| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Laboratorio di Biologia Cellulare e Neurobiologia, Dipartimento di Biologia Animale, Università di Pavia, Italia; and
Istituto di Genetica Molecolare del CNR, Sezione di Istochimica e Citometria, Università di Pavia, Italia
1Correspondence: Laboratorio di Biologia Cellulare e Neurobiologia, Dipartimento di Biologia Animale, Università di Pavia, Piazza Botta 10, Pavia 27100, Italy. E-mail: marcobig{at}unipv.it
ABSTRACT
We have followed at high resolution the ribosomal protein S6 entering the nucleus of HeLa cells, stopping in some (not all) interchromatin granules clusters and reaching, via Cajal bodies, the nucleolus. There, S6 is assembled with other proteins and rRNA into small ribosomal subunit (SSU), released in the nucleoplasm, and exported through the nuclear pores. We show for the first time the spatial association of nuclear myosin I (NMI) and actin with the SSU already at the nucleolar periphery to the nuclear pore. A blockade of NMI or actin induces an upstream accumulation of the S6 protein en route to the nucleolus, and a temperature lower than normal influences RNA export. Our data strongly suggest a functional relationship of SSU with NMI and actin. In our hypothesis, an active, myosin-driven movement of the small ribosomal subunit can be responsible for the export of 
10% of SSUs. This hypothesis is supported by ultrastructural, immunofluorescence, and biochemical analyses. The currently accepted model for the subunit release suggests a diffusive, temperature-independent mechanism. However, the advantage of the double mechanism would assure that the movement of a part of the subunits could be modulated, increased, or decreased according to the needs of the cell at a specific moment in the cell cycle.—Cisterna, B., Necchi, D., Prosperi, E., Biggiogera, M. Small ribosomal subunits associate with nuclear myosin and actin in transit to the nuclear pores.
Key Words: ribosomal protein S6 • ultrastructural immunocytochemistry • ribosome movement • nucleocytoplasmic transport • ribosome biogenesis
RIBOSOME BIOGENESIS TAKES place in the nucleolus, where the ribosomal proteins recruited from the cytoplasm are associated with rRNA (1
, 2)
. In this organelle, rRNA transcription occurs at the level of the dense fibrillar component (DFC) (3)
, and the pre-rRNA molecules are associated with proteins into processomes (4)
.
However, ribosomal proteins, later to be found in the ribosomal subunits, are very early associated with prerRNA; some have been shown to bind within the DFC (3
, 5)
, but the majority are present in the granular component (GC) (6)
. In the latter nucleolar component, the subunit assembly takes place before both the small subunit (SSU) and the large subunit (LSU) are released to the nucleoplasm for export. The ribosomal proteins in this way complete a circle, most of which is inside the nucleus. So far, the route followed by these proteins as well as the modality of movement of the subunits has been only partially clarified.
If we consider the complexity of the cell and the intricacy of the relationship between nucleus and nucleolus, it is difficult to exclude a priori that ribosomal proteins move without any interaction with nuclear structure on their way to the nucleolus. In this paper, we have tried to follow the nuclear movements of a SSU protein, S6, through the different nuclear compartments in HeLa cells in order to verify the possible involvement of nuclear constituents. Moreover, we have also used S6 as a marker of the SSU itself. It is not clear, in fact, whether movements of the ribosomal subunits depend purely on diffusion or on active processes (7)
, in particular for the SSU (8)
. It has recently been shown that a diffusive mechanism seems to be responsible for the export of the LSU and that an active process may account for only a negligible part of export (9
, 10)
. Here we have followed a different strategy, and we present a hypothesis of a different export mechanism for SSU.
MATERIALS AND METHODS
HeLa cells were grown in Dulbecco’s modified essential medium (DMEM) supplemented with 10% heat-inactivated FBS, 20 mM glutamine, and 100 U each of streptomycin and penicillin.
The cells were prepared for immmunofluorescence, biochemical analysis, and mainly for immunoelectron microscopy as follows.
Electron microscopy (EM)
For electron microscope immunocytochemistry, the cells were trypsinized, fixed with 4% paraformaldehyde in DMEM at 4°C for 2 h, and rinsed in phosphate buffer (pH 7.2). The specimens were then embedded in 2% low gelling agarose and placed into 0.5 M NH4Cl solution in buffer for 30 min at 4°C to block free aldehyde groups, dehydrated in ethanol at room temperature, and embedded in LR white resin.
Ultrathin sections on formvar carbon-coated nickel grids were incubated on a drop of normal goat serum (NGS) diluted 1:100 in PBS for 3 min. Incubation with the antibodies (Ab), diluted in PBS containing 0.05% Tween and 0.1% BSA, was performed at 4°C for 17 h. The following antibodies were used: polyclonal antibodies (pAb) against S6 (ab12864, Abcam) at 1:50, nuclear myosin Iß (NMI, Sigma, St. Louis, MO, USA) at 1:30, and monoclonal antibody (mAb) against actin (AC-40, Sigma) at 1:20. After rinsing with PBS-Tween and PBS, the grids were incubated with NGS as above. The grids were incubated with the appropriate secondary antibody (Ab) (Jackson Lab., Bar Harbor, ME, USA) coupled with colloidal gold, diluted 1:20 in PBS. All incubations were carried out for 30 min at room temperature. As a control, some grids were floated on the incubation mixture without the primary Ab, treated as above, and incubated with the appropriate secondary Ab. In the case of triple labeling, grids without formvar were used; on one side a double labeling with a monoclonal and a polyclonal Ab was carried out (anti-actin and anti-S6), then the appropriated secondary antibodies, conjugated with 6 and 12 nm gold, were applied. All grids were rinsed with PBS and distilled water. After drying for 30–40 min at room temperature, the second polyclonal (anti-NMI) was applied on the other side of the grid and the same protocol was followed. The secondary Ab was conjugated with 18 nm gold.
All the grids were rinsed with PBS and distilled water, and finally stained with the EDTA regressive technique of Bernhard (11)
preferential for RNP-containing nuclear components.
Some grids with the sections were floated onto 0.2 M terbium citrate, prepared according to Biggiogera and Fakan (12)
, for 1 h at room temperature, rapidly rinsed with water, and dried. Stained specimens were observed with a Zeiss EM900 electron microscope equipped with a 30 µm objective aperture and operating at 80 kV.
Fluorescence immunocytochemistry
Cells grown on coverslips were fixed with freshly made 2% formaldehyde in PBS for 20 min at room temperature and postfixed in 70% EtOH at –20°C. Cells were then washed in PBS and double labeled with anti-S6 (1:100) and anti-P105 PANA (1:3), followed by Alexa 488 and 594 conjugated antibodies. The cells were photographed using an Olympus BX51 fluorescence microscope. Alternatively, the slides were analyzed on a Leica DM IRBE confocal microscope.
Treatment with anti-myosin Iß antibodies and anti-actin
The culture medium was removed and replaced with fresh medium containing 40 µg/ml of L-
-lysophosphatidylcholine (LPC from yolk egg, Sigma). This LPC concentration allows the formation of transitory gaps in the cytoplasmic membrane, and the cells still adhere to the substrate (13)
. Previous experiments demonstrated that incorporation of antibodies in these conditions does not damage the cell both in terms of structural preservation and functionality (e.g., incorporation of BrU; see Fig. 1
A). The flasks were kept for 10 min at room temperature, then 500 µl of the antinuclear myosin Iß (at a 1:50 dilution) was added and the cells were kept at 37°C for 2 or 6 h. The efficiency of incorporation of anti-NMI antibodies was evaluated by counting, in a double blind test, thenumber of S6-positive spots on a sample of 100 cells both in control and 2 h after anti-NMI incorporation. The results were evaluated by the Student’s t test.
|
In parallel, other coverslips were also incubated with anti-actin. After treatment with LPC, anti-actin antibodies were added to the medium at a concentration of 1:50.
Samples of control cells (not shown) and samples of cells treated with anti-myosin Iß (not shown) were also incubated with an antifibrillarin (ab5821, Abcam) at a 1:200 dilution. Fibrillarin represents a good control since it has a nucleolar localization without being structurally related to the ribosomal subunits. After the blockade of NMI, no differences were observed in the distribution of fibrillarin.
Flow cytometry
To check for the influence of temperature on the export of the small ribosomal subunit from the nucleolus, HeLa cells were grown in flasks kept under A) normal culture conditions (37°C, 5% CO2); B) long-term exposure to low temperature (18 h at 23°C, 5% CO2); C) short-term exposure to low temperature (23°C for 90 min, 5% CO2). The cells were tripsynized, centrifuged, and each sample divided into two aliquots. The first was immediately fixed in 70% cold ethanol, then stained with a 500 nM solution of SYTO RNASelect (S32703, Molecular Probes, Eugene, OR, USA) in PBS at room temperature for 20 min. The second aliquot from the three samples was lysed according to Riva and co-workers (14)
in order to remove the soluble RNA. Briefly, the cells were rinsed in PBS, then treated with a hypotonic buffer (10 mM Tris-HCl pH 7.4, 2.5 mM MgCl2, 0.5% Nonidet P-40, and protease/phosphatase inhibitors) for 8 min at 4°C, rinsed in Tris-HCl, then fixed and stained as above. As a control of the specificity of the RNA stain, additional cell aliquots were digested with either 4 mg/ml RNase A for 20 min at 37°C or with RNase A, as before, combined with 20 U DNase I for 20 min at 37°C. The cells were then analyzed with a Epics XL flow cytometer.
Immunoprecipitation
To analyze the interaction of nuclear myosin and actin with S6, cells were lysed in hypotonic lysis buffer (10 mM Tris-HCl at pH 7.4, 2.5 mM MgCl2, 0.5% Nonidet P-40, 1 mM DTT, 1 mM PMSF, 0.2 mM sodium vanadate, 0.5 µM okadaic acid (OA), 100 µl/107cells of protease inhibitor cocktail), then centrifuged at 300 g for 1 min. The supernatant containing the detergent-soluble protein fraction (cytoplasmic fraction, C) was collected. The pellet was washed twice in isotonic buffer (10 mM Tris-HCl at pH 7.4, 150 mM NaCl, 1 mM DTT, 1 mM PMSF, 0.2 mM sodium vanadate, 0.5 µM OA, 100 µl/107cells of protease inhibitor). The pellet was finally resuspended in the same buffer and sonicated to obtain the nuclear fraction, N (15)
. C and N fractions were incubated for 3 h at 4°C with anti-S6 and protein A-Sepharose beads, centrifuged, and the pellet was washed three times in isotonic buffer. Samples were resuspended in loading buffer and analyzed by Western blot for anti-S6 (1:1000), antimyosin Iß (1:2000), anti-actin (1:500), or anti-histone H3 (1:10000), followed by peroxidase-conjugated secondary antibodies.
Statistical analysis
The data for the statistical analysis were obtained by EM. We evaluated the S6 immunolabeling in the nucleoplasmic area of a sample of 50 nuclei, considering the signal for the ribosomal protein labeled by a gold grain (S6), the signal for S6 associated with RNP particle stained by EDTA (representing SSU), and the number of SSU colocalizing with either nuclear myosin I (SSU-NMI) or actin (SSU-ACT). The number of gold grains labeling S6, as above, was taken as an indication of the amount of the proteins entering the nucleus, and the percentages of S6, SSU, and SSU-NMI or SSU-ACT were calculated as fractions of this total.
RESULTS
In Fig. 1
, an example at low magnification of cells stained with EDTA (Fig. 1A
) or terbium (Fig. 1B
) after labeling, is shown. In both cases, the condensed chromatin at the periphery of the cell nucleus is in less contrast than the RNP component. Labeling in Fig. 1A
refers to Br-uridine incorporation after LPC treatment; the morphology of the cell structures is well preserved and not damaged by the treatment. A double labeling for S6 and NMI is presented in Fig. 1B
.
At high resolution we can visualize S6 entering the nucleus. In Fig. 2
A, B, S6 labeling at the nuclear pore level is not associated with selectively stained RNA molecules. Consequently, S6 is not part of the small subunit and the labeling is likely to represent single proteins entering the nucleus. S6 moves in the nucleoplasm and is thereafter found in some (not all) interchromatin granules (IG) clusters (Fig. 2C
), especially at the periphery of this nuclear compartment. As shown by double immunofluorescence for S6 and the IG marker protein p105-PANA (16)
, the IG clusters found to be labeled are, in fact, a minority (see
Fig. 4B
), and the colocalizations are only partial. Another nuclear structure involved in the nuclear movements of S6 is represented by the Cajal bodies (CBs). We can confirm by EM that CBs do contain S6, as already described (17)
by immunofluorescence. S6, in fact, can be visualized on CBs both in the vicinity of IG (Fig. 2D
) and approaching the nucleolus (Fig. 2E
). In the nucleolus, S6 is first found on the DFC. Then, when associated with other SSU proteins, S6 becomes a part of the beak, a substructure of the head of the small subunit (18)
. In the granular cortex at the periphery of the nucleolus (Fig. 2F
), S6 is found to colocalize with nuclear myosin I (NMI). In our cell model, this represents the earliest association of NMI and ribosomal proteins. Later on, the subunits are released from the nucleolus. In the nucleoplasm, if associated with terbium-stained RNA, S6 is obviously part of the released SSU (Fig. 2G, H
); in this form, S6 associated with RNA often colocalizes with NMI (Fig. 2I, J
). When a triple labeling is performed on the two sides of the thin section, a close association of S6 with NMI and actin can be seen in contact with an RNP particle that probably corresponds to the SSU (Fig. 2K
). The small subunits then reach the nuclear pore, where they can often be localized by a doublet (rarely a triplet) of gold grains (Fig. 2L
). Note the presence of RNA associated with S6 (Fig. 2M, N
) during the transit through the pore, and colocalization of S6 and NMI at the nuclear pore level (Fig. 2O, P
). We did not find any association NMI/actin without the presence of an RNP particle.
|
|
|
To give a semiquantitative analysis of S6 immunolabeling, we have evaluated the nucleoplasmic density of the signal for S6; the results are shown in Table 1
. The signal for S6 indicates the protein entering the nucleus, whereas when the labeling is accompanied by an RNP particle, the signal is likely to represent an SSU. Double labeling for S6-NMI or S6-actin is always associated with an RNP particle. As expected, the amount of SSUs exported by an active motor proteins mechanism constitutes only 10% (see Table 1
) of the total export for the small subunits.
|
To support the hypothesis of the association of S6 with NMI and actin, we carried out an immunoprecipitation assay (IP) on HeLa cell nuclear and cytoplasmic extract with the anti-S6 Ab (Fig. 3
). The IP nuclear fraction contains S6 as well as NMI and actin, suggesting an interaction (direct or indirect) between S6 and these two proteins. An S6 signal in the IP cytoplasmic fraction is absent, probably because the S6 protein epitope is masked in the complete ribosome, where S6 is located at the interface between the small and the large subunits. However, S6 protein, when in denaturing conditions, is detected in both cytoplasmic and nuclear fractions before IP.
We then tested in living cells whether the Ab-mediated blockade of NMI via LPC treatment could induce a redistribution of S6.
In the nucleoplasm of control cells, S6 labeling is concentrated in a few spots (Fig. 4
A), most of which are recognizable as CBs at the EM level. The signal partially overlaps on some IG clusters (Fig. 4B
, arrow) after double staining with the IG marker protein p105-PANA (16)
. Two hours after incorporation ofanti-NMI antibodies, the number of S6-positive (Fig. 4C
) spots increases, indicating an accumulation of the S6 protein en route to the nucleolus (Fig. 4D
, arrows). The double labeling S6/PANA shows that the S6-positive spots are localized near and often in contact with the IG clusters and the partial colocalizations remain in low number. These results are confirmed by confocal microscopy (not shown). The number of spots returns to the level of the control after 6 h (Fig. 4E
). Statistical analysis with the Student’s t test indicated that a significant difference (P<0.001) exists between the control condition and the anti-NMI 2 h condition; in fact, in the nuclei of treated cells there is a general increase in the S6 foci, indicating that most of the cells internalized anti-NMI antibodies. After LPC treatment alone, the number of S6-positive foci is similar to controls (not shown). Moreover, this treatment does not seem to alter the distribution of the target antigens, as seen after antifibrillarin incorporation (see Material and Methods).
We obtained similar results in cells after incorporation of the anti-actin antibodies. An increase of the S6 nucleoplasmic spots after 2 h of actin blockade was observed (not shown).
Flow cytometric analysis has confirmed the trend obtained from EM. The rationale behind this experiment is that if all the subunit export is dependent on diffusion, there should not be any important influence of temperature lowering, whereas if active movements are involved, then more RNA should remain in the nucleolus. In particular, considering the A, B, and C samples from nonlysed cells, the graph in Fig. 5
shows that a 10% increase in the RNA content can be seen at 23°C (sample B). In lysed cells, the amount of retained RNA is even greater, probably due to the fact that in these conditions most of the "soluble" RNA is removed from nuclei, thus leaving a cleaner background for the measurements. In both cases (total and lysed), lowering the temperature for 90 min has little or no influence on RNA content.
|
DISCUSSION
We have analyzed the movements of a ribosomal protein from the cytoplasm to the nucleolus and of the same protein within the SSU when exported toward the pores.
Our results show that, along the S6 pathway from the cytoplasm to the nucleoplasm, then to the nucleolus, at least two nuclear structures participate as transient storage sites, IG and CBs.
Ribosomal protein S6, once inside the nucleus, can stop in some IG clusters, but not in all. This is significant, since IG clusters (or speckles, as described in ref 19
) may play a role in the storage of the ribosomal proteins during the nucleoplasmic transit toward the nucleolus, and have been shown to contain several proteins from both subunits (20)
.
The presence of labeling for S6 within IG could suggest the possibility that SSU itself could be stored in these structures. However, this is unlikely since only ribosomal proteins have been detected in the IG (19)
, not rRNA, and consequently no ribosomal subunits. We cannot exclude the possibility that this compartment may also be involved in some protein modification before they reach the nucleolus.
The second step in S6 nuclear movements involves CBs. These organelles are often found labeled in the proximity of one or more clusters of IG. Therefore, CBs are likely to represent an S6 carrier from the IG clusters to the nucleolus, where they are known to associate (17
, 21)
and sequester/release proteins.
We believe that other nuclear compartments could be involved as intermediate steps in the S6 route. This idea seems to be confirmed by the blockade of NMI or actin: the number of S6-positive spots increases, suggesting an upstream accumulation of the S6 protein. After double labeling for S6 and PANA, we observed that if some colocalizations are present, they are only partial. Therefore, we can suggest that neither is there an S6 accumulation in IG cluster, nor that such an increase of CBs is possible in the nucleus. In the control cells as analyzed by EM, S6 labeling can sometimes be found associated with dense fibrillar structures that correspond to the previously described interchromatin granule-associated zones (IGAZ, 22
) (not shown). In light of this, we could suggest that the blockade of the two motor proteins may produce a transient accumulation of S6 protein in such a structure.
The third localization is the nucleolus; in this organelle, gold labeling for S6 can already be found in the DFC, (i.e., at the site of rRNA transcription), then in the GC. Thereafter, ribosomal subunits are released in the nucleoplasm to be exported. It is not yet clear whether the movement toward the nuclear pore is always mediated by an anomalous diffusion, as described by Politz et al. (9
, 10)
, for the LSU or involves another mechanism.
For the export of at least a part of the mature SSUs, we may hypothesize an active mechanism. We have found, at EM, an early association of S6 and NMI in the GC; when the SSU is released in the nucleoplasm, this association can still be visualized in the presence of actin and single RNA molecules. This finding is highly suggestive of an interaction of the motor proteins with SSU.
It must be underlined here that an important role for NMI (and actin) in transcription has recently been proposed and demonstrated (23
, 24
, 25)
. Both the transcript of Pol II (in the form of perichromatin fibrils, 19
) and Pol I (as preribosomal subunits) are smaller or in the range of 100 nm; according to a recent paper (26)
, active movement rather than diffusion should be the best choice for structures of that size.
To better understand the problem, it is worth recalling here that it has been shown (9
,10)
that the majority of 60S subunits move in an energy-independent manner, diffusing toward the nuclear pores. The authors, on the other hand, do not exclude the idea that a subset of the subunits may move by an energy-consuming mechanism. Politz and co-workers (9)
have performed the same tests at 23°C and 37°C, expecting to observe a reduced transport if an active mechanism were present. No differences between the apparent diffusion coefficients were found, demonstrating that a predominant active process was unlikely for the LSU. However, concerning another type of transcript, the authors pointed out that almost the double amount of poly(A)RNA remained in the nucleus at 23°C compared with 37°C samples, suggesting that a part of the poly(A) transport would require an active, energy-dependent process (9)
. This also seems to be the case for the 40S subunit: after several hours at 23°C, at least 10% of RNA is retained in the nucleus (as shown after lysis of the cells), mainly in the nucleolus. It should be pointed out that a lowering of temperature immediately before the RNA staining is almost ineffective, which could help explain the discrepancies between our data and those published by Politz and co-workers (9)
.
We have shown here for the first time the direct or indirect interaction among S6, NMI, and actin (indicated by our results in IP assay and with the blockade of NMI or actin) and the simultaneous occurrence of these proteins with single, selectively stained RNA molecules present within SSUs. As expected, the semiquantitative analysis at EM has confirmed that a subset of SSUs is exported with an energy-consuming process. In fact, we have observed that 10% of the SSUs colocalized with NMI or actin. On the basis of these data, we suggest that SSUs can be helped by motor proteins to leave the nucleolus till the nuclear pore. If so, this mechanism would probably concern the movement of a minority of SSU. Our findings agree with the lower occurrence of 40S in the nucleoplasmic fraction, described by Udem and Warner (27)
, as well as with the EM data demonstrating that the two subunits are exported by distinct pathway or by different kinetics (28)
. A graphic interpretation of these data is given in Fig. 6
.
|
Our results need not, however, be in conflict with the diffusion mechanism (9
, 10)
. One could hypothesize that both mechanisms, diffusive and active, may actually coexist for the subunit movements; the diffusive one, however, could guarantee the transport of the majority of the subunits, the active one a subset only. The advantage of the double mechanism is to assure that although most subunits would move without consuming energy, the movement of the remnant could be modulated—that is, increased or decreased according to the cell needs at a precise moment of the cell cycle.
ACKNOWLEDGMENTS
The authors are indebted to Drs. Y. Osheim and C. Pellicciari for critical reading of the manuscript and to Ms. Paola Veneroni for the cell cultures. This work is supported by Fondo di Ateneo per la Ricerca (2005) and Prin 2002. Confocal microscopy was performed at the Centro Grandi Strumenti of the University of Pavia.
Received for publication November 21, 2005. Accepted for publication April 26, 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
