5′ tRNA halves are present as abundant complexes in serum, concentrated in blood cells, and modulated by aging and calorie restriction
© Dhahbi et al.; licensee BioMed Central Ltd. 2013
Received: 6 February 2013
Accepted: 26 April 2013
Published: 2 May 2013
Small RNAs complex with proteins to mediate a variety of functions in animals and plants. Some small RNAs, particularly miRNAs, circulate in mammalian blood and may carry out a signaling function by entering target cells and modulating gene expression. The subject of this study is a set of circulating 30–33 nt RNAs that are processed derivatives of the 5′ ends of a small subset of tRNA genes, and closely resemble cellular tRNA derivatives (tRFs, tiRNAs, half-tRNAs, 5′ tRNA halves) previously shown to inhibit translation initiation in response to stress in cultured cells.
In sequencing small RNAs extracted from mouse serum, we identified abundant 5′ tRNA halves derived from a small subset of tRNAs, implying that they are produced by tRNA type-specific biogenesis and/or release. The 5′ tRNA halves are not in exosomes or microvesicles, but circulate as particles of 100–300 kDa. The size of these particles suggest that the 5′ tRNA halves are a component of a macromolecular complex; this is supported by the loss of 5′ tRNA halves from serum or plasma treated with EDTA, a chelating agent, but their retention in plasma anticoagulated with heparin or citrate. A survey of somatic tissues reveals that 5′ tRNA halves are concentrated within blood cells and hematopoietic tissues, but scant in other tissues, suggesting that they may be produced by blood cells. Serum levels of specific subtypes of 5′ tRNA halves change markedly with age, either up or down, and these changes can be prevented by calorie restriction.
We demonstrate that 5′ tRNA halves circulate in the blood in a stable form, most likely as part of a nucleoprotein complex, and their serum levels are subject to regulation by age and calorie restriction. They may be produced by blood cells, but their cellular targets are not yet known. The characteristics of these circulating molecules, and their known function in suppression of translation initiation, suggest that they are a novel form of signaling molecule.
Several classes of small RNAs have been found to mediate biological functions in animals and plants [1–5]. miRNAs, siRNAs, piRNAs, and others are bound by Argonaute proteins, and have the common property of directing protein complexes to nucleic acids with sequence complementarity, where they may cleave or otherwise alter the target . In both plants and animals, some small RNAs are able to travel between tissues within an organism, thus transferring their functions to other cells. In vertebrates, there has been much recent interest in the presence of specific miRNAs in the plasma and serum; there is some evidence that these can be taken up by cells and alter gene expression, and there is also interest in the possibility that they can be markers of specific disease states, including cancer [7–9].
There is also evidence for processing of non-coding RNAs into smaller RNAs, many with as yet poorly understood functions [10, 11]. Many of the non-coding RNAs that appear to undergo processing into smaller RNAs have well studied functions, although their smaller derivatives often do not. In particular, tRNA is processed into shorter forms termed tRNA fragments (tRFs) [12, 13]. The subject of this report is a tRNA fragment created by cleavage of tRNA near the anticodon loop to create a ″5′ tRNA half” (the term we will use here). Previous reports have described 5′ tRNA halves as intracellular molecules interacting with components of the translation initiation complex. 5′ tRNA halves have been shown to be induced by the ribonuclease angiogenin in response to stress in cultured cells, to promote assembly of stress granules carrying stalled preinitiation complexes, and to inhibit mRNA translation [14, 15]; little more is known about their function.
We have sequenced small RNAs present in mouse serum; when multiple reportable alignments of the sequencing reads to the mouse genome were allowed, we noted the presence of a class of tRNA-derived 30–33 nt fragments that closely resemble the 5′ tRNA halves previously described in stressed cell cultures. Investigation of these 5′ tRNA halves reveals a novel class of circulating small RNAs whose characteristics, including changes with age that are antagonized by calorie restriction, strongly suggest physiologic regulation and function.
Results and discussion
Sequencing and computational analysis of small RNAs circulating in mouse serum
Only if multiple reportable alignments are allowed during bowtie mapping does an unfamiliar second peak emerge at 30–33 nt (Figure 1A). The 30–33 nt peak persists when the bowtie alignment mode is changed from the Maq’s default policy (n option) to the end-to-end k-difference policy (v option), but again disappears when multiple reportable alignments are suppressed (Figure 1B). The same two-peak pattern was observed when the 9 individual sequenced serum small RNA samples were mapped to the mouse genome (Figure 1C). Dependence of the 30–33 nt peak on multiple reportable alignments indicates that the reads are encoded by repetitive DNA. Six percent of the 163,078,230 mapped reads, aligned to a group of RepeatMasker classes (DNA, LINE, LTR, Low complexity, RC, SINE, Satellite, and Simple repeat); these reads were mainly < 20 nt in size (Additional file 1: Figure S1) and were not considered for further analysis.
Characterization of circulating small RNAs derived from tRNAs
Total number and percentage of the different sizes of sequencing reads that map to tRNAs
Read size in nucleotides
Number of reads
Percentage of reads
It is unlikely that this result is a sequencing artifact: the full length of most tRNAs is 75–90 nt, and the sequencing runs used to generate these data were 50 cycles while the reads occupy a narrow size range of 30–33 nt. This pattern suggests that the tRNA reads were derived from processed fragments of full length tRNAs; the remainder of the tRNA was not significantly detected in the serum small RNA libraries. In support of this conclusion, tRNAs have been shown to undergo cleavage within anticodon loops to produce tRNA-derived stress-induced fragments (tiRNAs) when cultured cells are subjected to stresses such as arsenite, heat shock, or ultraviolet irradiation [17, 18]. Such cleavage of the anticodon loop does not seem to be part of a tRNA degradation process, because the generated 5′ tRNA fragments are stable in the cell. Our findings indicate that tRNA fragments highly similar to tiRNAs are present under normal (unstressed) conditions, and can remain stable even after they are released into the peripheral blood. 5′ but not 3′ tRNA fragments inhibit mRNA translation initiation in cultured cell lines .
Frequencies of circulating 5′ tRNA halves and the gene copy number of tRNAs from which the tRNA halves were derived
Gene copy number
% of circulating tRNA halves
This implies a tRNA type-specific biogenesis and/or release of the circulating 5′ tRNA halves.
Presence in circulating mouse blood of particles containing stable cell-free 5′ tRNA halves
We also probed RNA from mouse serum with a probe complementary to the 5′ end of tRNA-Asn-GTT to confirm the low abundance of circulating tRNA halves derived from tRNAs that were barely detected in the sequencing data. A 5-day exposure to X-ray film showed a very weak signal from tRNA-Asn-GTT probe compared to the strong signal from the tRNA-Gly-GCC probe obtained after a short (25 minute) exposure (Additional file 1: Figure S2). These results are consistent with the sequencing, and inconsistent with a sequencing bias. They imply a tRNA type-specific biogenesis and/or release of the circulating 5′ tRNA halves.
Because the tRNA halves we observe are stable in circulation but not encapsulated in exosomes, they are most likely complexed to carrying factors (e.g., proteins that protect them from degradation). To determine the size range of the putative complexes carrying the 5′ tRNA halves in the serum, we Northern blotted RNA extracted from concentrate or filtrate fractions after ultrafiltration of mouse serum samples through Vivaspin 2 columns with 30, 100, or 300 kDa MW cut-off. A probe for the 5′ end of tRNA-Gly-GCC detected a ~30 nt band in the concentrates of 30 and 100 kDa MW cut-off, and in the filtrate of 300 kDa MW cut-off (Figure 5G). Identical results were obtained for the 5′ end of tRNA-Val-CAC (Figure 6B).
Thus 5′ tRNA halves circulate as part of 100–300 kDa complexes, while the 5′ tRNA halves themselves are only ~10 kDa. This is reminiscent of reports that miRNAs can circulate in the bloodstream as components of RNA-protein/lipoprotein complexes. Stable argonaute2-miRNA complexes that are not part of microvesicles were recovered from plasma and serum, and high-density lipoprotein has been reported to carry and deliver miRNAs to recipient cells [20–22].
5′ tRNA halves are concentrated in hematopoietic and lymphoid tissues
More extensive studies will establish if 5′ tRNA halves are concentrated in particular blood cell types, although the very high levels in lymph nodes point to lymphocytes as one such type. The evidence does not establish whether the 5′ tRNA halves are concentrated in hematopoietic cells because they are produced there, or because they are preferentially taken up from the blood: neither the origin nor the destinations of the 5′ tRNA halves is certain. The low levels of 5′ tRNA halves present in non-hematopoietic tissues may indicate low levels in those tissues, but they may also be derived from residual blood cells in those tissues.
A chelating agent destabilizes circulating 5′ tRNA halves
Calorie restriction offsets age-associated changes in levels of specific circulating 5′ tRNA halves
Calorie restriction (CR) can delay, prevent, or reverse many age-associated changes in physiologic parameters. We used aging and CR as model physiologic states to explore the possibility that they are associated with changes in the levels of circulating 5′ tRNA halves. We performed pairwise comparisons between young and old control groups to measure the differential abundance in circulating 5′ tRNA halves associated with old age, and between old control and old CR groups to determine whether CR has an effect on any age-associated changes.
Age-associated changes in the levels of mouse circulating 5′ tRNA halves and the effects of CR on the age-associated changes
Young control †
Deep sequencing of small RNAs extracted from mouse serum identifies a population of tRNA-derived molecules, termed 5′ tRNA halves, previously described only as stress-induced inhibitors of translation initiation in cultured cells. 5′ tRNA halves are more abundant than miRNAs in mouse serum, and are derived from distinct subset of tRNAs by cleavage near the anticodon loop; the 3′ portion of the tRNA molecule is present in serum only in trace quantities. Ultracentrifugation and size fractionation establish that the 5′ tRNA halves circulate as part of a larger complex, but are not contained in exosomes or microvesicles; their sensitivity to the chelating agent EDTA provides further evidence that they exist as circulating nucleoprotein complexes. They are concentrated in hematopoietic and lymphoid tissues, and present in other tissues at very low levels that may reflect residual blood cells. The origin of the serum particles, and their destinations, are uncertain; however their concentration in blood cells suggest that they may be produced by these cells. Levels of serum 5′ tRNA halves are distinctly changed in aged mice, and calorie restriction inhibits these changes, indicating that they are subject to physiologic regulation. Taken together with the extant evidence that 5′ tRNA halves can regulate mRNA translation, the characteristics of the circulating 5′ tRNA halves we have discovered suggest that they function as signaling molecules with as yet unknown physiologic roles.
To date, the only known function of 5′ tRNA halves is inhibition of translation in cultured cells subjected to a variety of stressors; transfection of 5′ tRNA halves inhibits global translation in U2OS cells [14, 18]. A study published while this paper was in preparation reported induction of 5′ tRNA halves in human airway epithelial cells upon infection with respiratory syncytial virus (RSV). Induction involves cleavage at the tRNA anticodon loop by angiogenin, and at least one type, the 5′ tRNA-Glu-CTC half, promotes RSV replication . Our findings indicate that 5′ tRNA halves function on an organismal rather than merely a cellular level. Furthermore they are likely to function in a context much broader than cellular stress or infection: we find 5′ tRNA halves in unstressed conditions. Changes in their expression (either increased or decreased) with age are also consistent with a broader physiologic role, and it is particularly interesting that these changes are partially mitigated by calorie restriction.
The most extensively studied cellular tRNA halves are generated under stress conditions by angiogenin, which cleaves mature tRNAs within the anticodon loops . The stress-induced tRNA halves target the translation initiation machinery to reprogram protein translation in order to promote cell survival during stress [14, 26]. Pull-down and mass spectrometry analyses of RNA-protein complexes have identified several cellular proteins (YB-1, FXR-1, and PABP1) bound to intracellular 5′ tRNA halves . The nature of the proteins and/or other factors that bind and stabilize the extracellular form of 5′ tRNAs halves has yet to be elucidated. Understanding of the origin, composition, and destinations of these complexes will provide insights into their role in organismal physiology.
Serum collection, RNA isolation, and small RNA library construction
Male mice of the long-lived B6C3F1 strain were fed either control or calorie-restricted (CR) diet (∼40% fewer calories than the control). Three mice were studied from each of three groups: young (7-month) and old (27-month) mice fed the control diet, and old (27-month) mice fed the CR diet. Total RNA including small RNA was isolated from each serum sample with miRNeasy kit (Qiagen) and used to construct indexed sequencing libraries with the Illumina TruSeq Small RNA Sample Prep Kit. The libraries were pooled and sequenced on an Illumina HiSeq 2000 instrument to generate 50 base reads. Further details about the mice and diets, and library construction are provided in SI Methods and Figures.
Mapping and annotation of sequencing reads
Sequencing reads were pre-processed with FASTX-Toolkit (hannonlab.cshl.edu) to trim the adaptor sequences, and discard low quality reads. The obtained clean reads were mapped to the mouse reference genome (GRCm38/mm10) with bowtie version 0.12.8  using different combinations of alignment and reporting options. We used the option ”-n 0 -l 14“ to align the sequencing reads according to a policy similar to Maq‘s default policy and requiring no mismatches in the first 14 bases (the high-quality end of the read). In addition, this mode of alignment was combined with options that define which and how many alignments should be reported; the option “-k 1 --best” instructed bowtie to report only the best alignment if more than one valid alignment exists, while the option “-m 1” instructed bowtie to refrain from reporting any alignments for reads having multiple reportable alignments. The “-k 1 --best” and “-m 1” modes of alignment reporting were also used in combination with the end-to-end k-difference (−v) alignment mode. Varying the alignment and reporting modes allowed the differential detection of two predominant peak sizes of sequencing reads as described in the results section.
Annotation analysis of the mapped sequencing reads was performed with bedtools  using the following databases: the Genomic tRNA Database  (gtrnadb.ucsc.edu), miRBase 18 (mirbase.org), and rRNA, snRNA, scRNA, and srpRNA which were extracted from the RepeatMasker track (genome.ucsc.edu; mm10).
Analysis of differentially abundant circulating tRNA halves
The bowtie alignment files generated above from the young and old control and old CR serum sequencing samples were analyzed with bedtools  to obtain the coverage of the tRNA genes included in the Genomic tRNA Database  (gtrnadb.ucsc.edu), and to determine the read count for each tRNA in the database. The tRNA read counts were further analyzed with the Bioconductor package edgeR  to detect the changes in the levels of circulating 5′ tRNA halves in the different experimental groups. The algorithm of edgeR fits a negative binomial model to the count data, estimates dispersion, and measures differences using the generalized linear model likelihood ratio test which is recommended for experiments with multiple factors, such as the simultaneous analysis of age and diet in our study. The fitted count data was analyzed by performing pairwise comparisons between the different experimental groups: young and old control groups were compared to measure the differential abundance in circulating 5′ tRNA halves associated with old age; old control and old CR groups were compared to determine whether CR has an effect on any age-associated changes. The results were further filtered to keep only 5′ tRNA halves that achieved a minimum of 500 counts per million (cpm) in at least one of the 3 experimental groups.
Northern blot assays
RNAs analyzed with Northern blots were extracted from normal or sodium arsenite-treated U2OS and from a variety of tissues and sera harvested from one-year-old mice fed control diet. Before RNA extraction, some serum samples were centrifuged at 110000 g for 2 hrs, and supernatant and pellet fractions were separated, or were separated into concentrate and filtrate fractions by ultrafiltration through Vivaspin 2 columns with 30, 100, or 300 kDa MW cut-off. RNAs were separated on 15% denaturing polyacrylamide gels, transferred and fixed to a membrane by chemical cross-linking , and hybridized with probes complementary to 5’ and 3′ ends of tRNAs. Further details about Northern blot assays, and probe sequences are provided in SI Methods and Figures.
We thank Noel Guerrero for his help.
- Okamura K: Diversity of animal small RNA pathways and their biological utility. Wiley Interdiscip Rev RNA. 2012, 3 (3): 351-368. 10.1002/wrna.113.View ArticlePubMedGoogle Scholar
- Wery M, Kwapisz M, Morillon A: Noncoding RNAs in gene regulation. Wiley Interdiscip Rev Syst Biol Med. 2011, 3 (6): 728-738. 10.1002/wsbm.148.View ArticlePubMedGoogle Scholar
- Zhang C: Novel functions for small RNA molecules. Curr Opin Mol Ther. 2009, 11 (6): 641-651.PubMed CentralPubMedGoogle Scholar
- Zhang S, Sun L, Kragler F: The phloem-delivered RNA pool contains small noncoding RNAs and interferes with translation. Plant Physiol. 2009, 150 (1): 378-387. 10.1104/pp.108.134767.PubMed CentralView ArticlePubMedGoogle Scholar
- Esteller M: Non-coding RNAs in human disease. Nat Rev Genet. 2011, 12 (12): 861-874. 10.1038/nrg3074.View ArticlePubMedGoogle Scholar
- Joshua-Tor L, Hannon GJ: Ancestral roles of small RNAs: an Ago-centric perspective. Cold Spring Harb Perspect Biol. 2011, 3 (10): a003772-10.1101/cshperspect.a003772.PubMed CentralView ArticlePubMedGoogle Scholar
- Allegra A, Alonci A, Campo S, Penna G, Petrungaro A, Gerace D, Musolino C: Circulating microRNAs: New biomarkers in diagnosis, prognosis and treatment of cancer (Review). Int J Oncol. 2012, 41 (6): 1897-1912.PubMedGoogle Scholar
- Etheridge A, Lee I, Hood L, Galas D, Wang K: Extracellular microRNA: a new source of biomarkers. Mutat Res. 2011, 717 (1–2): 85-90.PubMed CentralView ArticlePubMedGoogle Scholar
- Zen K, Zhang CY: Circulating microRNAs: a novel class of biomarkers to diagnose and monitor human cancers. Medicinal research reviews. 2012, 32 (2): 326-348. 10.1002/med.20215.View ArticlePubMedGoogle Scholar
- Rother S, Meister G: Small RNAs derived from longer non-coding RNAs. Biochimie. 2011, 93 (11): 1905-1915. 10.1016/j.biochi.2011.07.032.View ArticlePubMedGoogle Scholar
- Tuck AC, Tollervey D: RNA in pieces. Trends Genetics: TIG. 2011, 27 (10): 422-432. 10.1016/j.tig.2011.06.001.View ArticlePubMedGoogle Scholar
- Sobala A, Hutvagner G: Transfer RNA-derived fragments: origins, processing, and functions. Wiley Interdiscip Rev RNA. 2011, 2 (6): 853-862. 10.1002/wrna.96.View ArticlePubMedGoogle Scholar
- Lee YS, Shibata Y, Malhotra A, Dutta A: A novel class of small RNAs: tRNA-derived RNA fragments (tRFs). Genes Dev. 2009, 23 (22): 2639-2649. 10.1101/gad.1837609.PubMed CentralView ArticlePubMedGoogle Scholar
- Ivanov P, Emara MM, Villen J, Gygi SP, Anderson P: Angiogenin-induced tRNA fragments inhibit translation initiation. Mol Cell. 2011, 43 (4): 613-623. 10.1016/j.molcel.2011.06.022.PubMed CentralView ArticlePubMedGoogle Scholar
- Saikia M, Krokowski D, Guan BJ, Ivanov P, Parisien M, Hu GF, Anderson P, Pan T, Hatzoglou M: Genome-wide identification and quantitative analysis of cleaved tRNA fragments induced by cellular stress. J Biol Chem. 2012, 287 (51): 42708-42725. 10.1074/jbc.M112.371799.PubMed CentralView ArticlePubMedGoogle Scholar
- Langmead B, Trapnell C, Pop M, Salzberg SL: Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 2009, 10 (3): R25-10.1186/gb-2009-10-3-r25.PubMed CentralView ArticlePubMedGoogle Scholar
- Thompson DM, Lu C, Green PJ, Parker R: tRNA cleavage is a conserved response to oxidative stress in eukaryotes. RNA. 2008, 14 (10): 2095-2103. 10.1261/rna.1232808.PubMed CentralView ArticlePubMedGoogle Scholar
- Yamasaki S, Ivanov P, Hu GF, Anderson P: Angiogenin cleaves tRNA and promotes stress-induced translational repression. J Cell Biol. 2009, 185 (1): 35-42. 10.1083/jcb.200811106.PubMed CentralView ArticlePubMedGoogle Scholar
- Chan PP, Lowe TM: GRNA: a database of transfer RNA genes detected in genomic sequence. Nucleic Acids Res. 2009, 37: 93-97.View ArticleGoogle Scholar
- Arroyo JD, Chevillet JR, Kroh EM, Ruf IK, Pritchard CC, Gibson DF, Mitchell PS, Bennett CF, Pogosova-Agadjanyan EL, Stirewalt DL, et al: Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in human plasma. Proc Natl Acad Sci U S A. 2011, 108 (12): 5003-5008. 10.1073/pnas.1019055108.PubMed CentralView ArticlePubMedGoogle Scholar
- Turchinovich A, Burwinkel B: Distinct AGO1 and AGO2 associated miRNA profiles in human cells and blood plasma. RNA Biol. 2012, 9: 8-View ArticleGoogle Scholar
- Vickers KC, Palmisano BT, Shoucri BM, Shamburek RD, Remaley AT: MicroRNAs are transported in plasma and delivered to recipient cells by high-density lipoproteins. Nat Cell Biol. 2011, 13 (4): 423-433. 10.1038/ncb2210.PubMed CentralView ArticlePubMedGoogle Scholar
- Fu H, Feng J, Liu Q, Sun F, Tie Y, Zhu J, Xing R, Sun Z, Zheng X: Stress induces tRNA cleavage by angiogenin in mammalian cells. FEBS Lett. 2009, 583 (2): 437-442. 10.1016/j.febslet.2008.12.043.View ArticlePubMedGoogle Scholar
- Spindler SR, Dhahbi JM: Conserved and tissue-specific genic and physiologic responses to caloric restriction and altered IGFI signaling in mitotic and postmitotic tissues. Annu Rev Nutr. 2007, 27: 193-217. 10.1146/annurev.nutr.27.061406.093743.View ArticlePubMedGoogle Scholar
- Wang Q, Lee I, Ren J, Ajay SS, Lee YS, Bao X: Identification and Functional Characterization of tRNA-derived RNA Fragments (tRFs) in Respiratory Syncytial Virus Infection. Mol Ther. 2013, 21 (2): 368-379. 10.1038/mt.2012.237.PubMed CentralView ArticlePubMedGoogle Scholar
- Li S, Hu GF: Emerging role of angiogenin in stress response and cell survival under adverse conditions. J Cell Physiol. 2012, 227 (7): 2822-2826. 10.1002/jcp.23051.PubMed CentralView ArticlePubMedGoogle Scholar
- Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics. 2010, 26 (6): 841-842. 10.1093/bioinformatics/btq033.PubMed CentralView ArticlePubMedGoogle Scholar
- Robinson MD, McCarthy DJ, Smyth GK: edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010, 26 (1): 139-140. 10.1093/bioinformatics/btp616.PubMed CentralView ArticlePubMedGoogle Scholar
- Pall GS, Hamilton AJ: Improved northern blot method for enhanced detection of small RNA. Nature protocols. 2008, 3 (6): 1077-1084. 10.1038/nprot.2008.67.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.