Pre-ribosomal particles from nucleoli to cytoplasm

ABSTRACT

The analysis of nucleocytoplasmic transport of proteins and messenger RNA has been the focus of advanced microscopic approaches. Recently, it has been possible to identify and visualize individual pre-ribosomal particles on their way through the nuclear pore complex using both electron and light microscopy. In this review, we focused on the transport of pre-ribosomal particles in the nucleus on their way to and through the pores.

Introduction

In eukaryotic cells, gene content is stored in the form of DNA in the cell nuclei. When required, this information is copied to the messenger RNA (mRNA). mRNA is packed by auxiliary proteins into messenger ribonucleoprotein (mRNP) particles, which move around in the cell nucleus until they encounter a nuclear pore complex by which they leave the nucleus. In the cytoplasm, stored information is translated into functional proteins by ribosomes. Ribosomes are formed by two subunits: the small 40S and the large 60S subunit. Both are built using highly structured RNA strands and numerous supporting proteins. In typical human cells, ribosomal proteins make up approximately 5% of the total protein mass, and each cell contains approximately 10 million ribosomes [1]. Ribosomal RNA constitutes approximately 80% of the total cellular RNA. How is this immense amount of large supramolecular complexes produced and transported into the cytoplasm?

Creation of pre-ribosomes in nucleoli

Ribosome biosynthesis is a colossal cellular task that requires the coordinated interplay of more than 200 proteins across three different compartments in eukaryotic cells. Pre – rRNA synthesis, cleavage, and maturation are accompanied by the integration of approximately 80 ribosomal proteins and multiple accessory factors resulting in pre-40S and pre-60S subunits. The biogenesis of subunits is a complex process and begins with the synthesis of rRNAs, which provide the backbone for both subunits (for review, see [2,3]. In eukaryotic cells, within the nucleolus, 18S, 5.8S and 28S rRNAs are co-transcribed as a single 47S precursor by RNA polymerase I (Pol I), while 5S rRNA is transcribed by Pol III. Transcription elongation by Pol I is highly efficient with approximately 100 polymerases transcribing each active gene at a rate of 95 nucleotides per second [4]. In order to provide the high number of required ribosomes, the ribosomal DNA (rDNA) is present in 200 to 700 copies per mammalian cell, distributed over several chromosomes (for review, see [5]. The 47S rRNA gene repeats are located on the chromosomes 13, 14, 15, 21 and 22 in human cells, whereas the human 5S rRNA gene repeats are located on chromosome 1. The transcription of the rDNA occurs in the nucleolus, which is composed of three morphologically distinct structures, the fibrillar centers (FC), the dense fibrillar component (DFC) and the granular components (GC) [6]. A human nucleolus consists of several dozen FC/DFC units, each containing two to three transcriptionally active rDNAs at the FC/DFC border. From here, the nascent pre-rRNA is directionally sorted into the DFC [7]. The transcribed rDNA and pre-rRNAs can be electron-microscopically visualized by the famous chromatin or Miller spreading. Then they can be seen forming a Christmas tree-like structure, in which increasingly longer rRNAs and pre-ribosomal intermediates are formed as Pol I is transcribed along the rDNA [8]. Processing and modification of pre-rRNAs occur preferentially in the DFC, and their assembly with ribosomal proteins and the Pol III-synthesized 5S rRNA occurs in the GC region [5], which surrounds the FC/DFC units. Thereby, the GC has electron microscopically a granular appearance, which is due to the emerging pre-ribosomal particles.

However, the pre-ribosomal particles cannot yet participate in protein translation. The transition from the pre-subunits to the mature subunits begins toward the end of nucleolar and at the beginning of nuclear maturation. The pre-60S intermediates already contain some functional centers, but they do not yet have their final, definitive conformation. To complete them, a number of nuclear assembly factors are required. For example, an important transition of pre-60S nuclear maturation is the rotation of the 5S RNP sub-complex, and only after this rotation the further processing leads to the mature 5.8S rRNA [9]. Nucleoli are formed by liquid-liquid phase separation in response to ongoing rRNA transcription [5,10,11]. Erdmann et al. [12] observed that pre-ribosomal subunits are successively processed, increase in size, and maturate as they move radially out of the nucleolus until structurally well-defined intermediates appear. In particular, the processome of the pre-40S subunit acted as a marker indicating the border region between nucleolus to nucleoplasm, thus defining the surface of the nucleoli [12].

In this manner, a growing mammalian cell can produce and export approximately 10.000 ribosomal subunits per minute [11,13]. The eukaryotic 40S subunit has an average diameter of 16 nm and a molecular weight of approximately 1.4 MDa [14], whereas the larger 60S particles are approximately 25 nm in diameter with a molecular weight of approximately 2.5 MDa [15]. The final subunits are among the largest transport substrates that must be exported out of the cell nuclei [16].

Transport to the nuclear pore complexes

After their production in the nucleoli, the pre-ribosomal particles travel through the nucleoplasm until they reach the NPC. Ritland-Politz et al. [17] used caged-fluorescein-labeled anti-28S rRNA oligonucleotides to label pre-60S particles in live L6 myoblasts. Upon uncaging the signal in the nucleolus, it reached the nuclear envelope within 3.6 seconds and filled the cell nuclei within 10 s. Once inside the nucleoplasm, the subunits move in the same random way, sometimes accidentally visiting a different nucleolus before leaving the nucleus. They did not observe any sign of directed or active transport, and concluded that the motion was due to constrained diffusion. Probably one of the structures hindering the movements of large intranuclear particles in the nucleoplasm is chromatin [18–20]. Ritland-Politz et al. [17] determined an experimental estimate for the diffusion coefficient of 0.3 µm²/s on the average, but stated explicitly that multiple populations of particles with different mobilities exist and particle movement is modified by frequent encounters of binders or obstructions. Actually, the fact that the nuclei were filled with fluorescent particles within 10 s after uncaging implies a particle diffusion coefficient of D ≈2 µm²/s considering a nuclear diameter of 15 µm (see Table 1).

Table 1.

Comparison of intra-nuclear diffusion coefficients of ribosomal RNA and subunits and respective mobile fractions.

Molecule	Organism	D fast (µm2/s)	D retarded (µm2/s)	D immobile (µm2/s)	Reference
28S rRNA	L6 myoblast	~2a	0.3	n.a.	[17]
28S rRNA	C.tentans	2.87 (60%)	0.41 (40%)	n.a.	[21]
40S subunit	HeLa	2.3 (28%)b	0.31 (39%)b	0.054 (33%)b	[22]
		2.3 (7%)c	0.31 (28%)c	0.054 (65%)c
60S subunit	HeLa	1.7	n.a.	n.a.	[23]

^aEstimated, ^bNucleus, ^cNucleolus, n.a.: not available.

A diffusion coefficient of this order of magnitude was directly demonstrated by Spille et al. [21], who followed trajectories of single 28S rRNA labeled by specific antisense oligonucleotides carrying one to three fluorophores to minimize artifacts due to the size of the label. Feedback tracking using astigmatic imaging allowed the observation of single 28S rRNA particles over extended periods of time in three dimensions (3D) in the nuclei of salivary gland cells of Chironomus tentans. The high number of localizations allowed for rigorous statistical analysis of the temporal sequence of jump distances in individual trajectories. Distinct, repeated binding events were observed for single subunits, and diffusion coefficients of D₁ = 0.41 µm²/s (40%) and D₂ = 2.87 µm²/s (60%) were determined. Certainly, the intranuclear structure of the salivary gland cell nuclei is different from that of normal eukaryotic cells, because very large nuclear regions are devoid of chromatin. Presumably, this is the reason for the comparatively high diffusion coefficients and high fractions of mobile particles observed in that system (Table 1).

Pre-40S and pre-60S particles are large RNPs that can indirectly be labeled by means of proteins that associate with the particles in the early stages of their biogenesis. This approach has been successfully used to perform transport experiments on mRNA particles [24]. Recently, Landvogt et al. [22] succeeded in labeling pre-40S particles using PNO1 (Partner of NOb1, Defective in DNA Methylation 2; DIM2 in yeast [25,26]) coupled C-terminally to a SnapTag. They created a stable cell line carrying an inducible SNAP-PNO1 fusion protein and observed the pre-40S subunits in live cells. Using single-molecule microscopy, it was possible to track single pre-40S particles in the nucleoli and nucleoplasms. In both nucleoli and nucleoplasm mobile (D_mob = 2.5 µm²/s), retarded (D_ret = 0.25 µm²/s) and immobilized pre-40S subunit fractions were identified, and as expected, the immobile fraction was largest in the nucleoli (65%) [22]. The fast mobility fraction matched perfectly the expected nuclear mobility of pre-40S subunits (D = 2.3 µm²/s), when assuming a five-fold enhanced viscosity in the nucleus compared to water as was previously determined [17,22,27–29].

To visualize pre-60S particles, Ruland et al. [23] targeted eIF6, which is stably incorporated into particles during rRNA transcription [30]. This protein is known to prevent premature association of the pre-60S and pre-40S subunits. It was previously demonstrated that eIF6-HaloTag is functionally incorporated into pre-60S particles, which represents a feasible approach for fluorescently labeling pre-60S particles in vivo [31]. Ruland et al. [23] generated a HeLa cell line stably expressing both eIF6-HaloTag and eGFP-NTF2 [32]. Unlike POM121, which has been used as an NPC marker in other studies, NTF2 is not stably incorporated into NPCs and undergoes repeated replacement, making this labeling approach more resistant to bleaching [32,33]. This allows for precise localization with nanometer precision [23]. HaloTag enabled the labeling of pre-60S subunits with the bright and photostable dye JF549 [34], either in high concentration as a bulk dye for confocal imaging or at ultra-low concentrations (<1 nM), to visualize individual pre-60S particles in living cells using highly inclined and laminated optical (HiLo) sheet microscopy.

Similar to pre-40S, the pre-60S particles showed very small mobility inside the nucleoli, but in the nucleoplasm, restricted diffusive motion prevailed. The tracking of single particles revealed a maximal nuclear diffusion coefficient of 1.7 µm²/s. As expected, this value was smaller than the maximal diffusion coefficient exhibited by the single pre-40S particles, because these particles have a smaller diameter and are therefore slightly more mobile. These results demonstrated, that the transport of pre-ribosomal particles in the nucleoplasm is similar to that of mRNPs and proteins [20]. The nucleoplasm is mostly liquid-like at the nanoscale and microscale, with viscosities of up to a few mPa·s [28]. On larger length scales, transport is moderately slowed [27,35]. The mode of transport of these substrates within the nucleoplasm is constrained diffusion. Notably, RNA particles often exhibit phases of mobility, in which they diffuse as fast as theoretically expected for particles of a corresponding size in the effective nuclear viscosity, which is in the range of 3 to 5 mPas (reviewed by [36]. In general, single RNA tracking data from recent years confirm the Pederson lab’s early observations [17] that RNA particles spread evenly by constrained diffusion in the nuclei of cells within approximately 10 seconds.

Obviously, the maturation and associated transport process of pre-60S subunits within the nucleoli observed by Erdmann et al. [12] takes place on time scales longer than 30 s, as individual labeled pre-40S and pre-60S particle in the nucleoli appeared largely immobile on the time scale of the single-molecule tracking experiments [22,23].

Translocation across the pore

As mentioned above 10.000 pre-ribosomes are produced per minute by each eukaryotic cell in the growth phase. Thus, given the overall number of NPCs in a mammalian cell [37,38], approximately three small and large pre-ribosomal subunits are exported by each single NPC per minute. Therefore, it is extremely unlikely that more than one pre-ribosome is translocated across the NPC simultaneously.

The NPC is made up of around thirty different proteins, which in turn are arranged in different sub-modules to form an overall eightfold symmetrical structure (reviewed by [39]. In the central transport channel, the NPC is filled with multiple copies of about 10 different nucleoporins with numerous repeats of phenylalanines and glycines (FG-repeats), which ultimately form the central diffusion barrier with their intrinsically disordered domains. These FG repeats were estimated to have concentrations in the mM range within the NPC [40,41]. The diameter of the central channel is as narrow as ~40 nm, but it is known that it can vary under different physiological conditions. Diameters ranging from ~ 40–70 nm were observed depending on the cell state [42,43]. It is known that artificial transport cargoes with diameters of 17–36 nm can be translocated through the central channel [44]. This does not seem to require dilatation of the NPC, but the transport speed of these cargoes scales with the number of bound transport receptors.

Both mRNA and rRNA are exported as complex RNPs. The export of mRNA particles in mammalian cells requires nuclear RNA export factor 1 (NXF1 or Tap, Mex67 in yeast). Its mRNP binding is mediated by the adaptor protein NTF2-related export protein-1 (NXT1 or p15, Mtr2 in yeast). The directionality of mRNA export is achieved by the ATP-dependent RNA helicase DDX19 (Dpb5 in yeast) [45], which is bound to NUP214 on the cytoplasmic side of the NPCs and, together with Gle1 and IP6, remodels the translocated mRNP so that the bound export factors dissociate and the particle is released into the cytoplasm [46].

In yeast, Mex67/Mtr2 is also involved in the export of ribosomal subunits [47] but acts together with additional export factors, such as Rrp12p, Arx1, Ecm1, Bud20, Npl3, Gle2 and Xpo1 (Crm1 in mammalian cells) for the pre-60S particles (for review, see [2]. The Mex67/Mtr2 heterodimer interacts with the mature pre-60S subunit via a specific positively charged interaction site within the 5S rRNA, which is not present in NXF1/Nxt1 [47]. NXF1/Nxt1 therefore does not play a role in ribosomal export in human cells. Although ribosomes are evolutionarily highly conserved, the only transport receptor uniquely identified in the nuclear export of both yeast and human pre-60S particles is Xpo1/Crm1, which binds a C-terminal NES sequence of ribosome-bound NMD3 in a RanGTP-dependent manner. In human cells, RanGTP-binding exportin-5 (Exp5) is also required for the export of pre-60S particles [48]. After export, both Crm1 and Exp5 are removed from the pre-60S particles by GTP hydrolysis of Ran at the cytoplasmic face of the NPC, which provides the directionality of the translocation (reviewed in [49]).

Less is known about the nuclear export of pre-40S particles. As mentioned, Xpo1/Crm1 also plays a key role for enabling pre-40S particle export in yeast and vertebrates [50]. In yeast Dim2, Rio2 and Ltv1 serve as redundant adaptors for Crm1 binding (reviewed by [51]. Possibly Slx9 mediates Crm1 binding to Rio2. Also, the mRNA export factor Mex67/Mtr2 and Rrp12 promote export via interactions with the nucleoporins of the NPC. In human cells the protein kinase Rio2 [52] and PDCD2L [53] also probably serve as Crm1 binding sites to the particle, as do Pno1 and LTV1, and possibly also further yet unidentified ribosomal binding factors [51].

Analysis of pre-60S particle export by electron microscopy

An elegant electron tomographic study of high-pressure frozen yeast cells provided a detailed spatial overview of the actual NPC transit of the pre-60S subunits [54]. Delavoie et al. [54] recognized pre-60S subunits that migrated through NPCs based on their characteristic size and morphology. Approximately 4–5% of all analyzed yeast NPCs contained a large subunit in transit. The ribosomal subunits were exported sequentially and not in parallel, as the observation of two subunits within one NPC was extremely rare. The NPCs accommodated the large 20 nm-sized subunits well, there was no sign of a spatial dilation of the NPCs. This was expected, since the transport channel of yeast NPCs measures about 25 to 30 nm. Like for mammalian cells, a sequential export can be expected for yeast cells considering the production rate of 2000 ribosomes per minute and about 110 NPCs existing in the exponential growth phase [54–56], which suggests a translocation rate of about one subunit in two seconds per NPC. It can be assumed that this low transport rate does not overload NPCs. Within the NPC, the pre-60S particles moved almost exactly along the middle of the nuclear basket and the symmetry axis of the central channel. This was expected, since the pore-like internal scaffold of the NPC is filled with the intrinsically disordered FG Nups and a substantial amount of FG Nup mass must be displaced in order to facilitate the translocation of the bulky pre-60S particles. This is easiest to do in the center of the pore. When reaching the cytoplasmic side, the particles were distributed over a wider conical space. Here, their off-center positions suggested that the subunits emerging from the central channel, probably associated with nuclear transport factors, interacted with the asymmetric Nups on the cytoplasmic side of the NPC. In some cases, elongated densities reaching to the pre-60S particles were observed, presumably due to extended cytoplasmic Nups interacting with the transported pre-60S particles. Electron tomography obviously does not allow direct determination of the kinetic information of the pre-60S particle export. However, using a Jackson network queueing model that considers the expected overall translocation rate, NPC occupancy fraction and number of NPCs per cell, Delavoie et al. [54] obtained an estimate for the NPC translocation time of 90 ± 50 milliseconds (ms) for pre-60S particles. A recent study on the structural identity of yeast NPCs revealed that not all NPCs in a cell exhibit the same structure [57]. Rather, at least three different NPC isoforms have been identified: the majority of NPCs exhibits a single cytoplasmic and a single nuclear outer ring that together frame the inner ring, a second form has a single cytoplasmic and a double nuclear outer ring, both isoforms have nuclear baskets. A third NPC variant exhibited two single outer rings, but lacked the nuclear basket. The latter isoform was enriched over the nucleolus. Akey et al. [57] speculated that these isoforms may fulfill different functions, such as preferential protein import or RNA particle export. In case only a fraction of the yeast NPCs is available for export of the pre-60S subunit, the estimate of the translocation time must be adjusted accordingly.

Li et al. provided a high-resolution electron microscopic view of the export of pre-60S particles through NPCs [58]. They examined the structures of native pre-60S particles caught in affinity purified NPCs from yeast cells using cryo-electron microscopy. Within the structures they identified multiple factors with export functions including three copies of the Mex67/Mtr2 heterodimer, Ecm1, Gle2, Tma16, Arx1, Alb1, Bud20, Crm1–RanGTP and Nmd3. Presumably, the binding of Crm1–RanGTP to the identified Nmd3 could be verified. Interestingly, Li et al. [58] found that export factors bound to either the flexible regions of the RNP or the subunit-subunit interface region, forming numerous binding sites for the NPC-FG-domains. The localization of export factors in these regions suggests that these factors may play a role in preventing premature maturation of pre-60S particles and in reducing the negative charge of rRNA to facilitate nuclear export. In contrast to Delavoie et al. [54], Li et al. [58] observed the majority of pre-60S particles 10 nm off the central axis of the NPCs. They found also that Mex67-Mtr2 localizes approximately 10–15 nm closer to the NPC wall than the pre-60S particle center, indicating a preferential orientation of the pre-60S particles to the NPC wall during nuclear export. However, it cannot be excluded, that the preparation of affinity purified NPCs altered the details of the internal pore architecture. This was indicated by the fact that they observed up to four pre-60S particles in the same NPC, an observation that indicates a non-native situation.

Kinetic analysis of pre-60S particle export by high resolution fluorescence microscopy

Ruland et al. [23] used the above mentioned eIF6-HaloTag labeled pre-60S particles to study their translocation across eGFP-NTF2 labeled NPCs in live HeLa cells by single particle tracking. By adding JF549 in sub-nanomolar concentrations, they succeeded in optically separating the pre-60S particles. To directly observe the NPC transit of single pre-60S particles, they combined fast narrow-field single molecule microscopy and super resolution confocal Airyscan microscopy [23,59]. This enabled them to trace the pathways of single pre-60S particles across the NPCs with a colocalization precision of 35 nm. The observed translocation events were analyzed using an elaborate automatic data processing pipeline.

Ruland et al. [23] identified nearly 80 complete export events, in which they observed a nuclear approach to an NPC, interaction with that NPC, and subsequent release into the cytoplasm. The results of these experiments showed a translocation time of 24 ± 4 ms and a distribution of binding sites along the transport direction. The translocation time agreed well with the expected time extrapolated from the facilitated cargo molecule transport in previous studies [60]. Since Delavoie et al. [54] and their estimation of the translocation rate per NPC showed that the subunits are transported sequentially, the translocation time can be used to determine a maximally achievable transport rate of approximately 35–50 large subunits per second and NPC. This relates to a maximum mass flux of 90 to 125 MDa s⁻¹·NPC⁻¹, which is remarkably consistent with previous estimates for the maximum transport rate of roughly 100 MDa·s⁻¹ for a single human NPC [38,61] and clearly higher than the suggested in vivo minimum rate of 10–40 MDa for a growing cell [38,61. The in vivo measured translocation duration (24 ± 4 ms) in human cells was somewhat smaller than the value estimated by Delavoie et al. [54,60] for the translocation time of pre-60S particles in yeast (~90 ± 50 ms). A reason for this discrepancy might be the principal differences in nucleocytoplasmic transport between yeast and human cells. Moreover, as indicated above, the discrepancy might be due to functional differences in yeast NPCs, or alternatively, because a fraction of NPCs is not available for the export of pre-60S subunits due to occupancy by other large transport cargos.

A second approach to analyze the transport of pre-ribosomal subunits through single NPCs was performed using single-point edge-excitation sub-diffraction (SPEED) microscopy [60]. SPEED microscopy aims to optically isolate a single NPC from neighboring NPCs in live cells by observing an inclined or vertical diffraction limited region in the focal plane [62]. This microscopy method attempts to obtain spatio-temporal information of molecules transiting through biological channels with rotational symmetry at resolutions of approximately 10–20 nm and 0.4–2 ms.

To visualize the subunits, Junod et al. [60] also used an indirect labeling strategy, however, employing two fluorescent ribosomal proteins. To ensure that they only observed correctly assembled subunits, they used the FRET signal between two adjacent ribosomal particles on the respective subunit. They reported a fraction of 2/3 successful export events with an average translocation time of 13 ± 1 ms for the pre-60S and 9 ± 1 ms for the pre-40S subunits. The translocation times, however, were not obtained based on the commonly used exponential distribution functions, but rather using a Hill function, which prevents a direct comparison with the result from [23]. Junod et al. [60] performed extensive experiments on the inhibition of subunit export by leptomycin B (LMB). Surprisingly, they concluded from their respective inhibition experiments and on the basis of a sequence analysis of associated ribosomal proteins that the association of multiple Crm1 molecules per subunit provides the efficiency of pre-ribosomal subunit export. This was not in agreement with the cryo-electron microscopic study by Li et al. [58] on pre-60S particles in purified yeast NPCs.

In yeast at least eight different transport receptors – Mex67/Mtr2, Rrp12p, Arx1, Ecm1, Bud20, Npl3, Gle2 and Xpo1 – are loaded on a single pre-60S subunit, but for the human pre-60S particle so far only two distinct export receptor types are known: Crm1 and Exp5. Crm1 recognizes the NES in the ribosomal protein NMD3. Exp5 recognizes its rRNA cargo in a sequence-independent manner via double-stranded stem loops [48,49]. Therefore, it is quite probable, that several Exp5 molecules are bound to a single human pre-60S particle to ensure the necessary ‘transport receptor density’ for such a large transport substrate. This is further supported by the observation of fast NPC translocation of the pre-60S particles. Several studies have shown that a greater number of transport receptors on the cargo leads to a shorter translocation speed [38,44,63,64].

Spatial analysis of pre-60S particle export by high resolution fluorescence microscopy

The binding site distribution of pre-60S particles of abortive and successful translocation events was studied in the aforementioned studies of pre-60S export with comparable and instructive results. Trajectories of failed or unsuccessful transports, that is, when a particle enters the NPC from the nucleoplasm but then returns to the nucleoplasm, preferentially returned from the nuclear basket and rarely reached the NPC center [23,60].

For successful export, single particle transport measurements by electron microscopy (EM) and single molecule tracking enabled the analysis of the binding site distribution of subunits along the translocation axis and topology of NPCs. The data obtained by Delavoie et al. [54], Ruland et al. [23] and Junod et al. [60] agreed with each other and allowed a comparison with mRNP export. The microscopically measured distributions of trajectory positions were inherently less well resolved than the EM data, but both showed a clear, single maximum in the central region of the NPC, which was slightly shifted toward the cytoplasm, i.e., to the cytoplasmic ring of the NPC. No accumulation of binding events is observed in the NPC basket region. Therefore, Delavoie et al. [54] suspected that interactions between the pre-60S particles and the nuclear basket are negligible. This was confirmed by light microscopy, which showed that the duration of unsuccessful translocations was clearly shorter than that of successful translocations. This could be due to the lack of a quality control for pre-ribosomes in the nuclear basket [54,65], as it has been demonstrated for mRNPs [66,67].

Furthermore, both fluorescence microscopy studies observed a single rate limiting step for pre-60S export. In view of the notion that several transport factors presumably cover the large pre-60S subunit, probably several Exp5 and one Crm1, both bound to RanGTP, a single rate-limiting step is surprising at first glance. One might expect the release of various factors to be linked to several kinetic processes. To explain the one-step kinetics, we assume the following reaction process (Figure 1):

Figure 1.

Pre-60S particle export mechanism in human cells. Human pre-60S particle export begins with the non-FG interaction of Rae1 with Nup98 in the NPC. Entry into the FG-repeat domain of the NPC channel is enabled by the presence of multiple Exp5 and Crm1, and the preferred binding location is the internal region of the cytoplasmic ring. Interaction between pre-60S particle-bound Crm1 and Nup214 on the cytoplasmic side of the NPC is required for its release from this region. Nup214 catches particles out of the inner cytoplasmic ring domain and transfers it to Nup358/RanBP2, where Crm1 is released by RanGAP1 activation. Eventually, the pre-60S particles escaped into the cytosol after Exp5 release was induced by RanBP1/RanGAP1. The maximum diameter of the ribosomal subunit is approximately 25 nm, the diameter of the central transport channel of the NPC is about 40 nm. This figure was modified from [23].

For Crm1-mediated protein export it has been shown that nucleoporin Nup214 (Nup159 in yeast) plays a decisive role in the final step of export receptor dissociation [68–70]. Similarly, a key role of Nup214 has also been demonstrated in pre-60S subunit export in yeast and mammalian cells [48,71]. Nup88, Nup214, and Nup358 (also designated as RanBP2) are located in the cytoplasmic exits of mammalian NPCs. The Nup88/Nup214 sub-complex connects Nup358 to the NPC [72,73]. Nup358 increases GTP hydrolysis of RanGAP acting on RanGTP in the export complex, resulting in the release of the export receptor Crm1 of human pre-60S subunits [49,54,68]. Ruland et al. [23] suggested that only the Nup214/Crm1 interaction is required for its release from the NPC (Figure 1). The direct interaction between the pre-60S subunit bound to Crm1-RanGTP and Nup214 would literally grab the pre-60S particles out of the central NPC channel and take them into close proximity to NUP358. This translocation step would correspond to the rate-limiting step. Then, proximity to NUP358 results in GTP hydrolysis and the subsequent release of the pre-60S particles into the cytoplasm. We also hypothesize that Exp5-RanGTP export receptors, which are still bound to the pre-60S particle in this situation, are eventually separated by cytosolic RanBP1 [74].

Thus, both Exp5 and Crm1 enabled entry into the NPC and passage of the pre-60S particle, but only the Nup214/Crm1 interaction led to its release from the NPC. As mentioned above, Delavoie et al. [54] observed elongated densities, presumably due to the extended structures of cytoplasmic Nups interacting with transported pre-60s particles. This might show exactly the interaction between Crm1 and Nup159, leading to subunit release from the pore. This view is further supported by the observation that the pre-60S particles travel along a narrow channel within the yeast NPC center but are distributed over a cone on the cytoplasmic face of the NPC.

Inhibited and abortive translocations

Delavoie et al. [54] inhibited pre-60S particle transport using a temperature sensitive mutant of the NMD3 gene (nmd3–2), which encodes the adapter for Crm1 in pre-60S particles. They found that Crm1 is not required for entry into the nuclear basket, but for entry into the central region of the NPC, and concluded that the permeability barrier overcome by Crm1 begins at the nuclear side of the central transporter. This result was confirmed by light microscopy studies, which used LMB to inhibit the recruitment of Crm1 to the pre-60S particles.

Live cell studies on the transport of pre-60S particles have made it possible to distinguish between successful and failed transport processes without introducing mutations. In this respect, the results of two microscopic studies differed. We attribute these different results to the different labeling approaches of the pre-60S particles. Ruland et al. [23] directly observed pre-60S particles labeled with a bright organic fluorophore, whereas Junod et al. [60] evaluated the Förster resonance energy-transfer signal between GFP and mCherry, which was presumably much less efficient. Junod et al. [60] found 2/3 successful transport events, while Ruland et al. [23] observed that approximately 1/3 of attempted export events were successful. Similar success rates have been reported for mRNPs [24,75,76], so the value of approximately 1/3 of successful exports probably represents a general rule for large cargo particles.

In both studies, the distribution of binding sites for failed exports was found to have a clear maximum at the nuclear basket, and the corresponding pre-60S particles did not reach the hydrophobic interior of the NPC. It can be assumed that this was either due to occupancy of the NPC by another large transport cargo or that the particles were not yet sufficiently shielded by the export factors Exp5 and Crm1.

Summary and open questions

The assembly of ribosomal subunits is a multifaceted process involving a large number of accessory factors and RNA and protein components that interact in an incredibly complex way over time. It has been a challenge for researchers to identify the numerous components, assign a function to the respective players and uncover their detailed molecular mode of action. Countless biochemical, structural and cellular studies have led to a detailed understanding of the molecular role that many of the players play during the stepwise process of ribosome maturation [51]. In the past, however, the dynamics of ribosome maturation have been much less in the focus of research. In this review, we have focused on this aspect of ribosomal biogenesis. Microscopic visualization of pre-ribosomal particles within cells, ideally in vivo and especially during nuclear export, is very fruitful to understand crucial steps of ribosome maturation. Electron and single particle fluorescence microscopy are experimentally very laborious, but provide fascinating insights into this key cellular process. Above we discussed recent results on the dynamics and transport of pre-ribosomal particles in the nucleus and through the NPCs, focusing on the pre-60S particles.

One open question we encountered concerns the number of export factors covering the pre-60S particle and enabling NPC translocation. It has been shown that large transport cargoes require a greater number of transport factors on their surface [44,61]. There is only one Nmd3-bound Crm1, but it has been discovered that also Exp5 plays a role in human pre-60S nuclear export [48]. Exp5 recognizes structurally different RNAs and identifies them in a sequence-independent manner via double-stranded stem-loops [49]. It is possible that more than one Exp5 is loaded onto a single human pre-60S particle in this way to achieve the necessary ‘receptor density’ for such a bulky cargo. We speculated that Exp5 is not directly released by RanBP2 at the cytoplasmic face of the pore. Rather, we hypothesize that Exp5 release is achieved by RanBP1, but there is no data on exactly where and when this occurs (Figure 1). Also, it has been shown that in yeast not all NPCs are structurally identical, but it is not yet clear whether this is related to different functions. Furthermore, such a structural diversity has so far not been shown in mammalian cells. The question of different functions of structurally different NPCs could probably be answered, if export and import processes could simultaneously be observed in a single NPC. Theoretically, SPEED microscopy should be able to produce such data; however, this method has been employed in the past 10 years by a single research group only. A new, very promising approach allowing the analysis of 3D transport routes through the NPC was introduced by Chowdhury et al. [77]. In addition, Yu et al. [78] introduced a synthetic biology-enabled site-specific small-molecule labeling approach combined with highly time-resolved fluorescence microscopy to scrutinize the internal molecular structure and dynamics of NPCs. A combination of these emerging techniques would open completely new views on the study of pre-ribosomal particle export, as well as nucleocytoplasmic transport in general.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

UK and JPS drafted the paper and revised it critically for intellectual content. Both finally approved the published version and agree to be accountable for all aspects of the work.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.