tRNA-derived RNAs: biogenesis and roles in translational control

Abstract

Transfer RNA (tRNA)-derived RNAs (tDRs) are a class of small non-coding RNAs that play important roles in different aspects of gene expression. These ubiquitous and heterogenous RNAs, which vary across different species and cell types, are proposed to regulate various biological processes. In this review, we will discuss aspects of their biogenesis, and specifically, their contribution into translational control. We will summarize diverse roles of tDRs and the molecular mechanisms underlying their functions in the regulation of protein synthesis and their impact on related events such as stress-induced translational reprogramming.

Graphical Abstract

1. Introduction

Many thousands of various non-protein coding (nc) RNAs exist. Two classes of such ncRNAs, transfer and ribosomal RNAs (tRNAs and rRNAs respectively), are well studied in the context of protein synthesis. Extensive research within last few decades uncovered a large number of small ncRNAs, which are engaged in numerous molecular processes and contribute to specific biological functions. Surprisingly, some of these small RNAs originate from tRNAs. They come in different lengths and can represent various regions of tRNAs, in turn suggesting complex biogenesis. Although initial studies utilizing high throughput RNA sequencing techniques hinted at a “junk” nature of their production as random tRNA degradation and/or aberrant transcription products, further studies challenged these views. More recent studies suggest that tRNA-derived RNAs (tDRs) are products of non-random and highly regulated RNA cleavage (Holmes et al., 2023), although it should be noted that other names besides tDRs are used in the literature. Current knowledge of their biogenesis as well as the molecular mechanisms of the regulation of tRNA cleavage will be discussed in detail in this review.

Moreover, we will discuss here details of tDR participation in various aspects of protein synthesis. The number of contributions tDRs add to translational control through optimization of protein synthesis in wide range of biological situations steadily continues to grow. Likewise, we continue to uncover further details and molecular mechanisms of tDRs-mediated modulation of mRNA translation and their interactions with translational machinery and RNA targets. Discussion of such mechanisms, structure-functional interactions and their consequences for translational control is also a focus of this review.

2. Overview of tRNA biogenesis

As tDRs can be generated from precursor tRNAs (pre-tRNAs) as well as mature tRNAs, it is important to understand how pre-tRNAs are processed before maturation. Therefore, we will briefly describe the biogenesis pathway of tRNAs (Figure 1). In bacteria and archaea, all RNAs are transcribed by a single RNA polymerase, while eukaryotes have three RNA polymerases. Among them, RNA polymerase III (Pol III) is responsible for the transcription of tRNAs (reviewed in (Arimbasseri & Maraia, 2016; Willis & Moir, 2018)). tRNA transcription begins with the recruitment of Pol III to transcription start sites by transcription factors such as TFIIIB-1 (Arimbasseri & Maraia, 2016). When transcribed by pol III, all pre-tRNAs have redundant sequences called 5’-leader and 3’-trailer, on their 5’- and 3’-termini, respectively (reviewed in (Hopper & Phizicky, 2003)). These sequences must be removed during maturation. In addition, some tRNA genes, such as tRNA^Arg-TCT, tRNA^Leu-CAA, tRNA^Ile-TAT and tRNA^Tyr-GTA in humans have introns (Chan & Lowe, 2016) that are removed by tRNA-specific splicing machinery (reviewed in (Yoshihisa, 2014)). pre-tRNA splicing begins with endonucleolytic cleavage of the exon-intron junctions by the tRNA splicing endonuclease (TSEN) complex (Paushkin et al., 2004). Intron-removed pre-tRNA exons are then ligated by RtcB (Popow et al., 2011; Popow et al., 2014), an RNA ligase that ligates a 2’,3’-cyclic phosphate residue to a 5’-hydroxyl residue (Chakravarty et al., 2012; Tanaka & Shuman, 2011). While the bacterial RtcB is a stand-alone enzyme, its human homolog RTCB works as part of the ligase complex comprising of RTCB, DDX1, CGI-99 and FAM98B (Kroupova et al., 2021). RTCB-ligase complex is also responsible for the ligation of XBP1 mRNA exons as a part of the endoplasmic reticulum (ER) stress response (Jurkin et al., 2014; Kosmaczewski et al., 2014; Lu et al., 2014). Like TSEN complex-mediated cleavage, both 5’-leader and 3’-trailer are also cleaved by endonucleases RNase P (reviewed in (Esakova & Krasilnikov, 2010)) and RNase Z (reviewed in (Rammelt & Rossmanith, 2016)), respectively. Finally, in humans, a CCA trinucleotide, the common sequence at the 3’-termini of all tRNAs, is post-transcriptionally added by the CCA adding enzyme (reviewed in (Rammelt & Rossmanith, 2016)) tRNA nucleotidyl transferase 1 (TRNT1) (Nagaike et al., 2001). During the complicated process of maturation described above, a variety of post-transcriptional modifications also occur at several positions of tRNAs (Boccaletto et al., 2022), which aid in the stabilization of tertiary structure and the regulation of codon-anticodon recognition (reviewed in (Lorenz et al., 2017)).

Figure 1.

The tRNA biogenesis pathway in human cells. Precursor tRNAs are transcribed by RNA polymerase III with the aid of some pol III-specific transcription factors such as TFIIIB-1. Redundant sequences, 5’-leaders, 3’-trailers and introns must be removed during maturation. After end-processing (5’-leader and 3’-trailer removal) and splicing (intron removal followed by exon-exon ligation), the CCA sequence is added to the 3’-termini to create functional mature tRNAs.

3. Biogenesis of tDRs

3.1. Biogenesis of mature tRNA-originated tDRs

3.1.1. Generation of tRNA halves is a largely stress-induced phenomenon.

A variety of tDRs are generated from mature tRNAs (Figure 2A). One of the most well-studied subclasses of tDRs can be defined as tRNA halves. tRNA halves are generated by an RNase-mediated cleavage of mature tRNAs in the anticodon loops, and consequently consist of either the 5’-side or 3’-side of mature tRNAs. Because tRNA cleavage in the anticodon loops mainly occurs in response to stress, stress-induced tRNA halves are also called tRNA-derived stress-induced RNAs (tiRNAs, and consequently 5’- or 3’-tiRNAs based on the tRNA half they harbor) (Yamasaki et al., 2009). As 3’-tiRNAs possess variable loops and CCA termini, they are generally longer than 5’-tiRNAs. 5’-tiRNAs and 3’-tiRNAs are approximately 30–35 nt and 40–50 nt in length, respectively (Yamasaki et al., 2009).

Figure 2.

The biogenesis and classification of tDRs. (A) The biogenesis of mature tRNA-originated tDRs. Representative cleavage sites and RNases responsible for the cleavage are shown. (B) The biogenesis of pre-tRNA-originated tDRs. tDRs generated during pre-tRNA processing are shown.

In bacteria, tRNA cleavage in the anticodon-loop is a part of a host defense response. In E. coli, some strains carry a plasmid encoding a variety of secreted proteins, called colicins, that kill competitor strains (reviewed in (Cascales et al., 2007)). Among colicin proteins, Colicin D and E5 were identified as secreted RNases able to enter competitor cells and target their tRNAs (Ogawa et al., 2006; Ogawa et al., 1999; Tomita et al., 2000). Colicin D specifically and potently cleaves tRNA^Arg (Tomita et al., 2000), while Colicin E5 targets tRNA^Tyr, tRNA^His, tRNA^Asn and tRNA^Asp (Ogawa et al., 1999). PrrC, another E. coli-specific nuclease, is activated during T4 bacteriophage infection, specifically cleaving tRNA^Lys in infected cells (reviewed in (Kaufmann, 2000)). This tRNA cleavage results in an altruistic suicide of infected E. coli cells to prevent further spread of bacteriophage infection within their population.

tRNA cleavage in the anticodon is also widely conserved in eukaryotes. For example, oxidative stress (such as H₂O₂ treatment) induces tiRNA production similarly in yeast, plants, and mammalian cells (Thompson et al., 2008). However, the RNases responsible for the cleavage are different depending on species. In the yeast Saccharomyces cerevisiae, Rny1, a member of RNase T2 family, is responsible for stress-induced tRNA cleavage (Thompson & Parker, 2009), while Angiogenin (ANG), a vertebrate-specific RNase belonging to RNase A superfamily, is mainly responsible for stress-induced tRNA cleavage in vertebrates (Fu et al., 2009; Yamasaki et al., 2009). Other RNase A superfamily enzymes, especially RNase 1, are also suggested to cleave tRNAs into tiRNAs/tiRNA-like molecules (Akiyama et al., 2022a; Li et al., 2022). It is worth noting that RNase A superfamily enzymes, including ANG, are secreted RNases. Once produced, they are secreted into body fluids, contributing to host-defense immunity through a variety of antimicrobial and immune-modulation effects (reviewed in (Koczera et al., 2016; Lu et al., 2018)). Only a small proportion of them are taken up into the cell, cleaving self-RNAs in response to stress stimuli.

Although receptors for the uptake of extracellular RNases remain largely unidentified, Syndecan-4 (Skorupa et al., 2012) and Plexin-B2 (Yu et al., 2017) have been identified as receptors responsible for ANG uptake. Under non-stress conditions, the cytoplasmic pool of ANG is held inactive by an endogenous RNase inhibitor RNH1, while the nuclear pool of ANG is suggested to exist as a free form (Pizzo et al., 2013). Once cells are exposed to stress, cytoplasmic ANG is activated through the dissociation of RNH1, cleaving mature tRNAs into tiRNAs (reviewed in (Lyons et al., 2017a)). In addition, nuclear ANG translocates to the cytoplasm (Pizzo et al., 2013), which also promotes tiRNA production. Although the enzymatic activity of ANG toward various denatured RNAs is much lower than that of RNase A in vitro (Lee & Vallee, 1989), ANG efficiently and specifically cleaves tRNAs in the anticodons under conditions where RNAs are physiologically folded (Akiyama et al., 2021), suggesting that ANG is specialized to cleave physiologically folded tRNAs in the cell. There is accumulating evidence that tiRNAs act as cell protective molecules in cellular stress response. The biological functions of tiRNAs as stress responsive molecules are described later.

tDRs can be also generated in response to viral infection. When infected by viruses, cells sense double-stranded RNAs (dsRNAs), a hallmark of viral infection, thus activating RNase L as an antiviral response (reviewed in (Sadler & Williams, 2008)). RNase L selectively cleaves specific tRNAs such as tRNA^His and tRNA^Pro at their anticodon loops, generating tiRNA-like tDRs (Donovan et al., 2017). RNase L-mediated tDR production is accompanied by global translation repression, which is considered as an adaptive protective mechanism against viral infection. Schlafen family proteins (SLFNs) are induced by interferon, contributing to tumor suppression, immune responses, and restriction of viral replication (reviewed in (Mavrommatis et al., 2013)). Some of these functions are suggested to be dependent on tRNA cleavage. SLFN11 cleaves specific tRNAs such as tRNA^Ser and tRNA^Leu, resulting in translational suppression of specific mRNAs that are enriched with codons corresponding to the cleaved tRNAs (Li et al., 2018). One such mRNA target is ATR mRNA which encodes a serine/threonine kinase responsible for DNA damage response. As this mRNA is enriched with TTA (Leu) codons, SLFN11-mediated tRNA cleavage sensitizes tumor cells to DNA-damaging agents through repressing the translation of ATR mRNA. A related RNase, SLFN13, was also reported to have endoribonucleolytic activity. It cleaves mature tRNAs at the acceptor stem, as well as rRNAs, resulting in translation repression (Yang et al., 2018). SLFN13 inhibits HIV replication, probably through its RNase activity, although the precise mechanism remains to be fully elucidated. It should be noted that in these studies, it has been speculated that the biological effects of tRNA cleavage are due to the loss of function of cleaved tRNAs. Whether the cleavage products have biological activity remains unclear.

3.1.2. Generation of smaller tDRs is mostly constitutive phenomenon.

Advances in high-throughput sequencing technology have revealed a variety of tDRs shorter than tRNA halves are also enriched in the cell (Figure 2A). These smaller tDRs are also called tRNA-derived small RNA fragments (tRFs) (reviewed in (Kumar et al., 2016)). These smaller tDRs, typically 18–25 nt in length, can be generated by cleavage in any of the tRNA loops (i.e., D-loops, anticodon loops or T-loops) (Kumar et al., 2014). Generally, smaller 5’-tDRs and 3’-tDRs contain 5’- and 3’-terminal sequences of mature tRNAs, respectively. Therefore, smaller 3’-tDRs possess CCA on their 3’-end like 3’-tiRNAs. In addition to 5’- and 3’-tDRs, an intermediate fragment, consisting of the sequence around the anticodon, has also been reported (Goodarzi et al., 2015). In contrast to tiRNA production, the biogenesis pathway of smaller tDRs remains largely unclear (reviewed in (Kumar et al., 2016)). This is partly because it is difficult to identify the factors of smaller tDR biogenesis, as these smaller tDRs are constitutively produced and stably expressed, irrespective of stress stimuli.

Like tiRNA production, ANG is partly involved in smaller tDR production. For example, although ANG knockout cells do not decrease the levels of smaller 3’-tDRs (Su et al., 2019), a specific subset of smaller 5’-tDRs 26–30 nt in length are likely generated from 5’-tRNA halves (Honda et al., 2017; Su et al., 2019). In the plant Arabidopsis, RNase T2 is responsible for not only tiRNA production but also smaller tDR production (Megel et al., 2019). These data suggest that ANG and other anticodon-targeting RNases play a role in producing such smaller tDRs. These findings also suggest that some specific tDRs still can be generated in response to stress. Another candidate RNase for smaller tDR biogenesis is Dicer, a cytoplasmic RNase III-type enzyme responsible for pre-miRNA processing. An early report suggested that Dicer was involved in the biogenesis of smaller tDRs (Cole et al., 2009). Another report showed that small 3’-tDRs derived from tRNA^Gly-GCC were produced in a Dicer-dependent manner (Maute et al., 2013). In addition, pre-tRNA^Ile can be processed by Dicer into a smaller tDR in the absence of the chaperone protein La/SSB (Hasler et al., 2016). However, knockout of Dicer did not affect the abundance of smaller tDRs (Kumar et al., 2014; Li et al., 2012), suggesting that Dicer does not play a major role in producing smaller tDRs. Another study also showed that in human cells, smaller 3’-tDRs are generated in a Dicer-independent manner (Kuscu et al., 2018). In addition, a recent study showed that Dicer-like proteins, DCL1–4, do not have a critical role in the biogenesis of smaller tDRs in Arabidopsis (Megel et al., 2019). Therefore, the current consensus is that Dicer is not the principal RNase responsible for global smaller tDR production, but it does seem to be involved in the biogenesis of specific smaller tDRs.

3.2. Biogenesis of pre-tRNA-originated tDRs

Biogenesis of pre-tRNA-originated tDRs is dependent on the processing of pre-tRNAs during their maturation, as expected (Figure 2B). During pre-tRNA processing, RNase Z-mediated pre-tRNA cleavage produces 3’-trailer fragments called tRF-1 (Lee et al., 2009). Similarly, TSEN complex-mediated cleavage results in intron-containing fragments. It is worth noting that intron fragments can be generated as circular RNAs, called tRNA intronic circular RNAs (tricRNAs) (reviewed in (Schmidt & Matera, 2020)). However, the existence of endogenous tricRNAs have so far been shown in limited species that have long (>100 nt) intron-containing tRNA genes such as the archaea Haloferax volcanii (Singh et al., 2004) and Drosophila melanogaster (Lu et al., 2015). In contrast, in humans, the length of introns is relatively short (12–24 nt) (Chan & Lowe, 2016). As it is difficult to detect such short tricRNAs by either small RNA sequencing or quantitative PCR-based methods, it is still unclear whether tricRNAs are physiologically generated from short introns in the cell. Although RNase P-mediated cleavage theoretically bears 5’-leader fragments, there are no reports identifying tDRs consisting of 5’-leader alone so far, which may be partly because 5’-leader sequences are generally shorter than 3’-trailer sequences (Hanada et al., 2013; Karaca et al., 2014; Lee et al., 2009).

In addition to 3’-trailer fragments and intron fragments, pre-tRNAs have been reported to generate a novel subset of tDRs, called 5’-leader-exon fragments (representing 5’-leader sequences followed by 5’-exon) and 3’-exon-trailer fragments (3’-exon followed by 3’-trailer sequences). It was originally reported that these tDRs are generated under oxidative stress conditions induced by mutation of CLP1 (Hanada et al., 2013; Karaca et al., 2014), one of the subunits of TSEN complex (Paushkin et al., 2004). Mutations in any subunit (TSEN2, TSEN15, TSEN34, TSEN 54 and CLP1) are linked to pontocerebellar hypoplasia (PCH), a rare neurodegenerative disease, suggesting a link between pre-tRNA splicing and neurodegenerative disorders (reviewed in (Hayne et al., 2022)). It has been recently reported that oxidative stress inhibits the enzymatic activity of the RTCB ligase complex (Asanovic et al., 2021), showing that these tDRs are generated due to the failure of RTCB-mediated exon-exon ligation during splicing. Recent studies have suggested that these tDRs have biological activities. pre-tRNA^Tyr-originated tDRs (both 5’-leader-exon and 3’-exon-trailer) have been reported to sensitize cells to oxidative stress-induced neural cell death (Hanada et al., 2013; Inoue et al., 2020; Schaffer et al., 2014).

4. Mechanisms of Translation Regulation by tDRs

There is accumulating evidence that tDRs play important roles in translational regulation (Figure 3). Here, we describe diverse mechanisms of tDR-mediated translational regulation by classifying them into two categories, selective and global translational regulation, which are also summarized in Table 1.

Figure 3.

Major molecular mechanisms of tDR-mediated translational regulation. (A) The miRNA-like mechanism. (B) Global translation repression by TOG-containing tiRNAs.

Table 1.

Examples of diverse mechanisms of tDR-mediated translational regulation.

Selectivity	Mechanism	Species	Type	tRNA gene	Target	Function	Refs.
Selective	miRNA-like	Human	smaller 3’-tDR	tRNALeu	JAG2 mRNA	Repress colorectal cancer growth by inactivating Notch signaling through targeting JAG2 mRNA.	Huang et al., 2017
		Human	smaller 3’-tDR	tRNAGly-GCC	RPA1 mRNA	Repress proliferation of lymphoma cells by targeting RPA1 mRNA.	Maute et al., 2013
		Human	5’-tiRNA	tRNAHis-GTG	LATS2 mRNA	Promote colorectal cancer progression by inhibiting LATS2 expression.	Tao et al., 2021
		Human	5’-tiRNA	tRNAVal-CAC	FZD3 mRNA	Repress breast cancer progression by inhibiting Wnt/β-catenin pathway through targeting FZD3 mRNA.	Mo et al., 2019
		Human	5’-tiRNA	tRNAGlu	APOER2 mRNA	Promote respiratory syncytial virus (RSV) replication by repressing APOER2 expression.	Deng et al., 2015
		Rhizobia (bacteria)	smaller 5’- and 3’-tDR	tRNAVal, tRNAGly, tRNAGln	host RHD3a, RHD3b and LRX5 mRNAs	Induce root nodule formation in legumens by targeting host mRNAs via hijacking host RNA-interference machinery.	Ren et al., 2019
	Others	Human	Intermediate	tRNAGlu, tRNAAsp, tRNAGly, tRNATyr	YB-1	Destabilization of pro-oncogenic mRNAs through YB-1 displacement.	Goodarzi et al., 2015
		Human	smaller 5’-tDR	tRNAGln	Multisynthetase complex (MSC)	Enhance ribosomal protein translation through interaction with MSC.	Keam et al., 2017
		Human	5’-tiRNA	tRNAGln-CTG	IGF2BP2	Destabilize c-Myc mRNA by sequestering IGF2BP2 from c-Myc mRNA.	Krishna et al., 2019
		Human	trailer	tRNASer-TGA	La/SSB	Inhibit viral replication by sequestering La/SSB from infected RNA viruses.	Cho et al., 2019
		Human	smaller 5’-tDR	tRNAAla, tRNACys, tRNAVal	eIF4G/PABPC1	Repress translation of mRNAs possessing pyrimidine-enriched sequences in their 5’-UTR by sequestering PABPC1 from its partner PAIP1.	Guzzi et al., 2018; Guzzi et al., 2022
		Human	smaller 3’-tDR	tRNAGlu	Nucleolin	Facilitate p53 translation by sequestering Nucleolin from p53 mRNA.	Falconi et al., 2019
		Human	smaller 3’-tDR	tRNALeu-CAG	RPS28 mRNA	Enhance ribosome biogenesis through interaction with RPS28 mRNA.	Kim et al., 2017; Kim et al., 2019
Global	Repression of translation initiation	Human	5’-tiRNA	tRNAAla, tRNACys	eIF4G/A	Repress cap-dependent translation initiation by displacing eIF4G/A from mRNAs.	Ivanov et al., 2011; Lyons et al., 2020
	Interacion with ribosome	H. volcanii (archaeon)	smaller 5’-tDR	tRNAVal	Ribosome	Repress translation by displacing mRNAs from translation initiation complex through interaction with small ribosomal subunit.	Gebetsberger et al., 2017
		T. brucei	3’-tiRNA	tRNAThr	Ribosome	Enhance translation by facilitating mRNA loading through interaction with Ribosome.	Fricker et al., 2019
		Human	5’-tiRNA	tRNAPro	Ribosome	Repress translation by causing ribosome stalling thorough interaction with Ribosome.	Gonskikh et al., 2020
		S. cerevisiae	5’- and 3’-tiRNAs	Group of tRNAs	Ribosome	Repress translation through interaction with Ribosome.	Bakowska-Zywicka et al., 2016
		Human	smaller 5’-tDR	tRNAGln	unknown	Repress translation via conserved “GG” dinucleotide in tRNAGln.	Sobala et al., 2013

4.1. Selective translation regulation

4.1.1. miRNA-like mechanism

MicroRNAs (miRNAs) are short (20–25 nt) non-coding RNAs responsible for the silencing of target mRNAs via complementary binding (reviewed in (Gebert & MacRae, 2019; Jonas & Izaurralde, 2015)). miRNA-induced gene silencing is mediated by the RNA-induced silencing complex (RISC), which consists of core Argonaute (AGO) proteins and additional factors. Once miRNAs are loaded onto AGO, the miRNAs guide the RISC to the complementary site of target mRNAs. AGO proteins further recruit their partners, such as GW182. Then gene silencing is exerted through a combination of 1) translation repression by inhibiting the initiation step of translation through dissociating eIF4As from target mRNAs (Fukao et al., 2014; Fukaya et al., 2014; Meijer et al., 2013), and 2) mRNA decay through deadenylation and decapping (reviewed in (Bartel, 2018) ).

As shorter tDRs have a similarity in length to microRNAs (miRNAs), it was hypothesized that smaller tDRs could be involved in gene silencing by a miRNA-like mechanism. Indeed, there is building evidence that smaller tDRs can regulate translation of target mRNAs in a miRNA-like manner (Figure 3A). Like mature miRNAs, smaller 5’- and 3’-tDRs can be loaded onto AGO proteins (AGO1, AGO3, and AGO4) (Kumar et al., 2014). Another report showed that 3’-trailer fragments, generated by RNase Z-mediated cleavage of pre-tRNAs, can be loaded onto AGO2 (Haussecker et al., 2010). In addition, smaller 3’-tDR-loaded RISC was associated with GW182, similar to canonical miRNA-loaded RISCs (Kuscu et al., 2018). A computational approach has suggested that AGO-loaded smaller tDRs mainly target either the coding sequence or 3’-untranslated region (3’-UTR) of protein-coding mRNAs similarly to AGO-loaded miRNAs (Guan et al., 2020). Interestingly, this study also predicted that AGO-loaded smaller tDRs can target miRNAs. As another study reported that AGO-loaded miRNAs can target tDRs (Helwak et al., 2013), it is speculated that miRNAs and smaller tDRs may mutually regulate their expression levels and/or functions.

Growing evidence suggests that tDR-mediated gene silencing has pivotal roles in cancer progression. In colorectal cancer, a smaller 3’-tDR originated from tRNA^Leu reduces tumor formation and metastasis through the inhibition of Notch signaling by targeting a Notch ligand, JAG2 (Huang et al., 2017). Smaller 3’-tDRs originating from tRNA^Gly-GCC repressed proliferation of lymphoma cells by targeting RPA1 gene, an essential gene involved in DNA metabolism including DNA replication and DNA damage response (Maute et al., 2013). It should be noted that tDRs involved in miRNA-like gene silencing are not limited to smaller tDRs. In colorectal cancer tissues, the expression of 5’-tiRNA^His is up-regulated in an ANG-dependent manner, and 5’-tiRNA^His promotes cancer progression by repressing the translation of a tumor suppressor gene LATS2 (Tao et al., 2021). In breast cancer cells, 5’-tiRNA^Val acts as a tumor suppressor through the inhibition of Wnt/β-Catenin signaling by targeting FZD3, a receptor of Wnt ligands (Mo et al., 2019). The expression profiles of 3’-trailer fragments, which are produced through pre-tRNA processing, are also disturbed in cancer cells (Balatti et al., 2017), implying a role of such tDRs in cancer pathology.

tDR-mediated translational regulation is widely observed across species. For example, AGO2-dependent smaller tDR-mediated gene silencing mechanisms are well conserved in Drosophila, likely contributing to energy starvation stress response (Luo et al., 2018). Rhizobia is the symbiotic bacteria in the roots of legumens. A study showed that rhizobial smaller tDRs induce nodule formation by targeting host mRNAs through hijacking the host AGO1 (Ren et al., 2019). Respiratory syncytial virus (RSV) infection induces the production of various tDRs in an ANG-dependent manner (Wang et al., 2013). One of the tDRs, a smaller 5’-tDR originated from tRNA^Glu-CTC, promotes RSV replication. It is proposed that this tDR acts in a miRNA-like manner to suppress the expression of host APOER2, the factor that has an inhibitory effect on RSV replication (Deng et al., 2015).

In addition to miRNA-like gene silencing, tDRs have been shown to interact with PIWI proteins, other gene-silencing proteins mainly responsible for repression of transposon activity in germ cells. Indeed, some PIWI-interacting RNAs (piRNAs) are derived from tRNAs in Bombyx (Honda et al., 2017). Smaller tDRs (mainly 5’-tDRs) have been shown to interact with Piwi-like protein 4 (PIWIL4) in breast cancer cells (Keam et al., 2014). It is worth noting that PIWIL4 is abundant in somatic cell-derived cancer cells, but not germ cells where piRNAs function. 3’-trailer fragments have also been shown to interact with Piwi-like protein 2 (PIWIL2) in chronic lymphocytic leukemia cells and lung cancer cells (Balatti et al., 2017; Pekarsky et al., 2016). A smaller 5’-tDR originating from tRNA^Glu has been shown to interact with PIWIL4, thus down-regulating CD1A transcription through binding to the CD1A promotor region in monocytes (Zhang et al., 2016).

4.1.2. Others: Interaction with RNA-binding proteins and more

tDRs can regulate the translation of selective mRNAs independently of miRNA-like mechanism, mainly through interactions with RNA-binding proteins. For example, hypoxia-induced intermediate tDRs repress the translation of pro-metastatic mRNAs through sequestering YB-1 (YBX1), an RNA-binding protein involved in RNA stabilization, from the mRNAs, resulting in the inhibition of metastatic properties of cancers (Goodarzi et al., 2015). A smaller 5’-tDR originating from tRNA^Gln enhances translation of ribosomal and poly(A)-binding proteins through interactions with the multisynthetase complex (MSC), a complex comprising aminoacyl-tRNA synthetases (ARSs) and ARS-interacting proteins (Keam et al., 2017). 5’-tiRNA^Gln-CTG represses the translation of c-Myc mRNA by binding to and sequestering an RNA-binding protein IGF2BP2 from c-Myc mRNA, resulting in mRNA destabilization (Krishna et al., 2019). A tRNA^Glu-originated 3’-tDR represses the translation of p53 by displacing Nucleolin from p53 mRNA in breast cancer cells (Falconi et al., 2019). Trailer fragments, such as pre-tRNA^Ser-derived 19-nt fragment, have been shown to bind to La/SSB protein in the cytoplasm, inhibiting viral replication by sequestering of La/SSB (Cho et al., 2019). Smaller 5’-tDRs originating from tRNA^Ala, tRNA^Cys and tRNA^Val, repress translation through the interaction with a poly(A)-binding protein, PABPC1, which is involved in translation initiation. This inhibitory effect is dependent on pseudouridine modification at position 8 in the tDRs (Guzzi et al., 2018). Mechanistically, these tDRs sequester PABPC1 from its binding partner PAIP1, leading to translational repression of mRNAs that possess pyrimidine-enriched sequences in their 5’-UTR (Guzzi et al., 2022). On the other hand, the interaction between tDRs and their target mRNAs can regulate translation independently of miRNA-like mechanisms. A tRNA^Leu-originated smaller (22 nt) 3’-tDR has been reported to promote the translation of ribosomal protein S28 (RPS28) mRNA by stabilizing the secondary structure of the mRNA through a base-pairing interaction, resulting in the promotion of ribosome biogenesis (Kim et al., 2017; Kim et al., 2019).

4.2. Global translational regulation: eIF4F sequestration and ribosome binding

tDRs also regulate translation in a sequence-independent manner, which can cause either promotion or repression of global translation. Two major translational targets, the eIF4F complex and ribosomes, are targets of tDRs. The mechanisms involving interactions between tDRs and translation machinery are described below.

4.2.1. Stress-induced global translation repression targeting the eIF4F complex

As described above, various stress stimuli induce the production of tiRNAs. An early study reported that only tDRs (mainly tiRNAs) showed global translation inhibition, while full-length tRNAs and nucleotides derived from fully digested tRNAs had no effect on translation (Zhang et al., 2009), showing that tDRs are not just degradative products, but biologically functional molecules. Yamasaki et. al. examined the effect of endogenous tiRNAs on cellular translation by transfection of gel-purified tiRNAs. Interestingly, only 5’-tiRNAs showed the inhibitory effect on translation, while 3’-tiRNAs showed no effect on translation (Yamasaki et al., 2009). Further mechanistic studies revealed that specific 5’-tiRNAs originating from tRNA^Ala and tRNA^Cys show translation repression. This inhibitory effect is dependent on five consecutive guanosines at the 5’-termini, named 5’-terminal oligoguanine (TOG) motif (Ivanov et al., 2011). Through their TOG-motifs, these tiRNAs form RNA G-quadruplex (rG4) structures (Akiyama et al., 2020; Ivanov et al., 2014; Lyons et al., 2017b), stable non-canonical four-stranded RNA structures (reviewed in (Kharel et al., 2020)). Mechanistically, as illustrated in Figure 3B, rG4-tiRNAs displace eIF4F complex from the cap structure of mRNAs through the interaction with eIF4G, one of the subunits of the eIF4F complex, leading to repression of the scanning step of translation initiation (Lyons et al., 2020). Protein synthesis is strictly regulated especially under stress conditions, which contributes to energy conservation to promote recovery and to prioritize the translation of stress-responsive genes (reviewed in (Advani & Ivanov, 2019)). In addition to the TOG-tiRNA-mediated mechanism, two major stress responsive pathways, the integrated stress response (ISR) (reviewed in (Costa-Mattioli & Walter, 2020; Pakos-Zebrucka et al., 2016)) and the mTORC1 pathway (reviewed in (Liu & Sabatini, 2020; Saxton & Sabatini, 2017)) are involved in translational regulation. In response to various stress stimuli, both pathways also repress global translation by targeting multiple eukaryotic initiation factors (eIFs), emphasizing the importance of the initiation step of translation for stress-induced translational regulation.

TOG-containing tiRNAs also induce the assembly of stress granules (SGs) (Emara et al., 2010), cytoplasmic membraneless granules comprising RNA and proteins (reviewed in (Hofmann et al., 2021; Riggs et al., 2020)). Mechanistically, TOG-containing tiRNAs interacted with YB-1 through the cold shock domain, which is required for tiRNA-induced SG assembly (Lyons et al., 2016). Although the precise mechanisms have yet to be fully elucidated, SGs seem to contribute to translational reprogramming. For example, mRNAs encoding house-keeping genes, such as glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and β-actin, are incorporated into SGs (Stohr et al., 2006), resulting in translational repression of such genes under stress conditions. In contrast, mRNAs encoding stress-responsive genes such as HSP70 and HSP90, which are necessary for cellular stress response, are excluded from SGs (Kedersha & Anderson, 2002; Stohr et al., 2006), likely prioritizing the translation of such genes.

In addition to TOG-containing tiRNAs, some other tDRs have been suggested to be involved in global translational regulation. For example, 5’-tiRNA^Gly-GCC, 5’-tiRNA^Gly-CCC, and 3’-tiRNA^Pro have been shown to induce global translation repression, likely independently of eIF4F displacement (Ivanov et al., 2011). Smaller 5’-tDRs have also been shown to induce global translational repression independently of a miRNA-like mechanism. Although the precise mechanism has yet to be elucidated, a conserved “GG” dinucleotide sequence is required for this inhibitory effect (Sobala & Hutvagner, 2013).

4.2.2. Interaction with ribosomes

Ribosome-associated non-coding RNAs (rancRNAs) are an evolutionary conserved class of ncRNAs found in all three domains of life (Pircher et al., 2014). While most rancRNAs are able to associate with ribosomes, only some of them are shown to modulate ribosomal functions. Some of those active rancRNAs are tDRs, which affect translation efficiency through their interaction with ribosomes. While this interaction can either facilitate or repress translation depending on sequences and species, the mechanism of translational regulation through tDR-ribosome interaction seems conserved across species. In the halophilic archaeon Haloferax volcanii, a stress-induced smaller 5’-tDR derived from tRNA^Val binds to the small ribosomal subunit, which results in global translational attenuation through displacing mRNAs from the translation initiation complex (Gebetsberger et al., 2017; Gebetsberger et al., 2012). Similarly, in the yeast S. cerevisiae, all six synthetic tiRNAs tested (both 5’-halves and 3’-halves) inhibited translation through binding to ribosomes in vitro (Bakowska-Zywicka et al., 2016). Human 5’-tiRNA^Pro also binds to the ribosome in several human cell lines, which induces global translation inhibition by ribosome stalling accompanied by the formation of peptidyl-tRNA (Gonskikh et al., 2020). In contrast, nutrient deprived Trypanosoma brucei produce a tRNA^Thr-derived 3’-tiRNA that stimulates translation through the interaction with the ribosomes, aiding in the recovery of the cells once adverse conditions are removed (Fricker et al., 2019).

4.3. Impact of tRNA modifications on the function of tDRs

tRNAs are characterized by the presence of a variety of post-transcriptional modifications (Boccaletto et al., 2022; Suzuki, 2021). Several lines of evidence suggest that tRNA modifications can affect the biological function of tDRs. The translational inhibitory effect of purified endogenous 5’-tiRNA^Gly-GCC was stronger than synthetic 5’-tiRNA^Gly-GCC that does not contain modified nucleotides (Akiyama et al., 2020). Gene-silencing activity of smaller 3’-tDRs is attenuated by 1-methyladenosine (m¹A) modification by TRMT6/61A at position 58 through inhibiting annealing of 3’-tDRs to target mRNAs (Su et al., 2022). In contrast, the translational inhibitory effect of smaller 5’-tDRs originating from tRNA^Ala, tRNA^Cys and tRNA^Val was dependent on a pseudouridine (ψ) modification at position 8 (Guzzi et al., 2018). Additionally in mice, Dnmt2-mediated RNA methylations (m⁵C and m²G) are required for tDR-containing sperm-mediated transmission of metabolic phenotypes to offspring (Zhang et al., 2018). These data imply that experiments using synthetic tDRs do not necessarily reflect the biological functions of endogenous tDRs.

4.4. Functions of tDRs independent of translational regulation

Growing evidence suggests that tDRs can also function beyond translational regulation. Here, we describe some examples of diverse biological functions of tDRs independent of translational regulation.

Some tDRs (mainly 5’-tiRNAs) enriched in sperm contribute to the inheritance of metabolic disorders to offspring in high-fat diet fed mice (Chen et al., 2016). A tRNA^Gly-GCC-originated smaller tDR accumulates in mouse sperm in response to protein restriction, repressing genes associated with the endogenous retroelement MERVL (Sharma et al., 2016). This tDR has been shown to regulate the biogenesis of various small non-coding RNAs, resulting in the up-regulation of histone mRNAs through the regulation of U7 small nuclear RNA (snRNA), which is essential for 3’-UTR processing of histone pre-mRNAs (Boskovic et al., 2020). In mouse stem cells, smaller 3’-tDRs repress long terminal repeat (LTR)-retrotransposons by two mechanisms: blocking reverse transcription, and inducing RNA interference (Schorn et al., 2017). Smaller 3’-tDRs originating from tRNA^Pro have been shown to act as a primer for the reverse transcriptase of human T-cell leukemia virus (HTLV-1) (Ruggero et al., 2014). Finally, tiRNAs prevent apoptosis by inhibiting the interaction between cytochrome c and Apaf-1 through binding to cytochrome c (Saikia et al., 2014).

4.5. Translational regulation by loss-of-function of tRNAs

tRNA cleavage can induce translational repression independently of tDR function if the cleavage causes significant reduction of functional tRNAs. For example, anticodon cleavage in single-celled organisms such as bacteria and yeast, which is mainly observed as a part of host-defense responses, generally induces loss-of-function of target tRNAs (further described in section 5.2.3.). Oxidative stress has been reported to induce tRNA^Tyr depletion, and produce pre-tRNA-originated tDRs, impairing translation of mRNAs enriched in tyrosine codons (Huh et al., 2021). In addition, the cleavage of the 3’-CCA termini can cause a loss of function in tRNAs because the 3’-CCA ends are essential to charge corresponding amino acids. Trypanosoma brucei cleaves CCA-termini of its tRNAs as a stress response. In this organism, CCA-termini are cleaved in response to nutritional stress by a nuclease LCCR4, resulting in global translational repression. These 3’-trimmed tRNAs are rapidly repaired once stress is removed (Cristodero et al., 2021). Similarly, in vitro experiments suggested human ANG efficiently cleaves CCA ends before anticodon cleavage, likely inducing global translation inhibition (Czech et al., 2013). However, recent studies based on next generation sequencing revealed most of the 3’-tiRNAs had intact CCA termini (Akiyama et al., 2022a; Li et al., 2022; Su et al., 2019), implying that ANG predominantly targets anticodon loops compared to CCA termini. Therefore, at least in human cells, the impact of CCA-shortening on translation seems small.

5. Factors Modulating tDR Biogenesis

Many factors have been suggested to be involved with tDR biogenesis so far (Figure 4). As there are marked differences in biogenesis pathway between pre-tRNA-originated and mature tRNA-originated tDRs, the factors modulating tDR biogenesis also can be classified into the two groups.

Figure 4.

Factors that modulate tDR biogenesis. As tRNAs are transcribed by RNA polymerase III (pol III), the pol III activity determines the biogenesis of tDRs, especially pre-tRNA-originated tDRs. Various stress stimuli affect pol III activity through its upstream regulators, mTORC1 and MAF1. On the other hand, mature tRNA-originated tDRs are produced by endonucleolytic cleavage by RNases. Various factors affecting the efficiency of tRNA cleavage are shown.

5.1. Factors modulating the biogenesis of pre-tRNA-originated tDRs

As pre-tRNA-originated tDRs are generated during tRNA maturation, transcription activity of tRNA genes by RNA polymerase III (pol III) is theoretically the primary determinant of their biogenesis. Pol III activity is known to be regulated by several factors. First, pol III transcription activity is repressed by MAF1 (Pluta et al., 2001; Upadhya et al., 2002), a conserved negative regulator of Pol III. As MAF1 is negatively regulated by mTORC1 (Michels et al., 2010; Shor et al., 2010), mTORC1 positively regulates pol III-mediated transcription. Under adverse conditions such as nutrient deprivation, oxidative stress, and hypoxia, mTORC1 is inactivated through its phosphorylation by upstream stress-sensing molecules (reviewed in (Liu & Sabatini, 2020; Saxton & Sabatini, 2017)). Therefore, various stress stimuli generally repress pol III-mediated pre-tRNA transcription, likely decreasing the production of pre-tRNA-originated tDRs. It should be noted that the biogenesis of mature tRNA-originated tDRs, in contrast, is not directly affected by stress-induced pol III inhibition, as they are generated from the cytoplasmic pool of mature tRNAs.

On the other hand, stress conditions can facilitate the production of a subset of pre-tRNA-originated tDRs. As 5’-leader-exon and 3’-exon-trailer fragments are produced due to the failure in exon-exon ligation during pre-tRNA splicing (Asanovic et al., 2021; Hanada et al., 2013), oxidative stress-induced inhibition of the RTCB ligase complex is the most important determinant of the biogenesis of these tDRs.

5.2. Factors modulating the biogenesis of mature tRNA-originated tDRs

5.2.1. RNases

As mature tRNA-originated tDRs, especially tiRNAs, are produced by cleavage of mature tRNAs by various RNases, the intracellular RNase levels directly affect the biogenesis of tDRs. On the transcriptional level, ANG expression is up-regulated under stress conditions such as endoplasmic reticulum (ER) stress (Pereira et al., 2010) and hypoxia (Kishimoto et al., 2012). Because ANG is a secreted RNase enriched in body fluids such as the plasma and cerebrospinal fluids (reviewed in (Sheng & Xu, 2016)), the efficiency of ANG uptake and secretion is likely to be an important factor that determines the intracellular ANG levels. Although ANG is considered to be the main RNase responsible for stress-induced tiRNA production, tiRNAs can also be generated in ANG knockout cells (Akiyama et al., 2022a; Su et al., 2019). It has been recently suggested that RNase 1, another RNase A superfamily enzyme, is a main RNase responsible for ANG-independent stress-induced tiRNA production (Li et al., 2022). In addition, RNase L also generates tiRNA-like tDRs in response to viral infection (Donovan et al., 2017).

5.2.2. RNase inhibitor

Under non-stress conditions, intracellular RNase A superfamily enzymes are inactivated by an endogenous RNase inhibitor, RNH1 (reviewed in (Dickson et al., 2005; Sarangdhar & Allam, 2021)). In turn, tiRNA production is induced by various stimuli, such as heat shock, cold shock, UV irradiation, and viral infection (Fu et al., 2009; Wang et al., 2013; Yamasaki et al., 2009). In each of these conditions, tiRNA production is induced by RNH1 dissociation from ANG and other RNase A superfamily enzymes. RNH1 is a cytosolic, thiol residue-enriched protein. Under oxidative stress conditions, thiol residues in RNH1 are oxidized, resulting in the dissociation of RNH1 from RNases (Ferreras et al., 1995). In addition, oxidative stress induces RNH1 degradation (Blazquez et al., 1996; Moenner et al., 1998), likely contributing to the activation of RNases. However, it is still unclear whether oxidation of RNH1 is the principal mechanism of RNH1 dissociation. We have recently reported that there seem to be both oxidative stress-dependent and -independent mechanisms of RNH1 dissociation (Akiyama et al., 2022b). Menadione-induced RNH1 dissociation from ANG was completely abolished by an antioxidant N-acetyl-L-cysteine, while sodium arsenite-induced RNH1 dissociation was not affected at all, suggesting that sodium arsenite induces the dissociation of RNH1 in an oxidative stress-independent manner. As RNH1 acts as a stress sensor, the efficiency of stress-induced RNH1 dissociation is an important determinant of tiRNA biogenesis. In RNH1 knockout cells, tiRNAs are constitutively produced because intracellular RNases are fully activated. We recently showed that sodium arsenite treatment does not further increase the amount of tiRNAs in RNH1 knockout cells, suggesting that the amount of stress-induced tiRNAs will be the greatest when RNH1 is completely dissociated from intracellular RNases (Akiyama et al., 2022a).

5.2.3. tRNA modifications

There is accumulating evidence that tRNA modifications can modulate the biogenesis of tDRs, particularly the ANG-mediated tiRNA production (reviewed in (Lyons et al., 2018)). Several modifications have been reported to affect ANG-mediated tRNA cleavage. It is worth noting that these modifications inhibit, not promote, ANG-mediated cleavage without exception so far, which seems reasonable because one of the roles of tRNA modifications is considered to stabilize the tertiary structure of tRNAs (reviewed in (Lorenz et al., 2017)). The modification 5-methylcytosine (m⁵C) is one of the most extensively studied modifications that inhibit ANG-mediated tRNA cleavage. m⁵C modification in tRNAs is catalyzed by methyltransferases such as NSUN2 or DNMT2. DNMT2 is responsible for the methylation of cytosines at position 38 in tRNA^Asp-GTC, tRNA^Gly-GCC and tRNA^Val-AAC (Schaefer et al., 2010; Tuorto et al., 2012), while NSUN2 targets cytosines at position 48–50 in the variable loop of specific tRNAs (Blanco et al., 2014). Deletion of either NSUN2 or DNMT2 facilitates ANG-mediated tiRNA production (Blanco et al., 2014; Durdevic et al., 2013; Schaefer et al., 2010). Dnmt2 and Nsun2 double knockout mice showed reduced overall protein synthesis rates accompanied by complete lack of m⁵C modifications in tRNAs (Tuorto et al., 2012), which suggests that m⁵C modifications are critical for the regulation of protein synthesis. A recent study showed that oxidative stress down-regulates NSUN2 expression, which facilitates ANG-mediated tiRNA production through inhibition of m⁵C modification (Gkatza et al., 2019). This further suggests that tiRNA production may be dynamically regulated via the regulation of tRNA modification status. Several other modifications, most of which are methylation modifications, have been reported to inhibit ANG-mediated tiRNA production. 5-methyluridine (m⁵U) modification at position 54 catalyzed by TRMT2A also decreases ANG-mediated tRNA cleavage (Pereira et al., 2021). ALKBH3 is a tRNA-specific demethylase responsible for demethylation of 1-methyladenosine (m¹A) and 3-methylcytosine (m³C). ANG-induced tiRNA production was reduced in ALKBH3 knockout cells, suggesting that ALKBH3-mediated demethylation facilitates ANG-mediated tRNA cleavage (Chen et al., 2019). Demethylation of m¹A at position 58 by ALKBH1 has also been reported to promote stress-induced tiRNA production (Rashad et al., 2020). Methylation of the ribose moiety of nucleosides can also inhibit tRNA cleavage. 2’-O-methylcytidine (Cm) modification at the wobble position (position 34) of human elongator tRNA^Met showed resistance to ANG-mediated cleavage (Vitali & Kiss, 2019). In addition to methylation modifications, queuosine modification, a hypermodified 7-azaguanosine-derivative at the wobble position in specific tRNAs (tRNA^His, tRNA^Asn, tRNA^Tyr and tRNA^Asp), protect these tRNAs from ANG-mediated cleavage (Wang et al., 2018). It is worth noting that these modifications change the biogenesis of specific tiRNAs generated from tRNAs with the corresponding modifications. Therefore, the expression profile of modification enzymes may contribute to tissue- or cell-specific translational regulation by tDRs.

Accumulating evidence suggests that tRNA modifications also affect the biogenesis of ANG-independent tDRs. Hypomodification of 1-methylguanosie at position 9 (m¹G₉) due to a loss-of-function mutation in TRMT10A facilitates the production of tRNA^Gln-originated smaller 5’-tDRs, resulting in apoptosis in human pancreatic β-cells (Cosentino et al., 2018). RNase L-mediated tRNA cleavage is affected by queuosine (Q) modifications. tRNA^His possesses three potential cleavage sites in the anticodon loop (e.g. position 33, 34 and 36). As a Q modification at the wobble position inhibits RNase L-mediated cleavage at position 33 and 34, RNase L preferentially cleaves the other site (at position 37) (Donovan et al., 2017). tRNA^His is characterized by an additional guanosine residue at the 5’-end (position −1) that is added post-transcriptionally by THG1L (tRNA^His guanylyltransferase 1 like), the human homolog of yeast thg1 (Gu et al., 2003). It was reported that a methyltransferase, BCDIN3D, methylates the 5’-phosphate of the guanosine residue at position −1, forming a 5’-methylphosphate cap structure (Martinez et al., 2017). BCDIN3D has recently been reported to inhibit the processing of tRNA^His into smaller 3’-tDRs, likely through 5’-methylphosphate cap formation (Reinsborough et al., 2019). In contrast to other modifications, pseudouridine (ψ) modifications may facilitate tDR biogenesis. Depletion of pseudouridine synthase PUS7 decreased the production of 18-nt smaller 5’-tDRs that originated from tRNA^Ala, tRNA^Cys and tRNA^Val, accompanied by a complete loss of ψ modifications at position 8 in these tRNAs (Guzzi et al., 2018).

In single-cell organisms, preferences of RNases to a specific modified nucleotide can be an advantage during competition against other species. In E. coli, a secreted RNase colicin E5 specifically cleaves tRNAs with queuosine (Q) modifications at position 34 in competitor strains (Ogawa et al., 2006). In contrast, in human cells, a Q modification in the anticodon inhibits the cleavage by ANG or RNase L (Donovan et al., 2017; Wang et al., 2018). Another E. coli-specific nuclease Prrc specifically cleaves tRNA^Lys-TTT possessing a 5-methylaminomethyl-2-thiouridine (mnm⁵s²U) at the wobble position (Jiang et al., 2001). γ-toxin, a secreted RNase produced by the yeast Kluyveromyces lactis, specifically cleaves tRNAs at the anticodon in which the wobble position is modified to 5-methoxycarbonylmethyl-2-thiouridine (mcm⁵s²U), killing competitor strains such as S. cerevisiae (Lu et al., 2008).

5.2.4. Others

Some other factors have been reported to affect tDR biogenesis. Schlafen genes are a subset of interferon-stimulated genes (ISGs) which contribute to cellular immune response (Mavrommatis et al., 2013) (also see section 3.1.1.). Within the group of Schlafen proteins, Schlafen 2 (SLFN2) does not cleave tRNAs or rRNAs, in contrast to SLFN13 and SLFN11 that function as RNases (Li et al., 2018; Yang et al., 2018). Instead, SLFN2 acts as a protector of tRNAs against ANG-mediated cleavage through a direct interaction with tRNAs (Yue et al., 2021). ANG-mediated tiRNA production was promoted in SLFN2 knockout mice, resulting in T cell inactivation through tiRNA-mediated translation inhibition, which suggests that SLFN2 is critical for regulating translation rates so that T cells can be activated when necessary. Another study has shown that activated T cells release extracellular vesicles (EVs) that are enriched with tDRs. Once EV-mediated tDR release was inhibited, T cells failed to be activated due to the accumulation of tDRs (Chiou et al., 2018). These data further suggest that tDR-mediated translational regulation is essential for proper T cell function.

Sex hormones, such as estrogen and androgen, promote ANG-mediated tRNA cleavage, although the underlying mechanism remains to be elucidated. In estrogen receptor-positive breast cancer cells and androgen receptor-positive prostate cancer cells, ANG-mediated cleavage is enhanced by sex hormones and their receptors, leading to cancer cell proliferation through the production of tiRNA-like tDRs called Sex Hormone-dependent tRNA-derived RNAs (SHOT-RNAs) (Honda et al., 2015).

Translational states also affect the efficiency of ANG-mediated tRNA cleavage. In eIF2α phosphorylation-deficient (S51A mutant) mouse embryonic fibroblasts (MEFs), in which translation is not repressed in response to stress stimuli, ANG-mediated stress-induced tiRNA production was promoted (Saikia et al., 2012). On the other hand, when treated with hippuristanol, an inhibitor of translation initiation, ANG-mediated tRNA cleavage was decreased in parallel with translation repression (Saikia et al., 2012). In addition, ANG-mediated tiRNA production is enhanced under conditions where eIF2α phosphorylation is inhibited (Yamasaki et al., 2009). These data suggest that ANG-mediated tiRNA production is promoted under conditions where cells actively synthesize proteins.

We have recently reported that the RTCB ligase complex (RTCB-LC) negatively regulates tiRNA production (Akiyama et al., 2022b). Under normal conditions, RTCB-LC inhibits tiRNA production likely by sealing a cleaved site before the cleaved (so-called “nicked”) tRNAs are separated into 5’- and 3’-tiRNAs. Under oxidative stress conditions, tiRNA production is boosted as RTCB-mediated tRNA repair is repressed. This boosting mechanism through RTCB inhibition seems reasonable for cellular stress response because cells can rapidly produce tiRNAs without transcriptional or translational regulation of RNases.

Finally, it was recently was shown by two groups that the majority of tRNA halves actually remain associated in the form of nicked tRNAs, at least within some timeframe after tRNA cleavage (Chen & Wolin, 2023; Costa et al., 2023). It should be noted that the separation of the nicked tRNAs into individual tDRs is not understood, but likely involves actions of RNA helicases (Drino et al., 2023). While some of the helicases (e.g., DDX3X) were shown to efficiently separate nicked tRNAs into tRNA halves in vitro and in cytoplasmic extracts, whether this is also the case in the live cells is still under investigation. It is however very clear that availability and activity of such factors is critical for the biogenesis of tDRs. Interestingly, nicked forms of tRNA are also found in extracellular environment, e.g. in biofluids. It is also important to note that tDRs greatly vary in stability, both within the cells and in the extracellular compartment (Gambaro et al., 2020) . The factors contributing to such variations in stabilities are likely structural and/or dependent on the presence of specific RNA modifications within a tDR.

6. Concluding Remarks

With the implementation of various high-throughput technologies in various experimental settings, it is clear that tRNA is a rich source of diverse ncRNAs called tDRs. The importance of tDRs was indicated by the first observations suggesting that biogenesis of tDRs is a regulated process. Since then, accumulating evidence has shown a variety of active roles of tDRs in different aspects of gene expression, and the list of functions continues to grow.

What has become apparent, even from our incomplete understanding of tDR biogenesis and functions, is that tDRs contribute to many regulatory activities. It is clear that specific tDRs can target specific mRNA targets, while others can contribute to more global regulation, at least in the context of translational control and stress response. Further studies utilizing -omics approaches will facilitate the identification of interaction patterns between tDRs and other regulatory pathways operating on transcriptional and post-transcriptional levels.

In this review, we focused on the details of biogenesis and known roles of specific tDRs in the modulation of proteins synthesis. It is clear that tDR functions need further investigation, with focus on their molecular mechanisms. Understanding such mechanisms will pave the way to the development of novel therapeutic approaches to treat human disease, as it is already clear that perturbations in the biogenesis of tDRs or uncontrolled presence of specific tDRs is associated with multiple pathological conditions. Taken together, the roles and potential applications of tDRs will require future intensive research, and we anticipate a more complete understanding of the biological functions of tDRs.