatory Role, eds. Stein, W. D. & Bronner, F. 417, 211-236. (Academic, New York), pp. 3-32. 47. Rastinejad, F. & Blau, H. M. (1993) Cell 72, 903-917. 5. Otsuka, H. & Moskowitz, M. (1975) J. Cell Physiol. 87, 213-219. 48. Rastinejad, F., Conboy, M. J., Rando, T. A. & Blau, H. M. (1993) 6. Folkman, J. & Moscona, A. (1978) Nature (London) 273, 345- Cell 75, 1107-1117. 349. 49. Pienta, K. J., Partin, A. W. & Coffey, D. S. (1989) CancerRes. 49, 7. Gospodarowicz, D., Greenburg, G. & Birdwell, C. R. (1978) 2525-2532. Cancer Res. 38, 4155-4171. 50. Boyd, J., Pienta, K. J., Getzenberg, R. H., Coffey, D. S. & 8. Ingber, D. E. (1990) Proc. Natl. Acad. Sci. USA 87, 3579-3583. Barrett, J. C. (1991) J. Natl. Cancer Inst. 83, 862-866. 9. Benecke, B.-J., Ben-Ze'ev, A. & Penman, S. (1978) Cell 14, 51. Nelson, W. G., Pienta, K. J., Barrack, E. R. & Coffey, D. S. 931-939. (1986) Annu. Rev. Biophys. Biophys. Chem. 15, 457-475. 10. Luna, E. J. & Hitt, A. L. (1992) Science 25, 955-964. 52. Cervera, M., Dreyfuss, G. & Penman, S. (1981) Cell 23, 113-120. 11. Lees-Miller, J. P. & Helfman, D. M. (1991) BioEssays 13, 429- 53. Taneja, K. L., Lifshitz, L. M., Fay, F. S. & Singer, R. H. (1992)J. 437. Cell Biol. 119, 1245-1260. 12. Leonardi, C. L., Warren, R. H. & Rubin, R. W. (1982) Biochim. 54. Jeffery, W. R. (1984) Dev. Biol. 103, 482-492. Bioplhys. Acta 720, 154-162. 55. Lawrence, J. B. & Singer, R. H. (1986) Cell 45, 407-415. Downloaded by guest on September 28, 2021