Late ribosomal protein localization in Arabidopsis thaliana differs to that in Saccharomyces cerevisiae

Ribosome biogenesis is essential for cellular function and involves rRNA synthesis, rRNA processing and modification, and ribosomal protein assembly. Ribosome biogenesis factors and small nucleolar RNA assist these events. Ribosomal maturation takes place in the nucleolus, the nucleoplasm, and the cytosol in a coordinated and controlled manner. For example, some ribosomal proteins are thought to be assembled in the cytoplasm based on the observations in Saccharomyces cerevisiae. Here, we used cellular fractionation to demonstrate that cleavage of the 20S intermediate, the precursor to mature 18S rRNA, does not occur in the nucleoplasm of Arabidopsis thaliana. It most likely occurs in the cytoplasm. Further, we verified the proposed localization of RPS10e, RPS26e, and RPL24a/b in the nucleus and RPP1 in the nucleolus of A. thaliana by ribosome profiling, immunofluorescence, and analysis of the localization of GFP fusion proteins. Our results suggest that the order of events during ribosomal protein assembly in the ribosome biogenesis pathway differs between plants and yeast.

Ribosome biogenesis is essential for cellular function and involves rRNA synthesis, rRNA processing and modification, and ribosomal protein assembly. Ribosome biogenesis factors and small nucleolar RNA assist these events. Ribosomal maturation takes place in the nucleolus, the nucleoplasm, and the cytosol in a coordinated and controlled manner. For example, some ribosomal proteins are thought to be assembled in the cytoplasm based on the observations in Saccharomyces cerevisiae. Here, we used cellular fractionation to demonstrate that cleavage of the 20S intermediate, the precursor to mature 18S rRNA, does not occur in the nucleoplasm of Arabidopsis thaliana. It most likely occurs in the cytoplasm. Further, we verified the proposed localization of RPS10e, RPS26e, and RPL24a/b in the nucleus and RPP1 in the nucleolus of A. thaliana by ribosome profiling, immunofluorescence, and analysis of the localization of GFP fusion proteins. Our results suggest that the order of events during ribosomal protein assembly in the ribosome biogenesis pathway differs between plants and yeast.
Transcription of 35S rRNA (yeast) or 45/47S (mammals) precursor is catalyzed by RNA polymerase I [1-4], while 5S rRNA is transcribed independently by RNA polymerase III [5]. The transcription of 35S rRNA in the nucleolus leads to the formation of the initial preribosomal complex, known as the 90S preribosomal particle [4]. The maturation of 35S pre-rRNA to 18S, 5.8S, and 25S rRNA; the rRNA modifications; and the assembly of ribosomal proteins (RP) are regulated by different ribosome biogenesis factors (RBFs) and small nucleolar RNA (snoRNA) [4,6].
Ribosome biogenesis culminates in the cytoplasm by the maturation of two ribosomal subunits followed by an ultimate quality control process. In yeast, the cytoplasmic events of 40S ribosomal subunit maturation include the assembly of RPS10e (herein S10e) and formation of the 80S-like complex for the quality control events [7]. The latter induces cleavage of the pre-18S-rRNA at site D (D-cleavage) by the endonuclease Nob1 [8]. The assembly of S26e is discussed as last step of the maturation of the small ribosomal subunit [7]. During 60S maturation, the RPs L10 [9], L24a/b [10], L29 [11], L40 [12], P0, P1, and P2 [13] are incorporated into the ribosomal subunit in the cytoplasm. The assembly of L10 and L40 occurs last and requires the dissociation of Nmd3 by the function of Lsg1 [9]. Thus, while in fungi, at least 12 RPs (S1, S4, S6, S7, S8, S9, S11, S13, S17, S28, L4, and L13) are assembled in the 90S preribosomal particle [4,14], nine RPs are thought to be associated with the cytoplasm as part of the final maturation and the quality control cycle.
However, ribosome assembly in plants [15][16][17][18] or mammals [19,20] was found to be in parts distinct from that in yeast. For example, Lsg1 involved in the release of Nmd3 in the cytosol of yeast is localized in the nucleoplasm of plants [21]. Even more remarkable, a recent study of the protein content of the nucleolus and the nucleus by fractionation and proteome analysis revealed a putative localization of all ribosomal proteins in the nucleus [22]. This observation suggests that a different quality control mode operates in plants.
To explore these differences between fungi and plants, we selected four ribosomal proteins from Arabidopsis thaliana, namely S10e, S26e, L24a/b, and P1, the counterparts of which in yeast are known to be assembled in the cytoplasm [17,22,23], and determined their subcellular localization by fluorescence microscopy. We confirmed the localization of S10e, S26e, and L24a in the nucleoplasm and of P1 in the nucleolus. The consequence for the regulation of ribosome assembly in plants is discussed.

RNA isolation and northern blot analysis
The cellular fractionation of cells was performed as described [21,22]. For this purpose, A. thaliana root cell suspension cultures were grown for 5 days in the dark at 24°C while shaking at 150 r.p.m. After separation of cytoplasmic and nuclear fractions, RNA was isolated and analyzed by northern hybridization using previously described probes [21,24,25]. The signal density was quantified with ImageJ (imagej.net).

Analysis of protein distribution by confocal laserscanning microscopy
Transformed protoplasts were spotted on an object slide and analyzed using a Leica SP5 confocal microscope [26]. GFP and mCherry were excited at 488 and 568 nm, respectively. GFP and mCherry fluorescence were detected at 505-525 nm and 580-610 nm, respectively. Chlorophyll autofluorescence was excited at 488 nm and fluorescence monitored at 660-710 nm.

Analysis of protein distribution by immunofluorescence
Indirect immunofluorescence of A. thaliana mesophyll protoplasts and isolated nuclei were performed as described [27]. Cells and isolated nuclei were incubated overnight At5g41520 S10_F_KpnI ATATTAGGTACCATGATCATATCAGAGACTAACCGCCG S10_R_SpeI ATTTACTAGTAGGAAGATCAGATCCAGCAGCACCACC S26e At3g56340 S26_F_KpnI ATATTAGGTACCATGACTTTCAAGCGCAGGAATGG S10_R_SpeI ATTTACTAGTGGCACGAGGAGCAGCAGGAGCACC with antibodies against S10e (aS10e) diluted 1 : 5000 in PBS-1% bovine serum albumin. After washing, cells and nuclei were incubated with the secondary fluorochromelabeled antiserum in a 1 : 1000 dilution for 1.5 h in the dark. As a control, protoplasts and isolated nuclei were stained with 4 0 ,6-diamidino-2-phenylindole (DAPI) for localization of nuclei.

Ribosome profiling
For the analysis of GFP-fused ribosomal protein association with ribosomal subunits, GFP fusion proteins were expressed for 16 h in the dark in protoplasts of A. thaliana cell suspension culture. Subsequently, protoplasts were lysed and loaded onto a continuous 10-50% (w/v) sucrose gradient and centrifuged for 18 h at 100 000 g in a TST41.14 rotor. Collection of fractions was performed as described [25].

D-cleavage in plants occurs in the cytoplasm
The final steps of ribosome biogenesis in fungi occur in the cytoplasm [28]. Major hallmarks of 40S maturation are cleavage of the 20S rRNA at site D yielding the 18S rRNA and assembly of S10e and S26e. Further, 60S maturation in the cytoplasm involves displacement of several RBFs and assembly of seven ribosomal proteins. The last steps of 60S maturation are the release of Nmd3 catalyzed by Lsg1 and the assembly of L40 and L10. Contrary to this, the orthologue of Lsg1 in A. thaliana involved in maturation of the large subunit, namely Lsg1-2, is targeted to the nucleus [21]. This might indicate that the mechanism of Nmd3 regulation in plants is different to that in yeast. Consequently, we analyzed the distribution of 18S rRNA between cytosol and nucleus in A. thaliana (Fig. 1). The efficiency of fractionation was confirmed by immunodecoration with antibodies against fibrillarin serving as nucleolar marker and Nob1 as cytoplasmic marker (Fig. 1A, [24]). We analyzed the distribution of 27S and 18S-A 3 transcripts by northern blot (Fig. 1B) with previously described probes [21]. Quantification yielded an at least ten-fold higher concentration of both rRNA precursors in the nuclear fraction than in the cytoplasmic fraction (Fig. 1C). Next, we analyzed the distribution of the mature 18S and 25S rRNA by agarose gel migration (Fig. 1B).
Quantification yielded an about ten-fold higher concentration of the two rRNA in the cytosolic fractions than in the nuclear fractions (Fig. 1C). This result demonstrated that final processing apparently does not occur in the nucleoplasm. This observation together with the cytosolic localization of Nob1 suggests that the final processing of 18S rRNA occurs neither in the nucleolus nor in the nucleoplasm, but most likely in the cytosol. This is similar to the cytosolic 18S rRNA maturation in yeast. However, the processing of 20S pre-rRNA at site D is likely coordinated with the export of small subunit, because the 18S-A 3 rRNA precursor is barely if at all detectible in the cytoplasm.

Ribosomal proteins are targeted to the nucleus of A. thaliana mesophyll cells
The nuclear proteome analysis of A. thaliana, M. truncatula, rice, barley, and tomato [22,[29][30][31][32] led to identification of ribosomal subunit RPs that are usually of cytoplasmic in nature in yeast. In line with these observations, western blot analysis revealed that the S10e protein was detectable in nuclear fractions of A. thaliana (Fig. 1A). We selected four RPs and their orthologues to confirm the observed nuclear localization, namely two RPs of the small (S10e and S26e) and two of the large ribosomal subunit (L24a/b and P1) that are known in yeast to be assembled into ribosomes in the cytoplasm. The coding sequence of S10e was fused to the 3 0 or 5 0 end of the coding region of GFP (GFP-RPS10e or RSP10e-GFP, respectively). Both proteins were expressed in mesophyll protoplasts isolated from A. thaliana leaves. We cotransformed these cells with a plasmid coding for atFIB2-mCherry [21] to visualize nucleolus. We detected a nuclear and cytosolic localization by confocal fluorescence imaging, irrespective of the construct used ( Fig. 2A). This is in line with the protein distribution observed in proteomic studies [22].
Inspecting the localization of GFP-RPS26e or RPS26e-GFP by the approach described for S10e, we observed a nucleolar and a nuclear localization, while fluorescence in the cytoplasm was not observed. Nevertheless, our result is in line with proteome analysis of the nucleus isolated from A. thaliana, rice, and tomato (Table 2).
Next, we used C-and N-terminal GFP fusion constructs of RPL24a and RPL24b for protoplast transformation, respectively. Remarkably, GFP fluorescence was detected in the nucleolus, nucleus, and cytoplasm (Fig. 2B). This shows that nuclear targeting was not dependent on the position of GFP. Again, this observation confirms the nuclear localization of the proteins determined by proteomics in A. thaliana and other plant species (Table 2). The nucleolar GFP fluorescence observed for both proteins is likely specific, because RPS10e fused to GFP did not yield nucleolar GFP fluorescence ( Fig. 2A).
Further, we analyzed the localization of the three orthologues of P1. Proteomic studies of A. thaliana, rice, and tomato identified this protein in the nucleus ( Table 2) and even a nucleolar localization was suggested [22]. Using GFP fusion proteins, we observed GFP fluorescence in the nucleoplasm and cytoplasm (Fig. 2B). On the example of the orthologue P1.3, we demonstrated that nuclear localization is not dependent on the position of the GFP fusion. Moreover, the transport appears to be active as a double GFP fusion exceeding the size limit of the nuclear translocation pore [33] confirming nuclear localization as well (Fig. 2B).
To confirm the association of S26e and L24a with ribosomal complexes, protoplasts were transformed with RPS26e-GFP or RPL24a-GFP. After expression, cells were solubilized and ribosome profiling was performed based on sucrose density centrifugation. We observed a fraction of RPS26e-GFP bound to 40S ribosomes, while a portion was observed in the soluble fraction as well (Fig. 2C, left). The latter observation and the absence of a strong GFP signal in the cytoplasm after expression of RPS26e-GFP might suggest that GFP-tagged S26e can be assembled into pre-40S ribosomes, but the GFP tag interferes with the assembly of functional ribosomes in the cytoplasm. Nonfunctional ribosomal units are degraded explaining the low abundance of GFP fluorescence [34]. In turn, RPL24a-GFP was largely assembled into ribosomes (Fig. 2C, right).
Next, we used antibodies against S10e to detect the endogenous protein in cells by immunofluorescence. Staining entire cells yielded a strong signal in the cytoplasm (Fig. 3, left), likely representing S10e in cytosolic ribosomes. When isolated nuclei were incubated with antibodies against S10e, we observed a signal in the nucleoplasm, but not in the nucleolus (Fig. 3,  right). These results are consistent with the localization of S10e fused to GFP as well.
Yeast S10 is assembled into ribosomes in the cytoplasm [7]. It is likely that the assembly occurs in the nuclear pore as yeast S10 is important for the export of the small subunit from the nucleus [14]. Similarly, Arabidopsis S26e is already present in the nucleoplasm (Fig. 2, Table 2), while in yeast, it is assembled into ribosomes of cytoplasm markedly after release of Pno1 [7,35] and D-cleavage. [14]. In line with a nuclear assembly of S26e into ribosomes in plants, mass spectrometry revealed that Pno1 was solely detected in nucleolar and nuclear fractions of A. thaliana [22].
Mature 18S rRNA is exclusively present in the cytoplasm of A. thaliana, while the 18S-A 3 transcript is observed in the nucleoplasm (Fig. 1). This suggests that pre-rRNA cleavage at site D likely is associated with the transport from nucleoplasm to cytoplasm. This would resemble the path in yeast [8,14]. However, it remains to be established whether D-cleavage occurs in the vicinity of the nuclear pore complex. Maturation of the 60S ribosomal subunit differs between plants and yeast as well (Fig. 4B). Yeast L26 is assembled into ribosomes in the nucleus [36], while the plant protein was only detected in the cytoplasm [22]. In turn, many ribosomal proteins assembled into ribosomes in the cytoplasm in yeast were identified in the nucleus of various plant species (Table 2, [22,[29][30][31][32]). At least L24, L29, and L10, as well as Lsg1-2, are targeted to the nucleolus (Fig. 2, [21,22]). P1 associates with the precursor of the 60S subunit in the nucleoplasm (Fig. 2). In contrast, yeast L10, L24, L29, and P1 are incorporated into ribosomes in the cytoplasm [9][10][11]13]. For example, L24 replaces RLP24 during yeast ribosomal maturation [10], while alternative pattern of L24 incorporation into ribosomes is evident as RLP24 is only found in the nucleolus of A. thaliana [22]. However, L40 is an example where association with the ribosome is similar between yeast and A. thaliana (Table 2; [12]).
Thus, we provide experimental evidence for differences in the order of events during final maturation of ribosomes in plants when compared to yeast. Whether the initial association of the ribosomal proteins with the rRNA occurs at the site of function in the mature ribosome remains to be established. Moreover, the existing results do not contradict the formation of 80S-like particle for quality control in the cytoplasm [7] although its formation remains to be shown for plant systems.