Yeast Proteins Related to the p40/Laminin Receptor Precursor Are Required for 20S Ribosomal RNA Processing and the Maturation of 40S Ribosomal Subunits¹

The M_r 67,000 high-affinity 67-LR³ was the first receptor for laminin identified and was used as a prototype of a cellular adhesion molecule involved in the metastatic cascade (1). Increased expression of 67-LR in invasive cells was thought to enhance their interaction with basement membranes and to promote their dissemination in the circulatory system to establish metastatic foci. The role of 67-LR in the adhesion of cells to basement membranes is controversial, however, inasmuch as it has been difficult to purify the amounts necessary for detailed characterization and structural analysis (2, 3). A cDNA reported to encode 67-LR has been used extensively in studies correlating the expression of 67-LR with carcinogenesis (4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18). However, this cDNA does not encode a protein with properties expected for a cell surface receptor because the deduced protein lacks a recognizable signal peptide and transmembrane spanning domain (2). It is now known that the cDNA from several species encodes an abundant intracellular protein of M_r ∼40,000, termed p40, that is a component of the small ribosomal subunit (19, 20, 21, 22). It has been suggested that p40 is a multifunctional protein that, in addition to being a ribosomal protein, also serves as a precursor to 67-LR (3, 23). In this context, p40 has also been referred to as the M_r 37,000 LRP, or 37-LRP. However, the pathway through which p40/37-LRP is processed and directed to the cell surface is unclear. Given the uncertain relationship between p40/37-LRP and 67-LR, it seems reasonable to seek alternative explanations for the linkage between the overexpression of the p40/37-LRP gene and carcinogenesis. One such possibility relates to the role of p40/37-LRP as a component of the protein synthetic machinery.

Many studies have implicated components of the protein synthetic machinery in carcinogenesis (24, 25). Changing the level or activity of certain initiation factors can lead to malignant transformation (26, 27, 28). These changes presumably influence the expression of key growth regulatory genes that give rise to the transformed phenotype. Several growth regulatory genes, including cdk4, p27^Kip1, and Cln3 are subject to translational control (29, 30, 31). In the yeast Saccharomyces cerevisiae, cell growth and division are coupled by the translational regulation of cyclin (CLN3) mRNA (31). Translation of CLN3 mRNA is critically sensitive to the level of initiation competent ribosomes. Changes in the amount of both the initiation factors and ribosomal subunits influence initiation complex formation and, thus, the translation of CLN3 mRNA (31). The role of ribosomal content in cell cycle progression is interesting in light of observations that several ribosomal proteins are overexpressed in tumors relative to the surrounding normal tissues (32, 33, 34, 35, 36). This overexpression is not the result of a general increase in protein synthetic capacity in tumor cells inasmuch as the pattern of ribosomal proteins overexpressed is tissue-specific (35, 36). Some of the ribosomal protein genes, such as the gene encoding p40/37-LRP, are overexpressed in a wide range of tumors, whereas others are restricted in distribution to specific tumor types. Despite the body of evidence linking the overexpression of ribosomal proteins to carcinogenesis, their specific roles in this process remain obscure.

To gain more insight into the function of the p40/37-LRP family of proteins as components of the protein synthetic machinery, we have turned to the yeast homologues, Yst1 and Yst2. The Yst proteins are closely related, with greater than 60% sequence identity, to the human p40/37-LRP protein (37). YST1 and YST2 are a duplicated pair of genes that encode essential protein components of the 40S ribosomal subunit. Recently, the names YST1 and YST2 have been changed to RPS0A and RPS0B, respectively, to conform to a new system of nomenclature for yeast cytoplasmic ribosomal proteins (38). Studies have shown that the disruption of either RPS0 gene caused a reduction in growth rate, and the disruption of both genes was lethal (37). Polysomes from cells disrupted in one or the other RPS0 gene were smaller than wild-type polysomes. This reduction was accompanied by a decrease in the level of 40S ribosomal subunits and an increase in the amount of free 60S ribosomal subunits. Thus, the decreased growth rate in cells with suboptimal amounts of Rps0 protein seemed to result from a reduced amount of 40S subunits. Here, we show that this decrease is due to a defect in the maturation of 40S subunit precursors. The Rps0 proteins are required for a late step in the maturation of 40S subunits that involves processing of the 20S rRNA precursor to 18S rRNA. This processing step, which also occurs in human cells, plays a key role in defining the level of active 40S subunits in yeast cells. If the function of these yeast proteins is representative of other members of the p40/37-LRP family of proteins, then p40/37-LRP proteins could also play a critical role in defining the protein synthetic capacity of human cells. Overexpression of p40/37-LRP proteins in tumors relative to surrounding tissues may, therefore, function to assure an adequate supply of active 40S subunits to meet the demands for protein synthesis during tumor cell growth and proliferation.

The yeast strains used in this study were:

• (a) W8–3d (MATa/MATα, ρ⁺, ade2–1/ade2–1, can1–100/can1–100, his3–11,15/his3–11,15, ura3–1/ura3–1, leu2–3,112/leu2–3,112, trp1–1/trp1–1, RPS0A/rps0A::URA3, RPS0B/rps0B::HIS3);

• (b) CLF(pRS314-GAL1/RPS0A)(ρ⁺, ade2–1,can1–100, his3–11,15, ura3–1, leu2–3,112, trp1–1, rps0A::URA3, rps0B::HIS3, pRS314-GAL1/RPS0A);

• (c) CLF(pRS315-RPS0A) (ρ⁺, ade2–1,can1–100, his3–11,15, ura3–1, leu2–3,112, trp1–1, rps0A::URA3, rps0B::HIS3, pRS315-RPS0A); and

• (d) CLF(pRS315-RPS0A cDNA) (ρ⁺, ade2–1,can1–100, his3–11,15, ura3–1, leu2–3,112, trp1–1, rps0A::URA3, rps0B::HIS3, pRS315-RPS0A cDNA).

Media used in cultivating yeast were: (a) YPD (1% w/v yeast extract, 2% w/v peptone, and 2% w/v glucose); (b) YPGal/Suc (1% w/v yeast extract, 2% w/v peptone, 2% w/v galactose, and 0.1% w/v sucrose); and (c) synthetic (0.67% w/v yeast nitrogen base without amino acids, 2% w/v glucose or 2% w/v galactose, and 0.1% w/v sucrose). Where appropriate, nutrients were added to synthetic media in amounts specified by Sherman (39). Diploids were sporulated on solid sporulation media (1% w/v potassium acetate, 0.1% w/v yeast extract, 0.05% glucose, and 2% w/v agar and—where appropriate—adenine, histidine, uracil, leucine, and tryptophan in 25% of the amounts used in synthetic media). The Escherichia coli strain used in this study was XL1-Blue (Stratagene).

The GAL1/GAL10 promoter region was excised as a 685-bp EcoRI-BamHI fragment from plasmid pBM272 and inserted into the unique EcoRI and BamHI sites of pRS316. The GAL1/GAL10 promoter region was subsequently moved as an SpeI/SalI fragment from pRS316 to pRS314. Plasmid pRS314 containing the GAL1/GAL10 promoter region was digested with BamHI, which is downstream of the GAL1 promoter, and SacII, which is within the polycloning region of pRS314. A DNA fragment containing the RPS0A open-reading frame, lacking promoter sequences and encoding sequences for a c-myc epitope tag at the COOH-terminal end of the protein, was created using the PCR. The RPS0A gene was amplified from plasmid pRS315-RPS0A(c-myc) using oligonucleotides complementary to 5′ and 3′ flanking regions of the RPS0A(c-myc) open-reading frame (37). In addition to sequences derived from the RPS0A gene, the 5′ and 3′ oligonucleotides contained recognition sequences for BamHI and SacII, respectively, for use in cloning the RPS0A open-reading frame downstream of the GAL1 promoter. Each oligonucleotide also had five additional bases at their 5′ ends to facilitate digestion of the amplified product with the appropriate restriction enzyme. The oligonucleotides used for the PCR amplification of RPS0A were: (a) 5′-TACGTGGATCCGTAGAGTGAGGTATAGCTTAGA-3′ (5′ oligonucleotide) and (b) 5′-TACGACCGCGGTCTACGCAATCACTACGTACCA-3′ (3′ oligonucleotide), with sequences complementary to RPS0A italicized. The amplified product was digested with BamHI and SacII and cloned into pRS314-GAL1/GAL10 downstream of the GAL1 promoter. The resulting plasmid was transformed into W8–3d cells, which are diploid and heterozygous for RPS0A and RPS0B disruptions (37). Transformants were sporulated, and doubly disrupted cells containing the plasmid were selected by prototropy for histidine, tryptophan, and uracil using synthetic media with 2% galactose/0.1% sucrose as carbon sources. The only source of Rps0 protein in these cells is from the RPS0A(c-myc) gene under the control of the GAL1 promoter.

CLF(pRS314-GAL1/RPS0A) and CLF(pRS315-RPS0A) cells were grown to late log or stationary phase in synthetic media containing galactose and sucrose as carbon sources. Synthetic media used for CLF(pRS314-GAL1/RPS0A) cells lacked tryptophan and methionine, whereas media used for culturing CLF(pRS315-RPS0A) cells lacked leucine and methionine. Cells from both cultures were concentrated by centrifugation, washed once with synthetic media containing glucose as the sole carbon source, and suspended in glucose-containing media. Cells used for an extended time course and Northern analysis after shifting cells to glucose were typically suspended at an absorbance of approximately 0.1 at 640 nm. Cells were maintained at an A_{640 nm} between 0.1 and 0.2 by diluting cells into fresh media to prevent nutrient deprivation from becoming a confounding factor in the analysis of the effects of the depletion of Rps0 protein on cell growth and rRNA metabolism. These dilution factors were factored into the growth curves.

Polysomes were prepared from yeast cell extracts and fractionated on 7–47% sucrose gradients as described by Baim et al. (40). Centrifugation was for 9 h at 20,000 rpm in an SW28.1 rotor. Gradients were fractionated, and the absorbance at 254 nm was monitored using an ISCO model 185 density gradient fractionator and a UA-5 absorbance detector. RNA was extracted from the individual fractions as outlined by Baim et al. (40).

RNA was derived from polysome fractions as described above or from whole cell extracts according to Schmitt et al. (41). RNA was fractionated on 1.5% formaldehyde-agarose gels and transferred to Zeta-Probe (Bio-Rad) membrane. The membrane was hybridized with oligonucleotides complementary to RPS0A/RPS0B mRNA, 5′-TTGAACGT-TTCTAGCACCTAA-3′; 18S/20S rRNA, 5′-TGATCCTTCCGCAGGTTCACCTACGGAAAC-3′; and 20S rRNA, 5′-GAAATCTCTCACCGTTT-GGAATAGC-3′. Oligonucleotide probes were radiolabeled at their 5′ ends with [γ³²P]ATP (New England Nuclear) using T4 polynucleotide kinase (New England Biolabs).

Pulse-chase labeling of pre-rRNA was carried out as described by Kressler et al. (42). CLF(pRS314-GAL1/RPS0A) or CLF(pRS315-RPS0A) cells were grown in synthetic media lacking methionine and either tryptophan or leucine, respectively. Cells were initially grown in synthetic media containing galactose and sucrose as carbon sources and then shifted to glucose. Cells were suspended in glucose-containing media at an A_{640 nm} of approximately 0.2–0.5 and grown for one to two generations before being used in pulse-chase experiments. Cells (40 ml) were concentrated to a total of 1 ml in the synthetic medium containing glucose. These cells were pulse-labeled for 2 min with 250 μCi [methyl-³H]methionine. The labeled cells were split into four aliquots of 250 μl each. Each aliquot was diluted into synthetic media containing 1 mg/ml methionine and chased for 0, 2, 5, or 15 min after which time total RNA was isolated. Total RNA was fractionated on 1.5% agarose-formaldehyde gels and transferred to Zeta-Probe membrane. The membrane was baked for 2 h at 80°C, sprayed with EN³HANCE (New England Nuclear), and exposed to X-ray film at -70°C with an intensifying screen.

Yeast DNA was prepared from 10 ml of cells grown to stationary phase. Cells were harvested by centrifugation and suspended in 1 ml of 1 m sorbitol, 100 mm KH₂PO₄ (pH 7.4), 2 mg/ml Zymolyase 20T (ICN Biochemicals) and digested for 2 h at 37°C. Spheroplasts were harvested by centrifugation and suspended in 200 μl of cell resuspension buffer supplied with a Promega Wizard Plus miniprep DNA purification system. Cells were lysed and the DNA purified according to the instructions provided with the Wizard plus minipreps. Yeast DNA was used as a template for PCR amplification using oligonucleotides derived from the RPS0A gene. The 5′ oligonucleotide was 5′-TTGTTGGCTGCTAACACTCA-3′, and the 3′ oligonucleotide was 5-TCACCTTACTTACCACTCGA-3′. These oligonucleotides are derived from sequences upstream and downstream of the RPS0A intron. The PCR reactions were carried out using Taq polymerase from Promega Corporation.

Previous studies (37) demonstrated that cells with disrupted alleles of either RPS0A or RPS0B had decreased amounts of 40S ribosomal subunits relative to wild-type cells. This reduction could be the result of either a defect in synthesis or a reduced stability of 40S subunits that lack Rps0 protein. To distinguish between these possibilities, we created a yeast strain, CLF(pRS314-GAL1/RPS0A), in which the only source of Rps0 protein was from a hybrid gene in which the RPS0A reading frame was placed downstream of the yeast GAL1 promoter. Transcription from the GAL1 promoter is regulated by the carbon source. Cells were maintained in galactose in which the GAL1 promoter was active and then shifted to glucose in which transcription from the GAL1/RPS0A hybrid gene was repressed. Fig. 1 shows that the Rps0A protein was gradually depleted from CLF(pRS314-GAL1/RPS0A) cells after the shift to glucose; by 12 h, no Rps0A protein was evident by immunoblot analysis.

Fig. 2 shows a growth curve of CLF(pRS314-GAL1/RPS0A) and CLF(pRS315-RPS0A) cells after the shift from galactose to glucose-containing media. Both strains have a lag period of between 1 and 2 h before they begin exponential growth. By 4 h, the 2 strains begin to show differences in growth rate. CLF(pRS315-RPSOA) cells, in which RPS0A expression is driven by its own promoter, have a doubling time of approximately 2.4 h, whereas CLF(pRS314-GAL1/RPS0A) cells have a doubling time closer to 5 h. Between the 9 and 12 h time points, the growth rate of CLF(pRS314-GAL1/RPS0A) cells was further reduced to a doubling time greater than 12 h. Microscopic analysis of cells depleted of Rps0 protein at the 25-h time point revealed an approximately 2:1 ratio of large unbudded cells relative to cells that had either a single large bud or multiple buds (data not shown). Less than 2% of the cells had newly emerging buds, which indicated that growth was arrested.

Total RNA was isolated from CLF(pRS314-GAL1/RPS0A) and CLF(pRS315-RPS0A) strains at points throughout the growth curves. The RNAs were fractionated on formaldehyde gels and stained with ethidium bromide (Fig. 3). In both strains, the steady-state levels of 18S and 25S rRNAs was increased at early time points after the shift to glucose, consistent with the known effects of nutritional upshifts on ribosome synthesis (43). In CLF(pRS315-RPSOA) cells, 18S and 25S rRNAs attained a new steady-state level by approximately 6 h after the shift to glucose. Similar results were observed for 25S rRNA in CLF(pRS314-GAL1/RPS0A) cells. In contrast, after an initial rise, the 18S rRNA in CLF(RS314-GAL1/RPS0A) cells began to decrease by 4–6 h after the shift to glucose.

To gain a better understanding of the basis for the decrease in 18S rRNA observed in cells depleted of Rps0 protein, the RNAs from Fig. 3 were hybridized with an oligonucleotide probe complementary to 18S rRNA. Fig. 4, top panel, shows that the decrease in 18S rRNA is accompanied by the appearance of a slightly larger RNA species that hybridized with the 18S probe. We reasoned that the larger RNA was probably the 20S precursor of 18S rRNA. This was confirmed using a probe from sequences downstream of the 18S rRNA coding region that recognized the 20S rRNA precursor but not mature 18S rRNA (Fig. 4, bottom panel). The 20S rRNA precursor rose rapidly in CLF(pRS314-GAL1/RPS0A) cells after the shift to glucose and reached steady-state levels that were a minimum of 3- to 5-fold higher than the levels of the 20S rRNA precursor in CLF(pRS315-RPS0A) cells. The steady-state level of the 20S precursor in CLF(pRS314-GAL1/RPS0A) cells began to decline by the 12-h time point, which coincided with a break in the growth curve in which the doubling time extended beyond 12 h. This decline is probably due to a growth rate-dependent decrease in global rRNA synthesis as indicated by lower levels of the 35S and 32S rRNA precursors at these later time points (Fig. 4, bottom panel).

We used pulse-chase labeling of rRNA precursors to study rRNA processing in cells depleted of Rps0 protein. Eight h after the shift to glucose, CLF(pRS314-GAL1/RPS0A) and CLF(pRS315-RPS0A) cells were pulse-labeled for 2 min with [methyl-³H]-methionine. During this pulse-labeling period, tritiated methyl groups were rapidly transferred to rRNA (44). At various times after the chase in cold methionine, RNA was extracted from the cells, and the rates of processing of rRNA precursors were analyzed. Fig. 5 shows that during the chase period the 27S rRNA precursor was rapidly processed to mature 25S rRNA in both CLF(pRS314-GAL1/RPS0A) and CLF(pRS315-RPS0A) cells. The processing of the 20S rRNA precursor, on the other hand, showed a dramatic difference between these two strains. In CLF(pRS315-RPS0A) cells, the 20S precursor was processed to mature 18S rRNA by 5 min into the chase. In CLF(pRS314-GAL1/RPS0A) cells, on the other hand, no detectable processing of 20S to 18S rRNA was observed even after a 15 min chase. These data confirm that efficient processing of the 20S rRNA precursor is dependent on Rps0A proteins.

Udem and Warner (44) reported that immature 40S subunits containing 20S precursors are inactive in protein synthesis. Here, we addressed whether 40S subunits containing 20S rRNA that accumulate in the absence of Rps0 protein were active in protein synthesis. Fig. 6 shows polysome profiles from CLF(pRS314-GAL1/RPS0A) and CLF(pRS315-RPS0A) cells after a shift from galactose to glucose-containing media. Fig. 6,A shows a polysome profile from CLF(pRS315-RPS0A) cells at 8 h after the shift to glucose. The relatively large amount of free 60S subunits in the CLF(pRS315-RPS0A) cells was probably the consequence of suboptimal levels of protein expressed from the RPS0A gene on the low copy number plasmid. Elevated levels of 60S subunits have been noted previously in cells with a disrupted copy of RPS0B in which a single genomic copy of the RPS0A gene was the only potential source of Rps0 protein (37). Fig. 6, B and C, are polysome profiles from CLF(pRS314-GAL1/RPS0A) cells at 8 and 14 h after the shift to glucose, respectively. At these time points, CLF(pRS314-GAL1/RPS0A) cells had reduced levels of 18S rRNA relative to CLF(pRS315-RPS0A) and also had considerably higher levels of the 20S rRNA precursor (Fig. 4). Relative to CLF(pRS315-RPS0A) cells, extracts from CLF(pRS314-GAL1/RPS0A) showed a pronounced shift from larger to smaller polysomes, indicating that protein synthesis had been adversely affected by the depletion of Rps0 protein. To determine whether the 40S ribosomal subunits containing 20S rRNA precursors that accumulated under these conditions were active in protein synthesis, the distribution of the 20S rRNA was examined in fractions from CLF(pRS314-GAL1/RPS0A) cells at 14 h after the shift to glucose (Fig. 6,D). Using a probe that recognizes both 20S and 18S rRNAs, we found that the bulk of the 18S rRNA was clearly in the 80S monosome and polysome fractions (Fig. 6, C and D). In contrast, the signal corresponding to the 20S rRNA was found preferentially in fractions corresponding to the 40S and 80S peaks. This finding was confirmed using a probe specific for the 20S rRNA precursor. These data indicate that 40S subunits containing the 20S rRNA have a dramatically reduced ability to function in protein synthesis relative to subunits containing mature 18S rRNA. However, because there are trace amounts of 20S rRNA that cosediment with polysomes in these experiments, we cannot exclude the possibility that subunits containing the 20S precursor are at least partially active in protein synthesis.

Several studies have shown that introns in polymerase II transcripts contain snoRNAs that participate in rRNA processing (45). Because the RPS0A transcript contains an intron, it is possible that an intron-encoded snoRNA is responsible for 20S rRNA processing rather than the Rps0A protein. To distinguish between these possibilities we screened a yeast cDNA expression library for an RPS0A cDNA (46). The screen was based on the ability of the library plasmids to complement the slow-growth phenotype of cells disrupted in the RPS0B gene (37). Because cDNA expression was under control of the GAL1 promoter, cell growth was monitored on synthetic medium using 2% galactose/0.1% sucrose. Plasmids that complemented the slow-growth phenotype were isolated and mapped by restriction enzyme digestion; those thought to contain the RPS0A cDNA were confirmed by DNA sequence analysis (data not shown). The RPS0A cDNA together with the GAL1 promoter was cloned into the plasmid pRS315, which has a LEU2 selectable marker suitable for transforming the cDNA into CLF(pRS314-GAL1/RPS0A) cells.

CLF(pRS314-GAL1/RPS0A) cells that were transformed with either pRS315-RPS0A cDNA or pRS315 alone were grown on synthetic galactose-containing media lacking leucine but containing tryptophan. Under these conditions, the pRS314-GAL1/RPS0A plasmid can be lost, as long as there is another functional RPS0 gene remaining in the cells. Of 14 colonies screened from the cells transformed with pRS315 alone, all still contained the pRS314-GAL1/RPS0A plasmid. In contrast, of 19 colonies screened that were transformed with pRS315 containing the RPS0A cDNA, 7 were identified that had lost pRS314-GAL1/RPS0A. These data indicate that expression from the RPS0A cDNA can maintain the viability of cells lacking other RPS0 genes. Because it is possible that the RPS0A gene derived from the pRS314 plasmid may have been transferred to the chromosome by recombination, we used the PCR to determine whether there was still an intron-containing RPS0A gene present in cells that had lost this plasmid. PCR with primers that flank the intron in the RPS0A gene was used to assess the presence of the RPS0A gene or cDNA in CLF cells containing the pRS315 plasmid alone (Fig. 7, Lane 1) or CLF cells containing pRS315-RPS0A cDNA (Fig. 7, Lanes 2–4). Fig. 7, Lane 1, shows a predominant PCR product of approximately 1200–1300 bp from CLF cells containing pRS314-GAL1/RPS0A and pRS315. This product is nearly the size (1174 bp) expected for a product from the RPS0A gene containing an intron. The less intense fragment above the major product seems to result from the amplification of the disrupted allele of RPS0A in the chromosome that still retains sequences complementary to the primers used in PCR. The slightly larger size is expected because the URA3 gene used to disrupt RPS0A was slightly larger than the fragment from RPS0A that it replaced. This minor band is also evident in Lanes 2–4. The major PCR product in Lanes 2–4 has a size of approximately 700 bp, consistent with the size predicted for a product from the RPS0A gene lacking an intron (718 bp). Thus, CLF(pRS315-RPS0A cDNA) cells that lack intron sequences derived from either of the RPS0 genes are viable, which demonstrates that it is the Rps0A protein and not an intron encoded within the RPS0A mRNA that is required for 20S rRNA processing and the maturation of 40S ribosomal subunits.

The results shown in this report demonstrate that the yeast Rps0 proteins are required for the maturation of 40S ribosomal subunits. Specifically, the Rps0 proteins are needed for the efficient processing of the 20S rRNA precursor to mature 18S rRNA. This processing step involves an endonucleolytic cleavage that removes 209 nucleotides from the 3′ end of the 20S rRNA precursor (47). 20S rRNA processing and the formation of mature 18S rRNA is one of the final steps in the maturation of the 40S ribosomal subunit (44, 48). On the assumption that the immature subunits containing 20S rRNA are degraded at a faster rate than mature 40S subunits, the reduced rate of rRNA processing observed in Rps0-depleted cells would explain the overall reduction in the level of 40S subunits reported previously in cells in which one or the other of the RPS0 genes was disrupted (37).

Udem and Warner (44) have shown previously that 40S subunits containing 20S rRNA are excluded from polysomes and are presumably inactive in protein synthesis. Their experiments were carried out with exponentially growing cells under conditions in which the immature 40S subunits are transient intermediates found in relatively low amounts compared with mature 40S subunits. The low amounts of subunits containing 20S rRNA under those conditions may contribute to their inability to compete effectively with mature subunits for incorporation into polysomes. In contrast, cells depleted of Rps0 protein have relatively high levels of immature subunits containing the 20S precursor. Yet even when protein synthesis is limited by the availability of 40S subunits, subunits containing 20S rRNA that accumulate in Rps0A-deficient cells are still preferentially excluded from polysomes, which indicates that they may be inactive in protein synthesis. However, because small amounts of 20S rRNA were found to cosediment with 80S monosomes and with polysomes, the possibility exists that these subunits are at least partially active in protein synthesis. Nonetheless, these data demonstrate that the assembly of Rps0 proteins into immature 40S subunits, and the subsequent processing of the 20S rRNA precursor represent key steps in defining the translational capacity of yeast cells.

In addition to the RPS0 genes, two other yeast genes are known to be required for the processing of 20S rRNA. These genes are UBI3 and DRS2. UBI3 encodes a ribosomal protein S31/ubiquitin fusion protein. Finley et al. (49) showed that the S31 protein encoded by the UBI3 gene was needed for 20S rRNA processing. The ubiquitin encoded by UBI3, on the other hand, was dispensable if a gene encoding S31 was present in multiple copies within cells. These data indicate the ubiquitin may function as a chaperone to facilitate the incorporation of S31 into the assembling subunit that is then required for 20S rRNA processing. The function of the DRS2 gene is unknown, but it encodes a protein that belongs to a family of membrane-associated ATPases involved in divalent cation transport (50). Exactly how a putative membrane-associated ATPase is required for 20S rRNA processing is unknown. But its involvement raises the intriguing possibility that this step in the maturation of 40S subunits may be sensitive to changes in cation concentrations and, thus, subject to regulation.

Because many aspects of ribosome structure and function are conserved between humans and yeast, studies on yeast ribosomal proteins can potentially provide insight into the functions of their mammalian counterparts. A processing step similar to 20S rRNA processing in yeast also occurs during the maturation of the 3′ end of 18S rRNA in mammalian cells (51). Moreover, both the Rps0 and Ubi3 proteins that are required for this processing step in yeast have human homologues. Interestingly, p40/37-LRP and UbA80, homologues of the Rps0 and Ubi3 proteins, respectively, are overexpressed in colon cancer cells (5, 6, 32, 36). Although the expression of many genes encoding ribosomal proteins is increased in colon cancer, p40/37-LRP and UbA80 are noteworthy because their levels of expression correlate with tumor aggressiveness (6, 32). These observations point to a possible linkage between 20S rRNA processing and tumor aggressiveness in colon carcinomas. If, as is observed in yeast, 20S rRNA processing is required for subunit activity, the level of expression of p40/37-LRP and UbA80 could potentially play a critical role in defining the translational capacity of cells. Overexpression of these proteins in colon cancer cells may contribute to the transformed phenotype by increasing the amount of active 40S ribosomal subunits, which could then affect the synthesis of key growth regulatory proteins.

In contrast to p40/37-LRP which is overexpressed in a wide range of tumors, UbA80 overexpression has not been observed in gastric or hepatocellular carcinomas (35, 36). Whether this indicates that UbA80 is not limiting for 40S production in tumors derived from gastric and hepatic tissues has not been determined. The finding that p40/37-LRP protein is one of the few ribosomal proteins that is consistently overexpressed in a wide range of tumors suggests that it can play a pivotal role in meeting the demand for protein synthesis in many types of proliferating cells.

The basis for the link between expression of the p40/37-LRP gene and cell growth may reside not only with the protein encoded by this gene but also with an RNA product derived from the p40/37-LRP pre-mRNA. An intron in the pre-mRNA for mammalian p40/37-LRP has been shown to encode the small nucleolar RNA, E2 snoRNA (52). Several snoRNAs have been identified within the introns of mRNAs for either ribosomal proteins or nucleolar proteins thought to function in ribosome assembly (45). This arrangement may help coordinate the expression of the various proteins and RNAs responsible for ribosome synthesis and function. Interestingly, the E2 snoRNA has been shown to function in rRNA processing, specifically at the cleavage step that gives rise to the 3′ end of mature 18S rRNA (53). This processing event is equivalent to the processing of the 20S rRNA precursor in yeast that is dependent on expression from the RPS0A gene. Because the RPS0 genes have introns, it was possible that an intron-encoded snoRNA—rather than the Rps0 proteins—was required for 20S rRNA processing. Our results, which showed that a cDNA encoding the Rps0A protein supported growth of a strain doubly disrupted for the RPS0 genes, demonstrated that, in yeast, it was the Rps0A protein and not an intron of the RPS0A pre-mRNA that was required for 20S rRNA processing. Although our data indicate that intronic sequences within the RPS0A gene are not required for 20S rRNA processing, they do not rule out the possibility that another snoRNA encoded elsewhere in the genome may be required for this processing reaction.

Clearly, more studies are necessary to determine the role of p40/37-LRP gene products in the structure and function of human ribosomes. Of particular interest is the potential role for the p40/37-LRP protein in the processing of the 20S rRNA precursor and the formation of active 40S ribosomal subunits. If, like the Rsp0 proteins, mammalian p40/37-LRP proteins are required for the processing step that generates the 3′ end of 18S rRNA, then the p40/37-LRP pre-mRNA would encode two products required for the same step in maturation of 40S subunits, the E2 snoRNA and the p40/37-LRP protein. It would then be expected that the level of active 40S subunits would be particularly sensitive to changes in the level of p40/37-LRP pre-mRNA. This sensitivity may explain the widespread linkage between the overexpression of the p40/37-LRP gene and tumor cell growth and proliferation.

We thank Drs. Vilius Stribinskis, Lisa R. Williams, Robert Gray, and Thomas Geoghegan for critically reading the manuscript. We also thank Dorothy Struck and Leroy Heron for assistance in the preparation of the article.