Puf6 and Loc1 Are the Dedicated Chaperones of Ribosomal Protein Rpl43 in Saccharomyces cerevisiae

Ribosomal proteins are highly expressed, and the quality of ribosomal proteins must be rigorously controlled to build up a functional ribosome. Rpl43, ribosomal protein large subunit 43, is located nearby the E-site of ribosomes. In our previous study, we found that Puf6, Loc1, and Rpl43 form a trimeric complex in Saccharomyces cerevisiae. Rpl43 protein levels are under-accumulated in the absence of PUF6 or LOC1. However, why the loss of Puf6 or Loc1 decreased the protein levels of Rpl43 remained unclear. In the present study, we further dissected the connections among these three proteins and found that the processing defects of pre-ribosomal RNA in puf6Δ and loc1Δ are similar to those of the mutant with depletion of Rpl43. The stability of newly synthesized Rpl43 protein decreased slightly in puf6Δ and significantly in loc1Δ. We also found that Puf6 and Loc1 could interact with nascent Rpl43 co-translationally via the N-terminus of Rpl43. While the association and dissociation of Rpl43 with karyopherins did not depend on Puf6 and Loc1, Puf6 and Loc1 interacted with nascent Rpl43 in collaboration. While the N-terminus of Puf6 contained nuclear localization signals for transport, the PUF (Pumilio) domain was essential to interaction with Loc1, Rpl43, and 60S subunits. The C-terminus of Loc1 is more important for interaction with Puf6 and Rpl43. In this study, we found that Puf6 and Loc1 are the dedicated chaperones of ribosomal protein Rpl43 and also analyzed the potential interaction domains among the three proteins. Correct formation of the Puf6, Loc1, and Rpl43 ternary complex is required to properly proceed to the next step in 60S biogenesis.


Introduction
Ribosomes are massive complexes composed of ribosomal proteins and RNAs. Ribosome biogenesis is necessary for cell growth and proliferation and represents one of the major energy cost pathways of cells. Rapidly growing yeast cells generate about 2000 ribosomal subunits per minute [1]. Indeed, about 60% of the total transcription in a cell is devoted to ribosomal RNA synthesis to support such large-scale synthesis. The multiple steps of ribosome biogenesis allow cells to precisely regulate the balance between supply and demand. Cells can regulate at the RNA transcriptional and ribosomal protein synthesis levels [1][2][3][4].
Over 200 transacting (non-ribosomal) factors have been identified in the ribosome assembly process; these factors present various and important functions, such as maintenance of the stability of nascent ribosomal proteins, functional proofreading of ribosomes, assurance of correct interaction between ribosomal proteins and rRNAs, and processing and folding of rRNAs. Transacting factors join in the ribosome synthesis process sequentially to produce ribosomes assembled through a hierarchical mechanism (see refs [5][6][7][8][9][10][11]). Ribosomes in yeast are composed of 79 ribosomal proteins encoded by 138 ribosomal protein genes. Many of the ribosomal proteins are encoded by two genes to generate isoforms with identical or highly similar amino acid sequences [12]. To generate this enormous amount of ribosomal proteins, cells in the logarithmic phase devote approximately 50% of their total transcription activities of the RNA polymerase II [13].
Ribosomal protein large subunit 43 (Rpl43) and its human homolog, L37A, have been renamed L43e in the new nomenclature system [34]. Rpl43 is essential and uniquely present in Archaea and eukaryotes but not in prokaryotes. Rpl43 is a small ribosomal protein in the large subunit with only 94 amino acids. It is located in the vicinity of the E-site of 60S subunits and in close contact with Rpl2 [35]. Rpl2 and Rpl43 are found close to the peptidyl transferase center and required for 7S rRNA production [36]. Two homologs of Rpl43 have been identified in yeast, namely, RPL43A and RPL43B, and these homologs share 90% identity in DNA sequence and 100% identity in amino acid sequence. Puf6 and Loc1 have previously been shown to recognize the 3 untranslated region (UTR) and control the asymmetric localization of ASH1 mRNA in yeast [37]. In our previous study, we found that Puf6 and Loc1 form a complex with Rpl43. In the absence of Puf6 or Loc1, the loading and protein levels of Rpl43 are decreased [38]. Puf6 and Loc1 are loaded onto 60S together with Rpl43. If Rpl43 is depleted, no effect on loading is observed, but the release of Puf6 and Loc1 is impaired. Thus, recruitment of Rpl43 is required for the release of Puf6 and Loc1 [38]. However, the detailed molecular mechanism behind these events remains unclear. In the present work, we found that Puf6 and Loc1 are critical to maintaining the stability of Rpl43. Mutation of any one protein results in blockages at similar steps during biogenesis. We further dissect how these three proteins interact with each other and conclude that Puf6 and Loc1 are the dedicated chaperones of Rpl43.

The Analysis of Protein Stability and Western Blotting
Yeast cells were grown in liquid culture until OD 600~0 .5, cycloheximide (200 µg/mL) was added, and the cells were harvested after a 10 min interval. The cell lysates were normalized and the samples were centrifuged for 1 h at 4 • C and 80,000 rpm in an MLA130 rotor (Beckman, Indianapolis, IN, USA) to separate the ribosomes and nascent (unincorporated) ribosomal proteins to the pellets and supernatant, respectively. The pellets (ribosome fraction) were re-dissolved in 1× Laemmli buffer. The proteins in the supernatant (free form) were precipitated by TCA and re-dissolved in 1× Laemmli buffer. The proteins were detected and analyzed by Western blotting.

Immunoprecipitation
Cultures were grown to an OD 600 of~0.5 in selective medium. For the preparation of protein extracts, cells were resuspended in IP buffer (20 mM Tris pH 7.5, 50 mM NaCl, 6 mM MgCl 2 , 10% glycerol, 1 mM PMSF (phenylmethylsulfonyl fluoride), and 1 mM leupeptin), lysed by vortexing with glass beads, and clarified by centrifugation. α-c-myc antibody (9E10) and protein-A agarose beads (GE Healthcare, Chicago, IL, USA) were added to normalized samples and rocked at 4 • C. The beads were washed three times with IP buffer and the proteins were eluted in 1× Laemmli sample buffer. Subsequently, proteins were separated in SDS-PAGE and observed by Western blotting.
To identify whether Puf6 and Loc1 associate with nascent Rpl43 co-translationally, we followed the procedures described in the previous study [32]. Briefly, cells were treated with cycloheximide and lysed in IP buffer containing RNasin ® ribonuclease inhibitors (Promega, Madison, WI, USA). IgG (Immunoglobulin G) sepharose beads (GE Healthcare) were added to immunopurify TAP-Tag proteins. After washes, samples were eluted in elution buffer (100 mM Tris pH 8, 10 mM EDTA, 1% SDS) and RNA was extracted using TRI reagent (Sigma, St. Louis, MO, USA). cDNA was synthesized using Reverse Transcriptase (Ambion, city, Waltham, MA, USA). The levels of translating RPL43 mRNA were determined by quantitative real-time PCR (Applied Biosystems, Waltham, MA, USA) with Power SYBR Green (Thermo, Waltham, MA, USA) detection. RPL5 and RPL10 mRNA were included as controls.

Polysome Profile Analysis
Yeast cells were grown to an OD 600 of about 0.2-0.3. Cycloheximide was added to a final concentration of 50 µg/mL and incubated continuously for 10 min. Cells were lysed with polysome lysis buffer (20 mM Tris HCl pH7.5, 8 mM MgCl 2 , 12 mM mercaptoethanol, 100 mM KCl, 50 µg/mL cycloheximide, 1 mM PMSF, 1 mM leupeptin), and nine OD 260 units of protein extracts were loaded onto linear 7-47% sucrose gradients. After 2.5 h of centrifugation at 4 • C and 40,000 rpm in an SW41 Ti rotor (Beckman, Indianapolis, IN, USA), gradient fractions were continuously measured for their absorbance at 254 nm and sucrose fractions were collected (BR-188, Brandel, Gaithersburg, MD, USA). For detection of the RNA distribution, fractions were precipitated with 500 µL Trizol and 200 µL chloroform to extract RNA, then isopropanol was added to precipitate. The levels of RNA were quantitated by qPCR.

Northern Blotting
Northern blotting was used to analyze the static state levels of pre-rRNAs. The total RNA prepared with TRI reagent (Sigma, St. Louis, MO, USA) was resolved on a formaldehyde agarose gel for large-molecular-weight rRNAs or urea polyacrylamide gel for small-molecular-weight rRNAs. The RNAs were transferred to a nitrocellulose membrane. The probes were labeled with a Biotin 3 end labeling kit (Thermo, Waltham, MA, USA) and continually hybridized and detected with North2South ® Chemiluminescent hybridization and detection kits (Thermo), respectively. Table 3 lists the probe sequences.

In Vitro Interaction Experiments
Various PUF6, LOC1, and RPL43 mutants were cloned into the pET28a or pGEX-4T3 vectors and overexpressed in Escherichia coli BL21 (DE3) as recombinant proteins with His 6 or GST tags. Cell extracts were prepared by sonication (Chrom Tech, Taipei, Taiwan) in TEN100 lysis buffer (20 mM Tris, pH 7.4, 0.1 mM EDTA, and 100 mM NaCl). Whole-cell extracts containing GST or GST-tagged proteins were first incubated with glutathione beads for 1 h at 4 • C and washed with chilled TEN100 buffer five times. Protein extracts containing prey protein were subsequently incubated with glutathione beads for 1 h at 4 • C. Beads were washed with buffer five times and eluted with 1× SDS sample buffer. Samples were separated in SDS-PAGE and stained with Coomassie blue. The interacting proteins were further confirmed with anti-Puf6, anti-Loc1, and anti-GST antibodies in Western blotting.

Mutations of Puf6, Loc1, and Rpl43 Show Similar rRNA Processing Defects
In our previous study, Rpl43 formed a complex with Puf6 and Loc1 [38]. Moreover, deletion of Puf6 and Loc1 impacted the protein level and loading of Rpl43 onto 60S [38]. Pre-90S and pre-60S ribosomal subunits were immunoprecipitated by transacting factors at different stages to examine whether they are loaded onto the corresponding ribosomal subunits at similar stages ( Figure 1A). Nop56 and Noc4 were used for purification of pre-90S; Ssf1, Noc2, Nop15, Cic1, Brx1, and Tif6 were used for purification of early pre-60S from the nucleolus to the nucleus; Rix1 and Arx1 were used for purification of late pre-60S in the nucleus; and Rei1 was used for purification of pre-60S in the cytoplasm [5,7,[40][41][42]. Puf6, Loc1, and Rpl43 could be detected at the 90S stage. However, the major peak of these three proteins in pre-60S was observed at the nucleolar stage. Puf6 and Loc1 were not present at Rix1, Arx1, and Rei1. Although depletion of Rpl43 majorly influences the late stage of 7S rRNA processing [43], it has been shown to copurify with 35S, 32S, and 27S rRNAs [44]. This is consistent with the finding that Rpl43 is loaded onto the 90S pre-ribosome. The relative intensities compared to initial state (C) and WT strain (D) were calculated.

Puf6 and Loc1 Are Important for the Protein Stability of Rpl43
Ribosomal proteins are highly unstable because their structures feature highly positive charges and disordered extensions [14]. These extensions are inserted into ribosomes for contact with rRNAs and other ribosomal proteins to coordinate the conformational changes of the giant complex [14]. Several dedicated chaperones have recently been identified to maintain the protein stability of ribosomal proteins [14]. Since Puf6 and Loc1 could form a trimeric complex with Rpl43 ( Figure 1 and [38]), we tested how Puf6 and Loc1 impacted the protein levels of Rpl43. The status of rRNA processing in rpl43 mutated, puf6∆, and loc1∆ strains was evaluated. The probes used to detect the intermediates of rRNA processing are shown in Figure 1B. While deletion of RPL43A or RPL43B resulted in slight growth defects, deletion of both homologs caused lethality [45]. RPL43 was constructed under a GAL-driven promoter and complemented in an rpl43A∆rpl43B∆ strain. When glucose was added to this system to switch off the gene expression of GAL::RPL43, accumulation of 35S, 27S, 27SA2, and 23S was detected, which may be due to the blockage of C2 cleavage ( Figure 1C).

Puf6 and Loc1 Are Important for the Protein Stability of Rpl43
Ribosomal proteins are highly unstable because their structures feature highly positive charges and disordered extensions [14]. These extensions are inserted into ribosomes for contact with rRNAs and other ribosomal proteins to coordinate the conformational changes of the giant complex [14]. Several dedicated chaperones have recently been identified to maintain the protein stability of ribosomal proteins [14]. Since Puf6 and Loc1 could form a trimeric complex with Rpl43 ( Figure 1 and [38]), we tested how Puf6 and Loc1 impacted the protein levels of Rpl43.
The levels of Rpl43A and Rpl43B were examined in puf6∆ and loc1∆ cells to analyze how Puf6 and Loc1 affect the protein level of Rpl43. While Rpl43A and Rpl43B decreased slightly in puf6∆, both decreased significantly in loc1∆ cells (Figure 2A). Overexpression of Puf6 or Loc1 alone did not or slightly increased the protein level of Rpl43, respectively. Overexpression of Puf6 and Loc1 together increased the protein levels of Rpl43 about 2.1-fold ( Figure 2B). The levels of Rpl43A and Rpl43B were examined in puf6∆ and loc1∆ cells to analyze how Puf6 and Loc1 affect the protein level of Rpl43. While Rpl43A and Rpl43B decreased slightly in puf6∆, both decreased significantly in loc1∆ cells (Figure 2A). Overexpression of Puf6 or Loc1 alone did not or slightly increased the protein level of Rpl43, respectively. Overexpression of Puf6 and Loc1 together increased the protein levels of Rpl43 about 2.1-fold ( Figure 2B).
The protein stability of nascent Rpl43 was further tested in puf6∆ and loc1∆. Cycloheximide was added to inhibit protein synthesis, and samples were collected at various time points. The majority of the ribosomal proteins are incorporated into mature ribosomal subunits and have long half-lives. Therefore, samples were spun on the ultracentrifuge to separate nascent ribosomal proteins into their free form and ribosomal subunits into pellets ( Figure 2C, upper panel). Within the 20-minute treatment, the levels of Rpl43A and Rpl43B in the ribosome fractions remained constant ( Figure 2C, lower panel). In contrast, the stability of nascent Rpl43 proteins was decreased significantly in puf6∆ and loc1∆ ( Figure 2D). The data above suggest that Puf6 and Loc1 are the chaperones of Rpl43, important for maintaining the protein stability of newly synthesized Rpl43.  The protein stability of nascent Rpl43 was further tested in puf6∆ and loc1∆. Cycloheximide was added to inhibit protein synthesis, and samples were collected at various time points. The majority of the ribosomal proteins are incorporated into mature ribosomal subunits and have long half-lives. Therefore, samples were spun on the ultracentrifuge to separate nascent ribosomal proteins into their free form and ribosomal subunits into pellets ( Figure 2C, upper panel). Within the 20-minute treatment, the levels of Rpl43A and Rpl43B in the ribosome fractions remained constant ( Figure 2C, lower panel). In contrast, the stability of nascent Rpl43 proteins was decreased significantly in puf6∆ and loc1∆ ( Figure 2D). The data above suggest that Puf6 and Loc1 are the chaperones of Rpl43, important for maintaining the protein stability of newly synthesized Rpl43.

Puf6 and Loc1 Can Associate with Nascent Rpl43 but Are Not Required for Kap Interaction or Release
Rpl43 is a highly conserved protein. Several Rpl43 sequences were chosen from several model organisms for comparison, including yeast (Saccharomyces cerevisiae), nematode (Caenorhabditis elegans), human (Homo sapiens), mouse (Mus musculus), and fly (Drosophila melanogaster). The yeast Rpl43 sequence shares about 70% identity with other sequences. For instance, yeast Rpl43 shares 66.3% identity and 79.3% similarity with human Rpl43 ( Figure 3A).
The N-terminus of Rpl43 is a random coil followed by an alpha helix, a conserved C4-type Zn finger domain composed of four beta-sheets, and then another alpha helix ( Figure 3B adapted from PDB: 4V7F [46]). Zn finger domains are relatively small protein motifs important for interactions with DNA, RNA, proteins, or even lipids. Four conserved cysteines can chelate metals, such as Zn or Fe, or not bind ions. This motif is very important for the structure and functions of proteins. Four cysteines at amino acids 9, 42, 57, and 60 in Rpl43 form a C4-type Zn finger domain and are conserved across different species (Figure 3A,B).
Ribosomal proteins are translated in the cytoplasm and are required to be imported into the nucleus for assembly with rRNAs. Karyopherins are a group of proteins which play a central role in nucleocytoplasmic transport [47]. Some dedicated chaperones have been shown to be important for the ribosomal proteins to associate or disassociate with the karyopherins [14]. Kap121 is one of the major karyopherins responsible for the transport of ribosomal proteins [48]. Kap121 was fused with maltose-binding protein (MBP) and tested for its interaction with Rpl43 ( Figure 3C). Kap121 can interact with Rpl43 directly. Deletion of the N-terminal coil (∆aa1-21) decreased the interaction with Kap121 ( Figure 3C, lane 8). Further deletion of the N-terminal helix (∆aa1-34) ceased the interaction with Kap121 ( Figure 3C, lane 9). However, deletion of the C-terminal helix (∆aa74-92) did not change the interaction with Kap ( Figure 3C, lane 10). Therefore, the N-terminus is important for Rpl43 to interact with Kaps. Puf6, Loc1, and Rpl43 were added separately or in combination for binding assays with Kap121 to test whether Puf6 or Loc1 was required for the interaction or dissociation of Rpl43 with Kap. Puf6, Loc1, and Rpl43 can interact with Kap individually, and the addition of Puf6 or Loc1 did not change the interaction of Rpl43 with Kap ( Figure 3D).
Dedicated chaperones may interact with nascent ribosomal proteins when they have just been translated. In this situation, the dedicated chaperones recognize the N-terminus and capture nascent ribosomal proteins co-translationally [32]. The dedicated chaperones can then pull down the mRNA of its associated ribosomal protein ( Figure 3E, upper panel). To test this idea, cells were treated with cycloheximide to halt translation, and Puf6 and Loc1 were applied to test whether they can interact with translating RPL43 mRNA. Compared with the controls, RPL5 and RPL10 mRNAs, Puf6 and Loc1 showed specificity toward RPL43 mRNA ( Figure 3E). Since Puf6, Loc1, and Rpl43 have been shown to form a trimer [38], we further tested this association in puf6∆ or loc1∆ mutants and found that mutation of either one significantly dropped the interaction with RPL43 mRNA ( Figure 3F). The results thus far imply that Puf6 and Loc1 can associate with nascent Rpl43 in a cooperative manner and are not required for the Kap interaction or release.

Determination of the Potential Interaction Sites among the Three Proteins
Several mutants were constructed to further dissect the interaction domains among the three proteins. The Rpl43 mutants described above ( Figure 3B

Determination of the Potential Interaction Sites among the Three Proteins
Several mutants were constructed to further dissect the interaction domains among the three proteins. The Rpl43 mutants described above ( Figure 3B), deletion of the N-terminal coil (∆N: ∆1-21), deletion of the N-terminal helix (∆NH: ∆1-34), deletion of the middle (∆M: ∆39-60), deletion of the C-terminal helix (∆C: ∆74-92), and mutation of the four conserved cysteines (4C) were used to determine the potential interaction site of Rpl43 with Pu6 and Loc1. All of these mutants could not support the growth of yeast ( Figure 4A). To carry out the in vitro interaction assay, Rpl43 proteins with GST (glutathione S-transferase) fusion were used to examine the interaction with His 6 -tagged Puf6, ∆N100Puf6 (PUF domain only), and Loc1 ( Figure 4B,C). Because the molecular weight of recombinant Loc1 is a similar size to that of GST-Rpl43, the signals could only be detected by Western blotting. The data revealed that FL and rpl43∆C can interact with Puf6 ( Figure 4B) and Loc1 ( Figure 4C), but deletion of the N-terminus (rpl43∆N and rpl43∆NH) destroyed this interaction ( Figure 4B,C). This is consistent with the observation that Puf6 and Loc1 can capture nascent Rpl43 co-translationally ( Figure 3E): once the N-terminus emerges from the polypeptide exit tunnel (PET) of the ribosome, the dedicated chaperone loads and protects its accompanying ribosomal protein immediately [14]. Next, we tested whether these mutants could be incorporated into 60S subunits by using Brx1 to immunoprecipitate pre-60S. All of the mutants, except rpl43∆C, could be found in 60S ( Figure 4D). The C-terminal helix is buried in 60S and interacts with Rpl2 ( Figure 4E). Therefore, the N-terminus is important for the associations of Rpl43 with Puf6 and Loc1, and the C-terminal helix is important for its interaction with 60S. growth of yeast ( Figure 4A). To carry out the in vitro interaction assay, Rpl43 proteins with GST (glutathione S-transferase) fusion were used to examine the interaction with His6-tagged Puf6, ΔN100Puf6 (PUF domain only), and Loc1 ( Figure 4B,C). Because the molecular weight of recombinant Loc1 is a similar size to that of GST-Rpl43, the signals could only be detected by Western blotting. The data revealed that FL and rpl43ΔC can interact with Puf6 ( Figure 4B) and Loc1 ( Figure  4C), but deletion of the N-terminus (rpl43ΔN and rpl43ΔNH) destroyed this interaction ( Figure 4B,C). This is consistent with the observation that Puf6 and Loc1 can capture nascent Rpl43 cotranslationally ( Figure 3E): once the N-terminus emerges from the polypeptide exit tunnel (PET) of the ribosome, the dedicated chaperone loads and protects its accompanying ribosomal protein immediately [14]. Next, we tested whether these mutants could be incorporated into 60S subunits by using Brx1 to immunoprecipitate pre-60S. All of the mutants, except rpl43ΔC, could be found in 60S ( Figure 4D). The C-terminal helix is buried in 60S and interacts with Rpl2 ( Figure 4E). Therefore, the N-terminus is important for the associations of Rpl43 with Puf6 and Loc1, and the C-terminal helix is important for its interaction with 60S. Puf6 belongs to the PUF family of RNA-binding proteins, which are post-transcriptional regulators. Nearly all PUF homology domains contain eight Pumilio (PUM) repeats (~36 amino acids) that form an α-helical repeat rainbow-like structure [49][50][51]. Uniquely different from these proteins, Puf6 belongs to the PUF family of RNA-binding proteins, which are post-transcriptional regulators. Nearly all PUF homology domains contain eight Pumilio (PUM) repeats (~36 amino acids) that form an α-helical repeat rainbow-like structure [49][50][51]. Uniquely different from these proteins, Puf6 and its human homolog, Puf-A, contain 11 PUM repeats and assemble in an L shape. (Figure 5A-protein modeling with PDB (4WZR) as a template; the PUF domain of Puf6 spans from Met116 to the end of the sequence) [52]. The N-terminal 115 amino acids of Puf6 are not included in the PUF domain. It is a region rich with glutamic acids and aspartic acids (D/E) with an unidentified functional role and nuclear localization signal (NLS) to drive Puf6 into the nucleus ( Figure 5B). Several Puf6 mutants, including ∆DE region, ∆N, ∆NPUF, ∆PUF, ∆CPUF, and NLS mutants, were constructed. The mutants and wild-type Puf6 were transformed into the puf6∆ strain for growth tests ( Figure 5C). While the wild-type, ∆D/E, NLS, and ∆N mutants grew well, ∆NPUF, ∆PUF, and ∆CPUF did not complement the growth of puf6∆ cells ( Figure 5C).
The mutants and wild-type Puf6 were transformed into the puf6∆ strain for growth tests ( Figure 5C). While the wild-type, ∆D/E, NLS, and ∆N mutants grew well, ∆NPUF, ∆PUF, and ∆CPUF did not complement the growth of puf6∆ cells ( Figure 5C).
We monitored the distribution of different variations of Puf6 mutants in the cells ( Figure 5D). Wild-type and ∆D/E mutants were mainly localized in the nucleus and nucleolus. ∆N and NLS were mostly distributed in the cytoplasm, implying that this is genuine NLS signal. However, while they still can complement the growth of puf6∆ cells (Figure 5C), this suggests the redundancy of NLS. ∆NPUF and ∆PUF lost their nucleolar enhancement and localized only in the nucleus, suggesting that the PUF domain is important in ribosome synthesis ( Figure 5D). ∆D/E, ∆PUF, and ∆N were used to demonstrate interactions in vivo, and only ∆PUF lost its interaction with Loc1, Rpl43, and 60S ( Figure 5E,F). An intact PUF domain is required to interact with Rpl43 and Loc1 in vitro ( Figure 5G,  lanes 10 and 15). The N-terminus ( Figure 5G, lanes 9 and 14) or partial PUF domain ( Figure 5G, lanes 8 and 13) did not support these interactions. These findings support that the PUF domain is required for Puf6 to form a ternary complex with Rpl43 and Loc1 and to interact with pre-60S.  We monitored the distribution of different variations of Puf6 mutants in the cells ( Figure 5D). Wild-type and ∆D/E mutants were mainly localized in the nucleus and nucleolus. ∆N and NLS were mostly distributed in the cytoplasm, implying that this is genuine NLS signal. However, while they still can complement the growth of puf6∆ cells ( Figure 5C), this suggests the redundancy of NLS. ∆NPUF and ∆PUF lost their nucleolar enhancement and localized only in the nucleus, suggesting that the PUF domain is important in ribosome synthesis ( Figure 5D). ∆D/E, ∆PUF, and ∆N were used to demonstrate interactions in vivo, and only ∆PUF lost its interaction with Loc1, Rpl43, and 60S ( Figure 5E,F). An intact PUF domain is required to interact with Rpl43 and Loc1 in vitro ( Figure 5G,  lanes 10 and 15). The N-terminus ( Figure 5G, lanes 9 and 14) or partial PUF domain ( Figure 5G, lanes 8 and 13) did not support these interactions. These findings support that the PUF domain is required for Puf6 to form a ternary complex with Rpl43 and Loc1 and to interact with pre-60S.
Preliminary sequence analysis revealed that Loc1 does not contain an obvious functional domain, and its structure is not yet solved. Sequence analysis also indicated that two potential NLSs could be found at the N-and C-termini and one short stretch of the DE-rich region of the protein ( Figure 6A). The NLS1, ranging over amino acids 63-94, is similar to a classical bipartite NLS sequence, i.e., KKDKKGKYSEKDLNIPTLNRAIVP GVKIRRGK, revealing that two clusters of basic amino acids are separated by a 10-12-amino-acid linker [53]. We constructed N-and C-terminal truncation variants of Loc1 proteins for functional studies, i.e., ∆N50 deleted the first 50 amino acids, ∆N100 further deleted to eliminate NLS1, ∆C64 deleted the NLS2, ∆C74 deleted the NLS2 and partial D/E region, and ∆C84 deleted the NLS2 and the whole D/E region ( Figure 6A). Loc1 mutant plasmids were first transformed into the loc1∆ strain for complementation assays ( Figure 6B). Deletion of 64 amino acids from the C-terminus (loc1∆C64) did not impair the growth of Loc1, but further deletion of 10 or 20 more amino acids (loc1∆C74 and loc1∆C84) decreased cell growth ( Figure 6B). Deletion of 50 (loc1∆N50) and 100 (loc1∆N100) amino acids from the N-terminus impaired growth slightly and significantly, respectively ( Figure 6B). Next, we observed the localization of these loc1 mutants ( Figure 6C). Loc1 was predominantly localized in the nucleolus. Interestingly, Loc1 mutants with longer C-terminal truncations showed stronger signals in the nucleoplasm: Loc1∆C64 was present only in the nucleolus, loc1∆C74 was localized predominantly in the nucleolus and slightly in the nucleoplasm, and loc1∆C84 was distributed only in the nucleus. Loc1∆N50 was localized in the nucleus, but loc1∆N100 was localized in the cytoplasm. Therefore, NLS1 seems to be the major sequence driving Loc1 into the nucleus.

Discussion
Rpl43 is a small ribosomal protein (about 10 KDa) with only 92 amino acids; it contains a conserved C4-type Zn finger domain made up of four beta-sheets at the center and two helices connected at the Nand C-termini. We found that Rpl43 contains Zn at a 1:1 ratio. Mutations of four conserved cysteines to serine can abolish the chelation of Zn (unpublished data). Although the rpl43(4C) mutant could still be incorporated into pre-60S, it did not complement growth. The half-life of Rpl43 was very short and decreased by 50% within 10 min after translation arrest. This instability was even more severe in puf6∆ and loc1∆ mutants (Figure 2).

Discussion
Rpl43 is a small ribosomal protein (about 10 KDa) with only 92 amino acids; it contains a conserved C4-type Zn finger domain made up of four beta-sheets at the center and two helices connected at the N-and C-termini. We found that Rpl43 contains Zn at a 1:1 ratio. Mutations of four conserved cysteines to serine can abolish the chelation of Zn (unpublished data). Although the rpl43(4C) mutant could still be incorporated into pre-60S, it did not complement growth. The halflife of Rpl43 was very short and decreased by 50% within 10 min after translation arrest. This instability was even more severe in puf6Δ and loc1Δ mutants (Figure 2).
RNA-binding proteins have been shown to be important in ribosome biogenesis and development [54][55][56][57]. Puf6 and Loc1 have been shown to regulate the cellular distribution of ASH1 mRNA. The loss of these two proteins decreases the levels of 60S subunits. However, the molecular functions of Puf6 and Loc1 in 60S biogenesis remain unclear. We have shown that these proteins can form a hetero-trimer with Rpl43 and are critical to maintaining the protein level of Rpl43 and its assembly to pre-60S subunits [38]. Here we provide further evidence that Puf6 and Loc1 are important to maintaining the protein stability of Rpl43 ( Figure 2). The N-terminal 21 amino acids of Rpl43 are important for Kap, Puf6, and Loc1 interactions (see the rpl43∆N mutant in Figures 3  and 4). This region is less structured ( Figure 3B) and may be more easily targeted for degradation than other regions. We also demonstrated that Puf6 and Loc1 can bind nascent Rpl43 co-translationally in a cooperative manner (Figure 3). Therefore, once the N-terminus of Rpl43 emerges from the PET site, Puf6 and Loc1 may majorly associate with this region and protect its stability.
The present study supports the supposition that Puf6 and Loc1 are associated with nascent Rpl43 for transport but do not depend on Rpl43 for import [38]. While Puf6 lost its PUF domain to interact with Loc1, Rpl43, or 60S, the ∆PUF mutant could still be localized in the nucleus ( Figure 5). A similar finding was observed in Loc1 mutants. Truncations of the N-or C-terminus of Loc1 decreased its interactions with Puf6 or Rpl43 but were still retained in the nucleus ( Figure 6). This situation is likely because Puf6 and Loc1 have additional functions with ASH1 mRNA in the nucleus and could therefore be transported with Rpl43 or as individual proteins.
The loss of Rpl43 did not influence the loading of Puf6 or Loc1 but affected their release [38]. Similarly, although deletion of puf6 affected the interaction of Loc1 with nascent Rpl43 (Figure 3), Loc1 could still be loaded onto 60S [38]. However, the protein could not be released properly [38]. The positions of Puf6 and Loc1 on pre-60S have not been solved, but our data suggest that they are bound in the vicinity of Rpl43. This supposition supports the hypothesis that trimer formation is important for the accuracy of this local domain. Only when Rpl43, Puf6, and Loc1 are loaded as a ternary complex onto pre-60S can the process to the next step begin properly.