Deletion of L4 domains reveals insights into the importance of ribosomal protein extensions in eukaryotic ribosome assembly

Abstract

Numerous ribosomal proteins have a striking bipartite architecture: a globular body positioned on the ribosomal exterior and an internal loop buried deep into the rRNA core. In eukaryotes, a significant number of conserved r-proteins have evolved extra amino- or carboxy-terminal tail sequences, which thread across the solvent-exposed surface. The biological importance of these extended domains remains to be established. In this study, we have investigated the universally conserved internal loop and the eukaryote-specific extensions of yeast L4. We show that in contrast to findings with bacterial L4, deleting the internal loop of yeast L4 causes severely impaired growth and reduced levels of large ribosomal subunits. We further report that while depleting the entire L4 protein blocks early assembly steps in yeast, deletion of only its extended internal loop affects later steps in assembly, revealing a second role for L4 during ribosome biogenesis. Surprisingly, deletion of the entire eukaryote-specific carboxy-terminal tail of L4 has no effect on viability, production of 60S subunits, or translation. These unexpected observations provide impetus to further investigate the functions of ribosomal protein extensions, especially eukaryote-specific examples, in ribosome assembly and function.

INTRODUCTION

Elucidation of high-resolution ribosome structures revealed that the ribosomal RNA (rRNA) is the heart of the peptidyl transferase activity of the ribosome, and ribosomal proteins (r-proteins) are organized largely on the subunit exterior (Ban et al. 2000; Nissen et al. 2000; Wimberly et al. 2000; Yusupov et al. 2001; Schuwirth et al. 2005). Many r-proteins have a striking bipartite topology; they possess a globular domain that contacts rRNA on the subunit surface, as well as long extended domains that penetrate deep into the rRNA core. Therefore, it has been hypothesized that r-proteins participate in the dynamic folding and stabilization of rRNA during ribosome assembly and function (Klein et al. 2004). This has been supported by thermodynamic and kinetic studies on in vitro reconstitution of bacterial ribosomes (Recht and Williamson 2004; Holmes and Culver 2005; Ramaswamy and Woodson 2009a,b; Woolstenhulme and Hill 2009; Calidas and Culver 2010; Mayerle and Woodson 2013). In eukaryotes, conserved r-proteins have also evolved extra amino acid sequences that are largely found in either carboxy- or amino-terminal domains (Jenner et al. 2012; Yusupova and Yusupov 2014). Recent atomic structures of eukaryotic ribosomes showed that in contrast to the conserved internal r-protein extensions that penetrate the core, eukaryote-specific tails mainly thread across the surface of the ribosome (Ben-Shem et al. 2011). These internal loops and external tails of r-proteins are intrinsically disordered; hence, the current model is that r-protein extensions chaperone co-folding of rRNA and/or stabilize tertiary interactions during assembly (Wilson and Nierhaus 2005; Timsit et al. 2009; Peng et al. 2014). However, the overarching function of r-protein extensions in the eukaryotic ribosome, especially regions not present in bacteria, remains to be established.

To date, a limited number of r-protein extensions have been shown to be important in bacterial ribosome assembly: those present in S4 and S12 in the 30S subunit (Mayerle and Woodson 2013; Calidas et al. 2014), and the extension of L20 in the 50S subunit (Guillier et al. 2005). Similarly, only a handful of mutant alleles have been identified demonstrating that the conserved extended domains of eukaryotic r-proteins are important for ribosomal subunit formation (Jakovljevic et al. 2004; Bussiere et al. 2012; Garcia-Gomez et al. 2014). Although flexible extensions are prevalent in r-proteins, there is a lack of information about their biological significance, which has largely been inferred computationally (Peng et al. 2014). Here, we investigated two prominent extended domains of L4 in the yeast, Saccharomyces cerevisiae. Yeast L4 possesses two of the most dramatic extensions in the ribosome, which include an internal loop (∼70 amino acids) and a carboxy-terminal tail (∼100 amino acids) (Fig. 1A; Ben-Shem et al. 2011), representing both conserved and eukaryote-specific examples, respectively. L4 is thus a good model to begin to understand the functional role of these unstructured domains.

FIGURE 1.

Effects of removing the long intrinsically disordered extensions of L4 on cell viability. (A) Cartoon representation of yeast L4 showing the two disordered extensions investigated in this study: the conserved internal loop and the eukaryote-specific carboxy-terminal tail. (B) Growth of yeast cells solely expressing wild-type or mutant derivatives of L4 at permissive (30°C) and nonpermissive temperatures (37°C and 16°C).

L4 belongs to a conserved family of r-proteins with mixed α-helices and β-strands (Klein et al. 2004) and is essential for early steps of both bacterial and eukaryotic ribosome assembly (Rohl and Nierhaus 1982; Poll et al. 2009; Gamalinda et al. 2014). In mature ribosome structures, its globular body domain straddles the exterior of domains I and II, while its extended internal loop region penetrates deep into the same domains, also reaching parts of the peptidyl transferase center (PTC) in domain V (Schuwirth et al. 2005; Ben-Shem et al. 2011). The internal loop of L4 provides the largest surface area of any r-protein to the polypeptide exit tunnel (PET) wall (Ben-Shem et al. 2011). In bacteria, this extended loop is dispensable for formation of 50S subunits (Zengel et al. 2003), but it has been shown to be important for translation (Chittum and Champney 1994; Zaman et al. 2007), as it is involved in a dynamic gating mechanism during cotranslational folding of nascent peptides (Gabashvili et al. 2001; Fulle and Gohlke 2009). In addition to the conserved body and internal loop, yeast L4 contains a eukaryote-specific carboxy-terminal tail that threads more than half of the width of the solvent-exposed interface of the 60S subunit. This eukaryotic tail interacts with several essential eukaryote-specific r-proteins (e.g., L18e, L20e, and L21e; formerly called L18, L20, and L21, respectively) and rRNA sequences, referred to as expansion segments (e.g., ES7 and ES15) (Ben-Shem et al. 2011). Hence, the carboxy-terminal tail of L4 has been hypothesized to be important for the assembly of eukaryote-specific ribonucleoprotein neighborhoods in 60S subunits (Melnikov et al. 2012).

Here, we show that in contrast to its dispensability in bacteria (Zengel et al. 2003), the internal loop of L4 is important for efficient formation of large subunits; deleting this loop affects very late stages in pre-60S ribosome assembly. This late assembly phenotype is distinct from the very early assembly defect observed when the entire L4 protein is depleted (Poll et al. 2009; Gamalinda et al. 2014). Thus, it appears likely that the globular domain of L4 initially binds to stabilize early intermediates of assembling large subunits while its internal loop region facilitates final steps of 60S subunit assembly, most likely to structure functionally important regions such as the PET and PTC. Surprisingly, truncating the entire eukaryote-specific carboxy-terminal tail of L4 has no effect on ribosome assembly and function. In view of these differences in phenotypic consequences of deleting conserved and eukaryote-specific r-protein extensions, it is imperative to examine more examples in order to reveal the functions of extended regions in ribosome assembly and function in vivo.

RESULTS AND DISCUSSION

To begin to understand the general biological function of r-protein extensions, we generated mutant versions of yeast L4 in which either the conserved internal loop or the eukaryote-specific carboxy-terminal tail is deleted (hereafter referred to as L4-LΔ or L4-CΔ, respectively) (Fig. 1A). Plasmids containing these constructs were introduced into a yeast strain engineered to conditionally express wild-type L4 by replacing its endogenous promoter with the galactose-inducible, glucose-repressible GAL1 promoter. Thus, in glucose-containing medium, plasmid-borne derivatives of RPL4 are the sole sources of L4 protein. We confirmed that these two mutated versions of L4 stably assemble into preribosomes (see below). As expected, the plasmid expressing wild-type L4 supported growth in glucose-containing medium while the empty vector did not (Fig. 1B). Surprisingly, complete truncation of the carboxy-terminal tail only found in eukaryotes supported cell viability at all temperatures tested. In contrast, deleting the entire internal extended loop led to a severely impaired growth at 30°C and complete lethality at 37°C and 16°C (Fig. 1B).

Macrolide antibiotics such as erythromycin disrupt translation elongation by binding to the constriction of the ribosome polypeptide exit tunnel, which is formed in part by the internal extension of L4 (Kannan and Mankin 2011). The carboxy-terminal tail of L4 extends toward the A-site finger (25S rRNA helix 38) on the solvent interface via eukaryotic ribosomal components (Ben-Shem et al. 2011), also suggesting its importance in eukaryotic translation. To initially test the importance of either the internal loop or carboxy-terminal tail of L4 in yeast ribosome function, we tested for hypersensitivity to drugs that inhibit various aspects of translation in vivo (Yonath 2005; Kannan and Mankin 2011; Blaha et al. 2012). Interestingly, the growth of both mutants in the presence of translation inhibitors was the same as when the drugs are absent (Fig. 2). Thus, deletion of either the internal loop or the carboxy-terminal tail of L4 does not appear to affect translation.

FIGURE 2.

Removing either the conserved loop or the eukaryotic tail of L4 does not lead to hypersensitivity to translation inhibitors. Defects in ribosome function were assayed by growing yeast cells in media containing various drugs that impair translation.

We next wanted to test if the long extended domains of L4 are specifically important for ribosome assembly. To do so, we first assayed for the levels of individual ribosomal subunits, 80S complexes, and translating ribosomes by sucrose gradient fractionation. Consistent with the role of L4 in large subunit assembly, the complete absence of L4 led to the reduction in the amounts of free 60S subunits relative to 40S subunits (Fig. 3A). In addition, halfmer polysomes were detected when L4 was depleted, indicative of a deficit in 60S subunits. Previously, the extended loop of L4 in Escherichia coli had been shown to be dispensable for 50S subunit assembly (Zengel et al. 2003). On the contrary, we observed that expressing the L4 loop deletion mutant (L4-LΔ) did affect the production of 60S subunits, albeit to a lesser extent than depleting the entire protein (Fig. 3A). The level of free 60S subunits in L4-LΔ was slightly lower than 40S subunits, nevertheless was still accompanied by formation of halfmer polysomes, indicating a defect in 60S subunit production. In support of this, we saw a decrease in the total amount of mature 25S rRNA relative to 18S rRNA (data not shown). On the other hand, L4-CΔ exhibited levels of 40S and 60S subunits, 80S ribosomes, and polyribosomes that are comparable to wild-type (Fig. 3A), consistent with its viability (Fig. 1B). Taken together, we conclude that unlike its bacterial complement, the extended internal loop of L4 is important for assembly of large ribosomal subunits while the eukaryote-specific tail of L4 is dispensable for 60S subunit production.

FIGURE 3.

Importance of the long extended domains of L4 in ribosome biogenesis. (A) Ribosomal subunits, 80S monosomes, and polyribosomes were fractionated across a 7%–47% sucrose gradient. Halfmers resulting from a deficit in 60S subunits relative to 40S ribosomal subunits are indicated by asterisks. (B) Primer extension analysis was used to detect the 5′ ends of large subunit pre-rRNA intermediates. Pre-rRNA species containing these ends are labeled. (C) Northern hybridization was used to detect downstream pre-rRNA intermediates and mature rRNAs.

In yeast, mature 18S, 5.8S, and 25S rRNAs are produced from a 35S pre-rRNA transcript via a series of nucleolytic processing events (Supplemental Fig. S1; Woolford and Baserga 2013). To examine which specific step of pre-60S assembly is affected leading to inefficient formation of 60S subunits in these two L4 mutants, we assayed the levels of pre-rRNA intermediates by primer extension and Northern blotting. Consistent with previous observations (Poll et al. 2009; Gamalinda et al. 2014), depleting L4 resulted in the accumulation of 27SA₂ and 27SA₃ pre-rRNAs, while levels of downstream intermediates were significantly reduced (Fig. 3B), indicating the role of L4 in early stages of pre-60S assembly. In agreement with the absence of observable growth defects and no reduction in 60S subunits (Figs. 1B, 3A), L4-CΔ exhibited wild-type levels of 27S pre-rRNA, as well as downstream 7S and 6S pre-rRNA processing intermediates (Fig. 3B,C). The carboxy-terminal extension of L4 is among the longest tails in the eukaryotic ribosome, stretching >100 Å across the surface of the 60S subunit (Ben-Shem et al. 2011). The finding that this eukaryote-specific tail is dispensable for efficient processing of pre-rRNAs is peculiar, considering that it appears to have coevolved with ES7B and ES15 with which it interacts (Ben-Shem et al. 2011). Eukaryote-specific r-proteins that bind to ES7B and ES15 (L6e, L20e, L32e, and L33e; L6, L20, L32, and L33 in the former nomenclature) are essential and required for early steps of pre-rRNA processing (Poll et al. 2009; Gamalinda et al. 2014).

Interestingly, L4-LΔ showed a pre-rRNA processing phenotype distinct from the complete depletion of L4. When the internal loop was removed, levels of pre-rRNAs containing B_L and B_S 5′ ends increased relative to levels of A₂- and A₃-containing ends (Fig. 3B). Northern blotting confirmed the accumulation of downstream 7S and 6S pre-rRNA intermediates (Fig. 3C). This indicates that the internal loop of L4 is important for 60S subunit formation by mediating late steps of pre-rRNA processing. The amino acid sequence of the L4 extended loop is largely conserved, but eukaryotic L4 has 10 additional residues interspersed throughout the extension. Nevertheless, the length of both the bacterial and eukaryotic L4 loop is similar (∼60 Å), and they occupy identical positions in the large ribosomal subunit. One noticeable structural difference, however, is that the distal end of the loop in yeast is bifurcated in 60S subunits, providing more extensive interactions with rRNA. This might explain why the extended loop of L4 is required for large subunit biogenesis in yeast but not in bacteria. In support of this, deleting only this branched tip of the L4 loop in yeast resulted in the reduced levels of 60S subunits and a late pre-rRNA processing defect, identical to deleting the entire loop (data not shown). We conclude that, in contrast to depleting L4 that blocks early stages in assembly, the internal loop of L4 is required for efficient progression of late 60S assembly steps. These observations therefore reveal that L4 is important for two different steps in the biogenesis of large ribosomal subunits. One possible scenario is that the globular domain of L4 binds first to pre-rRNA cotranscriptionally, primarily in domain II of 25S rRNA, stabilizing early assembly intermediates during which processing of 27SA pre-rRNAs occur. Interactions of the loop region with 5.8S rRNA, as well as domains I, II, and V, might facilitate the folding and bringing together of these rRNA sequences later in assembly, most likely to form the primary constriction of the PET and a part of the PTC. In the process, the unstructured loop of L4 possibly transitions to a folded state similar to what is found in the structures of mature 60S subunits (Ben-Shem et al. 2011). Our results support prevailing models that depict coupled binding and folding of r-protein extensions with rRNA during ribosome biogenesis (Klein et al. 2004; Timsit et al. 2009; Peng et al. 2014). Moreover, consistent with the late pre-rRNA processing defect observed when the internal loop of L4 is removed, recent studies demonstrate that the functionally important centers of ribosomal subunits are assembled during later stages in assembly (Bussiere et al. 2012; Strunk et al. 2012; Karbstein 2013; Gamalinda et al. 2014; Garcia-Gomez et al. 2014).

In yeast, ∼300 trans-acting assembly factors facilitate the step-wise remodeling of preribosomes to form mature ribosomal subunits (Maxwell and Fournier 1995; Karbstein 2011; Kressler et al. 2012; Woolford and Baserga 2013). To further assess the role of the extended regions of L4 in large ribosomal subunit assembly, we purified preribosomes and assayed their composition by SDS-PAGE followed by silver staining or Western blotting. Similar to depleting other early-acting r-proteins in the large subunit (Jakovljevic et al. 2012; Gamalinda et al. 2014), SDS-PAGE profiles of these ribosome assembly intermediates that are present when L4 is depleted showed accumulation of assembly factors present in early pre-60S ribosomes (high-molecular weight bands), accompanied by the reduction of many r-proteins (low-molecular weight bands) (Fig. 4A). We assayed for more specific changes by Western blotting. Upon depletion of L4, levels of tested early-assembling factors necessary for processing of 27SA (Nop7, Cic1, and Has1) and 27SB (Nog1, Tif6, and Rlp24) pre-rRNAs, as well as late-assembling factors (Nsa2 and Nog2) were reduced in preribosomes (Fig. 4B). Western blotting confirmed that many large subunit r-proteins, notably L17 and L25 that surround the exit tunnel, were also diminished in assembling ribosomes when L4 is depleted. These observations are consistent with the observed early block in pre-rRNA processing (Fig. 3B,C; Poll et al. 2009; Gamalinda et al. 2014) and in accordance with previous observations upon depleting other r-proteins functioning early during 60S subunit assembly (Jakovljevic et al. 2012; Gamalinda et al. 2014).

FIGURE 4.

Deletion of either the conserved internal loop or eukaryotic tail of L4 does not significantly affect the formation of preribosomes. (A) Pre-60S components isolated by affinity purification were separated by SDS-PAGE and detected by silver staining. In contrast to the complete depletion of L4, the overall preribosome profile is not affected by the absence of either extension of L4. (B) Specific proteins in preribosomes were assayed by Western blotting. Depletion of L4 leads to the reduction of many assembly factors and r-proteins in preribosomes. These factors and r-proteins in preribosomes remain unaffected when shorter derivatives of L4 are expressed. Mutant versions of L4 are indicated by an asterisk.

On the contrary, no significant changes relative to wild-type were detected by SDS-PAGE followed by silver staining or Western blotting when deleted variants of either the internal loop or carboxy-terminal tail of L4 is present in preribosomes (Fig. 4A,B). This is similar to our previous observations in r-protein depletion mutants that block late steps in 60S assembly (Ohmayer et al. 2013; Gamalinda et al. 2014). Importantly, we show that these phenotypes are caused specifically by the removal of either the internal loop or carboxy-terminal tail of L4 because these mutant proteins (L4-CΔ and L4-LΔ) can stably associate with preribosomes (Fig. 4B). In summary, key assembly factors required for early and middle steps are capable of stable association with preribosomes containing a version of L4 lacking the internal extended loop. The presence of these factors enables the proper progression of upstream pre-rRNA maturation events, consistent with the late pre-rRNA processing defect observed in the L4-LΔ mutant. Finally, because of the viability and apparent wild-type SDS-PAGE profile of yeast strains expressing L4-CΔ, we infer that the carboxy-terminal tail of L4 is not required to load essential r-proteins L18e, L20e, L21e, and L30 into assembling ribosomes despite their extensive contacts with the tail of L4.

CONCLUDING REMARKS

Almost all of the ∼80 r-proteins in eukaryotic ribosomes have extended loops or termini that are intrinsically disordered (Peng et al. 2014). Missense mutations that alter amino acid sequences in r-protein extensions are associated with developmental and hematopoietic disorders, as well as cancer progression (Gazda et al. 2008; De Keersmaecker et al. 2012; Watkins-Chow et al. 2013). However, the functional significance of intrinsically disordered r-protein domains in the ribosome in vivo is yet unclear. In this study, we have investigated the conserved internal loop and the eukaryote-specific carboxy-terminal tail of L4 in yeast. In contrast to bacteria, we found that the internal extended loop of yeast L4 is important for large ribosomal subunit assembly. Moreover, we showed that distinct from complete depletion of L4 that affects early assembly steps, the absence of just the internal loop specifically affects only late stages of pre-60S assembly. This inner loop is unstructured in the unbound form of L4 (Worbs et al. 2000), but switches to an ordered form upon binding rRNA. Hence, while binding of the globular domain of L4 likely facilitates early rRNA folding events, the internal loop may be specifically important in bringing together and stabilizing long-range interactions between domain I, II, and V, which ultimately organizes the functionally active sites of the 60S subunit such as the PET and PTC (Ben-Shem et al. 2011). Investigating the extended regions of other r-proteins therefore promises to uncover more specific functions of these flexible domains in ribosome assembly.

Universally conserved r-proteins such as L4 often have extra amino acid residues at their termini in eukaryotes that intimately interact with rRNA and r-proteins (Ben-Shem et al. 2011; Peng et al. 2014). Thus, the conventional paradigm is that the eukaryotic tail of L4 acts as an RNA chaperone and/or mediates the assembly of r-proteins during the formation of 60S subunits (Melnikov et al. 2012). Paradoxically, we found that this carboxy-terminal tail of L4 that is conserved throughout eukaryotes is dispensable for ribosome assembly and function. Because of the ubiquity and strategic locations of eukaryote-specific tails in the ribosome, our observation with L4 might be an exception rather than the rule. We therefore propose that more examples of eukaryote-specific extensions in r-proteins have to be investigated to unravel their evolutionary significance.

MATERIALS AND METHODS

Plasmid construction

A plasmid containing intact RPL4A obtained from the Yeast genomic Tiling Collection (Open Biosystems) was used as a template for a two-step PCR reaction to incorporate the restriction sites BamHI ∼1000 nucleotides upstream of and SacII ∼500 nucleotides downstream from the r-protein gene (oligonucleotide sequences available upon request). Amplification products were inserted into pRS315 (LEU2) using standard cloning procedures in E. coli DH5α. Incorporation of correct wild-type RPL4A gene sequences was validated by sequencing (Genewiz). Deletion and truncation mutations in RPL4A were generated by the QuikChange Site-directed Mutagenesis Kit (Stratagene). Mutant constructs were verified by sequencing.

Construction and growth of yeast strains

Empty pRS315 vector, or plasmids bearing derivatives of the RPL4A gene were individually introduced into a yeast strain conditional for the expression of RPL4 (GAL-RPL4; MATa ura3-52 trp1-1 lys2-801 his3-Δ200 leu2-Δ1 rpl4b::KANMX6 rpl4a::GAL-HA-RPL4A TRP1) (Gamalinda et al. 2014). For growth on solid medium, conditional strains expressing wild-type or mutant derivatives of RPL4 were grown in galactose-containing liquid medium and diluted to an OD610 of 0.8 (∼4 × 10⁷ cells/mL). Ten microliters of serially diluted yeast cultures (10- to 100,000-fold) were spotted onto glucose-containing medium and then incubated at 16°C, 30°C, or 37°C. For growth in liquid medium, strains were either grown in galactose-containing medium or shifted from galactose- to glucose-containing medium to deplete endogenous wild-type L4 (OD610 of 0.6–0.8).

In vivo translation assay

Hypersensitivity to different translation inhibitors was tested by growing strains expressing wild-type or mutant alleles of RPL4 in galactose-containing medium and diluted to an OD610 of 0.8. Ten- to 100,000-fold serial dilutions (10 μL each) were spotted on glucose-containing media with the following antibiotics: 3 mg/mL erythromycin, 20 μg/mL hygromycin, 5 mg/mL neomycin, and 2 mg/mL paramomycin. Cells were incubated at 30°C for 3 d.

Yeast polysome profile analysis

Sucrose gradient centrifugation was used to fractionate free ribosomal subunits, 80S ribosomes, and polyribosomes, as previously described (Gamalinda et al. 2013). Before harvesting, logarithmically growing yeast cultures in 100 mL glucose-containing liquid medium were grown to OD 610 of 0.7 and then treated with 5 mg cycloheximide. Cells were lysed and 40 A₂₅₄ units of whole-cell lysates were centrifuged through 7%–47% (w/v) sucrose gradients at 27,000 rpm (Sorvall AH-629 swinging bucket rotor) for 4 h. A₂₅₄ of fractions was monitored using a density-gradient fraction collector (Teledyne ISCO Foxy R1).

Analysis of pre-rRNA intermediates and mature rRNAs

Primer extension and Northern blotting was carried out to assay the steady-state levels of pre-rRNAs and mature rRNAs, as described in Gamalinda et al. (2014).

Affinity purification of preribosomes

Pre-60S ribosomes associated with TAP-tagged Rpf2 were isolated from IgG-coated magnetic beads using a modified single-step affinity purification technique (Sahasranaman et al. 2011). Rpf2 is a suitable bait to purify preribosomes because it is present in all pre-60S complexes and its association with preribosomes is unaffected when various steps in 60S assembly are blocked (Gamalinda et al. 2014).

Analysis of preribosome composition

Components of purified preribosomes were separated by SDS-PAGE (4%–20% Tris–glycine or 4%–12% Bis–Tris precast gels, Invitrogen). Polypeptide bands were detected by standard silver-staining or Western blotting, using a protocol adapted from Sahasranaman et al. (2011).

SUPPLEMENTAL MATERIAL

Supplemental material is available for this article.