Paralogous Ribosomal Protein L32-1 and L32-2 in Fission Yeast May Function Distinctively in Cellular Proliferation and Quiescence by Changing the Ratio of Rpl32 Paralogs

Fission yeast cells express Rpl32-2 highly while Rpl32-1 lowly in log phase; in contrast, expression of Rpl32-1 raises and reaches a peak level while Rpl32-2 is downregulated to a low basic level when cells enter into stationary phase. Overexpression of Rpl32-1 inhibits cell growth while overexpression of Rpl32-2 does not. Deleting rpl32-2 impairs cell growth more severely than deleting rpl32-1 does. Cell growth impaired by deleting either paralog can be rescued completely by reintroducing rpl32-2, but only partly by rpl32-1. Overexpression of Rpl32-1 inhibits cell division, yielding 4c DNA and multiple septa, while overexpressed Rpl32-2 promotes it. Transcriptomics analysis proved that Rpl32 paralogs regulate expression of a subset of genes related with cell division and stress response in a distinctive way. This functional difference of the two paralogs is due to their difference of 95th amino acid residue. The significance of a competitive inhibition between Rpl32 paralogs on their expression is discussed.


Introduction
Paralogous genes exist after the gene duplication event and usually code for proteins with similar function and/or structure. Gene duplication is thought to supply raw genetic material, allowing functional divergence and rapid biological evolution [1][2][3]. In yeast, most of cytoplasmic ribosomal proteins are retained in duplication. Fission yeast Schizasaccharomyces pombe has 80 different ribosomal proteins encoded by 143 different genes, 56 of which are encoded by two or more duplicated genes (http://ribosome. med.miyazaki-u.ac.jp). For example, S. pombe ribosomal protein L32 (Rpl32) paralogs are encoded by two paralogous genes, rpl32-1 (SPBC16C6.11) and rpl32-2 (SPAC3H5.10). These ribosomal protein paralogs have many common and distinct properties, such as (a) a very similar amino acid sequence among paralogs, (b) a high mRNA expression correlation among paralogs, and (c) the whole functional class, required the whole genomic duplication [4] or small-scale duplications [5], implying a low level of functional differentiation and possibly an mRNA dosage increase as an explanation for the retention of duplicates in ribosomal proteins [6]. However, recent studies showed that duplicated ribosomal proteins have various functional divergences [7] [8]. Komili et al. (2007) proposed that different combinations of RP paralogs even generate ''ribosome codes'' which are involved in translational regulation of specific mRNAs [9][10][11].
An essential function of ribosomal proteins is to interact with rRNA to constitute protein synthesis machinery-ribosomes [12]. Whereas many studies have revealed that some ribosomal proteins have ''extraribosomal functions'' [10] [13][14][15][16][17]. Our lab reported previously that Rpl32-2 specifically bound to DNA sequence containing GTTGGT, activating transcription of reporter genes in GAL4-base hybrid system in S. cerevisiae [18]. Recently, we have reported that deletion of Rpl32 paralogs causes reduction of ribosome level which may trigger flocculation of fission yeast cells [19]. In the present study, we further report that Rpl32 paralogs, Rpl32-1 and Rpl32-2 are involved differently in the regulation of the transition between proliferation and quiescence which are two distinct cell states for all organisms. When the nutrient is sufficient, yeast cells are normally in proliferation state, with high metabolism rate and active cell division, while under nutrition deprivation conditions such as a stationary phase culture, cells are usually in quiescence state with low metabolism rate and inactivated division, as well as high resistance to environment stress [20]. Yeast cells can shift between proliferation and quiescence in response to environmental cues. This work focuses on the functional divergence of Rpl32 paralogs in proliferation and the transition from proliferation to quiescence, as well as their molecular mechanisms.

Rpl32-1 and Rpl32-2 Expressed Distinctively in Fission Yeast Cells in Proliferation or Proliferation to Quiescence Transition
As shown in Fig. 1A upper panel, during log phase fission cells expressed rpl32-1 at a lower basic level while expressed rpl32-2 highly. The rpl32-2 expression reached to a peak level before midlog phase and then slowed down to a lower basic level before end of log phase. In contrast, when fission yeast cells were entering into stationary phase they expressed rpl32-2 at a lower basic level while raised expression of rpl32-1 rapidly. The rpl32-1 expression reached to a peak level when cells just entered into stationary phase (at ,36 h, early stationary phase). To further confirm differential expression patterns of these two paralogs during the course of cell growth, we constructed double-labeled mutant strain rpl32-1-6his-rpl32-2-HA. Western blot on rpl32-1-6his-rpl32-2-HA cells respectively using antibodies against 6His or HA also confirmed that in log phase Rpl32-2 was highly expressed and Rpl32-1 was lowly expressed in cells; in contrast, in early stationary phase protein level of Rpl32-1 was upregulated and Rpl32-2 was downregulated (Fig. 1A, lower panel). Since heterogeneous molecular weight of Rpl32-1-6His and Rpl32-2-HA, the Rpl32-2 antibodies against Rpl32 paralogous proteins was used for Western Blot on Rpl32 in unlabeled WT cells and results showed that total protein of Rpl32 remained at the same level in WT cells in both log phase and early stationary phase (Fig. 1A lower panel).
We hypothesize that after exponential growth, changes of expression patterns of Rpl32 paralogous genes are related with need for cells to adjust metabolism status to transit from proliferation to quiescence state when cells sense a shortage of nutrients in the medium [21][22][23][24]. Since cells can be induced to enter into quiescence state by nitrogen stress, carbon stress or stationary phase culturing [20], proliferating cells in log phase were transferred into fresh EMM2, cell-free log phase medium (LP), cell-free stationary phase medium (SP), nitrogen deficient EMM2-N medium and carbon deficient EMM2-C medium for further cultivation for 12 h. QPCR analysis showed that in cells grown in rich media such as fresh EMM2 and LP, rpl32-2 mRNA level was higher than rpl32-1, whereas, in cells cultured in SP, EMM2-N and EMM2-C, the mRNA levels of rpl32-2 and rpl32-1 were in reverse ( Fig. 1B upper panels). Western blot on rpl32-1-6his-rpl32-2-HA cells grown in above various media confirmed similar expression patterns of Rpl32-2 paralogs at protein level to mRNA level. However, the total protein expression level of Rpl32 stayed nearly at the same level in WT cells grown in all tested media ( Fig. 1B lower panel).
There is only a Single Amino Acid Difference at 95 th Position between the Mature Protein Rpl32-1 and Rpl32-2 Alignment between Rpl32-1 and Rpl32-2 in S. pombe ( Fig. 2A) shows that they share 96.85% similarity in amino acids sequence. There are only 4 different residues, Ile 4 Val, Val 7 Ile, Leu 21 Arg and Ser 95 Gly, in 127 amino acids of Rpl32 paralogs [5]. It is known that after ribosome proteins are synthesized in the cytosol, they need to enter into the nucleus to be assembled into ribosomal subunits, and then move back to the cytosol to assist protein synthesis [25] [26]. Analysis by PSORT II software for recognition of signal sequence (http://psort.hgc.jp/form2.html) suggests that the first 23 amino acids of Rpl32-1 or Rpl32-2 may be a nuclear localization signal sequence ( Fig.2A). Thus, we constructed mutants harboring overexpression plasmids with a gene, respectively, coding Rpl32 paralogs or Rpl32 paralogs deleted of nuclear localization signal peptide, all of which were labeled with EGFP. Fluorescent microscopic images showed that Rpl32-1-EGFP and Rpl32-2-EGFP were located in the nucleus, while Nterminal 23 amino acids lacking Rpl32-1-23-EGFP and Rpl32-2-23-EGFP were found in the cytosol (Fig. 2B). This result confirmed the function of 1-23 amino acids as a nuclear localization signal sequence. In mature Rpl32 proteins without nuclear localization signal sequence, the only different amino acid between two paralogs is Ser 95 Gly. So a site-directed mutagenesis method was used to replace Ser 95 on Rpl32-1 with Gly and Gly 95 on Rpl32-2 with Ser to create Rpl32-1M and Rpl32-2M mutant proteins, respectively, for determination of their functional differences in the following experiments.
Overexpression or Deletion of Rpl32-1 or Rpl32-2 Exerted Different Effects on Cell Growth We constructed mutants harboring plasmid pREP3X overexpressing Rpl32-1, or Rpl32-2, or Rpl32-1M, or Rpl32-2M, respectively, in order to mimic the expression pattern of Rpl32 paralogs for determination of their effects on cell growth states. The growth curves of rpl32-1 and rpl32-2 cells showed that Rpl32-1 overexpression slowed down apparently cell growth while Rpl32-2 overexpression had no significant effect on it, compared to that of WT cells (Fig. 3A). Interestingly, like Rpl32-1, Rpl32-2M overexpression inhibited cell growth, while, like Rpl32-2, Rpl32-1M overexpression did not exert effect on cell growth, suggesting the ability of Rpl32-1 overexpression to cause growth defects is mediated by serine-95. QPCR analysis (Fig. 3C) showed that as compared to WT cells in log phase, rpl32-1 mRNA level increased by 40 times and rpl32-2 mRNA level decreased by 81.5% in rpl32-1 cells, while rpl32-1 mRNA level decreased by 50% and rpl32-2 mRNA level increased by 4.5 time in rpl32-2 cells; as compared to WT cells in early stationary phase, rpl32-1 mRNA level increased by 30 times and rpl32-2 mRNA level decreased only by 10% in rpl32-1cells, while rpl32-1 mRNA level decreased by 90% and rpl32-2 mRNA level increased by 3 times in rpl32-2 cells. Apparently, overexpression of either rpl32-1 or rpl32-2 reduced mRNA level of its paralog, whereas overexpression of rpl32-2 reduced mRNA level of rpl32-1 less in rpl32-2 cells in log phase than in early stationary phase and overexpression of rpl32-1 reduced mRNA level of rpl32-2 less in rpl32-1 cells in early stationary phase than in log phase. It is notable that rpl32-1M mRNA level in rpl32-1M cells was similar to rpl32-1mRNA level in rpl32-1 cells, and rpl32-2M mRNA level in rpl32-2M cells was similar to rpl32-2 mRNA level in rpl32-2 cells, implying that the single site mutagenesis did not change the mRNA level of corresponding rpl32 paralogs. Therefore we can exclude the possibility that differential growth phenotypes reported above were resulted from different transcript levels of either rpl32 paralog rather than 95 th amino acid. Western blot showed that protein expression level of Rpl32-1 was decreased in rpl32-1-6his/rpl32-2 and rpl32-1-6his/rpl32-1M cells as compared to rpl32-1-6his cells harboring empty plasmid ( Fig. 3D upper panel) and protein expression level of Rpl32-2 was reduced in rpl32-2-HA/rpl32-1 and rpl32-2-HA/rpl32-2M cells compared to rpl32-2-HA cells harboring empty plasmid ( Fig. 3D middle panel), whereas total protein level of Rpl32 were almost unchanged in all of rpl32-1, rpl32-2, rpl32-1M, and rpl32-2M cells as compared to WT cells ( Fig. 3D lower panel), implying that expression of Rpl32 paralogs is also regulated at translational level, and the protein expression level of corresponding Rpl32 paralogs was not influenced by the single site mutation at 95 th position.

Cellular Analysis Explores the Physiological Significance and Mechanism of Differential Expression of Rpl32-1 and Rpl32-2
In order to elucidate the physiological significance and mechanism of Rpl32 paralogs differential expression, we compared phenotypes between WT and Rpl32 paralogs overexpression cells in different growth states.
FACS analysis (Fig. 4A) showed that in log phase WT cells exist essentially as 2c DNA cells; in early stationary phase many cells contained 4c DNA which reflected retardation of their cell separation after nucleus division. Interestingly, in log phase rpl32-1 cell population had many 4c DNA cells as seen in WT cells in early stationary phase, while in early stationary phase rpl32-2 cells were not found to contain 4c DNA, implying that highly expressed Rpl32-1 inhibited cell separation and prolonged cell separation time while high expressed Rpl32-2 promoted cell separation. Elutriation of cells which were previously synchronized at G1/S phase with HU ( Fig. 4B) showed that cells from all strains doubled their DNA and entered into the G2 phase after 30 min of elutriation, indicating null effect of highly expressed Rpl32-1 or Rpl32-2 on DNA duplication. After 2 h of elutriation WT and rpl32-2 cells showed significant population with 4c DNA, while cells with 4c DNA occurred less in rpl32-1 cell populations, implying that nucleus division in rpl32-1 cells was retarded [28]. After 4 h of elutriation, WT and rpl32-2 cells essentially had no 4c DNA, while many rpl32-1 cells stayed still in the stage with 4c DNA. These results suggested that high expression of Rpl32-1inhibits nucleus division and cell separation.
DIC (differential interference contrast microscope) observation showed that in log phase, WT cells were rod-like in morphology, rpl32-2 cells had no significant morphological difference from WT cells, while rpl32-1 cells were longer equivalent to approximately 46 the length of WT cells (Fig. 4D1); in early stationary phase, all of rpl32-2 cells, and most of WT and rpl32-1 cells showed the smaller pear-like shape, but some of WT and rpl32-1 cells exhibited a little longer pear-like shape (Fig. 4D3). Thus, DIC results also suggest that highly expressed Rpl32-1 suppressed cell separation while highly expressed Rpl32-2 promoted cell separation.
DAPI nucleus staining showed that in log phase smaller mononuclear morphology of rpl32-2 cells was not different from WT cells, while rpl32-1 cells often exhibited a larger nucleus which reflected interruption of nucleus division (Fig. 4D2), consisted with 4c DNA cells explored by FACS analysis. In early stationary phase, 16% of WT and rpl32-1 cells contained two nuclei, while rpl32-2 cells with doubled nuclei were seldom found (Fig. 4D4), consistent with DIC data.
CW (calcofluor white) staining (Fig. 4E, F) explored that in log phase both WT and rpl32-2 septum positive cells had only one septum each, while 66.7% rpl32-1 septum positive cells had multiple septa each, suggesting that highly expressed Rpl32-1 inhibited cell separation, resulting in accumulation of septa. In early stationary phase, the nutrient stress withheld cells from reaching division threshold in size, so cells became smaller and formed no septa (Fig. 4F). Interestingly, in early stationary phase, though some cells of WT and rpl32-1 had a little larger size and two nuclei they lacked septum, while some rpl32-2 cells exhibited septum-like structure though they were smaller (Fig. 4D), indicating a promotive role of Rpl32-2 on formation of cell septum.
It is notable that rpl32-1M mutant cells shared the same cellular phenotype with rpl32-2 cells, and rpl32-2M mutant cells shared the same cellular phenotype with rpl32-1 cells (Fig. 4A-F), again, indicating that the functional differences of Rpl32 paralogs are determined by their difference of the 95 th amino acid.
Significantly enriched Gene Ontology (Fig. 5), which is based on GeneDB (http://www.genedb.org/genedb/pombe/index.jsp), clearly shows that 16 genes among above 64 genes are involved in mitotic cell cycle, cell separation during cytokinesis, DNA replication and nucleus division. These genes were highly expressed in WT cells in log phase while were downregulated when cells entered into early stationary phase. Overexpression of Rpl32-2 upregulated mRNA levels of above genes in cells in early stationary phase, especially the mRNA level of fta2, which codes a protein crucial for segregation of the duplicated sister chromatids into two equal sets [29], and ace2, which codes a protein crucial to cell separation [30] [31]. In contrast, overexpression of Rpl32-1 downregulated mRNA levels of above some genes, especially, apparently reduced mRNA levels of both fta2 and ace2 in cells in log phase. In addition to the regulation of cell proliferation, the two paralogs are also involved in the response to environmental nutrient status. Expression of 25 genes related to cellular response to stress were upregulated in WT cells in early stationary phase as compared to log phase, whereas expression of these genes were downregulated in rpl32-2 cells in early stationary phase and were upregulated in rpl32-1 cells in log phase. It is known that the meiotic cell cycle as a process of yeast sexual reproduction is also a cellular response to negative environments [32]. We found that expression of 12 genes involved in meiotic cell cycle were more highly expressed in stationary phase than log phase, however overexpression of Rpl32-2 downregulated transcription of these genes in stationary phase, and overexpression of Rpl32-1 upregulated transcription of these genes in log phase.
From the results of microarray analysis, we randomly determined 5 genes including fta2 and ace2, with an average transcription fold change .2, which are regulated oppositely by Rpl32 paralogs. We performed QPCR analysis and confirmed that, consistent with that of microarray analysis, these genes were differentially expressed in cells in log phase and stationary phase, and upregulated or downregulated in cells overexpressing either Rpl32 paralog in different growth phases as compared to WT cells (seen noted in Fig. 5).

The Two Paralogs of Ribosomal Protein L32 Function Differently in Proliferation and Quiescence by Regulation of Different Genes Expression
This study first explores that expression patterns of RPL32 paralogous genes in S. pombe were different from each other: Rpl32-2 had higher expression level than Rpl32-1 in log phase, while Rpl32-1 had a higher expression level than Rpl32-2 in early stationary phase. Further study found that overexpression of Rpl32-1 inhibited cell growth while overexpression of Rpl32-2 did not. Deleting Rpl32-2 impaired cell growth more severely than Rpl32-1. Cell growth impaired by deleting either paralog could be rescued completely by reintroducing Rpl32-2, but only partly by Rpl32-1. We propose that Rpl32-2 plays a leading role in proliferation, while Rpl32-1 functions in transition from proliferation to quiescence. The reason why overexpression of Rpl32-2 did not promote cell division in log phase might be due to that endogenous expression level of Rpl32-2 in log phase was already up some threshold for cell proliferation so extra overexpressed Rpl32-2 would bind to its own transcripts, regulating their splicing to modulate the protein level of Rpl32-2 itself [33] [34]. Western blot analysis also supported that overexpression of Rpl32-2 did not raise total protein level of Rpl32 in cells. Since endogenous expression level of Rpl32-2 was very low in early stationary phase, recombinant overexpressed Rpl32-2 in cells in this phase was used as a model to determine functions of Rpl32-2. For the same reason, recombinant overexpressed Rpl32-1 in cells in log phase could be used as a model to determine functions of Rpl32-1. Following this lead, we further proved that Rpl32-2 favored for cell division and septum formation, while Rpl32-1 suppressed nucleus division and cell separation. Previous research work reported that the hyper-active SIN (septation initiation network) pathway also could cause multiple septa formation [35]. If the SIN pathway was activated in asynchronous cells, the cells stopped polarized growth and did not elongate, arresting as mononucleate or binucleate cells with multiple septa [36][37][38]. But in our observations, rpl-32-1 cells with multiple septa were much longer than WT cells so the elongation seemed not to be ceased. And the transcriptomic data (NCBI Series Entry: GSE43827) also shows that the expression profiles of genes related to SIN pathway have no difference between WT and Rpl32 paralogs overexpression cells. So we believe that the multiple septa we observed are more likely to be resulted from the suppression of cell separation by Rpl32-1 rather than the hyper-activation of SIN pathway.
Transcriptomics analysis suggested a molecular mechanism based on which Rpl32 paralogs function distinctly in proliferation and transition from proliferation to quiescence by regulation of expression of a subset of genes related with cell division and stress response in a distinctive way. First, compared to WT cells, overexpression of Rpl32-2 promoted expression of some genes related to mitotic cell cycle, cell separation and nuclear division, such as ace2 and fta2, in rpl32-2 cells in early stationary phase while overexpression of Rpl32-1 inhibited expression of ace2 and fta2 in rpl32-1 cells in log phase. Second, overexpression of Rpl32-2 downregulated expression of many genes related to cellular responses to environmental nutrient stress in rpl32-2 cells in early stationary phase, while overexpression of Rpl32-1 upregulated expression of these genes in rpl32-1 cells in log phase, compared to WT cells, which is consistent with that transition from proliferation to quiescence is a cellular response to stress. Third, high expression of Rpl32-2 also downregulated, while high expression of Rpl32-1 upregulated those genes involved in meiotic cell cycle which is also one kind of the responses to environmental stress [32], compared to WT cells. The study concerning how Rpl32 paralogs regulate their target gens expression is in proceeding, and Figure 5. Transcriptomic patterns of WT, rpl32-1 or rpl32-2 cells respectively in log or stationary phase. Color panel indicates relative increase (red), decrease (blue) and median (white) of transcription level for 64 genes regulated by Rpl32-1 and Rpl32-2 in a distinctive way (at least 2fold changes). Bold type means genes are regulated oppositely by both Rpl32 paralogs (Fold change .2) and nonbold type means genes are regulated by either of Rpl32 paralogs (Fold change .2). * Changes in transcripts level were confirmed by QPCR. doi:10.1371/journal.pone.0060689.g005 before it is elucidated we may not exclude the possibility of indirect effects due to cell-cycle arrest yet.

Expression of Proliferation and Quiescence Related Genes may be Regulated by Changing the Ratio of RPL32 Paralogs
Although Rpl32 paralogs have different functions in cell proliferation and transition from proliferation to quiescence, they do share the same basic ribosomal protein function. (1) Deletion of either of the two paralogs was not vital to cells, but the double deletion cells could not survive (Data not shown), consistent with the report by Kim et al. (2010) [33]; (2) single deletion of either paralogous gene led to the shortfall of total Rpl32 protein and this shortfall could be rescued by overexpression of either Rpl32 paralog; (3) deletion of either of the two paralogs impaired cell growth, and overexpression of Rpl32-2 could fully complement the growth defect of either rpl32-1gor rpl32-2g deletion mutants. These suggest that Rpl32 paralogous genes are complementary on their ribosomal function, therefore, Rpl32 paralogous proteins have dosage benefit [6] [27] [34]. With regards to that overexpression of Rpl32-1 only partly complement the growth defect of either rpl32-1g or rpl32-2g deletion mutant, this may be due to its extraribosomal proliferation inhibition besides the basic ribosomal protein function.
In other hand, overexpression of either paralog did not lead to accumulation of total Rpl32 protein. This suggests S. pombe Rpl32 has a self-feedback inhibition, as did by Rpl30 in S. cerevisiae [39][40][41]. In S. cerevisiae Rpl30 can regulate the splicing and the subsequent translation of its own mRNA, and the pre-rRNA processing to control expression of itself [42]. However, unlike S. cerevisiae Rpl30, S. pombe Rpl32 has two paralogs. We have proved that overexpression of Rpl32-1 inhibited expression of Rpl32-2 and vice versa in order to maintain the total RPL32 protein level unchanged. Since the change of Rpl32-1 and Rpl32-2 relative expression level had a distinctive regulation to a subset of genes related to cell proliferation, it is reasonable to conclude that specific genes expression may be regulated by changing the ratio of Rpl32 paralogs, which is considered as a ribosomal code [9] [10]. This competitive inhibition between Rpl32 paralogs on their expression provides a new evidence for the transcriptional regulation with the ribosome code in S. pombe [43], and also in S. cerevisiae [11] (Fig. 6). Certainly we have not excluded the possibility that Rpl32 paralogs individually regulate cell growth when reaching up a higher level rather than as a ribosome code as proposed.
The Functional Differences of Rpl32 Paralogs is due to One Amino Acid Difference at Position 95 th Although S. pombe Rpl32 paralogs have distinct expression patterns and functions in different cellular states, there are only 4 different amino acids in their sequence. Three of them are located within the 1-23 amino acid residues signal sequence. In mature protein (24-127 amino acids), only one amino acid residue is different: Ser 95 for RPL32-1 and Gly 95 for Rpl32-2. By replacing Ser 95 on Rpl32-1 with Gly, we generated Rpl32-1M mutant and vice versa we generated Rpl32-2M mutant. Cells overexpressing Rpl32-2M had the similar features observed on cells overexpressing Rpl32-1, with impaired proliferation, and inhibited nucleus division and cell separation. Cells overexpressing Rpl32-1M exhibited similar characteristics shown on cells overexpressing Rpl32-2, with promoted proliferation, nucleus division, septa formation and cell separation. This suggests that the functional difference of these two paralogues proteins mainly is due to the divergence of the 95 th amino acids, rather than the other 3 different residues in the 1-23 amino acid residues signal sequence.
It is thought that protein paralogs result from a small-scale duplication including single gene or segmental duplications, or from a whole-genomic duplication [4]. Protein paralogs resulting from the whole-genome duplication are more likely to share interaction partners and biological functions than smaller-scale duplicates [44]. But the whole-genome duplications did not occurred in S. pombe [5], different from S. cerevisiae which have undergone some whole-genome duplication approximately 100 mln years ago [45]. Our results provide a new evidence of the derivation of functional diversity in paralogous proteins from small-scale duplication during evolution.

S. pombe Strains, Media and Culture Conditions
Fission yeast strains used in this work are described in Table S1. Edinburgh Minimal Medium 2 (EMM2), EMM2-N, and EMM2-C media were made as described by Su et al. (1996) [46]. The cell-free stationary phase medium (SP) and the cell-free log phase medium (LP) were made respectively from stationary phase culture and log phase culture in EMM2. Cells were removed by centrifugation and the supernatant was sterilized by vacuum filtration through 0.45-mm (pore-size) cellulose nitrate filters.
Fission yeast cells were generally cultured in EMM2 at 30uC on a shaker with 220 rpm. When necessary, cells were cultured to log phase (10 7 cells/ml, OD 600 0.5), or early stationary phase (2610 8 cells/ml, OD 600 10). For the determination of nutrient effects, cells cultured in EMM2 to log phase were harvested and washed twice with phosphate-buffered saline (PBS, pH 7.4), then transferred into same volume of EMM2, EMM2-N, EMM2-C, Figure 6. Schematic representation of a possible regulation of cell growth states by changing the ratio of Rpl32 paralogs. In rich nutrient environment, cells highly express Rpl32-2 while lowly express Rpl32-1 to promote cell division for proliferation. Nutrition deficiency induces cells to upregulate Rpl32-1 expression and downregulate Rpl32-2 expression to inhibit cell division for quiescence. doi:10.1371/journal.pone.0060689.g006 LP, or SP for incubation. For synchronous growth [47], cells cultured in EMM2 to log phase were collected, transferred to EMM2 with 11 mM hydroxyurea (HU) and incubated for 6 h. The synchronized cells were harvested, washed 3 times with PBS (pH 7.4), and then transferred to EMM2 for elutriation.

Gene Deletion
The rpl32-1 or rpl32-2-targeting DNA fragments containing kanmx6 were amplified by PCR from WT cells genomic DNA and plasmid pFA6a-kanmx6 using deletion primers sets (Table S2), and transformed into WT cells to generate deletion mutants by gene replacement [49].
Above plasmids or empty plasmids (as control) were transformed into WT cells or deletion mutants (in deletion-rescuing experiments) to produce transformants using LiAc method [51].

Quantitative PCR (QPCR)
Total RNA was isolated from cells using TRIZOL reagent (Invitrogen), and treated with RNase-free DNase I (TaKaRa) to eliminate any genomic DNA. First-strand cDNA was synthesized from DNA-free RNA using the SuperScript III First-Strand Synthesis System (Invitrogen) and oligo-dT primers. QPCR was done using amplification mixtures containing TaKaRa SYBR Premix Ex Taq II (Tli RNaseH Plus), reverse-transcribed RNA and primers (Table S2). The ACT1 was used to standardize mRNA level. The experiments were repeated for three biological replicates. Reactions were run on a Rotor-Gene Q 5plex HRM System (Qiagen) using a fluorescent threshold manually set to OD 0.100 for all run [53]. The Cycles Threshold (C T ) values were used to calculate the mean fold change of the reactions via the 2 2DDCT method [54].

FACS Analysis
Indicated cells were collected by centrifugation and fixed overnight in 70% cold EtOH at 4uC, followed by a 2-hr incubation at 37uC with 0.1 mg/ml RNase A in 50 mM sodium citrate buffer, pH 7.0. The treated cells were collected and washed, then resuspended in 2.5 mg/ml propidium iodide in 50 mM Na citrate buffer, pH 7.0, at 2610 6 cells/ml, followed by sonicating for 45s for staining. The DNA content was analyzed on Becton-Dickinson FACSCalibur TM and BD CellQuest TM Pro software [55].

Microarray Analysis
Indicated cells cultured to their log phase or early stationary phase were pelleted by centrifugation, and immediately frozen in liquid nitrogen. Total RNA isolation, reverse-transcription and synthesis of cDNA, and microarray analysis were carried out by Gene-Tech Company Limited (Shanghai, China) as a feebased service. GeneChip Yeast Genome 2.0 Array (Affymetrix) was used. Biotin labled amplified cDNA was stained with Streptavidin, R-phycoerythrin conjugate (SAPE). Microarray chips processed through the FS-450 fluidics station were scanned with the 30007 G scanner (Affymetrix) and analyzed with Partek Genomics Suite 6.5. Microarray analysis was performed on three independent biological replicates. A t-test was applied to detect differences in gene expression between each experimental group and control group. Two criteria were used to determine whether a gene was differentially expressed: fold change of 62.0 and p value ,0.05 using a two-tailed distribution, according to current reports [8] [23] [47]. The value of 2.0 is an accepted cut-off with statistical significance, and likely to be validated by QPCR.