In vitro ribosomal ribonucleoprotein transport. Temperature-induced "graded unlocking" of nuclei.

We examine the effect of temperature on the export of ribosomal precursor particles from nuclei isolated from Tetrahymena. A new phenomenon is observed. Temperature does affect not only the export rate, but also the maximal portion of particles exported. At 8 degrees C, for example, the export kinetics reveals a significantly lower saturation plateau which does not equilibrate with the higher plateau at 28 degrees C even after 3 h. This nonequilibration is not due to (i) a different physical quality of the exported particles, (ii) a degradation of the nuclear rRNA, (iii) a backward import of exported particles into nuclei, (iv) an irreversible inactivation of potentially transportable nuclear ribosomal ribonucleoprotein (rRNP) particles, or (v) a thermodynamic equilibrium between transportable rRNP particles associated with nuclei and those exported from nuclei. We conclude, therefore, that potentially transportable rRNP particles are somehow "locked" in nuclei at low temperature and temperature raising induces a "graded unlocking."

In eukaryotes, the transport of the different RNP' species from nucleus to cytoplasm remains poorly understood to date despite its great regulatory potential for gene expression. The mechanisms underlying RNP transport can be reasonably explored in cell-free systems, i.e. in isolated nuclei. Only under such conditions is the virtual transport process decoupled from transcription and processing of RNA as well as from nucleocytoplasmic feedback mechanisms (for recent reviews, see . In recent years, several groups tried to get more insight into the transport mechanisms by analyzing the temperature dependence of the RNP transport. It turned out, however, that the different in vitro systems responded differently to temperature, thus actually resulting in a controversial discussion ( c f e.g. Refs. 4-8 with 9-11). Unfortunately, the interest concentrated almost exclusively on the temperature dependence of the rate of R N P export. The present paper, however, provides evidence that temperature does not only affect the rate, but also the total number of exported RNP molecules. Specifically, we can show that nuclei isolated from the unicellular eukaryote Tetrahymena export only a definite portion of ribosomal precursor particles at a given temperature, although enough transportable particles are present in nuclei.
_~_ _ _ _ _ _ _ _ _ __.__ * This work was supported by the Deutsche Forschungsgemeinschaft and Stiftung Volkswagenwerk. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "adoertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Cells a n d Labeling with ("H/Uridine-Axenic 3-liter cultures of' the ciliate protozoan Tetrahyrnenapyriformis (amicronucleate strain (GL) were grown under continuous aeration in the logarithmic growth phase (15,OOO-40,OOO cells/ml) at 28 "C as described previously (12). For labeling, the cultures were incubated with 1-2 mCi of [5,fi-"H] uridine (specific activity, 50 Ci/mmol, NEN chemicals, Frankfurt, Federal Republic of Germany) at 28 "C for 30-60 min. Then, about 700 g of crushed ice were added to the culture which was immediately centrifuged (4 X 1-liter swinging bucket rotor in a cryofuge, Christ, Osterode, Federal Republic of Germany) at 3,000 rpm and about 4 "C for 10 min. The pelleted cells were recentrifuged (4 X 100-ml swinging bucket rotor in a Christ junior centrifuge) at 3,500 rpm and about 4 "C for 5 min.
Isolation of Nuclei-The cells were gently broken by our glycerol technique (13, 14). Nuclei were then isolated and purified according to our previous method (15) with the modification described by Herlan et al. (16). The final nuclear fractions were always kept a t 0-4 "C in ST (0.2 M sucrose, 3 mM Ca&, 2 mM MgC12, 20 mM Tris-HC1 a t pH 7.4).
In Vitro Export of["H]RNA-Nuclei prelabeled with [ 'Hluridine were incubated to export RNP by pipetting 0.2-ml aliquots of a given nuclear fraction into 2.8 ml of RM (0.2 M sucrose, 0.75 mM CaCli, 0.50 mM MgCla, 20 mM Tris-HC1 a t pH 7.4) or HRM (0.2 M sucrose, 0.56 mM CaClz, 0.37 mM MgC12, 20 mM Tris-HC1 at pH 7.4), respectivelv. These media were pre-equilibrated in 10-ml centrifuge tubes at respective temperatures for a t least 15 min. RNP export was stopped by adding 1 ml of ice-cold SM (0.2 M sucrose, 8.4 mM CaCl?, 5.6 mM MgCL, and 20 mM Tris-HC1 at pH 7.4) under gentle shaking and immediately putting the tubes on ire for 5 min before centrifugation at 900 X g for 7 min. Supernatants and pellets were mixed with 0.5 ml of bovine serum albumin (2 mg/ml) and then precipitated with 5% trichloroacetic acid, hydrolyzed in 0.5 ml of 5% trichloroacetic acid at 90 "C for 30 min, mixed with 10 ml of Unisolve (Zinsser, Frankfurt, Federal Republic of Germany), and counted in a liquid scintillation counter (Packard model 3380 or Berthold model BF 8000). HNP export is expressed in terms of per cent trichloroacetic acid-precipitable radioactivity in supernatants from total trichloroacetic acidprecipitable radioactivity in a given sample.
Backulard Transport-Unlabeled nuclei were induced to export RNP at 28 "C for 30 min. The export was stopped as usual and nuclei were kept a t 0 "C. Exported ['HJRNP was prepared by incubating prelabeled nuclei in HRM at 28 "C for 30 nlin before nuclei were removed by centrifuging twice at 800 X g and 4 "C for 10 min. 2.8-ml aliquots of the radioactive supernatant were incubated at 28 "C for at least 15 min before adding the unlabeled nuclei. The reaction was stopped by adding SM. Nuclei were centrifugally removed, and both nuclei and supernatants were precipitated with trichloroacetic acid and counted as described above.

RESULTS
Export at 28 "C-We have previously shown that nuclei isolated from Tetrahymena can be induced to export ribosomal precursor particles simply by lowering the final Ca2+/ Mg'+ (3:2) concentration from 5 to 1.5 mM (17). The export can be quantitatively followed in terms of the export of radioactivity from nuclei prelabeled with ["Hluridine. Fig. 1 shows the kinetics of the rRNP export at the cells' optimal 13 1 I n Vitro Ribosomal Ribonucleoprotein Export After a 15-min export at 8 "C, temperature raising from 8-28 "C in an aliquot portion induces the export of additional RNP (arrow). Total counts/sample, -4700 cpm.
growth temperature of 28 "C. The export increases almost linearly during the f i s t 3 min and reaches saturation after about 15-20 min. The saturation plateau does not decline even after 3 h. Also, the absolute radioactivity in the sample remains almost constant during the export. This indicates that a possible RNA degradation does not take place in our in vitro assay for R N P export.
The maximal level of rRNP export at 28 "C varies from a given nuclear fraction to the next between 4 and 8% for unknown reasons. In a given nuclear fraction, it can be experimentally manipulated. For instance, the maximum export can be increased or decreased, respectively, when the RNP export is induced instead by 1.5 mM Ca2+/Mg2+, by 1.2 mM Ca'+/Mg'+, or 1.7 mM Ca'+/Mg2+, respectively. A new and rather puzzling fact is that the maximal export level becomes significantly reduced upon preincubation of the nuclei at 28 "C under nonexport conditions, i.e. a t a final Ca"/Mg'+ (3:2) concentration of 5 mM, before induction of export. Preincubations for 8 and 40 min, for example, result in a decline of the net portion of exported particles at about 35 and 8070, respectively (Fig. 2). These data indicate that a process occurs in nuclei under nonexport conditions which ultimately leads to an inactivation of the rRNP export. In control experiments, we could reveal that this inactivation phenomenon is not due to a degradation of RNA and proteins in nuclei (data not shown).
In order to find out a possible backward import of once exported ribosomal precursor particles into nuclei, unlabeled nuclei are first incubated under export conditions at 28 "C. After 30 min, these nuclei are then incubated, again under export conditions, with that material released from labeled nuclei. Approximately 30% of the original radioactivity can be recovered in nuclei (Fig. 3 ) . This amount remains relatively constant during 60 min, thus indicating an unspecific absorption of the ribosomal precursor particles to nuclei rather than a specific nuclear import.
Export at 8 "C-In Fig. 1, one can compare the rRNP export from the same nuclei a t 8 and 28 "C. Not only the rate of the export is lowered at 8 "C, but also the maximal amount of the exported rRNP particles. Most interesting is the fact that the maximal export at 8 "C never equilibrates with that at 28 "C in a given experiment even after 3 h.
At 8 "C, nuclei also export ribosomal precursor particles. These are physically indistinguishable from those exported at 28 "C (data not shown). Moreover, the nuclear adsorption of once exported ribosomal precursor particles is significantly lower at 8 than at 28 "C (Fig. 3 ) .
Furthermore, Fig. 1 shows that raising the temperature from 8-28 "C induces the export of additional rRNP particles reaching the normal 28 "C plateau after about 5-10 min. This indicates that low temperature does not induce any irreversible inactivation of transportable rRNP molecules and/or of

In Vitro
possible transport sites. This finding, however, could indicate that there is a temperature-dependent equilibrium between transportable rRNP molecules associated with nuclei and those already exported. In order to examine this possibility, we induce export in nuclei at 28 and 8 "C (Fig. 4). After 15 min, the export from nuclei at 8 "C is stopped in an aliquot portion. This portion is then divided in two further aliquots. Nuclei in these aliquots are induced again to export rRNP particles at 8 and 28 "C. While no significant export is observed at 8 "C in comparison with the corresponding 8 "C control, additional rRNP particles are exported at 28 "C ( Fig.  4). This indicates that nuclei export only a definite portion of rRNP particles at a given temperature.
Export between 28 and 8 "C-At 20 "C, for example, nuclei export rRNP particles at a lower rate than at 28 "C; export saturation reveals a lower plateau (Fig. 5).
In Fig. 6, one can see that the maximal portion of exported rRNP particles decreases with falling temperatures. This fact indicates that the initial export rate cannot be used as a measure of the virtual rate of the export reaction, as done Total counts/sample, -3000 cpm previously (8). The effect of temperature on the export rate can be, therefore, reasonably compared only under the assumption that the export follows first order reaction kinetics at the different temperatures. Then, the rate constants can be evaluated in terms of initial rates normalized to the corresponding maximum levels. Fig. 7 shows an Arrhenius plot of the rate constants of export. The export does not decrease linearly, but rather is downward curved with falling temperatures.

DISCUSSION
Our present study on the export of ribosomal precursor particles from nuclei isolated from Tetrahymena indicates that the temperature dependence of the nucleocytoplasmic RNA export is much more complicated than assumed to date. Not only the rate but also the total number of the exported rRNP particles become reduced at lower temperatures. At 8 "C, for example, the export reaches a plateau after about 20 min, which does not equilibrate with that a t 28 "C. This nonequilibration is obviously not due to (i) a different physical quality of the exported rRNP particles, (ii) degradation of the nuclear rRNA by endogenous RNases, (iii) nuclear backward import of exported rRNP particles, (iv) irreversible inactivation of potentially transportable nuclear rRNP particles, or (v) thermodynamic equilibrium between transportable rRNP particles associated with nuclei and those exported from nuclei. We can show that nuclei export only a defined portion of rRNP particles at a given temperature, although enough transportable rRNP particles are present in nuclei. Obviously, the nuclei possess a specific mechanism which determines how many rRNP particles can be maximally exported at a given temperature. In operational terms, potentially trans.. portable rRNP particles are somehow "locked' in nuclei a t low temperature and temperature raising leads to a "graded unlocking" of nuclei.
Such a temperature-induced "unlocking" has never been reported in any other in vitro system of RNA transport developed to date, although the temperature dependence of the RNA export has been intensively studied in recent years (cf: e.g. Refs. 5-11). The most intensive study was performed by Clawson and Smuckler (10) who investigated the release of the overall RNA from rat liver nuclei driven by exogenous ATP. These nuclei release maximally about two-thirds of the total nuclear RNA. The same maximum is reached at all temperatures between 35 and 0 "C. Puzzling is the fact that after reaching the maximal level, the percentage of transported RNA begins to decline. This decline is ascribed to RNA degradation by RNases and backward import of released RNA into nuclei. These findings in rat liver nuclei are completely different from our observations in Tetrahymena. It is possible that this discrepancy can be attributed to differences in the exported RNA species, in the inductors of RNA export, and in the phylogenetic origin of the cells from which nuclei have been isolated. As a major source for this discrepancy, one should also take into account artificial damaging of the nuclei during isolation so that the in vitro RNA export does not reflect specific nuclear transport mechanisms, but rather perturbed processes, as e.g. a leakage.
In Tetrahymena, the existence of the graded unlocking mechanism per se indicates that the export of rRNP particles cannot he driven by a leakage process. For thermodynamic reasons, a leakage would always result in an equilibrium of the maximum transport to the same level at different temperatures. Moreover, our finding speaks against leakage, namely that nuclear unlocking can be endogenously inactivated under nonexport conditions without any degradation of nuclear RNA or proteins. In addition, there already exist several findings indicating that the export of rRNP particles involves specific nuclear processes. (i) Unlocking can be inhibited by Be')+ (17). This normally blocks only the ATP-driven release but not the leakage-driven release (18, 19). (ii) Unlocking is sensitive to the sulfhydryl blocking reagent p-hydroxymercuribenzoate (20). (iii) Unlocking is significantly reduced by protease inhibitors such as phenylmethylsulfonyl fluoride and sodium tetrathionate (20).

A further difference from other in vitro systems is that
Tetrahymena nuclei have been shown to be surrounded by a structurally well preserved nuclear envelope which still exhibits typical in situ properties such as temperature sensitivity under export conditions (13,14). It appears, therefore, attractive to speculate that the nuclear envelope, in particular the pore complexes, is predominantly "closed" at low temperature, thus preventing the export of free nuclear rRNP particles. An involvement of nuclear membranes in unlocking is indicated by the nonlinear, i.e. downward curved, dependence of the export rate constants from temperature. This is namely in line with our previous suggestion that the transport capacity of the nuclear envelope pore complexes is affected by the lipid fluidity of nuclear membranes (21,22). Specifically, we suggested that thermotropic changes in the nuclear membrane fluidity induce the pore complexes to change from a more "open" into a more "closed" state. This suggestion was based on the finding that both the fluidity and the overall R N P export revealed an abrupt change in the same temperature range and on the assumption that these changes relate causally and not fortuitously. It should be stressed, however, that our present study can neither prove nor disprove such a causal correlation in vitro. The reason is the graded unlocking phenomenon which signalizes namely a temperature-dependent change in the "reactivity" of at least one of the reactants involved in the rRNP export (cf: also Refs. 23 and 24). On the other hand, it is also quite conceivable that the nuclear envelope is not involved in unlocking at all. For instance, it could be possible that the potentially transportable rRNP particles are not free, but rather "bound" inside nuclei. In line with this view is our previous finding that the major portion of the nuclear rRNA is tightly associated with the nuclear matrix (16).