tRNAs as primers and inhibitors of retrotransposons

ABSTRACT

The functional relationship between tRNAs and retrotransposons have been known for more than 35 years. tRNAs are used as primer molecules to guide the reverse transcription of retrotransposons. Recently, tRNAs have also emerge as important players in the postranscriptional regulation of retrotransposons by means of tRNA-derived small RNAs. This surprisingly new layer of regulation indicates that tRNAs are used both in the promotion and the suppression of the reverse transcription of retrotransposons indicating their primary role in the life cycle of LTR retrotransposons. This adds another level of translational control to tRNAs. Here we review the different known levels of interactions of tRNAs and retrotransposons and highlight the unknown parts of this interaction.

KEYWORDS:

• RNAi
• retrotransposons
• tRNAs
• small RNAs
• tRNA-derived fragments

Transposable elements (TEs) are considered obligate genomic parasites that have contributed to genome structure and the transcriptional control of its content.^1,2 According to their requirement for a reverse transcription step, TEs are broadly classified as retrotransposons (if they produce an RNA intermediate) or DNA transposons.³ Retrotransposons were first discovered in Drosophila melanogaster by analogy to retroviral repetitive sequences. Usually retrotransposons are abundantly present in eukaryotic genomes making, for example, reverse transcriptase one of the coding genes more broadly present in eukaryotic genomes.⁴

Retrotransposons are classified according to their sequence characteristics and transposition mechanism into LTR transposons and non-LTR transposons. LTR transposons harbor repeats of a few hundred base pairs at each end of their genomic sequence known as long terminal repeats (LTRs).⁵ During their transcription retrotransposons produce a polycistronic transcript that produces the proteins needed for retrotranscription: gag, pol and env. The gag protein is a structural protein that forms a virus-like structure housing reverse transcription, while the pol gene encodes a protease, a reverse transcriptase and an integrase. The env protein facilitates infection as a functional part of the viral envelope.^5,6 Generally, LTR transposons are transcribed by RNA polymerase II from a promoter located within the 5′ LTR and are translated in the cytoplasm into the proteins that form the virus-like structure. Within this structure (Fig. 1A), retrotranscription of the RNA molecule takes place primed by a captured tRNA or tRNA fragment from the cytoplasmic pool⁷ that pairs to a sequence near the 5′ end of the RNA termed the primer binding site (PBS) helped by a specific chaperone that targets the 5′ unique sequence (U5) and the tRNA.⁸ This structure copies a first short sequence located at the 5′ of the PBS that includes a short repeat sequence (R) and U5 sequence. This short repeat sequence pairs with another homologous repeat sequence present in the 3′ extreme of the transposon (U3) and allows the reverse transcription of the whole element. The RNA template is partially degraded leaving a fragment at the poly-purine tract that primes the second DNA strand synthesis. First, a partial copy of the new 3′LTR is copied, then it switches positions to the 5′ LTR to finally copy the whole length element.^4,9 This process (summarized on Fig. 2) gives rise to the two LTRs that are formed by the U3, R and U5 sequences,

Figure 1. Life cycle of a LTR retrotransposon and inhibition mediated by tRFs. A. Schematic representation of the main steps in the life cycle of a LTR retrotransposon. B. Schematic representation of the known interactions of tRFs with the life cycle of retrotransposons and the potential inhibitions at the posttranscriptional and reverse transcription levels. The gag, pol and env genes are depicted in the genomic sequence of the LTR retrotransposon. GAG: GAG protein produced from the gag gene that forms the virus-like particles inside which retrotranscription takes place. RT: retrotranscriptase protein encoded by the pol gene that mediates the retrotranscription of the LTR-retrotransposon RNA into DNA. INT: integrase protein coded by the pol gene that mediates the reintegration of the retrotranscribed DNA into the genome.

Figure 2. LTR retrotransposon structure and reverse transcription process. A. Structure of a model LTR retrotransposon depicting the two long terminal regions (5′ and 3′ LTR), the gag, pol [with its three different enzymatic functions: protease (prot), retrotransciptase (rt) and integrase (int)] and env genes, the primer binding site (PBS) and the poly-purine tracts (PPT). B. Model of LTR retrotransposon reverse transcription process: (1) for most LTR retrotransposons and retroviruses 18 nts of the 3′ extreme of a mature tRNA interact with the PBS region of the Pol II transcript from the genomic LTR retrotransposon. (2) This interaction primes the synthesis of the unique 5′ sequence (U5) and a repeat sequence (R). (3) This initial transcript pairs with the R region on the 3′ extreme of the LTR retrotransposon transcript and (4) primes the transcription of the whole element. (5) The original RNA template is then degraded leaving only a fragment in the PPT region that (6) primes the second strand synthesis, first of the 3′ extreme which then moves to the 5′ region (7) where it pairs with the R and U5 regions and primes the final synthesis of the whole element. The final process gives rise to a dsDNA with two LTR that contain both the U3, R and U5 regions (8). The integrase coded by the pol gene inserts this dsDNA fragment within the chromosomal DNA.

The priming process by tRNAs is key for the retrotranscription.¹⁰ This priming by tRNAs is not stochastic, but rather uses specific tRNAs or their derivative fragments.^9,11 The tRNA priming of retrotranscription has been studied in depth for retroviruses at the molecular level, what has led to a more or less complete view of the whole process and the different mechanisms implied. First, 18 nucleotides at the 3′ end of the tRNA bind to the viral RNA PBS through strict complementarity followed by the interaction of a portion of the tRNA TΨC arm with a primer activation signal.^9,11,12 Annealing to both regions requires the chaperone activity of the retroviral nucleocapsid protein that frees these regions from intramolecular interactions allowing molecular interactions with the tRNA.⁸ Three different strategies have been followed in order to identify the tRNAs interacting with retroviruses and LTR-retrotransposons: 1) direct identification of tRNAs interacting with retroviral/LTR-retrotransposon transcripts, 2) isolation and direct sequencing of the tRNAs contained within the virus-like structure content and 3) bioinformatic prediction of the tRNAs interacting with the predicted PBS.¹³ The later has been used for the identification of tRNA-Lys as the primer for HIV-1 retrotranscription.¹⁴

Altogether, these analyses have allowed concluding that the tRNAs used for viral and transposon retrotranscription have subtle differences. All retroviruses and most retrotransposons use the 3′ end of specific mature tRNAs to prime retrotranscription. Nevertheless, some retrotransposons use different fragments derived from tRNAs as primers.¹⁵ This is the case for example of the S. cerevisae Ty5 element and the D. melanogaster copia element, which use a fragment from the 5′ region of the tRNA-Met. In general, retrotransposons of the copia group use tRNA-Met predominantly while the Gypsy group uses a more diverse array of tRNAs for retrotranscription that include tRNA-Met, but also tRNA-Asp, tRNA-Lys, tRNA-Arg, tRNA-Ser and tRNA-Leu.^9,11,16 The array of tRNAs used by retroviruses goes from tRNA-Trp, tRNA-Pro, tRNA-Lys and tRNA-Gln. Pararetroviruses (a class of plant virus similar to retrovirus that lack the ability to integrate in the genome) use tRNA-Met exclusively. Interestingly, the PBS size also differs between retrotransposons and retroviruses. In all retroviruses studied, the PBS is 18 nts in size, while for LTR-retrotransposons it varies from 8 to 18 nts as a direct consequence of the presence of modified nucleotides present in the primer tRNA.¹⁵

In a cellular context, retrotransposon transcription is posttranscriptionally controlled by RNAi through TE-derived small RNAs that degrade or epigenetically silence TEs in a sequence specific manner.⁶ Unexpectedly, tRNAs have been found to not only be primers of retrotransposon transcription, but also to play a role in the initiation of the RNAi response against them.^17,18 In the last years, increased sequence depth has help to identify that tRNAs can produce two different classes of sRNAs: tRNA-derived fragments (termed tRFs; 18–28 nt in length) and tRNA-halves (30–35 nt in length).^19,20 Both classes are produced massively during stress, but have two different biogenesis pathways: tRFs are dependent on RNAi while tRNA halves are dependent on angiogenin/Rny1 cleavage.^19–22 At the functional level, these two classes of sRNAs also have different modes of action: while tRNA halves interfere with translation,^23–25 tRFs are loaded into Argonaute proteins and exert their functions through RNAi.^17,18,26

We have recently identified that in Arabidopsis tRFs of 19 nts in length derived from the 5′ end of mature tRNAs (termed tRF-5s) target TEs mainly from the gypsy family and induce the cleavage of their transcripts. 19 nt tRF-5s accumulate to very high levels in the plant structure containing the male gametes, the pollen grain.¹⁷ Interestingly, in this same tissue there is a natural reactivation of transposable elements due to the epigenetic resetting that takes place during gametogenesis in plants.^27–30 The accumulation of this pollen exclusive tRFs is conserved in the pollen grain or analogous reproductive structures of other plant species, pointing to a potentially conserved function of these sRNAs.¹⁷ Intriguingly, tRFs and tRNA-halves derived from both the 5′ and 3′ regions of tRNAs are known to accumulate to high levels in mouse gametes and zygotes, what leads to interesting analogies about the conservation of this mechanism.³¹ Indeed, tRF-5s and 5′ tRNA halves have a transgenerational role and regulate the transcription of LTR-associated genes.³² A recent study in mouse shows that a similar class of tRFs derived instead from the 3′ end of mature tRNAs (tRF-3s) regulate the transcription of retrotransposons by directly targeting their PBS.¹⁸ Schorn and collegues (2017) show that two different tRFs of 18 or 22 nts inhibit or induce RNAi of retrotransposons respectively. Altogether, these two works point to a conserved role for eukaryotic tRFs in the protection of genomic integrity through the degradation or translational inhibition of retrotransposon activity. The known interactions of tRFs with retrotransposon's replicative cycle are shown in Figure 1B.

It is plausible to speculate that in case of massive epigenetic and transcriptional reprograming (as the one taking place in the tissues used in these two mentioned studies: the pollen grain and mouse preimplantation stem cells) the unused tRNA pool could be converted into tRFs in order to accomplish two objectives: I) avoid the priming of unwanted TE transcription and II) degrade the transcripts from TEs that can freely transcribe in the absence of repressive epigenetic marks. Surprisingly, the tRNA pool is highly dynamic under different cellular transcriptional needs.^33,34 Mature tRNAs can be incorporated into RNAi pathways when the stability of their secondary structure is compromised³⁵ and be degraded under stress.²²

Thus, the latest discoveries enumerated above indicate that the interaction between tRNAs and LTR retrotransposons does not only happen for transcriptional promotion purposes, but is also used by the cell to inhibit unwanted transcription of retrotransposons that can compromise genomic stability. Although extremely interesting for the appealing conservation that this posttranscriptional use of tRNA-derived small RNAs for retrotransposon control would mean, at this moment, there are more questions than answers. For example, from the evolutionary point of view, it is unknown how conserved this mechanism is in different species. Also, since the use of priming tRNAs is reduced to a very small number of tRNAs, it could be plausible to think that these tRNAs have special modifications or transcriptional profiles or even that they could have been evolutionary selected to interact with LTR retrotransposons. From the molecular and cellular points of view, it is still unknown how the cell recognizes that a TE is being actively transcribed and needs to be silenced through the activation of the production of tRFs. Furthermore, if tRF biogenesis involves the degradation of members of the tRNA pool this can have massive consequences for cellular translation. It is interesting to note that genomes still harbor remnants of the original tRNA halves that lead to the origin of tRNAs³⁶ so, potentially, those fragments could be used for production of tRFs independently of mature tRNAs. Moreover, it still needs to be further elucidated how tRNAs, which are known to interact with more than 30 proteins during their biogenesis and are rarely free cytoplasmic RNAs,³⁷ can enter into RNAi pathways in the cytoplasm and how this RNAi-derived tRFs can access the retrotransposon transcript.

The discovery that tRNAs are used for both priming and postranscriptional control of retrotransposons provides a very interesting innate immunity strategy in eukaryote organisms that needs to be explored in depth. Clearly, further studies are required to answer all these questions and that would allow the comparison of tRF based strategies for potential retrovirus and retrotransposon control and its obvious implications for genome stability maintenance.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgements

We thank Duarte Figueiredo and Clément Lafon-Placette for their helpful comments and ideas on this manuscript. We apologize to the many authors of important original research articles that could not be cited for lack of space.

Funding

Research in the Martinez laboratory is supported by SLU and a grant from the Swedish Research Council.