5′UTR Variants of Ribosomal Protein S19 Transcript Determine Translational Efficiency: Implications for Diamond-Blackfan Anemia and Tissue Variability

Background Diamond-Blackfan anemia (DBA) is a lineage specific and congenital erythroblastopenia. The disease is associated with mutations in genes encoding ribosomal proteins resulting in perturbed ribosomal subunit biosynthesis. The RPS19 gene is mutated in approximately 25% of DBA patients and a variety of coding mutations have been described, all presumably leading to haploinsufficiency. A subset of patients carries rare polymorphic sequence variants within the 5′untranslated region (5′UTR) of RPS19. The functional significance of these variants remains unclear. Methodology/Principal Findings We analyzed the distribution of transcriptional start sites (TSS) for RPS19 mRNAs in testis and K562 cells. Twenty-nine novel RPS19 transcripts were identified with different 5′UTR length. Quantification of expressed w.t. 5′UTR variants revealed that a short 5′UTR correlates with high levels of RPS19. The total levels of RPS19 transcripts showed a broad variation between tissues. We also expressed three polymorphic RPS19 5′UTR variants identified in DBA patients. The sequence variants include two insertions (c.-147_-146insGCCA and c.-147_-146insAGCC) and one deletion (c.-144_-141delTTTC). The three 5′UTR polymorphisms are associated with a 20–30% reduction in RPS19 protein levels when compared to the wild-type (w.t.) 5′UTR of corresponding length. Conclusions The RPS19 gene uses a broad range of TSS and a short 5′UTR is associated with increased levels of RPS19. Comparisons between tissues showed a broad variation in the total amount of RPS19 mRNA and in the distribution of TSS used. Furthermore, our results indicate that rare polymorphic 5′UTR variants reduce RPS19 protein levels with implications for Diamond-Blackfan anemia.


Introduction
Diamond-Blackfan anemia (DBA; OMIM #205900) is a rare congenital bone marrow failure characterized by decreased numbers or absence of erythroid precursor cells [1]. Approximately 50-60% of DBA patients carry a mutation in one of nine ribosomal protein (RP) genes of which RPS19 mutations account for 25% [2]. A series of .100 coding mutations in the RPS19 gene have been identified ranging from deletions and insertions of various sizes, single base substitutions resulting in both non-sense and missense mutations and splice site mutations [3]. A large proportion of mutations presumably result in functional haploinsufficiency for RPS19 by removing transcription from one allele (deletions or insertions) or by nonsense mediated mRNA decay (splice site, non-sense and missense mutations) [4]. A translated RPS19 protein variant can also mediate haploinsufficiency due to reduced stability, inappropriate localization to the nucleoli, reduced affinity to interacting partners and failure to assemble into the pre-ribosome (missense mutations) [5][6][7]. Haploinsufficiency for a ribosomal protein leads to perturbed ribosome subunit synthesis [8][9][10] followed by increased cellular stress, cell cycle arrest and apoptosis [11]. The precise mechanism by which RP mutations mediate the erythroid specific phenotype in DBA is still unclear. It has been hypothesized that erythropoiesis is particularly sensitive to ribosomal protein insufficiency and cellular stress because of a high proliferative and protein synthesis rate [12].
A few gene variants have been described in the non-coding 59UTR of the RPS19 gene, i.e. in the first exon upstream of the ATG start codon [13][14][15]. These variants were initially identified in a subset of DBA patients and later in healthy individuals but at a low frequency [14,15]. In addition, targeted resequencing of the entire RPS19 gene in DBA patients has revealed a number of noncoding sequence variants and rare polymorphisms localized to introns and flanking sequences [13,14,16,17]. One of the 59UTR variants (c.-147_-146insGCCA) has been associated with rRNA processing defects but the RPS19 protein levels appeared unchanged in erythroid cells from a patient with this variant [14]. However, it is still unclear if this non-coding sequence variant is transcribed and if it interferes with the translation of RPS19. In addition, the distribution of TSS of w.t. RPS19 have not been carefully analyzed.
The regulation of ribosomal protein expression is critical for cellular adaptation to different requirements. It is well established that the 59UTR of mRNAs is of importance for gene expression by influencing mRNA stability, subcellular localization, accessibility to the ribosomes and interaction with the translational machinery [18,19]. Thus, the 59UTR mediates the adjustment of protein levels to developmental stages, tissue types and growth rate [20]. Conversely, inappropriate expression of 59UTRs can contribute to abnormal developmental phenotypes and disease [21][22][23]. Furthermore, the 59UTR of mRNAs encoding ribosomal proteins contains a 59TOP sequence which enables fast up-or downregulation of RP levels [24,25].
The RPS19 coding sequence and its 59UTR are highly conserved [7,13]. However, the significance of the rare polymorphic 59UTR sequence identified in DBA patients is yet unknown. We examined the transcription of 59UTR variants and we hypothesized that they affect RPS19 protein synthesis rate as a possible contributing mechanism in DBA. We show herein that the expression of three structural RPS19 59UTR variants leads to a reduced translation into RPS19. Furthermore, RPS19 uses a broad range of TSS with effects on RPS19 translation and with tissue variations.

RPS19 expression constructs with different 59UTR variants
RPS19 cDNA clones with three w.t. variants of the 59UTR (figure 1) were amplified and cloned into the reporter vector pAcGFP-N1 (Clontech) downstream of the CMV promoter. The resulting constructs encode fusion proteins consisting of a full length RPS19 linked to a fluorescent reporter at the C-terminus ( Figure 2A). Three DBA associated 59UTR variants (c.-147_-146insGCCA, c.-147_-146insAGCC and c.-144_-141delTTTC) were generated by site directed mutagenesis from the pAcGFP-N1-382-S19-59UTR clone (INTERMEDIATE clone) with Quick change II site directed mutagenesis kit (Stratagene) according to manufacturers recommendations.

Cell culture and transfection
K562 [26], HeLa [27] and HEK293 [28] cells were cultured in RPMI1640 medium supplemented with 10% fetal calf serum, 2 mM L-glutamine and 20 IU penicillin/streptomycin (all Invitrogen) at 37uC with 5% CO 2 in a humidified environment. Cells were transfected in 10 cm dishes with 5 mg of the respective vector using Lipofectamine 2000 H following manufacturer's protocols. Transfected cells were checked for expression of GFP by fluorescence microscopy and harvested using a cell scraper. Cells were collected by centrifugation.

RNA isolation and quantitative RT/PCR
Total RNA was isolated from K562, HeLa and HEK293 cells using TrizolH reagent (Invitrogen). Quality of RNA was checked using the Agilent RNA 6000 nano kit and the Agilent 2100 bioanalyser according to manufacturer's instructions. RNA samples from a panel of different primary tissues were purchased (Human total RNA master panel #636643; Clontech). cDNA was synthesized with M-MULV reverse transcriptase (MBI Fermentas) using random hexamer primers and 2 mg of total RNA following manufacturer's recommendations. Quantitative real-time PCR was performed in triplicates using platinum SYBR green qPCR supermix UDG (Invitrogen) according to the protocol supplied by the manufacturer. Primer sequences and PCR conditions used to quantify w.t. RPS19 mRNA with different 59UTRs are available upon request.
59 Rapid Amplification of cDNA ends (59 RACE) 59RACE was performed with 1 mg of total RNA using the GeneRacerH kit (Invitrogen) according to manufacturer's recommendation. Initially, the RNA was treated with DNase I to clean the samples from any genomic DNA. First strand cDNA synthesis was carried out with GeneRacer Oligo(dT) primer and Superscript RT III RACE ready cDNA kit. For amplification of the cDNA end we used the 59GeneRacer forward primer included in the kit and RPS19 specific reverse primer. The PCR product was cloned into a TOPO-TA vector (Invitrogen) and randomly picked individual clones were sequenced.

Western blotting
HEK293T and K562 cells were lysed in RIPA buffer supplemented with MG132 proteasome inhibitor (SIGMA), phosphatase inhibitor cocktail 1 (SIGMA), 0.1 mM Sodium vanadate (SIGMA) and protease inhibitor cocktail (SIGMA). Cell lysates were separated on a 10% Bis-Tris SDS-PAGE (NuPage gel; Invitrogen), and transferred to PVDF Immobilon-FL membranes (Millipore). Membranes were hybridized with primary antibodies against GFP (Clontech) and b-actin (Abcam). Proteins detected by the antibodies were visualized using Alexa Fluor 680 (a-rabbit) and IRD 800 labeled (a-mouse) secondary antibodies (Molecular probes and LiCor Bioscience, respectively). Western blots were analyzed using the OdysseyH infrared imaging system determining integrated intensities, using b-actin as a normalization control as described previously [29].

Multiple transcript variants of the RPS19 gene
The RPS19 gene spans a genomic region of 11.5 Mb on chromosome 19q and consists of 6 exons [13]. The first exon is untranslated and the start codon is located in the immediate beginning of exon 2. Ribosomal protein S19 exists in one single form consisting of 145 amino acids. Six variants of the RPS19 transcript have been described with differences only in length of the 59UTR from the use of alternate transcription initiation sites [7,13,30,31]. This prompted us to search for additional RPS19 transcript variants and to analyse their effect on RPS19 expression. We performed cDNA 59RACE with RNA from K562 cells and testis and we determined the transcript sizes and their transcriptional start sites. We identified altogether 31 alternative RPS19 transcripts with 59UTRs ranging from 32 to 467 nucleotides. Twenty-nine transcript variants are yet undescribed ( figure 1A). The distribution of the RPS19 59UTR length appeared to be different when comparing randomly picked clones from K562 cells and testis, respectively (figure 1). Clones from testis had a distribution towards longer 59UTRs with the longest clone spanning 467 nucleotides, extending about 100 nucleotides beyond the previously reported longest RPS19 transcript (Gene bank #BC018616). Longer RPS19 59UTRs appear highly structured as predicted by the RNA secondary structure prediction program Mfold (data not shown).
We then analyzed the relative expression of three groups of RPS19 transcripts with different 59UTRs on mRNA from a panel of human tissues (Stratagene) as well as the cell lines K562 and HeLa by quantitative real time PCR. Three specific primer pairs were designed that generate amplicons from the RPS19 59UTR corresponding to nucleotide positions 203 to +89 (amplicon A), 2350 to 2239 (amplicon B) and 2449 to 2354 (amplicon C), respectively ( figure 1B). Quantification of the RT-PCR products revealed different patterns of 59UTRs when comparing different tissues and cell lines (1C and D). Transcripts detected by the shortest amplicon (amplicon A) are predominant in all tissues investigated. The expression of longer 59UTRs defined by the amplicons B and C constitute from ,1% to 17% of the total amount of RPS19 mRNA (figure 1D). No strong correlation was observed between the total amount of RPS19 mRNA and the relative proportions of longer and short 59UTR.

Short RPS19 59UTRs show high translational activity
We made three constructs with RPS19 59UTRs of 35 nucleotides (SHORT; containing the 59TOP sequence), 382 nucleotides (INTERMEDIATE) and 467 nucleotides (LONG), respectively. Each construct expresses a full-length RPS19. The constructs were fused with Green fluorescent protein at the carboxy-terminus (figure 2A) and analyzed when transiently transfected into HEK293T and K562 cells. The SHORT 59UTR variant is translated 4-6 fold more efficient than the variant with 59UTRs of 382 nt and .10 fold more efficient than RPS19 with the 467 nt 59UTR ( figure 2B-C). The results were similar for both cell lines.

DBA associated 59UTR variants affect translational activity
We next investigated the effect on translation of three distinct polymorphic sequence variants in the 59UTR of RPS19 found in a subset of patients with DBA. We introduced two insertions (c.-147_-146insGCCA, c.-147_-146insAGCC) and one deletion (c.-144_-141delTTTC) into the INTERMEDIATE construct ( Figure 3A) followed by a transient transfection into K562 and HEK293T cells. The three DBA associated 59UTR variants reduced the RPS19 levels by approximately 20-30% when compared to the w.t. INTERMEDIATE construct (figure 3B and C). A marked reduction in expression (32%) was observed for the TTTC deletion in both cell lines. The AGCC insertion showed a 30% reduced expression in HEK293 cells and a 26% reduction in K562 cells. The GCCA insertion was associated with a 20% reduction in expression in both cell lines ( figure 3B and C). The RPS19 mRNA levels were similar when comparing cells transfected with the rare variant constructs to cells transfected with the w.t. construct.

Discussion
The 59-untranslated region (59UTR) of an mRNA is an important regulator of translation by influencing e.g. mRNA stability, sub-cellular localization and translational efficiency [32][33][34]. Furthermore, multiple transcriptional start sites and 59UTRs expressed from a single gene encoding one and the same protein may regulate gene expression through differential expression with respect to developmental stages, tissue type and in response to stimuli [20,35,36]. One element that enables fast up-or downregulation of ribosomal proteins in response to nutrient supply is the 59TOP sequence, a stretch of 4 to 14 pyrimidines following a Cytidine as the first nucleotide located at the 59end of an mRNA [37]. This 59TOP sequence is contained in the SHORT 59UTR variant used for expression analysis in our study and possibly responsible for fast adaptation of RPS19 levels. The heterogeneous 59UTRs of mRNAs transcribed from a single gene arise from the use of alternate transcriptional initiation sites and differential RNA processing [38]. It has been estimated that 10-18% of genes express alternate 59UTRs by multiple promoter usage. Alternate untranslated regions determine tissue specific function and their inappropriate expression can contribute to the development of abnormal phenotypes and disease [39].
We have characterized the RPS19 59UTR variants with respect to expression levels, tissue specificity, and translation efficiency. RPS19 is ubiquitously expressed and mutations in this gene are associated with DBA. The precise molecular mechanisms behind the disease remain unknown but we hypothesized that the expression of different RPS19 59UTR variants may contribute to the regulation of RPS19 protein levels and, ultimately, to DBA. We determined the extent of the RPS19 mRNA 59UTR by 59-RACE on poly(A)+ purified mRNA from testis and K562 cells. The results show an extensive variation in the transcriptional start sites with .30 different 59UTRs of which 29 are novel. The total amount of RPS19 mRNA varied considerably between tissues and we observed up to 10-fold differences. Interestingly, bone marrow shows relatively low level of total RPS19 transcripts when compared to several other primary tissues analyzed. We then investigated the distribution of 59UTR variants in different tissues. Transcripts were divided into three groups containing a 59UTR of at least 3 nt, 350 nt and 449 nt, respectively. Our data indicate clear differences in the distribution of RPS19 59UTRs when comparing different tissues. The amplicon corresponding to the shorter 59UTR (amplicon A) constituted between 83% to .99% of RPS19 transcripts but without correlation to the variation in total amounts of RPS19 mRNA.
To get a better insight into the translational regulation of RPS19 59UTR variants we investigated the RPS19 levels expressed from constructs with three distinct 59UTRs length. The SHORT 59UTR variant, spanning a 35 bp 59UTR, is translated four to ten fold more efficiently than the two longer variants with 59UTRs of 382 bp and 467 bp, respectively. This is also consistent with the analysis of variable 59UTR length of other genes [34,36]. A possible explanation is that the SHORT variant exhibits a less complex secondary structure, facilitating scanning by the transla- Figure 1. Transcriptional start sites and tissue expression of RPS19 variants. (A) RPS19 59UTR variants in testis and K562 cells. Schematic presentation of 39 different RPS19 59UTRs identified of which 29 are yet undescribed. 59RACE was performed with 1 mg of total RNA using the GeneRacerH kit (Invitrogen) according to manufacturer's recommendation. The RNA was treated with DNase I to clean samples from genomic DNA. The 59RACE protocol selected full length G-capped mRNA and ruled out the possibility of partially degraded mRNA. PCR products were cloned into a TOPO-TA vector (Invitrogen) and 122 clones were picked randomly (83 from testis, 39 from K562 cells) and analyzed by bidirectional sequencing. The 59UTR variants identified are indicated and aligned to the first exon of RPS19 from databases with a known maximum 59UTR of 382 nt (bottom). (B) A schematic picture of the 59 region of RPS19 cDNA (horizontal line) with relative positions of the start codon and the amplicons generated for quantification. Primers used to generate amplicons A, B and C for quantitative PCR are shown as arrows (sequences available upon request). (C) Tissue distribution of total RPS19 as determined by qPCR of amplicon A showing relative expression of RPS19 normalized to b-actin on a panel of primary human tissues and cell lines. Analyses were run in triplicates and the average is shown for each tissue. (D) Expression of the amplicons B and C representing longer variants of 59UTR as determined by qPCR and expressed as a percentage of total RPS19 expression determined by amplicon A shown in (C). doi:10.1371/journal.pone.0017672.g001 tion machinery. The reduced translation from transcripts with longer 59UTRs may be related to the more complex secondary structures, making the transcripts less accessible for translation. The functional significance of the longer RPS19 59UTRs is unclear, but may be of importance for 59TOP independent translation providing baseline amounts of RPS19. The shorter variants could in this case be used for a fast adaptation of RPS19 levels in response to cellular needs. In combination, our observations suggest a large variation in RPS19 mRNA levels between tissues as well as in TSS used. The predominant and shorter 59UTRs are more efficiently translated and may directly reflect the levels of RPS19.
We then analyzed the effect on translation of rare sequence variants in the RPS19 59UTR found in subsets of DBA patients. We confirmed that the polymorphic 59UTR variants are indeed transcribed and we hypothesized that these transcripts affect translational efficiency. We therefore investigated the RPS19 levels expressed from constructs with each of the two insertions (c.-147_-146insGCCA, c.-147_-146insAGCC) or the deletion (c.-144_-141delTTTC), respectively. All three variants result in significantly  reduced RPS19 levels when compared to the corresponding wild type sequence. A possible explanation is that the mutation causes the mRNA to adopt a more complex secondary structure that represses translation. It is noteworthy that the observed reduction in RPS19 levels in our cell-systems is related to a relatively large proportion of the ''INTERMEDIATE'' RPS19 transcripts containing each of the specific 59UTR variants. Still, the observed effects of the three 59UTR variants associated with DBA do not result in haploinsufficiency and, accordingly, the impact on RPS19 levels in vivo would depend on the relative amounts of longer RPS19 mRNAs. Although these 59UTR variants may lead to suboptimal conditions for growth and differentiation of tissues sensitive to reduced RPS19 levels it is likely that additional factors are required for overt clinical forms of DBA. Our results are consistent with the increased ratio of 21S/18S pre-rRNAs associated with the c.-147_-146insGCCA variant observed previously [14]. Failure to detect reduced RPS19 levels in that study may be due to low abundance of longer transcripts and/or minor changes in RPS19 levels in the cells analyzed.
Our combined findings suggest complex regulatory mechanisms of RPS19. RPS19 uses a broad range of TSS with tissue specific differences and shorter 59UTRs are more efficiently translated. We also show that DBA associated 59UTR variants of RPS19 are less efficiently translated. Further investigations are now required to understand how RPS19 is regulated in different tissues both at the transcriptional and the translational level. These studies may clarify the distribution and levels of RPS19 59UTR variants as well as RPS19 protein levels at different stages of erythropoiesis. Thus, analysis of the RPS19 TSS used in erythroid precursor cells may provide valuable information in search for molecular mechanisms behind DBA.