- Email: [email protected]
Cytoskeleton involvement in the distribution of mRNP complexes and small cytoplasmic RNAs
Biochimica et Biophysica Acta 868 (1986) 215-225
215
Elsevier BBA 91611
Cytoskeleton involvement in the distribution of mRNP complexes and small cytoplasmic R N A s R. Curtis Bird a and Bruce H. Sells b " Department of Microbiology, College of Veterinary Medicine. Auburn University, A uburn, A L (U.S.A.) and I, College of Biological Science, University of Guelph, Guelph (Canada)
(Received 18 June 1986) (Revised manuscript received17 September1986) Key words: histone H4 mRNA; mRNA complex; Cytoskeleton;small cytoplasmicRNA These studies were designed to determine whether small cytoplasmic RNAs and two different mRNAs (actin mRNA and histone 114 mRNA) were uniformly distributed among various subcellular compartments. The cytoplasm of HeLa $3 cells was fractionated into four RNA-containing compartments. The RNAs hound to the cytoskeleton were separated from those in the soluble cytoplasmic phase and each RNA fraction was further separated into those hound and those not hound to polyribosomes. The four cytoplasmic RNA fractions were analysed to determine which RNA species were present in each. The 7 S RNAs were found in all cytoplasmic fractions, as were the 5 S and 5.8 S ribosomal RNAs, while transfer RNA was found largely in the soluble fraction devoid of polysomes. On the other hand a group of prominent small cytoplasmic RNAs (scRNAs of 105-348 nucleotides) was isolated from the fraction devoid of polysomes but hound to the cytoskeleton. Actin mRNA was found only in polyrihosomes hound to the cytoskeleton. This mRNA was released into the soluble phase by cytochalasin B treatment, suggesting a dependence upon actin filament integrity for cytoskeletai binding. A significant portion of several scRNAs was also released from the cytoskeleton by cytochalasin B treatment. Analysis of the spatial distribution of histone 1-14 mRNAs, however, revealed a more widely dispersed message. Although most (60%) of the H4 mRNA was associated with polyribosomes in the soluble phase, a significant amount was also recovered in both of the cytoskeleton bound fractions either associated or free of polyrihosome interaction. Treatment with cytochalasin B suggested that only cytoskeleton hound, untranslated H4 mRNA was dependent upon the integrity of actin filaments for cytoskeletal binding.
Introduction
The cytoplasm is a highly structured and compartmentalized region of the cell in which diverse biological processes occur. While attachment of organelles to membranes accounts for much of this compartmentalization, the portion of the cytoplasm outside these membranes has, until recently, been considered largely unstructured. A Correspondence: Dr. B.H. Sells, Collegeof Biological Science, University of Guelph, Guelph, Ontario, Canada, NIG 2Wl.
framework of filamentous elements called the cytoskeleton, which is found throughout the cytoplasm, is thought to contribute to the spatial separation of various processes. The cytoskeleton may also be involved in maintenance and fluctuations in cell shape [1-3], the organization and function of cytoplasmic organelles [4-6] and cell motility [7-9]. During the process of mRNA translation, the translational machinery (the polyribosome) is also thought to be attached to elements of the cytoskeleton [10-16]. While differences in opinion exist concerning the identity of the cytoskeletal
0167-4781/86/$03.50 © 1986 ElsevierSciencePublishers B.V. (BiomedicalDivision)
216
element to which the polyribosomes are bound, there is general agreement that most mRNA populations are bound to the cytoskeleton. The cytoskeleton and its components have been extensively characterized using both biochemical and microscopic techniques. This network of filaments is composed largely of (although it may not be limited to) three major elements. Extensive networks of microtubules [16-18] intermediate filaments [19,10] and actin filaments [21,11,7] have been observed and their component proteins have been characterized. The cytoskeleton also contains non-filamentous components of 'cytoskeleton bound' material which seems to include, at least in some cells, the polyribosomes and translation initiation factors [23]. A eucaryotic cell characteristically contains a wide variety of ribonucleic acid molecules in its cytoplasm, which may include many species of mRNA, tRNA and a group of polynucleotides designated small cytoplasmic RNAs (scRNAs). The mRNAs are normally associated with protein as messenger ribonucleoprotein (mRNP) complexes and exist either in polysome arrays when actively being translated or in free m R N P complexes. The present studies were designed to determine the distribution of these RNAs throughout the cytoplasm and to determine whether specific species or populations are preferentially located on cytoskeletal elements. These studies investigated: (i) whether the distribution of histone H4 mRNA and actin mRNA in free and polysomal mrNAs, between the cytoskeleton and soluble fractions, is the same for each message and (ii) whether scRNAs are uniformly distributed. The results obtained reveal that (i) the distribution differs for each message and (ii) the bulk of scRNAs are located predominantly with the m R N P fraction - free of polysomes - associated with the cytoskeleton. Materials and Methods
Cell culture and labelling HeLa $3 cells were grown in suspension culture (500 ml) to a density of ( 1 - 3 ) . 105 cells/ml in minimal essential medium supplemented with 10% fetal bovine serum at 37°C and 5% CO 2. In cases where cells were to be labelled, identi-
cal cultures were treated with either 2 F C i / m l [3H]lysine or 2 / t C i / m l [3H]uridine for 14 to 16 h. Prior to cell fractionation the cells were washed in ice cold Hepes-buffered saline (0.9% NaC1/10 mM Hepes (pH 7.2)). Cells to be treated with cytochalasin B were first incubated in the presence of 100 t~g/ml emetine for 5 min to preserve polysome structure, and then were treated with 10 /~g/ml cytochalasin B for 30 min [11]. Subsequently, the cell suspensions were processed for subcellular fractionation. Fractionation into cytoskeletal bound and soluble R N P compartments HeLa $3 cells were fractionated into cytoskeletal bound and soluble compartments according to a modification of the procedure of Fey et al. [24]. Suspension cultures of HeLa $3 cells were treated with 50 /~g/ml cycloheximide and harvested by centrifugation at 200 × g (1000 rpm) for 3 min (Beckman TJ-6 fitted with a TH-4 rotor). The cells were gently resuspended in 50 ml ice-cold Hepes-buffered saline with a Pasteur pipette. The cell suspension was centrifuged again ( 2 0 0 x g for 3 min), resuspended in 15 ml of ice-cold lysis buffer number 1 (300 mM sucrose, 20 mM Pipes (pH 6.8), 10 mM KC1, 5 mM MgSO 4, 0.1% (v/v) Triton X-100, 10 mM c-amino caproic acid, 0.005% (w/v) phenylmethylsulphonyl fluoride, 0.1 mM aurintricarboxylic acid, 10 mM dithiothreitol and 5 units/ml placental ribonuclease inhibitor) and very gently mixed for approx. 45 s. Following another centrifugation (200 x g for 3 min) the supernatant was removed and designated 'soluble fraction'. The cytoskeleton bound material, in the pellet, was further extracted by resuspending it in 15 ml of ice-cold lysis buffer number 2 (same composition as lysis buffer number 1 but lacking KC1 and containing 0.25 M ammonium sulfate and Triton X-100 at 0.5 % (v/v)) for 8 min, during the initial part of which the suspension was homogenized by ten passes with a pre-chilled Dounce homogenizer. The homogenized cytoskeletal fraction was centrifuged once again (10000 x g for 20 rain) and the supernatant was designated 'cytoskeleton bound fraction'. The residual pellet contained largely nuclei and insoluble cell debris. Aliquots of these samples (50 #l), in which acid-precipitable radioactive incorpora-
217 tion was to be measured, were dried onto Whatman 540 filter discs and processed as previously described [25]. Once the cells had been fractionated into cytoskeleton-bound and soluble compartments, each fraction was further separated into free and polyribosome-bound RNP complexes. RNA was extracted from the polysomal and free RNP fractions by a modification of a procedure previously described [26,27]. Polyribosomes were prepared by centrifugation of the resulting supernatant (made 100 /~g/ml poly(vinyl sulfate)) through a 30% ( w / v ) sucrose cushion at 178000 x g (1 h at 2°C). The free RNP fractions were prepared from both of the decanted supernatants (both cytoskeletonbound and soluble) by further centrifugation at 178000 x g (16-18 h at 2°C). The RNP pellets from each of the four fractions were dissolved in 2 ml of a solution containing 10 mM Tris-HC1 (pH 7.6), 1 mM N a 2 E D T A and 1% (w/v) sodium dodecyul sulfate (SDS) at room temperature and digested with proteinase K (EC 3.4.21.14) for 1 h at 37°C. To obtain RNA, the solutions were p h e n o l / c h l o r o f o r m extracted, precipitated with 2 vol. of ethanol, and stored overnight at - 2 0 ° C . The four m R N P fractions were designated cytoskeleton-bound polysomes and free m R N P complexes, as well as the soluble polysomes and free m R N P complexes.
Polysome profiles Following preparation of polysome-bound and free RNP pellets from both the cytoskeleton and soluble fractions, polysome profiles were prepared to analyse polysome integrity and size (Bagchi, Larson and Sells, unpublished results). Each pellet was resuspended in 0.3 ml of resuspension buffer (20 mM Tris-HC1 (pH 7.5), 200 mM KC1, 10 mM MgC12 and 0.2 m g / m l heparin) by 10 passes in a 1 ml microhomogenizer on ice. 200 #1 of each sample was loaded onto a 15-50% ( w / v ) linear sucrose gradient (4.8 ml) made up in resuspension buffer and centrifugated in a Beckman 50.1 rotor for 30 min at 40000 rpm and 4°C. Eachgradient was subsequently fractionated and absorbance at 260 nm was measured continuously while maintaining the gradient at approx. 4°C. Electrophoresis and northern blot analysis Total RNA samples (50 /xg/lane) were frac-
tionated on one of two different gel-electrophoresis systems. Samples in which larger molecular weight RNAs were to be resolved were run on 1.5% agarose gels, containing 50 mM Mops (pH 7.0), after denaturation in a solution containing 50% (v/v) formamide, 7% (v/v) formaldehyde and 25 mM Mops (pH 7.0) as previously described [26]. Alternatively, samples in which lower-molecular-weight RNAs were to be resolved were separated on 6% polyacrylamide gels containing 8.3 M urea following denaturation in formamide as previously described [25,27]. RNA gels to be photographed were stained for 15 n'fin in 1 # g / m l ethidium bromide. All gels were subsequently prepared for Northern electroblot analysis to diazotized aminophenylthioether paper by sequential washing in N a O H and then K H z PO 4 as previously described [25,27]. The blots were prehybridized and hybridized as previously described [25,27] to 2.106 cpm of nick-translated plasmid [28] at an approximate specific activity of ( 2 - 3 ) . 108 cpm//~g of plasmid DNA, for 3 days. Excess probe was washed from the blots as described [29] except that each wash lasted 20 min. Autoradiograms were prepared and quantitation was performed by scanning original unsaturated autoradiograms with a microdensitometer. As previously described, the response was linear with respect to the number of micrograms of RNA loaded on the gel [27]. The specific D N A probes employed were a cloned human genomic sequence encoding an entire histone H4 gene [30] and a cloned heterologous cDNA containing most of the a-actin coding sequence. Histone and actin mRNAs were identified in total RNA samples by electrophoretic mobility in comparison to other RNAs of known size (28 S, 18 S, 5.8 S and 5 S rRNAs, 7 S RNA and tRNA) as well as actin mRNAs identified in poly(A) ÷ RNA samples prepared as described previously [27]. Results
Fractionation of HeLa $3 cells into cytoskeleton bound and soluble mRNA fractions Initial studies were performed to define the distribution of m R N P complexes between the cy-
218
toskeleton and soluble fractions, in HeLa $3 cells, employing the lysis conditions of Fey et al. [24]. The cells were gently lysed to solubilize only the plasma membrane. As a consequence of this treatment only those particles which were unattached to the cytoskeletal elements, including the m R N P complexes were liberated into the soluble fraction. After thorough washing of the cellular remnants to remove the remaining soluble material, a highsalt extraction was used to release the bound m R N P complexes from the cytoskeleton. In our hands, the conditions of Fey et al. [24] consistently over-extracted the cytoskeletal fraction of HeLa $3 cells and resulted in an excessively large amount of A260 material (isolated as polysomes) in the soluble fraction (Fig. 1). A modification of this procedure allowed us to obtain a similar distribution of polysomes between the cytoskeletal and soluble fractions as that reported by Fey et al. [24]. The modified lysis procedure involved a reduction in the concentration of both the KC1 and the Triton X-100 concentrations in the first lysis solution. Analysis of the effects of varying KC1 concentration (Fig. la) demonstrates that above 10 mM KC1 the soluble fraction contains increasing amount of polysomal material. At 100 mM KC1 the amount of A260 material (isolated from polysomes) released into the soluble fraction exceeded that found in the cytoskeletal bound material by approx. 100% (69% of the polysomes appeared soluble). In contrast Cervera et al. [12] using HeLa cells reported that no polyribosomes and only 25-27% of the mrNA were recovered in the soluble phase. Reduction of the KCI concentration in our lysis solution, to 10 mM, resulted in only about 24% of the polysomes being recovered in the soluble fraction. The Triton X-100 concentration was also optimized by monitoring the extraction during initial cell lysis. Variations in Triton X-100 concentration between 0.01% and 1.0% produced little change in the amount of material obtained in the soluble phase at 100 mM KC1 (data not shown). At the more moderate 10 mM KC1 concentration, however, as little as 0.1% Triton X-100 in the first lysis solution was sufficient to extract the soluble material fully (Fig. lb). Analysis of polysome profiles of the soluble and cytoskeleton bound fractions at optimal KC1
A
i
0
I
20
,
,
40 KCl
i
GO O0 (mid)
,
I(]0
I00i
I10" "-0
e040.
O-
6
oh
o'.z o~ o~, Triton X..IO0 (%, V/v)
o'.~
Fig. 1. Optimization of KCI and Triton X-100 concentrations for the fractionation of cytoskeleton-bound and soluble cytoplasmic polysomal fractions. (A) HeLa cells were lysed in lysis buffer 1, containing increasing concentrations of KC1 and 0.5% (v/v) Triton X-100, into cytoskeleton-bound and soluble cytoplasmic polysomal fractions. The nucleotide component (RNA) of each fraction was measured as absorbance at 260 nm and relative amount (percentage) in each fraction determined for various KC1 concentrations. (B) HeLa cells were lysed in lysis buffer 1, containing increasing concentrations of Triton X-100 and 10 mM KCI, into cytoskeleton-bound and soluble cytoplasmic polysomal fractions. The relative RNA levels in each fraction were determined for various Triton X-100 concentrations. Each pair of curves was compiled from either three or four independent experiments. Determination of percent distribution of A 260 material at optimal concentration was independently determined six times (maximum sample variation _+6%). Cytoskeleton-bound (O); soluble phase (0).
(10 mM) and Triton X-100 (0.1%) concentrations revealed that 80% of the polysomes were associated with the cytoskeletal fraction (Fig. 2). These results demonstrate that the modified lysis procedure leaves most of the polysomes bound to the cytoskeleton in HeLa cells as previously reported [121. Analysis of the RNA and protein concentrations in each of the soluble and cytoskeleton bound
219
Z
go nO
~
ELETON
oO gO
SEDIMENTATION
Fig. 2. Polysomc profiles from cytoskeleton bound and soluble cytoplasmic polysomal fractions from HeLa cells. Polysome
suspensions were fractionated by centrifugation through a 15-50% (w/v) linear sucrose gradient followed by fractionation and analysis of absorbance at 260 nm.
fractions was performed following long-term labelling with either [3H]lysine or [3H]uridine. Of the labelled macromolecules which were recovered in the cytoskeleton-bound and soluble fractions as acid-precipitable material, 59% of the labelled R N A was localized in the soluble fraction and 41% was bound to the cytoskeleton (Table I). Approx. 70% of the labelled protein was found in the soluble fraction, while 30% was isolated in the cytoskeletal fraction. These results are similar to those reported by Bonneau et al. [31] for African green monkey kidney cells extracted for 1 - 2 rain and by Cervera et al. [12] for H e L a cells.
Qualitative differences in RNAs found in different cytoplasmic fractions We have examined the R N A species present in each of the four cytoplasmic fractions to determine whether there were any qualitative difTABLE I PROPORTION OF RNA AND PROTEIN IN CYTOSKELETON-BOUND AND SOLUBLE CYTOPLASMIC FRACTIONS Mean values for both RNA and protein represent the percentage of acid-precipitable counts observed in each fraction isolated from cells labelled with either [3H]uridine (RNA) or [3H]lysine (protein). Maximum sample variation is noted and each experiment was done twice. Fractions Cytoskeleton Soluble
RNA 41 + 7% 59 + 7%
Protein 70 + 6% 30 + 6%
ferences. Both the cytoskeleton-bound and soluble R N P compartments, as defined above, have been inspected and both the polysomal and free m R N P complexes isolated from each compartment by differential ultracentrifugation. To determine the distribution of scRNAs the R N A s extracted from each of these four cytoplasmic fractions were analysed on 6% polyacrylamide gels containing 8.3 M urea (Fig. 3). The smaller R N A s were thus resolved into four unique populations by isolating the R N A from (1) the cytoskeletal polysomes, (2) the cytoskeletal free RNPs, (3) the soluble polysomes and (4) the soluble free RNPs. Most of the messenger RNAs, 18 S and 28 S ribosomal RNAs were located near the top of the gel or failed to penetrate the gel. Only the smaller R N A species migrated into the middle and lower regions of the gel. The 7 S RNAs, thought to be involved in the signal sequence recognition particle [32] were observed in all of the cytoplasmic fractions as were the 5 S and 5.8 S ribosomal RNAs. The transfer RNAs, while present in all four fractions, were concentrated mainly in the soluble free R N P fraction. Seven additional distinct species of small cytoplasmic RNA, identified as prominent hands, were found only in the cytoskeleton-bound free R N P fraction. These small cytoplasmic RNAs (scRNAs) have approximate sizes of 348 (No. 1), 284 (No. 2), 253 (No. 3), 214 (No. 4), 160 (No. 5), 123 (No. 6), and 105 (No. 7), nucleotides, scRNA1, -4 and -7 were also represented, to a much lesser extent, in the soluble free R N P fraction. The function of these small RNAs is uncertain. They
220 d e v o i d of p o l y ( A ) ( B i r d a n d Sells, u n p u b l i s h e d results). T o d e t e r m i n e w h e t h e r a n y o f the s c R N A s , isol a t e d in the c y t o s k e l e t o n - b o u n d free R N P frac-
Fig. 3. Electrophoresis analysis of low molecular weight RNAs on 6% polyacrylamide gels containing 8.3 M urea. Total RNA (50 ~g/lane) was isolated from the cytoskeleton polysome bound (CP) and free (CF) RNP fractions as well as from the soluble polysome bound (SP) and free (SF) RNP fractions and subjected to denaturing electrophoresis followed b~' staining with ethidium bromide. Right side: RNA species of known size (7 S, 5.8 S, 5 S, 4 S). Left side: scRNA species (numbered 1-7) located largely in the CF fraction. a r e all s m a l l e r t h a n h i s t o n e H 4 m R N A , at 3 6 4 - 4 3 9 n u c l e o t i d e s l o n g in h u m a n [33] a n d in rat [27] cells, a n d d o n o t a p p e a r to b e m e s s a g e s since, in a d d i t i o n to t h e i r s m a l l size, t h e y a r e a p p a r e n t l y
Fig. 4. Electrophoretic analysis of low-molecular-weight RNAs in cytoplasmic fractions isolated from HeLa cells treated with cytochalasin B. RNA was isolated from rapidly proliferating (control) HeLa cells and HeLa cells which had been pretreated with emetine (100/~g/ml) for 5 rain and subsequently treated with cytochalasin B (10 /~g/ml) for 30 min to depolymerize actin filaments. Total R N A (50 /~g/lane) samples, isolated from the cytoskeleton polysome bound (CP) and free (CF) fractions as well as from the soluble polysome bound (SP) and free (SF) RNP fractions were analyzed on a 6% polyacrylamide gel containing 8.3 M urea followed by staining with ethidium bromide. RNA species of known size: (7 S, 5.8 S, 5 S, 4 S). scRNA species (numbered 1-4, 7) located in both the CF and SF fractions, and scRNAs (numbered 5, 6) which are still located largely in the CF fraction.
221 tion, were associated with the actin filament component of the cytoskeleton, H e L a cells were incubated with cytochalasin B prior to fractionation. Several scRNAs were partitioned into both the soluble free and cytoskeleton-bound free R N P fractions after treatment (Fig. 4). These results suggest that a portion of scRNA-1, -2, -3, -4 and -7 may be associated with actin filaments, since cytochalasin B treatment of H e L a cells has been shown to rapidly depolymerize actin filaments (Ref. 11; Bagchi, Larson and Sells, unpublished results).
Spatial distribution of histone H4 and actin mRNAs Polysomes are found principally with the cytoskeleton fraction. Distributions of actin and histone H4 m R N A s were analysed to determine whether two individual m R N A s having different characteristics are distributed in a manner similar to the total population of polysomal mRNAs. Actin m R N A was chosen as an example of a typical poly(A) + m R N A and because its association with the cytoskeleton has been reported in a number of different cell types [14,15,31]. The family of histone H4 m R N A s were examined because these messages are a m o n g the best characterized examples of poly(A)- m R N A s [3446,30]. Among the four cytoplasmic compartments actin m R N A was detected in only the cytoskeleton bound polysomal m R N A fraction (Fig. 5). This result supports the observations by others [14,15,31] and suggests that the cytoskeleton bound fraction was not over extracted during the first lysis step. To determine whether the actin m R N A might be associated with the microfilaments of the cytoskeleton, H e L a cells were incubated with cytochalasin B. Actin-mRNA-containing polysomes were released from the cytoskeleton following treatment (Fig. 5). Since microfilaments in H e L a cells and L6 myoblasts rapidly disassemble in the presence of cytochalasin B [11,47], these results imply that actively translating actin m R N A s are probably dependent upon the integrity of microfilaments either directly or indirectly, for their binding to the cytoskeleton. In this analysis 50/~g of total R N A was loaded in each lane, to permit us to asses both poly(A) ÷ and poly(A)- m R N A species. The limiting
Fig. 5. Northern blot analysis of actin mRNAs from rapidly proliferating HeLa cells, and from HeLa cells which had been pretreated with emetine (100 /~g/ml) for 5 rain and subsequently treated with cytochalasin B (10 /~g/ml) for 30 min. Total RNA (50/~g/lane) samples isolated from the cytoskeleton polysome bound (CP) and free (CF) RNP fractions, and from the soluble polysome bound (SP) and free (SF) RNP fractions, were denatured with formamide and formaldehyde and subjected to electrophoresis on denaturing 1.5% agarose gels, followed by electroblotfing to DPT-celhilose paper and hybridization to labelled actin and H4 DNA probes. An X-ray film was exposed to the blot for about 5 days, at -70 o C, for autoradiographic detection.
amounts of homologous sequence present in each lane made necessary long exposures (approx. 7 days) and a reduction in stringency of the last wash of the blot (conducted at room temperature) prior to autoradiography, to enhance as much as possible hybridization signal. The large mass of non-messenger R N A present as well as the reduction in stringency are responsible for the higher background hybridization in the lanes and the smearing observed in some of the bands. Since actin hybridization intensity was not correlated with the amount of 16 S r R N A observed on the gels, it is unlikely that there was a significant amount of cross-reactivity with this slightly larger R N A species. The density appearing at higher electrophoretic mobility is most likely due to cross-hybridization with unidentified R N A s migrating between the 16 S and 28 S r R N A bands.
222
Fig. 6. Northern blot analysis of histone H4 mRNAs from HeLa cells which were rapidly proliferating, and which had been pretreated with emetine (100 t~g/ml) for 5 min and subsequently treated with cytochalasin B (10 /~g/ml) for 30 min. Total RNA (50 /~g/lane) samples isolated from the cytoskeleton polysome bound (CP) and free (CF) RNP fractions as well as from the soluble polysome bound (SP) and free 9SF) RNP fractions, were denatured by boiling in formamide and subjected to electrophoresis on denaturing 6% polyacrylamide gels containing 8.3 M urea followed by electroblotting onto DPT-cellulose paper and hybridization to a labelled histone H4 DNA probe. X-ray film was exposed to the blot for about 5 days, at - 7 0 o C, for autoradiographic analysis.
TABLE II PROPORTION OF HISTONE H4 mRNA IN CYTOPLASMIC COMPARTMENTS FROM CELLS TREATED WITH CYTOCHALASIN B Mean values from two experiments represent the percentage of total histone H4 mRNA, quantified on autoradiograms, observed in each cytoplasmic fraction. Maximum samf01e variation is noted for two independent experiments. Fractions
Untreated
Cytochalasin Btreated
Cytoskeleton polysomal Cytoskeleton free RNP Soluble polysomal Soluble free RNP
24 + 7 16 +_ 3 58 + 11 2+ 1
30 _+5 3 _+1 59 + 8 8+ 2
A n e x a m i n a t i o n of the d i s t r i b u t i o n of histone H 4 m R N A s a m o n g the four c y t o p l a s m i c fractions revealed a situation m a r k e d l y different from that for actin m R N A (Fig. 6). M o s t of the histone H 4 m R N A (approx. 58%, T a b l e II) was located in the soluble p o l y s o m a l - b o u n d fraction, free of association with the cytoskeleton. Significant a m o u n t s of H 4 m R N A were also recovered in the cytoskelet o n - b o u n d p o l y s o m e s a n d in the cytoskeletalb o u n d free m R N P s . A l m o s t 24% of the H4 m R N A was recovered in the c y t o s k e l e t o n - b o u n d polys o m a l fraction, while 16% was recovered in the c y t o s k e l e t o n - b o u n d free R N P fraction. O n l y trace a m o u n t s were f o u n d in the soluble free R N P fraction. A l t h o u g h m o s t of the p o l y s o m e s were associated with the cytoskeleton fraction, H4 m R N A was f o u n d in all four c y t o p l a s m i c fractions. F u r t h e r m o r e , the largest p o r t i o n of the H 4 m R N A was l o c a t e d in the soluble p o l y s o m a l p h a s e (approx. 60%) r a t h e r t h a n associated with the cytoskeleton. T r e a t m e n t with c y t o c h a l a s i n B, a n d p r e s u m a b l y the d i s a s s e m b l y of microfilaments, p r i o r to cell fractionation, resulted in an a p p a r e n t release of histone H 4 m R N A f r o m only the cytoskeletal free R N P fraction (Fig. 6). The a m o u n t of histone H 4 m R N A in this fraction decreased f r o m 16% to 3% ( T a b l e II). M o s t of the H 4 m R N A released was p a r t i t i o n e d equally between cytoskeletonb o u n d p o l y s o m a l a n d soluble free R N P fractions. T h e d a t a i m p l y that a p p r o x . 13% of the H4 m R N A is d e p e n d e n t u p o n the integrity of actin filaments for c y t o s k e l e t o n b i n d i n g a n d that it is released f r o m the c y t o s k e l e t o n - b o u n d free R N P fraction d u r i n g m i c r o f i l a m e n t disassembly. W h i l e a b o u t half of the H 4 m R N A was released into the solub l e free R N P fraction, the r e m a i n d e r of the cytos k e l e t o n - b o u n d free R N P fraction H 4 m R N A app e a r e d in the c y t o s k e l e t o n - b o u n d p o l y s o m a l fraction. Discussion O u r results d e m o n s t r a t e that different R N A species, i n c l u d i n g various messages a n d small cyt o p l a s m i c R N A s , are differentially d i s t r i b u t e d in the c y t o p l a s m of H e L a cells. W e have o b s e r v e d that actin m R N A behaves m u c h like viral m R N A s [10,11] in that it is largely associated with the
223 polysomes bound to cytoskeleton. In contrast, histone H4 mRNAs are found principally in the soluble polyribosome fraction unassociated with the cytoskeleton. Cytoskeleton association does not appear to be a prerequisite for histone H4 m R N A assembly into polysomes, although some polysome-bound H4 m R N A was also found associated with the cytoskeleton. These results argue against the notion that a message must first become associated with the cytoskeleton before translation can occur [12,31]. Although eucaryotic translation initiation factors have been found associated with the cytoskeleton in HeLa cells [23], our results imply that this association may not be obligatory. Histone H3 mRNAs remain uniformly distributed during periods of major cytoplasmic rearrangement in Xenopus embryos [48], suggesting that they are independent of these mobile elements. Small RNA species also display a non-uniform distribution in the cytoplasm. The largest proportion of tRNA was located in the soluble free RNP fraction unassociated with either ribosomes or the cytoskeleton. This finding is similar to that reported by Cervera et al. [12] where 80% of the tRNA was isolated from the soluble RNA fraction in HeLa cells. Our results also reveal the presence of at least seven other prominent small RNA species which are associated almost ' exclusively with the cytoskeleton-bound free RNP fraction. In some cases a portion of each scRNA population may be dependent upon the integrity of actin filaments for cytoskeleton binding. While we do not believe that these RNAs are messages (although this conclusion Js tentative) their marked segregation suggests that they are functional. A possible role may be to interact with untranslatable mRNA in the cytoskeleton-bound free RNP fraction. Indeed, there are claims for such functions among small cytoplasmic RNAs. A series of translational control RNAs (tcRNAs) of approx. 4 S have been reported which inhibit translation and may be associated with the free m R N P complexes [49-53]. It is conceivable that the large number of scRNAs associated with the cytoskeleton-bound free RNP fraction could contribute to RNA transport, RNA binding to the cytoskeleton, or RNA stability. A study of subcellular compartmentalization,
especially one involving subcellular fractionation, is predicated on a careful analysis of the fractions generated, including an effort to demonstrate that they are not cross-contaminated. We have presented such an analysis of the four cytoplasmic fractions obtained during this investigation. The results demonstrate both quantitative and qualitative differences in these fractions with respect to RNA and protein content. Over-extraction of the cytoskeleton bound fraction would have resulted in a proportion of the actin mRNA and all of the scRNAs being represented in the soluble cytoplasmic fraction. Since this was not observed, we conclude that over-extraction of the cytoskeleton bound fraction did not occur. The profound effect which salt has upon the p o l y s o m e / R N P extraction from the cytoskeleton suggests that these complexes are bound to cytoskeletal elements by electrostatic interaction. Since we have reduced the KC1 concentration during the initial extraction by 10-fold, and the total ionic strength below physiologic levels, it is possible that the cells were relatively under-extracted due to artifactual protein-protein or protein-nucleic acid interactions at reduced ionic strength. Under-extraction of the soluble cytoplasmic fraction would have resulted in a proportion of those RNAs, isolated specifically in the soluble phase, being also represented in the cytoskeleton bound fraction. A portion (approx. 10%) of the tRNA recovered was found in the cytoskeleton bound fraction and is comparable to the amount of tRNA observed in HeLa cells previously [12]. The small amount of tRNA in the cytoskeleton bound fraction is likely associated with bound polysomes and is not present due to under-extraction. Since others have reported a similar distribution of rRNAs, we believe only a small minority of the cytoskeleton bound RNAs might be attributed to under-extraction of the soluble fraction. Our results and those of others suggest a nonuniform distribution of some mRNAs within the cytoplasm [48,54]. It is therefore important to determine which elements might be involved in mRNA-cytoskeleton interactions. Mammalian cells in culture, derived from different species, release the bound polyribosomes into the soluble fraction when treated with cytochalasin B [11,55,13,56]. In the ascidian egg however, cyto-
224
skeletal binding of specific mRNAs, including actin and histone messages, appears to be dependent upon elements other than actin filaments, since they are insensitive to either cytochalasin or DNAase I treatment [14]. In addition, heat shock protein translation and transport appears to be independent of cytoskeleton interaction [57]. It seems clear that message identity, cell type and species affect profoundly the manner in which any particular message will be compartmentalized in the cytoplasm. Our results demonstrate that two unrelated messages can be independently segregated into separate compartments of the same cytoplasm. The differences we observe in the partitioning of actin and histone H4 m R N A may reflect not only a spatial separation of the two message populations, but also may reflect a fundamental difference in the way the expression of each of these messages is regulated in the cytoplasm. These differences may have implications for the final destination of the proteins produced and in the case of histone mRNAs may be a reflection of their remarkable sensitivity to posttranscriptional regulation [25,36,39,37,58-60]. Histone mRNAs are also characterized by relatively small size and absence of 3'-poly(A) tails. Since at least some H4 mRNAs were found associated with the cytoskeleton, it seems unlikely that poly(A) tails are required or involved in this interaction of mRNAs and the cytoskeleton. Acknowledgements We would like to thank Dr. G. Stein for the generous gift of the cloned human histone H4 gene and Dr. J. Haron for the generous gift of the cloned actin gene used in this study. We would also like to thank Dr. D. Larson and Mr. T. Bagchi for helpful comments and discussion throughout the course of this work. This investigation was supported by grants from the Medical Research Council, NSERC (Canada), National Cancer Institute, and the Canadian Arthritis Society. References 1 Holtzer, H., Croop, J., Dienstman, S., Ishikawa, H. and Somlyo, P. (1975) Proc. Natl. Acad. Sei. USA 72, 513-517
2 Sheets, M.R., Painter, R.G. and Singer, S.J. (1976) Cell Motility in (Goldman, R., Pollard, T. and Rosebaum, J., eds.), pp. 651-664, Cold Spring Harbor Laboratory, New York 3 Albrecht-Buehler, G. (1977) J. Cell Biol. 72, 595-603 4 Wessels, N.K., Spooner, B.S., Ash, J.F., Bradley, M.D., Luduens, M.A., Taylor, E.L., Wrenn, J.T. and Yamada, K.M. (1971) Science 171,135-143 5 Hynes, R.O. and Destree, A.T. (1978) Cell 13, 151-163 6 Heggeness, M.H., Simon, M. and Singer, S.J. (1978) Proc. Natl. Acad. Sci. USA 75, 3863-3866 7 Lazarides, E. (1976) J. Cell Biol. 68, 202-219 8 Fuziwara, K. and Pollard, T. (1976) J. Cell Biol. 71,848-875 9 Albrecht-Buehter, G. (1977) Cell 12, 333-339 10 Lenk, R. and Penman, S. (1979) Cell 16, 289-301 11 lenk, R., Ransom, L., Kaufman, Y. and Penman, S. (1977) Cell 10, 67-78 12 Cervera, M., Dreyfuss, G. and Penman, S. (1981) Cell 23, 113-120 13 Ramaekers, F.C.S., Benedetti, E.L., Dunia, I., Vorstenbosch, P. and Bloemendal, H. (1983) Biochim. Biophys. Acta 740, 441-448 14 Jeffery, W.R. (1984) Devel. Biol. 103, 482-492 15 Jeffery, W.R. (1985) Devel. Biol. 110, 217-229 16 Brinkley, B.R., Fuller, G.M. and Highfield, C.P. (1975) Proc. Natl. Acad. Sci. USA 72, 4981-4985 17 Osborn, M. and Weber, K. (1976) Proc. Natl. Acad. Sci. USA 73, 867-871 18 Osborn, M. and Weber, K. (1977) Exp. Cell Res. 106, 339-349 19 Jorgensen, A.O., Subrahmanyan, L., Turnbull, C. and Kalnins, V.L. (1976) Proc. Natl. Acad. Sci. USA 73, 3192-3196 20 Osborn, M., Franke, W.W. and Weber, K. (1977) Proc. Natl. Acad. Sci. USA 74, 2490-2494 21 Weber, K. and Groeschel-Stewart, U. (1974) Proc. Natl. Acad. Sci. USA 71, 4561-4564 22 Heggeness, M.H., Wang, K. and Singer, S.J. (1977) Proc. Natl. Acad. Sci. USA 74, 3883-3887 23 Howe, J.G. and Hershey, J.W.B. (1984) Cell 37, 85-93 24 Fey, E.G., Wan, K.M. and Penman, S. (1984) J. Cell Biol. 98, 1973-1984 25 Bird, R.C., Jacobs, F.A., Stein, G., Stein, J. and Sells, B.H. (1985) Biochim. Biophys. Acta 824, 209-217 26 Jacobs, F.A., Bird, R.C. and Sells, B.H. (1985) Eur. J. Biochem. 150, 255-263 27 Bird, R.C., Jacobs, F.A., Stein, G., Stein, J. and Sells, B.H. (1985) Proc. Natl. Acad. Sci. USA 82, 6760-6764 28 Maniatis, T., Jeffery, A. and Kleid, D.G. (1975) Proc. Natl. Acad. Sci. USA 72, 1184-1188 29 Thomas, P.S. (1980) Proc. Natl. Acad. Sci. USA 77, 2391-2410 30 Plumb, M., Stein, J. and Stein, G.S. (1983) Nucleic Acids Res. 11, 2391-2410 31 Bonneau, A.-M., darveau, A. and Sonenberg, N. (1985) J. Cell Biol. 100, 1209-1218 32 Walter, P. and Blobel, G. (1982) Nature 299, 691-698 33 Lichtler, A.C., Sierra, F., Clark, S., Wells, J.R.E., Stein, J.L. and Stein, G.S. (1982) Nature 298, 195-198
225 34 Borun, T.W., GabrieUi, F., Ajiro, K., Zweidler, A. and Baglioni, C. (1975) Cell 4, 59-67 35 Delegeane, A.M. and Lee, A.D. (1982) Science 215, 79-81 36 Gallwitz, D. and Mueller, G.C. (1969) J. Biol. Chem. 244, 5947-5952 37 Heintz, N., Sive, H.L. and Roeder, R.G. (1983) Mol. Cell. Biol. 3, 539-550 38 Hereford, L.M., Osley, M.A., Ludwig, J.R. and McLaughlin, C.S. (1981) Cell 24, 367-375 39 Hereford, L.M., Bromley, S. and Osley, M.A. (1982) Cell 30, 305-310 40 Plumb, M., Marashi, F., Green, L., Zimmerman, A., Zimmerman, S., Stein, J. and Stein, G. (1984) Proc. Natl. Acad. Sci. USA 81,434-438 41 Rickles, R., Marashi, R., Sierra, F., Clark, S., Wells, J., Stein, J. and Stein, G. (1982) Proc. Natl. Acad. Sci. USA 79, 749-753 42 Sittman, D.B., Graves, R.A. and Mal-zluff, W.F. (1983) Proc. Natl. Acad. Sci. USA 80, 1849-1853 43 Tarnowka, M.A., Baglioni, C. and Basilico, C. (1978) Cell 15, 163-171 44 Waithe, W.I., Renand, J., Nadeau, P. and Pallotta, D. (1983) Biochemistry 22, 1778-1783 45 Wu, R.S. and Bonner, W.M. (1981) Cell 27, 321-374 46 Wu, R.S., Tsai, S. and Bonner, W.M. (1982) Cell 31, 367-374 47 Bagchi, T., Larson, D.E. and Sells, B.H. (1986) Exp. Cell Res., in the press
48 Phillips, C.R. (1985) Devel. Biol. 109, 299-310 49 McCarthy, T.L., Siegel, E., Mroc-zkowski, B. and Heywood, S.M. (1983) Biochemistry 22, 935-941 50 Pluskal, M.G. and Sarkar, S. (1981) Biochemistry 20, 2048-2055 51 Sarkar, S., Mukherjee, A.K. and Guha, C. (1981) J. Biol. Chem. 256, 5077-5086 52 Bag, J., Hubley, M. and Sells, B. (1980) J. Biol. Chem. 255, 7055-7058 53 Bag, J., Maundrell, K. and Sells, B.H. (1982) FEBS Lett. 143, 163-167 54 Rebagliati, M.R., Weeks, D.L., Harvey, R.P. and Melton, D.A. (1985) Cell 42, 769-777 55 Adams, A., Fey, E.G., Pike, S.F., Taylorson, C.J., White, H.A. and Rabin, B.R. (1983) Biochem. J. 216, 215-226 56 Ornelles, D.A., Fey, E.G. and Penman, S. (1986) Mol. Cell. Biol. 6, 1650-1662 57 Welch, W.J. and Feramisco, J.R. (1985) Mol. Cell. Biol. 5, 1571-1581 58 Bird, R.C., Jacobs, F.A. and Sells, B.H. (1985) Biochem. Cell. Biol. 64, 99-105 59 Baumbach, L.L., Marashi, F., Plumb, M., Stein, G. and Stein, J. (1984) Biochemistry 23, 1618-1625 60 Helms, S., Baumbach, L., Stein, G. and Stein, J. (1984) FEBS Lett. 168, 65-69
215
Elsevier BBA 91611
Cytoskeleton involvement in the distribution of mRNP complexes and small cytoplasmic R N A s R. Curtis Bird a and Bruce H. Sells b " Department of Microbiology, College of Veterinary Medicine. Auburn University, A uburn, A L (U.S.A.) and I, College of Biological Science, University of Guelph, Guelph (Canada)
(Received 18 June 1986) (Revised manuscript received17 September1986) Key words: histone H4 mRNA; mRNA complex; Cytoskeleton;small cytoplasmicRNA These studies were designed to determine whether small cytoplasmic RNAs and two different mRNAs (actin mRNA and histone 114 mRNA) were uniformly distributed among various subcellular compartments. The cytoplasm of HeLa $3 cells was fractionated into four RNA-containing compartments. The RNAs hound to the cytoskeleton were separated from those in the soluble cytoplasmic phase and each RNA fraction was further separated into those hound and those not hound to polyribosomes. The four cytoplasmic RNA fractions were analysed to determine which RNA species were present in each. The 7 S RNAs were found in all cytoplasmic fractions, as were the 5 S and 5.8 S ribosomal RNAs, while transfer RNA was found largely in the soluble fraction devoid of polysomes. On the other hand a group of prominent small cytoplasmic RNAs (scRNAs of 105-348 nucleotides) was isolated from the fraction devoid of polysomes but hound to the cytoskeleton. Actin mRNA was found only in polyrihosomes hound to the cytoskeleton. This mRNA was released into the soluble phase by cytochalasin B treatment, suggesting a dependence upon actin filament integrity for cytoskeletai binding. A significant portion of several scRNAs was also released from the cytoskeleton by cytochalasin B treatment. Analysis of the spatial distribution of histone 1-14 mRNAs, however, revealed a more widely dispersed message. Although most (60%) of the H4 mRNA was associated with polyribosomes in the soluble phase, a significant amount was also recovered in both of the cytoskeleton bound fractions either associated or free of polyrihosome interaction. Treatment with cytochalasin B suggested that only cytoskeleton hound, untranslated H4 mRNA was dependent upon the integrity of actin filaments for cytoskeletal binding.
Introduction
The cytoplasm is a highly structured and compartmentalized region of the cell in which diverse biological processes occur. While attachment of organelles to membranes accounts for much of this compartmentalization, the portion of the cytoplasm outside these membranes has, until recently, been considered largely unstructured. A Correspondence: Dr. B.H. Sells, Collegeof Biological Science, University of Guelph, Guelph, Ontario, Canada, NIG 2Wl.
framework of filamentous elements called the cytoskeleton, which is found throughout the cytoplasm, is thought to contribute to the spatial separation of various processes. The cytoskeleton may also be involved in maintenance and fluctuations in cell shape [1-3], the organization and function of cytoplasmic organelles [4-6] and cell motility [7-9]. During the process of mRNA translation, the translational machinery (the polyribosome) is also thought to be attached to elements of the cytoskeleton [10-16]. While differences in opinion exist concerning the identity of the cytoskeletal
0167-4781/86/$03.50 © 1986 ElsevierSciencePublishers B.V. (BiomedicalDivision)
216
element to which the polyribosomes are bound, there is general agreement that most mRNA populations are bound to the cytoskeleton. The cytoskeleton and its components have been extensively characterized using both biochemical and microscopic techniques. This network of filaments is composed largely of (although it may not be limited to) three major elements. Extensive networks of microtubules [16-18] intermediate filaments [19,10] and actin filaments [21,11,7] have been observed and their component proteins have been characterized. The cytoskeleton also contains non-filamentous components of 'cytoskeleton bound' material which seems to include, at least in some cells, the polyribosomes and translation initiation factors [23]. A eucaryotic cell characteristically contains a wide variety of ribonucleic acid molecules in its cytoplasm, which may include many species of mRNA, tRNA and a group of polynucleotides designated small cytoplasmic RNAs (scRNAs). The mRNAs are normally associated with protein as messenger ribonucleoprotein (mRNP) complexes and exist either in polysome arrays when actively being translated or in free m R N P complexes. The present studies were designed to determine the distribution of these RNAs throughout the cytoplasm and to determine whether specific species or populations are preferentially located on cytoskeletal elements. These studies investigated: (i) whether the distribution of histone H4 mRNA and actin mRNA in free and polysomal mrNAs, between the cytoskeleton and soluble fractions, is the same for each message and (ii) whether scRNAs are uniformly distributed. The results obtained reveal that (i) the distribution differs for each message and (ii) the bulk of scRNAs are located predominantly with the m R N P fraction - free of polysomes - associated with the cytoskeleton. Materials and Methods
Cell culture and labelling HeLa $3 cells were grown in suspension culture (500 ml) to a density of ( 1 - 3 ) . 105 cells/ml in minimal essential medium supplemented with 10% fetal bovine serum at 37°C and 5% CO 2. In cases where cells were to be labelled, identi-
cal cultures were treated with either 2 F C i / m l [3H]lysine or 2 / t C i / m l [3H]uridine for 14 to 16 h. Prior to cell fractionation the cells were washed in ice cold Hepes-buffered saline (0.9% NaC1/10 mM Hepes (pH 7.2)). Cells to be treated with cytochalasin B were first incubated in the presence of 100 t~g/ml emetine for 5 min to preserve polysome structure, and then were treated with 10 /~g/ml cytochalasin B for 30 min [11]. Subsequently, the cell suspensions were processed for subcellular fractionation. Fractionation into cytoskeletal bound and soluble R N P compartments HeLa $3 cells were fractionated into cytoskeletal bound and soluble compartments according to a modification of the procedure of Fey et al. [24]. Suspension cultures of HeLa $3 cells were treated with 50 /~g/ml cycloheximide and harvested by centrifugation at 200 × g (1000 rpm) for 3 min (Beckman TJ-6 fitted with a TH-4 rotor). The cells were gently resuspended in 50 ml ice-cold Hepes-buffered saline with a Pasteur pipette. The cell suspension was centrifuged again ( 2 0 0 x g for 3 min), resuspended in 15 ml of ice-cold lysis buffer number 1 (300 mM sucrose, 20 mM Pipes (pH 6.8), 10 mM KC1, 5 mM MgSO 4, 0.1% (v/v) Triton X-100, 10 mM c-amino caproic acid, 0.005% (w/v) phenylmethylsulphonyl fluoride, 0.1 mM aurintricarboxylic acid, 10 mM dithiothreitol and 5 units/ml placental ribonuclease inhibitor) and very gently mixed for approx. 45 s. Following another centrifugation (200 x g for 3 min) the supernatant was removed and designated 'soluble fraction'. The cytoskeleton bound material, in the pellet, was further extracted by resuspending it in 15 ml of ice-cold lysis buffer number 2 (same composition as lysis buffer number 1 but lacking KC1 and containing 0.25 M ammonium sulfate and Triton X-100 at 0.5 % (v/v)) for 8 min, during the initial part of which the suspension was homogenized by ten passes with a pre-chilled Dounce homogenizer. The homogenized cytoskeletal fraction was centrifuged once again (10000 x g for 20 rain) and the supernatant was designated 'cytoskeleton bound fraction'. The residual pellet contained largely nuclei and insoluble cell debris. Aliquots of these samples (50 #l), in which acid-precipitable radioactive incorpora-
217 tion was to be measured, were dried onto Whatman 540 filter discs and processed as previously described [25]. Once the cells had been fractionated into cytoskeleton-bound and soluble compartments, each fraction was further separated into free and polyribosome-bound RNP complexes. RNA was extracted from the polysomal and free RNP fractions by a modification of a procedure previously described [26,27]. Polyribosomes were prepared by centrifugation of the resulting supernatant (made 100 /~g/ml poly(vinyl sulfate)) through a 30% ( w / v ) sucrose cushion at 178000 x g (1 h at 2°C). The free RNP fractions were prepared from both of the decanted supernatants (both cytoskeletonbound and soluble) by further centrifugation at 178000 x g (16-18 h at 2°C). The RNP pellets from each of the four fractions were dissolved in 2 ml of a solution containing 10 mM Tris-HC1 (pH 7.6), 1 mM N a 2 E D T A and 1% (w/v) sodium dodecyul sulfate (SDS) at room temperature and digested with proteinase K (EC 3.4.21.14) for 1 h at 37°C. To obtain RNA, the solutions were p h e n o l / c h l o r o f o r m extracted, precipitated with 2 vol. of ethanol, and stored overnight at - 2 0 ° C . The four m R N P fractions were designated cytoskeleton-bound polysomes and free m R N P complexes, as well as the soluble polysomes and free m R N P complexes.
Polysome profiles Following preparation of polysome-bound and free RNP pellets from both the cytoskeleton and soluble fractions, polysome profiles were prepared to analyse polysome integrity and size (Bagchi, Larson and Sells, unpublished results). Each pellet was resuspended in 0.3 ml of resuspension buffer (20 mM Tris-HC1 (pH 7.5), 200 mM KC1, 10 mM MgC12 and 0.2 m g / m l heparin) by 10 passes in a 1 ml microhomogenizer on ice. 200 #1 of each sample was loaded onto a 15-50% ( w / v ) linear sucrose gradient (4.8 ml) made up in resuspension buffer and centrifugated in a Beckman 50.1 rotor for 30 min at 40000 rpm and 4°C. Eachgradient was subsequently fractionated and absorbance at 260 nm was measured continuously while maintaining the gradient at approx. 4°C. Electrophoresis and northern blot analysis Total RNA samples (50 /xg/lane) were frac-
tionated on one of two different gel-electrophoresis systems. Samples in which larger molecular weight RNAs were to be resolved were run on 1.5% agarose gels, containing 50 mM Mops (pH 7.0), after denaturation in a solution containing 50% (v/v) formamide, 7% (v/v) formaldehyde and 25 mM Mops (pH 7.0) as previously described [26]. Alternatively, samples in which lower-molecular-weight RNAs were to be resolved were separated on 6% polyacrylamide gels containing 8.3 M urea following denaturation in formamide as previously described [25,27]. RNA gels to be photographed were stained for 15 n'fin in 1 # g / m l ethidium bromide. All gels were subsequently prepared for Northern electroblot analysis to diazotized aminophenylthioether paper by sequential washing in N a O H and then K H z PO 4 as previously described [25,27]. The blots were prehybridized and hybridized as previously described [25,27] to 2.106 cpm of nick-translated plasmid [28] at an approximate specific activity of ( 2 - 3 ) . 108 cpm//~g of plasmid DNA, for 3 days. Excess probe was washed from the blots as described [29] except that each wash lasted 20 min. Autoradiograms were prepared and quantitation was performed by scanning original unsaturated autoradiograms with a microdensitometer. As previously described, the response was linear with respect to the number of micrograms of RNA loaded on the gel [27]. The specific D N A probes employed were a cloned human genomic sequence encoding an entire histone H4 gene [30] and a cloned heterologous cDNA containing most of the a-actin coding sequence. Histone and actin mRNAs were identified in total RNA samples by electrophoretic mobility in comparison to other RNAs of known size (28 S, 18 S, 5.8 S and 5 S rRNAs, 7 S RNA and tRNA) as well as actin mRNAs identified in poly(A) ÷ RNA samples prepared as described previously [27]. Results
Fractionation of HeLa $3 cells into cytoskeleton bound and soluble mRNA fractions Initial studies were performed to define the distribution of m R N P complexes between the cy-
218
toskeleton and soluble fractions, in HeLa $3 cells, employing the lysis conditions of Fey et al. [24]. The cells were gently lysed to solubilize only the plasma membrane. As a consequence of this treatment only those particles which were unattached to the cytoskeletal elements, including the m R N P complexes were liberated into the soluble fraction. After thorough washing of the cellular remnants to remove the remaining soluble material, a highsalt extraction was used to release the bound m R N P complexes from the cytoskeleton. In our hands, the conditions of Fey et al. [24] consistently over-extracted the cytoskeletal fraction of HeLa $3 cells and resulted in an excessively large amount of A260 material (isolated as polysomes) in the soluble fraction (Fig. 1). A modification of this procedure allowed us to obtain a similar distribution of polysomes between the cytoskeletal and soluble fractions as that reported by Fey et al. [24]. The modified lysis procedure involved a reduction in the concentration of both the KC1 and the Triton X-100 concentrations in the first lysis solution. Analysis of the effects of varying KC1 concentration (Fig. la) demonstrates that above 10 mM KC1 the soluble fraction contains increasing amount of polysomal material. At 100 mM KC1 the amount of A260 material (isolated from polysomes) released into the soluble fraction exceeded that found in the cytoskeletal bound material by approx. 100% (69% of the polysomes appeared soluble). In contrast Cervera et al. [12] using HeLa cells reported that no polyribosomes and only 25-27% of the mrNA were recovered in the soluble phase. Reduction of the KCI concentration in our lysis solution, to 10 mM, resulted in only about 24% of the polysomes being recovered in the soluble fraction. The Triton X-100 concentration was also optimized by monitoring the extraction during initial cell lysis. Variations in Triton X-100 concentration between 0.01% and 1.0% produced little change in the amount of material obtained in the soluble phase at 100 mM KC1 (data not shown). At the more moderate 10 mM KC1 concentration, however, as little as 0.1% Triton X-100 in the first lysis solution was sufficient to extract the soluble material fully (Fig. lb). Analysis of polysome profiles of the soluble and cytoskeleton bound fractions at optimal KC1
A
i
0
I
20
,
,
40 KCl
i
GO O0 (mid)
,
I(]0
I00i
I10" "-0
e040.
O-
6
oh
o'.z o~ o~, Triton X..IO0 (%, V/v)
o'.~
Fig. 1. Optimization of KCI and Triton X-100 concentrations for the fractionation of cytoskeleton-bound and soluble cytoplasmic polysomal fractions. (A) HeLa cells were lysed in lysis buffer 1, containing increasing concentrations of KC1 and 0.5% (v/v) Triton X-100, into cytoskeleton-bound and soluble cytoplasmic polysomal fractions. The nucleotide component (RNA) of each fraction was measured as absorbance at 260 nm and relative amount (percentage) in each fraction determined for various KC1 concentrations. (B) HeLa cells were lysed in lysis buffer 1, containing increasing concentrations of Triton X-100 and 10 mM KCI, into cytoskeleton-bound and soluble cytoplasmic polysomal fractions. The relative RNA levels in each fraction were determined for various Triton X-100 concentrations. Each pair of curves was compiled from either three or four independent experiments. Determination of percent distribution of A 260 material at optimal concentration was independently determined six times (maximum sample variation _+6%). Cytoskeleton-bound (O); soluble phase (0).
(10 mM) and Triton X-100 (0.1%) concentrations revealed that 80% of the polysomes were associated with the cytoskeletal fraction (Fig. 2). These results demonstrate that the modified lysis procedure leaves most of the polysomes bound to the cytoskeleton in HeLa cells as previously reported [121. Analysis of the RNA and protein concentrations in each of the soluble and cytoskeleton bound
219
Z
go nO
~
ELETON
oO gO
SEDIMENTATION
Fig. 2. Polysomc profiles from cytoskeleton bound and soluble cytoplasmic polysomal fractions from HeLa cells. Polysome
suspensions were fractionated by centrifugation through a 15-50% (w/v) linear sucrose gradient followed by fractionation and analysis of absorbance at 260 nm.
fractions was performed following long-term labelling with either [3H]lysine or [3H]uridine. Of the labelled macromolecules which were recovered in the cytoskeleton-bound and soluble fractions as acid-precipitable material, 59% of the labelled R N A was localized in the soluble fraction and 41% was bound to the cytoskeleton (Table I). Approx. 70% of the labelled protein was found in the soluble fraction, while 30% was isolated in the cytoskeletal fraction. These results are similar to those reported by Bonneau et al. [31] for African green monkey kidney cells extracted for 1 - 2 rain and by Cervera et al. [12] for H e L a cells.
Qualitative differences in RNAs found in different cytoplasmic fractions We have examined the R N A species present in each of the four cytoplasmic fractions to determine whether there were any qualitative difTABLE I PROPORTION OF RNA AND PROTEIN IN CYTOSKELETON-BOUND AND SOLUBLE CYTOPLASMIC FRACTIONS Mean values for both RNA and protein represent the percentage of acid-precipitable counts observed in each fraction isolated from cells labelled with either [3H]uridine (RNA) or [3H]lysine (protein). Maximum sample variation is noted and each experiment was done twice. Fractions Cytoskeleton Soluble
RNA 41 + 7% 59 + 7%
Protein 70 + 6% 30 + 6%
ferences. Both the cytoskeleton-bound and soluble R N P compartments, as defined above, have been inspected and both the polysomal and free m R N P complexes isolated from each compartment by differential ultracentrifugation. To determine the distribution of scRNAs the R N A s extracted from each of these four cytoplasmic fractions were analysed on 6% polyacrylamide gels containing 8.3 M urea (Fig. 3). The smaller R N A s were thus resolved into four unique populations by isolating the R N A from (1) the cytoskeletal polysomes, (2) the cytoskeletal free RNPs, (3) the soluble polysomes and (4) the soluble free RNPs. Most of the messenger RNAs, 18 S and 28 S ribosomal RNAs were located near the top of the gel or failed to penetrate the gel. Only the smaller R N A species migrated into the middle and lower regions of the gel. The 7 S RNAs, thought to be involved in the signal sequence recognition particle [32] were observed in all of the cytoplasmic fractions as were the 5 S and 5.8 S ribosomal RNAs. The transfer RNAs, while present in all four fractions, were concentrated mainly in the soluble free R N P fraction. Seven additional distinct species of small cytoplasmic RNA, identified as prominent hands, were found only in the cytoskeleton-bound free R N P fraction. These small cytoplasmic RNAs (scRNAs) have approximate sizes of 348 (No. 1), 284 (No. 2), 253 (No. 3), 214 (No. 4), 160 (No. 5), 123 (No. 6), and 105 (No. 7), nucleotides, scRNA1, -4 and -7 were also represented, to a much lesser extent, in the soluble free R N P fraction. The function of these small RNAs is uncertain. They
220 d e v o i d of p o l y ( A ) ( B i r d a n d Sells, u n p u b l i s h e d results). T o d e t e r m i n e w h e t h e r a n y o f the s c R N A s , isol a t e d in the c y t o s k e l e t o n - b o u n d free R N P frac-
Fig. 3. Electrophoresis analysis of low molecular weight RNAs on 6% polyacrylamide gels containing 8.3 M urea. Total RNA (50 ~g/lane) was isolated from the cytoskeleton polysome bound (CP) and free (CF) RNP fractions as well as from the soluble polysome bound (SP) and free (SF) RNP fractions and subjected to denaturing electrophoresis followed b~' staining with ethidium bromide. Right side: RNA species of known size (7 S, 5.8 S, 5 S, 4 S). Left side: scRNA species (numbered 1-7) located largely in the CF fraction. a r e all s m a l l e r t h a n h i s t o n e H 4 m R N A , at 3 6 4 - 4 3 9 n u c l e o t i d e s l o n g in h u m a n [33] a n d in rat [27] cells, a n d d o n o t a p p e a r to b e m e s s a g e s since, in a d d i t i o n to t h e i r s m a l l size, t h e y a r e a p p a r e n t l y
Fig. 4. Electrophoretic analysis of low-molecular-weight RNAs in cytoplasmic fractions isolated from HeLa cells treated with cytochalasin B. RNA was isolated from rapidly proliferating (control) HeLa cells and HeLa cells which had been pretreated with emetine (100/~g/ml) for 5 rain and subsequently treated with cytochalasin B (10 /~g/ml) for 30 min to depolymerize actin filaments. Total R N A (50 /~g/lane) samples, isolated from the cytoskeleton polysome bound (CP) and free (CF) fractions as well as from the soluble polysome bound (SP) and free (SF) RNP fractions were analyzed on a 6% polyacrylamide gel containing 8.3 M urea followed by staining with ethidium bromide. RNA species of known size: (7 S, 5.8 S, 5 S, 4 S). scRNA species (numbered 1-4, 7) located in both the CF and SF fractions, and scRNAs (numbered 5, 6) which are still located largely in the CF fraction.
221 tion, were associated with the actin filament component of the cytoskeleton, H e L a cells were incubated with cytochalasin B prior to fractionation. Several scRNAs were partitioned into both the soluble free and cytoskeleton-bound free R N P fractions after treatment (Fig. 4). These results suggest that a portion of scRNA-1, -2, -3, -4 and -7 may be associated with actin filaments, since cytochalasin B treatment of H e L a cells has been shown to rapidly depolymerize actin filaments (Ref. 11; Bagchi, Larson and Sells, unpublished results).
Spatial distribution of histone H4 and actin mRNAs Polysomes are found principally with the cytoskeleton fraction. Distributions of actin and histone H4 m R N A s were analysed to determine whether two individual m R N A s having different characteristics are distributed in a manner similar to the total population of polysomal mRNAs. Actin m R N A was chosen as an example of a typical poly(A) + m R N A and because its association with the cytoskeleton has been reported in a number of different cell types [14,15,31]. The family of histone H4 m R N A s were examined because these messages are a m o n g the best characterized examples of poly(A)- m R N A s [3446,30]. Among the four cytoplasmic compartments actin m R N A was detected in only the cytoskeleton bound polysomal m R N A fraction (Fig. 5). This result supports the observations by others [14,15,31] and suggests that the cytoskeleton bound fraction was not over extracted during the first lysis step. To determine whether the actin m R N A might be associated with the microfilaments of the cytoskeleton, H e L a cells were incubated with cytochalasin B. Actin-mRNA-containing polysomes were released from the cytoskeleton following treatment (Fig. 5). Since microfilaments in H e L a cells and L6 myoblasts rapidly disassemble in the presence of cytochalasin B [11,47], these results imply that actively translating actin m R N A s are probably dependent upon the integrity of microfilaments either directly or indirectly, for their binding to the cytoskeleton. In this analysis 50/~g of total R N A was loaded in each lane, to permit us to asses both poly(A) ÷ and poly(A)- m R N A species. The limiting
Fig. 5. Northern blot analysis of actin mRNAs from rapidly proliferating HeLa cells, and from HeLa cells which had been pretreated with emetine (100 /~g/ml) for 5 rain and subsequently treated with cytochalasin B (10 /~g/ml) for 30 min. Total RNA (50/~g/lane) samples isolated from the cytoskeleton polysome bound (CP) and free (CF) RNP fractions, and from the soluble polysome bound (SP) and free (SF) RNP fractions, were denatured with formamide and formaldehyde and subjected to electrophoresis on denaturing 1.5% agarose gels, followed by electroblotfing to DPT-celhilose paper and hybridization to labelled actin and H4 DNA probes. An X-ray film was exposed to the blot for about 5 days, at -70 o C, for autoradiographic detection.
amounts of homologous sequence present in each lane made necessary long exposures (approx. 7 days) and a reduction in stringency of the last wash of the blot (conducted at room temperature) prior to autoradiography, to enhance as much as possible hybridization signal. The large mass of non-messenger R N A present as well as the reduction in stringency are responsible for the higher background hybridization in the lanes and the smearing observed in some of the bands. Since actin hybridization intensity was not correlated with the amount of 16 S r R N A observed on the gels, it is unlikely that there was a significant amount of cross-reactivity with this slightly larger R N A species. The density appearing at higher electrophoretic mobility is most likely due to cross-hybridization with unidentified R N A s migrating between the 16 S and 28 S r R N A bands.
222
Fig. 6. Northern blot analysis of histone H4 mRNAs from HeLa cells which were rapidly proliferating, and which had been pretreated with emetine (100 t~g/ml) for 5 min and subsequently treated with cytochalasin B (10 /~g/ml) for 30 min. Total RNA (50 /~g/lane) samples isolated from the cytoskeleton polysome bound (CP) and free (CF) RNP fractions as well as from the soluble polysome bound (SP) and free 9SF) RNP fractions, were denatured by boiling in formamide and subjected to electrophoresis on denaturing 6% polyacrylamide gels containing 8.3 M urea followed by electroblotting onto DPT-cellulose paper and hybridization to a labelled histone H4 DNA probe. X-ray film was exposed to the blot for about 5 days, at - 7 0 o C, for autoradiographic analysis.
TABLE II PROPORTION OF HISTONE H4 mRNA IN CYTOPLASMIC COMPARTMENTS FROM CELLS TREATED WITH CYTOCHALASIN B Mean values from two experiments represent the percentage of total histone H4 mRNA, quantified on autoradiograms, observed in each cytoplasmic fraction. Maximum samf01e variation is noted for two independent experiments. Fractions
Untreated
Cytochalasin Btreated
Cytoskeleton polysomal Cytoskeleton free RNP Soluble polysomal Soluble free RNP
24 + 7 16 +_ 3 58 + 11 2+ 1
30 _+5 3 _+1 59 + 8 8+ 2
A n e x a m i n a t i o n of the d i s t r i b u t i o n of histone H 4 m R N A s a m o n g the four c y t o p l a s m i c fractions revealed a situation m a r k e d l y different from that for actin m R N A (Fig. 6). M o s t of the histone H 4 m R N A (approx. 58%, T a b l e II) was located in the soluble p o l y s o m a l - b o u n d fraction, free of association with the cytoskeleton. Significant a m o u n t s of H 4 m R N A were also recovered in the cytoskelet o n - b o u n d p o l y s o m e s a n d in the cytoskeletalb o u n d free m R N P s . A l m o s t 24% of the H4 m R N A was recovered in the c y t o s k e l e t o n - b o u n d polys o m a l fraction, while 16% was recovered in the c y t o s k e l e t o n - b o u n d free R N P fraction. O n l y trace a m o u n t s were f o u n d in the soluble free R N P fraction. A l t h o u g h m o s t of the p o l y s o m e s were associated with the cytoskeleton fraction, H4 m R N A was f o u n d in all four c y t o p l a s m i c fractions. F u r t h e r m o r e , the largest p o r t i o n of the H 4 m R N A was l o c a t e d in the soluble p o l y s o m a l p h a s e (approx. 60%) r a t h e r t h a n associated with the cytoskeleton. T r e a t m e n t with c y t o c h a l a s i n B, a n d p r e s u m a b l y the d i s a s s e m b l y of microfilaments, p r i o r to cell fractionation, resulted in an a p p a r e n t release of histone H 4 m R N A f r o m only the cytoskeletal free R N P fraction (Fig. 6). The a m o u n t of histone H 4 m R N A in this fraction decreased f r o m 16% to 3% ( T a b l e II). M o s t of the H 4 m R N A released was p a r t i t i o n e d equally between cytoskeletonb o u n d p o l y s o m a l a n d soluble free R N P fractions. T h e d a t a i m p l y that a p p r o x . 13% of the H4 m R N A is d e p e n d e n t u p o n the integrity of actin filaments for c y t o s k e l e t o n b i n d i n g a n d that it is released f r o m the c y t o s k e l e t o n - b o u n d free R N P fraction d u r i n g m i c r o f i l a m e n t disassembly. W h i l e a b o u t half of the H 4 m R N A was released into the solub l e free R N P fraction, the r e m a i n d e r of the cytos k e l e t o n - b o u n d free R N P fraction H 4 m R N A app e a r e d in the c y t o s k e l e t o n - b o u n d p o l y s o m a l fraction. Discussion O u r results d e m o n s t r a t e that different R N A species, i n c l u d i n g various messages a n d small cyt o p l a s m i c R N A s , are differentially d i s t r i b u t e d in the c y t o p l a s m of H e L a cells. W e have o b s e r v e d that actin m R N A behaves m u c h like viral m R N A s [10,11] in that it is largely associated with the
223 polysomes bound to cytoskeleton. In contrast, histone H4 mRNAs are found principally in the soluble polyribosome fraction unassociated with the cytoskeleton. Cytoskeleton association does not appear to be a prerequisite for histone H4 m R N A assembly into polysomes, although some polysome-bound H4 m R N A was also found associated with the cytoskeleton. These results argue against the notion that a message must first become associated with the cytoskeleton before translation can occur [12,31]. Although eucaryotic translation initiation factors have been found associated with the cytoskeleton in HeLa cells [23], our results imply that this association may not be obligatory. Histone H3 mRNAs remain uniformly distributed during periods of major cytoplasmic rearrangement in Xenopus embryos [48], suggesting that they are independent of these mobile elements. Small RNA species also display a non-uniform distribution in the cytoplasm. The largest proportion of tRNA was located in the soluble free RNP fraction unassociated with either ribosomes or the cytoskeleton. This finding is similar to that reported by Cervera et al. [12] where 80% of the tRNA was isolated from the soluble RNA fraction in HeLa cells. Our results also reveal the presence of at least seven other prominent small RNA species which are associated almost ' exclusively with the cytoskeleton-bound free RNP fraction. In some cases a portion of each scRNA population may be dependent upon the integrity of actin filaments for cytoskeleton binding. While we do not believe that these RNAs are messages (although this conclusion Js tentative) their marked segregation suggests that they are functional. A possible role may be to interact with untranslatable mRNA in the cytoskeleton-bound free RNP fraction. Indeed, there are claims for such functions among small cytoplasmic RNAs. A series of translational control RNAs (tcRNAs) of approx. 4 S have been reported which inhibit translation and may be associated with the free m R N P complexes [49-53]. It is conceivable that the large number of scRNAs associated with the cytoskeleton-bound free RNP fraction could contribute to RNA transport, RNA binding to the cytoskeleton, or RNA stability. A study of subcellular compartmentalization,
especially one involving subcellular fractionation, is predicated on a careful analysis of the fractions generated, including an effort to demonstrate that they are not cross-contaminated. We have presented such an analysis of the four cytoplasmic fractions obtained during this investigation. The results demonstrate both quantitative and qualitative differences in these fractions with respect to RNA and protein content. Over-extraction of the cytoskeleton bound fraction would have resulted in a proportion of the actin mRNA and all of the scRNAs being represented in the soluble cytoplasmic fraction. Since this was not observed, we conclude that over-extraction of the cytoskeleton bound fraction did not occur. The profound effect which salt has upon the p o l y s o m e / R N P extraction from the cytoskeleton suggests that these complexes are bound to cytoskeletal elements by electrostatic interaction. Since we have reduced the KC1 concentration during the initial extraction by 10-fold, and the total ionic strength below physiologic levels, it is possible that the cells were relatively under-extracted due to artifactual protein-protein or protein-nucleic acid interactions at reduced ionic strength. Under-extraction of the soluble cytoplasmic fraction would have resulted in a proportion of those RNAs, isolated specifically in the soluble phase, being also represented in the cytoskeleton bound fraction. A portion (approx. 10%) of the tRNA recovered was found in the cytoskeleton bound fraction and is comparable to the amount of tRNA observed in HeLa cells previously [12]. The small amount of tRNA in the cytoskeleton bound fraction is likely associated with bound polysomes and is not present due to under-extraction. Since others have reported a similar distribution of rRNAs, we believe only a small minority of the cytoskeleton bound RNAs might be attributed to under-extraction of the soluble fraction. Our results and those of others suggest a nonuniform distribution of some mRNAs within the cytoplasm [48,54]. It is therefore important to determine which elements might be involved in mRNA-cytoskeleton interactions. Mammalian cells in culture, derived from different species, release the bound polyribosomes into the soluble fraction when treated with cytochalasin B [11,55,13,56]. In the ascidian egg however, cyto-
224
skeletal binding of specific mRNAs, including actin and histone messages, appears to be dependent upon elements other than actin filaments, since they are insensitive to either cytochalasin or DNAase I treatment [14]. In addition, heat shock protein translation and transport appears to be independent of cytoskeleton interaction [57]. It seems clear that message identity, cell type and species affect profoundly the manner in which any particular message will be compartmentalized in the cytoplasm. Our results demonstrate that two unrelated messages can be independently segregated into separate compartments of the same cytoplasm. The differences we observe in the partitioning of actin and histone H4 m R N A may reflect not only a spatial separation of the two message populations, but also may reflect a fundamental difference in the way the expression of each of these messages is regulated in the cytoplasm. These differences may have implications for the final destination of the proteins produced and in the case of histone mRNAs may be a reflection of their remarkable sensitivity to posttranscriptional regulation [25,36,39,37,58-60]. Histone mRNAs are also characterized by relatively small size and absence of 3'-poly(A) tails. Since at least some H4 mRNAs were found associated with the cytoskeleton, it seems unlikely that poly(A) tails are required or involved in this interaction of mRNAs and the cytoskeleton. Acknowledgements We would like to thank Dr. G. Stein for the generous gift of the cloned human histone H4 gene and Dr. J. Haron for the generous gift of the cloned actin gene used in this study. We would also like to thank Dr. D. Larson and Mr. T. Bagchi for helpful comments and discussion throughout the course of this work. This investigation was supported by grants from the Medical Research Council, NSERC (Canada), National Cancer Institute, and the Canadian Arthritis Society. References 1 Holtzer, H., Croop, J., Dienstman, S., Ishikawa, H. and Somlyo, P. (1975) Proc. Natl. Acad. Sei. USA 72, 513-517
2 Sheets, M.R., Painter, R.G. and Singer, S.J. (1976) Cell Motility in (Goldman, R., Pollard, T. and Rosebaum, J., eds.), pp. 651-664, Cold Spring Harbor Laboratory, New York 3 Albrecht-Buehler, G. (1977) J. Cell Biol. 72, 595-603 4 Wessels, N.K., Spooner, B.S., Ash, J.F., Bradley, M.D., Luduens, M.A., Taylor, E.L., Wrenn, J.T. and Yamada, K.M. (1971) Science 171,135-143 5 Hynes, R.O. and Destree, A.T. (1978) Cell 13, 151-163 6 Heggeness, M.H., Simon, M. and Singer, S.J. (1978) Proc. Natl. Acad. Sci. USA 75, 3863-3866 7 Lazarides, E. (1976) J. Cell Biol. 68, 202-219 8 Fuziwara, K. and Pollard, T. (1976) J. Cell Biol. 71,848-875 9 Albrecht-Buehter, G. (1977) Cell 12, 333-339 10 Lenk, R. and Penman, S. (1979) Cell 16, 289-301 11 lenk, R., Ransom, L., Kaufman, Y. and Penman, S. (1977) Cell 10, 67-78 12 Cervera, M., Dreyfuss, G. and Penman, S. (1981) Cell 23, 113-120 13 Ramaekers, F.C.S., Benedetti, E.L., Dunia, I., Vorstenbosch, P. and Bloemendal, H. (1983) Biochim. Biophys. Acta 740, 441-448 14 Jeffery, W.R. (1984) Devel. Biol. 103, 482-492 15 Jeffery, W.R. (1985) Devel. Biol. 110, 217-229 16 Brinkley, B.R., Fuller, G.M. and Highfield, C.P. (1975) Proc. Natl. Acad. Sci. USA 72, 4981-4985 17 Osborn, M. and Weber, K. (1976) Proc. Natl. Acad. Sci. USA 73, 867-871 18 Osborn, M. and Weber, K. (1977) Exp. Cell Res. 106, 339-349 19 Jorgensen, A.O., Subrahmanyan, L., Turnbull, C. and Kalnins, V.L. (1976) Proc. Natl. Acad. Sci. USA 73, 3192-3196 20 Osborn, M., Franke, W.W. and Weber, K. (1977) Proc. Natl. Acad. Sci. USA 74, 2490-2494 21 Weber, K. and Groeschel-Stewart, U. (1974) Proc. Natl. Acad. Sci. USA 71, 4561-4564 22 Heggeness, M.H., Wang, K. and Singer, S.J. (1977) Proc. Natl. Acad. Sci. USA 74, 3883-3887 23 Howe, J.G. and Hershey, J.W.B. (1984) Cell 37, 85-93 24 Fey, E.G., Wan, K.M. and Penman, S. (1984) J. Cell Biol. 98, 1973-1984 25 Bird, R.C., Jacobs, F.A., Stein, G., Stein, J. and Sells, B.H. (1985) Biochim. Biophys. Acta 824, 209-217 26 Jacobs, F.A., Bird, R.C. and Sells, B.H. (1985) Eur. J. Biochem. 150, 255-263 27 Bird, R.C., Jacobs, F.A., Stein, G., Stein, J. and Sells, B.H. (1985) Proc. Natl. Acad. Sci. USA 82, 6760-6764 28 Maniatis, T., Jeffery, A. and Kleid, D.G. (1975) Proc. Natl. Acad. Sci. USA 72, 1184-1188 29 Thomas, P.S. (1980) Proc. Natl. Acad. Sci. USA 77, 2391-2410 30 Plumb, M., Stein, J. and Stein, G.S. (1983) Nucleic Acids Res. 11, 2391-2410 31 Bonneau, A.-M., darveau, A. and Sonenberg, N. (1985) J. Cell Biol. 100, 1209-1218 32 Walter, P. and Blobel, G. (1982) Nature 299, 691-698 33 Lichtler, A.C., Sierra, F., Clark, S., Wells, J.R.E., Stein, J.L. and Stein, G.S. (1982) Nature 298, 195-198
225 34 Borun, T.W., GabrieUi, F., Ajiro, K., Zweidler, A. and Baglioni, C. (1975) Cell 4, 59-67 35 Delegeane, A.M. and Lee, A.D. (1982) Science 215, 79-81 36 Gallwitz, D. and Mueller, G.C. (1969) J. Biol. Chem. 244, 5947-5952 37 Heintz, N., Sive, H.L. and Roeder, R.G. (1983) Mol. Cell. Biol. 3, 539-550 38 Hereford, L.M., Osley, M.A., Ludwig, J.R. and McLaughlin, C.S. (1981) Cell 24, 367-375 39 Hereford, L.M., Bromley, S. and Osley, M.A. (1982) Cell 30, 305-310 40 Plumb, M., Marashi, F., Green, L., Zimmerman, A., Zimmerman, S., Stein, J. and Stein, G. (1984) Proc. Natl. Acad. Sci. USA 81,434-438 41 Rickles, R., Marashi, R., Sierra, F., Clark, S., Wells, J., Stein, J. and Stein, G. (1982) Proc. Natl. Acad. Sci. USA 79, 749-753 42 Sittman, D.B., Graves, R.A. and Mal-zluff, W.F. (1983) Proc. Natl. Acad. Sci. USA 80, 1849-1853 43 Tarnowka, M.A., Baglioni, C. and Basilico, C. (1978) Cell 15, 163-171 44 Waithe, W.I., Renand, J., Nadeau, P. and Pallotta, D. (1983) Biochemistry 22, 1778-1783 45 Wu, R.S. and Bonner, W.M. (1981) Cell 27, 321-374 46 Wu, R.S., Tsai, S. and Bonner, W.M. (1982) Cell 31, 367-374 47 Bagchi, T., Larson, D.E. and Sells, B.H. (1986) Exp. Cell Res., in the press
48 Phillips, C.R. (1985) Devel. Biol. 109, 299-310 49 McCarthy, T.L., Siegel, E., Mroc-zkowski, B. and Heywood, S.M. (1983) Biochemistry 22, 935-941 50 Pluskal, M.G. and Sarkar, S. (1981) Biochemistry 20, 2048-2055 51 Sarkar, S., Mukherjee, A.K. and Guha, C. (1981) J. Biol. Chem. 256, 5077-5086 52 Bag, J., Hubley, M. and Sells, B. (1980) J. Biol. Chem. 255, 7055-7058 53 Bag, J., Maundrell, K. and Sells, B.H. (1982) FEBS Lett. 143, 163-167 54 Rebagliati, M.R., Weeks, D.L., Harvey, R.P. and Melton, D.A. (1985) Cell 42, 769-777 55 Adams, A., Fey, E.G., Pike, S.F., Taylorson, C.J., White, H.A. and Rabin, B.R. (1983) Biochem. J. 216, 215-226 56 Ornelles, D.A., Fey, E.G. and Penman, S. (1986) Mol. Cell. Biol. 6, 1650-1662 57 Welch, W.J. and Feramisco, J.R. (1985) Mol. Cell. Biol. 5, 1571-1581 58 Bird, R.C., Jacobs, F.A. and Sells, B.H. (1985) Biochem. Cell. Biol. 64, 99-105 59 Baumbach, L.L., Marashi, F., Plumb, M., Stein, G. and Stein, J. (1984) Biochemistry 23, 1618-1625 60 Helms, S., Baumbach, L., Stein, G. and Stein, J. (1984) FEBS Lett. 168, 65-69