Proc. Natl. Acad. Sci. USA Vol. 86, pp. 1791-1795, March 1989 Biochemistry Nucleocytoplasmic transport of in a eukaryotic system: Is there a facilitated transport process? (Tetrahymena ribosomal subunits/ribosomal RNA/Xenopus laevis oocyte) ARATI KHANNA-GUPTA AND VASSIE C. WARE* Department of Biology and Center for Molecular Bioscience and Biotechnology, Mountaintop Campus, Building A, Lehigh University, Bethlehem, PA 18015 Communicated by Clement L. Markert, December 5, 1988

ABSTRACT We have examined the kinetics of the process Studies involving the transport of tRNAs from microin- by which ribosomes are exported from the nucleus to the jected Xenopus oocyte nuclei have revealed that a carrier- cytoplasm using Xenopus laevis oocytes microinjected into the mediated tRNA nuclear transport mechanism exists and that germinal vesicle with radiolabeled ribosomes or ribosomal the rate of tRNA movement is dramatically affected by subunits from X. laevis, Tetrahymena thermophila, or Esche- nucleotide changes in a conserved domain of the tRNA richia coli. Microinjected eukaryotic mature ribosomes are structure itself (6, 7). Therefore, tRNA may interact directly redistributed into the oocyte cytoplasm by an apparent carrier- with constituents to promote its transport. mediated transport process that exhibits saturation kinetics as To gain an understanding of the process for translocating ribosomal subunits from the nuclei of eukaryotic cells, we increasing amounts of ribosomes are injected. T. thermophila have introduced radiolabeled homologous (Xenopus laevis) ribosomes are competent to traverse the Xenopus nuclear and heterologous (Tetrahymena thermophila and Esche- envelope, suggesting that the basic mechanism underlying richia coli) ribosomes or ribosomal subunits into the nucleus transport is evolutionarily conserved. Microinjected of individual Xenopus oocytes by microinjection and have E. coli ribosomes are not transported in this system, indicating followed the partitioning of the injected ribosomes between that prokaryotic ribosomes lack the "signals" required for the nucleus and cytoplasm as a function of time. We report transport. Surprisingly, coinjected small (40 S) and large (60 S) that the nuclear export of ribosomal subunits, like that of subunits from T. thermophila are transported significantly tRNA species (6), is saturable, indicative of a carrier- faster than individual subunits. These observations support a facilitated transport process. Our data suggest that the basic facilitated transport model for the translocation of ribosomal ribosomal subunit transport mechanism in eukaryotes has subunits as separate units across the nuclear envelope whereby been conserved through evolution since ribosomes from a the transport rate of 60S or 40S subunits is enhanced by the diverse species such as Tetrahymena can be transported in presence of the partner subunit. Although the basic features of Xenopus oocytes. Translocation of subunits across the nu- the transport mechanism have been preserved through evolu- clear envelope is, however, a property specifically limited to tion, other aspects of the process may be mediated through eukaryotic subunits, as prokaryotic ribosomes are not trans- species-specific interactions. We hypothesize that a species- ported over the time interval studied. The kinetics of heter- specific nuclear 40S-60S subunit association may expedite the ologous eukaryotic ribosomal subunit transport suggest a transport of individual subunits across the nuclear envelope. model in which the transport rate of individual subunits may be facilitated through nuclear subunit associations. The assembly ofribosomal subunits and their eventual export into the cytoplasm is a complex process. Our knowledge is MATERIALS AND METHODS more advanced for earlier events in this process, such as rRNA gene transcription (1) and the maturation pathways for Synthesis of 3H/32P-Labeled Ribosomes. E. coli cells (strain rRNA ref. than for later steps in ribosome HB101) in early logarithmic phase were labeled with 2 mCi of precursor (e.g., 2), 32P-labeled H3PO4 in water (ICN; 1 Ci = 37 GBq) or 0.5 mCi maturation and nucleocytoplasmic transport. of [3H]uridine (ICN) in low-phosphate M9 medium. Cells Several investigators have hypothesized that nuclear ribo- were harvested after 16 hr at 37°C. T. thermophila cells grown nucleoproteins (RNPs) are associated with the nuclear matrix at 25°C for 16 hr were transferred to medium 357 (American and that the matrix may play a role in processing and in the Type Culture Collection) containing 32P-labeled H3PO4 as movement ofRNPs toward the nuclear pore complexes (3, 4). above and allowed to grow with gentle shaking for 16 hr at Although electron microscopy studies (5) have implicated 25°C. X. laevis stage VI oocytes were excised from the ovary nuclear pore complexes as the sites from which ribosomal of a mature female and transferred to Holtfreter's medium particles as well as other RNPs emerge into the cytoplasm, (8). Radiolabeling and harvesting were as above. the mechanism for transport remains unknown. According to Isolation of 3H/32P-Labeled Ribosomes and RNA. 32p- a general "gating" mechanism proposed by Wunderlich and labeled E. coli cells were disrupted by the NaDodSO4 lysis Speth (5), nuclear RNPs bind to pore complex constituents method (9). The crude lysate was centrifuged at 30,000 rpm until a threshold of bound RNP particles is reached. Above (SW 41 rotor) for 2 hr at 4°C. The transparent pellet was a critical concentration the RNPs are translocated through resuspended in 10 mM Tris (pH 8). Ribosome and RNA nuclear pores and released into the cytoplasm. The possible preparations from all sources were treated with RNase-free contribution of other nuclear factors to the transport process DNase (Promega Biotec) for 30 min at 25°C. 32P-labeled T. is unclear. Presumably specific molecular interactions must thermophila cells were pelleted and disrupted by freezing and occur involving both the RNP and pore complex constitu- thawing three times at -70°C. Cells were resuspended in ents. ice-cold 10 mM Tris (pH 8) and centrifuged at 10,000 rpm (Ti 60 rotor) for 15 min. The supernatant was recentrifuged at The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviation: RNP, ribonucleoprotein. in accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed. 1791 Downloaded by guest on September 26, 2021 1792 Biochemistry: Khanna-Gupta and Ware Proc. Natl. Acad Sci. USA 86 (1989) 30,000 rpm (SW 41 rotor) for 2 hr. The pellet was resuspended either in 10 mM Tris (pH 8) or in dissociation buffer (20 mM 100_ Tris HCl, pH 7.4/16 mM MgCl2/1.0 M KCl/12 mM 2- mercaptoethanol/0.2 mM EDTA). Radiolabeled X. laevis stage VI oocytes were washed in Holtfreter's solution (8) and homogenized at 40C. Centrifugation was as described. RNA was prepared by phenol/chloroform extraction of crude ribosomes followed by ethanol precipitation. 32P-labeled I-C RNA was resuspended in 10 mM Tris (pH 8). Isolation of Ribosomal Subunits. Ribosomes in dissociation '0 buffer were loaded onto a 15-30% (wt/vol) linear sucrose gradient in dissociation buffer and centrifuged at 10'C for 17 hr at 20,000 rpm (SW 41 rotor). The radioactivity in each 0 15 30 45 60 0.5-ml fraction was assayed by liquid scintillation counting. Incubation time, min or small ribosomal subunits were Fractions containing large FIG. 1. Kinetics ofXenopus ribosome transport as a function of pooled and pelleted at 30,000 rpm (SW 41 rotor) for 16 hr at the concentration of ribosomes injected into the nuclei of X. laevis 10'C. Pellets were resuspended in 10 mM Tris (pH 8). oocytes. 32P-labeled X. laevis ribosomes were injected in a vol of 20 Microinjection and Microdissection of X. laevis Oocytes. nl into the nuclei of X. laevis oocytes. After incubation of injected Microinjections into the germinal vesicle of stage V and VI oocytes for 0, 5, 15, 30, 45, and 60 min in Holtfreter's medium at oocytes and manual dissections of oocytes at various times 250C, the oocytes (average number 10 per time point per experiment) after injection were carried out as described (6). An average were fixed in ice-cold 1% CC13COOH and manually dissected, and of 10 oocytes was injected for each time point. In some the radioactivity present in either pooled or individual nuclear and experiments nuclear and cytoplasmic fractions for each time cytoplasmic fractions was determined by liquid scintillation count- was In other ing. A range of 0.2-2.8 ng of rRNA per nucleus (0.4-6 x 103 cpm) point were pooled and radioactivity quantitated. was used in this experiment. Several concentrations are not shown. experiments the radioactivity in nuclear and cytoplasmic For clarity, SDs between 2 and 10%6 have been omitted from the fractions of each individual oocyte was determined. Meth- figure. The best-fit first-order exponential plots were generated. ylene blue (0.1%) was added to the injected material in most Statistically significant differences (at a minimum 95% confidence experiments as an indicator of the success of nuclear injec- level) are noted as follows: 0.2 ng/0.4 ng; 0.2 ng/0.6 ng; 0.2 ng/0.8 tion. Only those oocytes with dye remaining in the nucleus ng; and 0.4 ng/0.6 ng. v, 0.6 ng; 0, 0.8 ng; A, 0.4 ng; o, 0.2 ng. were analyzed. The dye has been shown to have no measur- able effect on the transport of ribosomes or ribosomal In a parallel set of oocytes, radiolabeled Xenopus ribosomes subunits (unpublished observations). On the average only 1 were injected. After incubation of the oocytes for periods up nucleus in 20-30 oocyte nuclei was not targeted. to 30 min, no detectable transport of the labeled DNA was Measurement of Ribosome Specific Activity. An aliquot of noted while the Xenopus ribosome radioactivity was redis- each ribosome preparation was phenol extracted and ethanol tributed into the cytoplasmic fractions (data not shown). precipitated. The pellet was resuspended in 10 mM Tris (pH To estimate the maximal transport rate for the injected 8). The A260 and amount of radioactivity were determined. material, we calculated the transport rate of the injected Activities are expressed as cpm/,ug of RNA. ribosomes as a function of the Xenopus ribosome concen- Statistical Analysis. Sample measurements of radioactivity tration (expressed as ng of rRNA per min) injected per (expressed as percentage in the nuclear fraction) were taken nucleus. An apparent maximal rate of 1.2 x 10-2 ng/min was over time. Data from multiple trials (at least three) of each obtained. Saturation of the component(s) involved in trans- experiment were compiled for statistical evaluation. A least- locating ribosomal subunits across the nuclear envelope squares regression line fit was performed on the data assum- strongly favors the existence of a carrier-mediated mecha- ing first-order kinetics. Comparative statistical tests, using a nism for the transport of ribosomal subunits. software package called SAS (owned by the SAS Institute), It should be noted, however, that the "apparent" maximal were performed on the slopes of the lines to determine if the rate calculated based on the data from Fig. 1 and pooled observed slopes were significantly different (a minimum of fraction experiments (data not shown) does not take into 95% confidence). A homogeneity of slopes model was used consideration the endogenous synthesis and transport of for the SAS general linear model (GLM) procedure. ribosomes in oocytes [rRNA synthesis is high in stage VI (11)] at this stage of development. The amount of injected material is probably relatively low compared to the amount RESULTS of rRNA synthesized and transported in this system at this Ribosome Transport Across the Nuclear Envelope Exhibits time in development. That the apparent maximal rate is an Saturation Kinetics. Carrier-mediated transport phenomena underestimation of the true capability of mature oocytes is are usually distinguishable from diffusion-limited processes evident from the fact that at a maximal transport rate of 1.2 in that they exhibit saturation kinetics (10). To distinguish x 10-2 ng of rRNA per min, it would take a mature oocyte between these two possibilities, increasing amounts of 32P- 9.6 months instead of the normal 3-month period to accumu- labeled X. Iaevis ribosomes (0.2-2.8 ng of RNA) were late 5 ,ug of RNA (the majority of which is rRNA) in the injected into the nuclei of X. laevis oocytes and the intracel- oocyte cytoplasm. lular partitioning of radioactivity was determined. The til2 Tetrahymena Ribosomes Traverse the X. laevis Oocyte Nu- value for a given ribosome concentration was equivalent in clear Envelope. 32P-labeled ribosomes (0.2-1.1 ng of RNA, 3 pooled oocyte and individual oocyte experiments. The time x 103 cpm/,g) were isolated from T. thermophila and required for half the injected ribosomes to be exported from injected into the nuclei ofXenopus oocytes. Our data (Fig. 2) the nucleus increased with increasing concentrations of show that Tetrahymena ribosomes are transported out of injected material (Fig. 1). The experimental t1/2 values range Xenopus nuclei. from 6.8 min for 0.2 ng of RNA injected per nucleus to 260 To confirm the transport of these foreign ribosomes in the min for 2.8 ng of RNA injected per nucleus. Xenopus oocyte system, we coinjected into the same batch of To distinguish between the transport of ribosomes and the oocytes 32P-labeled ribosomes from Tetrahymena (0.7 ng of possible leakage of injected material, we injected radiola- RNA, 6 x 103 cpm/gg of RNA) and DNA (23 kilobase pairs) beled DNA fragments (2-4 kilobase pairs) into oocyte nuclei. labeled at the 3' end by using T4 DNA polymerase and Downloaded by guest on September 26, 2021 Biochemistry: Khanna-Gupta and Ware Proc. Natl. Acad. Sci. USA 86 (1989) 1793

LI)I[00' ..... __v_7 ence of the E. coli ribosome preparation in no way hindered the translocation of Tetrahymena ribosomes across the L...... o 0 nuclear envelope, as the kinetics of transport were the same _ ...... as for experiments shown in Fig. 2, and thus discounted the ._ o presence of a transport inhibitor in our E. coli ribosome preparations. r. 0 Our results favored the hypothesis that Xenopus oocytes could distinguish between eukaryotic and prokaryotic ribo- CZ 0 somes, only exporting the former into the cytoplasm. We tested this hypothesis directly by coinjecting equal concen- trations (1.0 ng of RNA) of 3H-labeled E. coli ribosomes and 32P-labeled Tetrahymena ribosomes. As shown in Fig. the 0 45 3, 15 30 60 32P-labeled Tetrahymena ribosomes emerged from the nu- Incubation time, min cleus into the cytoplasm with a t1/2 of 60 min whereas the 3H-labeled E. coli ribosomes remained confined to the Fi G. 2. Kinetics of nucleocytoplasmic transport of Tetrahymena nucleus for the entire length of the experiment. The slopes of ribosomes. 32P-labeled Tetrahymena ribosomes (0.2-1.1 ng of RNA; the best-fit lines are statistically different at a confidence level 0.6-1.5 x 10 cpm/,ug) were injected into the nucleus of Xenopus oocyttes as described in Fig. 1. In a control experiment, 32P-labeled of >99%. Clearly there was selective transport of heterolo- Xenonpus total rRNA (0.7 ng of RNA; 105 cpm/,mg) was injected into gous eukaryotic ribosomes. oocyltes and the partitioning of radioactivity was measured. SDs Transport of Individually Injected Heterologous Subunits Is betw4een 3 and 11% are absent from the figure. Statistically significant Altered in the Presence of Heterologous Partner Subunits. To differ-ences (at a minimum 95% confidence level) are noted: 0.2 determine if the transport rates for heterologous 40S and 60S ng/0..8 ng; 0.2 ng/1.1 ng; 0.2 ng per Xenopus rRNA; 0.8 ng/1.1 ng; subunits were similar, we examined the kinetics of transport 0.8 njg per Xenopus rRNA; 1.1 ng per Xenopus rRNA. V, Xenopus of heterologous 40S or 60S subunits injected individually rRNAk;Q, 1.1 ng; A, 0.8 ng; o, 0.2 ng. compared to the kinetics of transport when heterologous subunits were coinjected. Data from both pooled nuclear and [5_-3] []dCTP (105 cpm/,ug). Quantitation of each isotope pooled cytoplasmic fractions as well as those obtained from reve,aled that the 3H activity remained in the nucleus while individual oocyte nuclear and cytoplasmic fractions were the 32p activity was partitioned into the cytoplasm (data not used in this experiment (no significant differences between sho"in). As expected DNA remained intranuclear while the the til2 values obtained by the two meihods were evident). ribossomes segregated into the cytoplasm. In parallel sets of Fig. 4 indicates that the til2 value for the 60S subunit (0.16 ng contirol experiments, we injected 32P-labeled Xenopus rRNA of RNA, 8 x 102 cpm/,ug) from Tetrahymena is 94 min; the into one batch of oocytes and labeled Tetrahymena ribo- t1/2 for the 40S subunit (0.16 and 0.1 ng of RNA, 5 x 102 som(es into a separate batch. Xenopus rRNA remained cpm/1xg) is 162 min. These values, however, are not statis- confiined to the nucleus, indicating that naked RNA is not tically significant. For the coinjection experiments only one transsported across the nuclear envelope whereas the Tet- subunit was radiolabeled. An example of a coinjection rahymena ribosomes were transported (Fig. 2). experiment is shown (Fig. 4) in which radiolabeled 40S En((perimental t1/2 values of 17.6, 37.4, and 60.2 min for subunits were coinjected with unlabeled 60S subunits. In the conc-entrations of 0.2-1.1 ng of injected ribosomes were presence of the 60S subunit from Tetrahymena, the 40S observed. Each tl/2 value is statistically different from all subunit from Tetrahymena was transported from the Xeno- otheir values at a minimum 95% confidence level. It is also pus oocyte nucleus much more rapidly; the 40S subunit ti12 apparent that the til2 values for Tetrahymena increased with decreased from 162 min to 19 min-a statistically significant incre-asing concentrations of injected ribosomes, following a change (at a minimum 95% confidence level). In fact the t1/2 patte-rn similar to that of the Xenopus ribosome injection value for 40S subunits in this coinjection experiment was experiments (Fig. 1). No significant difference in the apparent approximately the same as for Tetrahymena ribosomes at a maxiimal rates of transport of Tetrahymena or Xenopus concentration of 0.2 ng of RNA injected (see Fig. 2). This ribossomes was observed. effect was not limited to the Tetrahymena 40S subunit; the Ass a measure of the integrity of the transported material, kinetics of transport were similar when radiolabeled 60S total KNA was extracted from nuclear and cytoplasmic fractions and analyzed by PAGE. Our data corroborated the 100l i A- A -- ---IA quantitative radioactivity data and showed that the labeled Ct Tetrahymena rRNA remained intact (data not shown). C) Prokaryotic Ribosomes Are Not Transported in Xenopus C Oocytes. Since we had demonstrated that foreign eukaryotic ribosomes could be transported across the Xenopus nuclear envelope, it was of interest to determine the extent of 10l evolutionary conservation of elements ofthe transport mech- ._C) anism. In initial experiments equal concentrations of 32P- 0- labeled ribosomes (from several different preparations) from ox the eubacterium E. coli and from Xenopus were injected into separate batches of oocytes and Our results indi- analyzed. 0 15 30 45 60 cated that while the Xenopus ribosomes were transported into the cytoplasm, the E. coli ribosomes failed to leave the Incubation time, min oocyte nucleus over the period studied (data not shown). Based on these experiments, the transport phenomenon in FIG. 3. Kinetics of 32P-labeled E. coli ribosome transport in this system appeared to be limited to eukaryotic ribosomes. Xenopus oocytes. Equal concentrations (1.0 ng of RNA per nucleus) of 3H-labeled E. coli 5 x and 32P-labeled To rule out the presence of a inhibitor (A, 104cpm) Tetrahymena nonspecific transport ribosomes (D, 4 x 103 cpm) were coinjected into the same set of in our E. ribosome we coli preparations, coinjected equal oocytes. Individual oocytes were assayed as in Fig. 1. SDs between concentrations ofunlabeled E. coli ribosomes and 32P-labeled 1 and 10%have been omitted from the figure. The slopes of the two Tetrahymena ribosomes. The results showed that the pres- lines are statistically different at a confidence level of >99%. Downloaded by guest on September 26, 2021 1794 Biochemistry: Khanna-Gupta and Ware Proc. Natl. Acad. Sci. USA 86 (1989)

w 1001 A,,._...... 100' u~~u ~ ~n ~ no Ra 0) ~~~~ AA

0

._ .D100 A~ 10 U( 0

0 ce C) A C) 1 I 0 15 30 45 60 .2 15 30 45 60 75 Ct Incubation time, min

C._ :0( CZ FIG. 4. Kinetics of Tetrahymena 32P-labeled 40S, 60S, and 40S 0 mA..,.. E plus 60S subunit transport. Equal concentrations (0.16 ng of RNA) 1-o of 32P-labeled Tetrahymena 40S subunits (El, 5 x 102 cpm), 60S Ce )l subunits (0, 8 x 102 cpm), or a mixture of unlabeled 60S and 32P-labeled 40S subunits (A, 103 cpm) were injected into separate 0.1 ng of batches of oocytes. In addition 32P-labeled 40S subunits (v, 1( RNA) were injected. After incubation the oocytes were microdis- sected as described in Fig. 1. SDs range between 1 and 12%. Statistically significant differences (at a minimum 95% confidence level) are as follows: 32P-labeled 40S (0.16 ng)/32P-labeled 40S (0.16 ng) + 60S (0.16 ng); 32P-labeled 60S (0.16 ng)/32P-labeled 40S (0.16 B A ng) + 60S (0.16 ng); 32P-labeled 40S (0.1 ng)/32P-labeled 40S (0.16 ng) 1 + 60S (0.16 ng). 1E5 30^ 45A- 60n 1A75 9u% subunits were coinjected with unlabeled 40S subunits (data Incubation time, min not shown). The transport of individually injected heterolo- gous subunits is slower than that of coinjected heterologous FIG. 5. Kinetics of Tetrahymena ribosomal subunit transport subunits. The coinjection experiments clearly show a syner- after delayed injection of Tetrahymena partner subunits. (A) X. laevis getic effect on the transport of heterologous subunits when oocytes were microinjected with 32P-labeled 60S subunits (o, 0.7 ng the heterologous partner subunit is present in the Xenopus of rRNA per nucleus; 1.5 x 103 cpm). At times 0, 15, 30, 45, and 60 oocytes were fixed and microdissected as described in Fig. 1. system; the transport of one subunit appears to be facilitated min, At 60 min of incubation (arrow), the batches of oocytes were divided presence by the of the other subunit. into two groups: one group was mock-injected and the second group The subunit experiments were also undertaken using was injected with unlabeled Tetrahymena 40S subunits (A, 0.27 ng radiolabeled 40S subunits (0.2 ng of RNA)'from Xenopus. In per nucleus). At 5, 10, 15, and 20 min after mock-injection or 40S this case when the 40S subunits were injected alone, the subunit injection, oocytes were treated as described. SDs range from kinetics of transport were identical to that for injected 2 to 15%. The slopes of the two lines are statistically different at a Xenopus ribosomes at 0.2 ng of RNA injected (Fig. 1). This confidence level of >99%o. (B) 32P-labeled Tetrahymena 40S subunits result is consistent with the heterologous results in that the (n, 0.13 ng of rRNA per nucleus; 2 x 103 cpm) and unlabeled transport rate of the injected 40S subunit may be enhanced in Tetrahymena 60S subunits (A, 0.28 ng of rRNA per nucleus) were injected as described in A at the arrow. SDs range from 2 to 18%. The the presence of the endogenous 60S subunit. slopes of the two lines are statistically different at a confidence level To investigate further the apparent facilitation phenomenon, of 98%. we injected oocytes with 32P-labeled Tetrahymena 60S subunits at (Fig. 5A) or 32P-labeled Tetrahymena 40S subunits (Fig. 5B) subunits across the nuclear envelope into the cytoplasmic t = 0. At 60 min afterthefirst injection, the oocytes were divided compartment. It is unknown what relationship, if any, exists into two groups: one group was mock-injected (nucleus was between the ribosome transport system described here and was with unlabeled repricked) and the other group injected the tRNA transport mechanism (6). Whether or not a com- 40S or 60S subunits, respectively. We reasoned Tetrahymena mon carrier system exists in eukaryotic cells to transport all that we might observe a change in the slope of the line cytoplasmic-bound /RNPs is unclear. the rate oftransport of40S or 60S subunits at a time representing The rate of transport of injected ribosomes appears to be interval to the injection of partner subunits. Our subsequent the concentration of a nuclear carrier(s) that ti, values for subunits decreased limited by results indicate that the 60S subunits to effect nuclear export. a 10-fold increase in the interacts with ribosomal from 131 min to 13 min, representing is of exporting heterologous transport rate of 60S subunits after 40S subunits were intro- The transport system capable duced into the nucleus. In reciprocal experiments the transport eukaryotic ribosomes from an organism as evolutionarily as T. This of40S subunits increased 9-fold in the presence of60S subunits distant from Xenopus the ciliate thermophila. (to1, values decreased from 89 min to 10 min). The magnitude of observation supports the previous claim that the basic mech- the change in the rates of transport for the delayed injection anism underlying ribosomal RNP nuclear export is essen- experiments (Fig. 5) is comparable to that observed for the tially the same in all eukaryotes (12). Prokaryotic ribosomes coinjection experiments. In all experiments the transport rate of were not translocated over the period analyzed. In fact, the heterologous ribosomal subunits is dramatically increased in the Xenopus oocyte transport system could distinguish heterol- and presence of heterologous partners in the nucleus. ogous eukaryotic ribosomes from prokaryotic ribosomes selectively export the former across the nuclear envelope into the cytoplasm, leaving the prokaryotic ribosomes intranu- DISCUSSION clear. The inability of prokaryotic ribosomes to be trans- Our results demonstrate that a carrier-mediated translocation ported indicates that these prokaryotic counterparts lack the process exists in Xenopus oocytes to transport ribosomal "signals" for interaction with the carrier mechanism. Downloaded by guest on September 26, 2021 Biochemistry: Khanna-Gupta and Ware Proc. Natl. Acad. Sci. USA 86 (1989) 1795 The similarity in the kinetics ofribosome transport for such in the process (e.g., ref. 16). Clearly the processes are tightly diverse species as Tetrahymena and Xenopus suggests that regulated, and it is likely that the transport rate of ribosomal the biochemical nature of the interaction of the carrier with subunits is not constant but is subject to changes in other steps the ribosomal subunit has been conserved throughout the in ribosome biogenesis. The proposed facilitation mechanism evolution of the eukaryotes. It is likely that the carrier may contribute to posttranscriptional regulatory processes recognizes some feature(s) unique to eukaryotic ribosomes that ultimately control the pool sizes of nuclear and cytoplas- that is essential for nucleocytoplasmic transport. Whether mic ribosomes. The exact nature ofthe facilitation effect must this common characteristic is related to the overall shape of await further experimentation. eukaryotic ribosomes and/or to a specific or nucleic These experiments also address the question of putative acid sequence awaits determination. conformational changes that may occur in ribosomal subunits A surprising finding is that the transport rate of heterolo- as they are translocated. Recent data describe (17) a nuclear gous ribosomal subunits apparently is accelerated by the protein called ribocharin that associates with 65S precursor presence of both subunits. The basic transport mechanism particles. This protein dissociates from the 60S subunit, has been evolutionarily conserved; yet, there seems to be a requiring a rearrangement in rRNA-associated prior species-specific component to the process, as the rate en- to transport. It is assumed that the majority of the injected hancement effect is only observed in the presence of homol- ribosomes or subunits in our were Is the a experiments mature, ogous partners. species-specific component transport cytoplasmic particles, based on the knowledge that in loga- factor(s) that cosediments with each Tetrahymena subunit rithmic-phase cells of T. pyriformis, 98% of the ribosome fraction or is the component the partner subunit itself? We have that a facilitated mechanism is in population is cytoplasmic (13), and certainly the majority of proposed transport the 1012 ribosomes assembled in mature operation which is either based on a direct or indirect nuclear Xenopus oocytes interaction between ribosomal subunits. Among the possible accumulate in the cytoplasm (11). Therefore, cytoplasmic interactions, we favor two hypotheses that are not mutually ribosomes, when microinjected into the nucleus, retain the exclusive: (i) a specific noncanonical interaction between 40S capacity to interact with the nuclear components that effect and 60S subunits mediated through species-specific subunit ribosomal subunit transport. This suggests that ribosomal RNA and/or ribosomal protein interactions to enhance subunits do not undergo an irreversible rearrangement in carrier-mediated transport across the nuclear envelope structure as they are translocated. and/or (ii) a concerted interaction of the subunits with The elucidation of the complex interactions that are in- another nuclear component(s) located either in the nuclear volved in the nucleocytoplasmic transport of ribosomal pore complex or in the nuclear matrix, to produce the subunits will have a significant impact on our understanding facilitation effect. In this latter possibility, the additional of the regulation of protein synthesis in eukaryotic cells. component(s) may or may not be the carrier itself. In either case a direct or indirect subunit interaction facilitates the We thank Dr. Steve McKnight for help in teaching oocyte injection movement of ribosomal subunits across the nuclear enve- to V.C.W., Dr. J. Hoblitzell and W. Taylor for statistical analyses lope; transport, however, does occur in the absence of the and preparation of figures. We also thank Drs. S. Gerbi and A. albeit more Dahlberg for stimulating discussions. We acknowledge J. Garvey, J. partner slowly. Heil-Male, and D. Weaver for manuscript preparation and Dr. J. The nature of these interactions remains to be determined; Burton for computer instruction. This work was supported in part by however, the possibility that the subunits interact directly at National Institutes of Health Grant GM38574 to V.C.W. some stage in the transport process is supported by the observation that in experiments where only one heterologous 1. Perry, R. P. (1967) Progr. Nucleic Acid Res. Mol. Biol. 6, 219- subunit was injected, the rate ofexport was much slower than 257. when the heterologous subunit partner was present in the 2. Hadjiolov, A. A. (1984) The Nucleolus and Ribosome Biogen- nucleus as well. It would appear that Xenopus 40S subunits esis (Springer, New York). were not as effective as Tetrahymena 40S subunits in facil- 3. Wunderlich, F., Brezney, R. & Kleinig, H. (1976) in Biological itating the transport of Tetrahymena 60S subunits. We Membranes, eds. Chapman, D. & Wallach, D. F. H. (Aca- that the subunit demic, New York), Vol. 3, pp. 241-333. speculate Tetrahymena interaction was a 4. Herlan, G., Eckert, W. A., Kaffenberger, W. & Wunderlich, F. requisite step for further interaction with the Xenopus carrier (1979) Biochemistry 18, 1782-1788. system. 5. Wunderlich, F. & Speth, V. (1972) J. Microscopie 13, 361-382. Several studies (e.g., refs. 13-15) in which pulse-chase 6. Zasloff, M. (1983) Proc. Natl. Acad. Sci. USA 80, 6436-6440. experiments were performed concluded that newly synthe- 7. Tobian, J. A., Drinkard, L. & Zasloff, M. (1985) Cell 43, 414- sized 40S subunits appear in the cytoplasm prior to the 422. appearance of newly synthesized 60S subunits. Our data do 8. Hamberger, V. (1960) in A Manual ofExperimental Embryol- not address this question directly. The pulse-chase experi- ogy, (Univ. of Chicago Press, Chicago), Rev. Ed., p. 85. ments plus data on the cytoplasmic pool sizes of subunits (13) 9. Maniatis, T., Fritsch, E. F. & Sambrook, S. (1982) Molecular argue against the cotransport of subunits as 80S ribosomes; Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold the Spring Harbor, NY). however, previous studies do not rule out the possibility 10. Christensen, H. N. (1975) Biological Transport (Benjamin, for other types of subunit associations that may result in an New York), pp. 107-165. accelerated rate of transport of individual units across the 11. Scheer, U., Trendelenburg, M. F., Krohne, G. & Franke, nuclear envelope. Although unexpected, our data do not W. W. (1977) Chromosoma (Berlin) 60, 147-167. contradict any previous claims about the nature of the nucle- 12. Hadjiolov, A. A. & Nikolaev, N. (1976) Prog. Biophys. Mol. ocytoplasmic transport of ribosomal subunits as independent Biol. 31, 95-144. units across the nuclear envelope. Instead our heterologous 13. Leick, V. & Andersen, S. B. (1970) Eur. J. Biochem. 14, 460- studies may have identified a mechanism that serves to 464. modulate the rate of of as 14. Wunderlich, F. (1972) J. Membr. Biol. 7, 220-230. transport subunits ribosome bio- 15. W. genesis progresses. For several temperature-sensitive muta- Eckert, A., Franke, W. W. & Scheer, U. (1975) Exp. Cell tions that Res. 94, 31-46. affect ribosome synthesis in yeast, the evidence is 16. Warner, J. R. & Udem, S. A. (1972) J. Mol. Biol. 65, 243-257. compelling that inhibition of any step in the synthesis of 17. Hugle, B., Scheer, U. & Franke, W. W. (1985) Cell 41, 615- ribosomes has pleiotropic inhibitory ramifications on all steps 627. Downloaded by guest on September 26, 2021