Retrotransposon targeting to RNA polymerase III-transcribed genes

Abstract

Retrotransposons are genetic elements that are similar in structure and life cycle to retroviruses by replicating via an RNA intermediate and inserting into a host genome. The Saccharomyces cerevisiae (S. cerevisiae) Ty1-5 elements are long terminal repeat (LTR) retrotransposons that are members of the Ty1-copia (Pseudoviridae) or Ty3-gypsy (Metaviridae) families. Four of the five S. cerevisiae Ty elements are inserted into the genome upstream of RNA Polymerase (Pol) III-transcribed genes such as transfer RNA (tRNA) genes. This particular genomic locus provides a safe environment for Ty element insertion without disruption of the host genome and is a targeting strategy used by retrotransposons that insert into compact genomes of hosts such as S. cerevisiae and the social amoeba Dictyostelium. The mechanism by which Ty1 targeting is achieved has been recently solved due to the discovery of an interaction between Ty1 Integrase (IN) and RNA Pol III subunits. We describe the methods used to identify the Ty1-IN interaction with Pol III and the Ty1 targeting consequences if the interaction is perturbed. The details of Ty1 targeting are just beginning to emerge and many unexplored areas remain including consideration of the 3-dimensional shape of genome. We present a variety of other retrotransposon families that insert adjacent to Pol III-transcribed genes and the mechanism by which the host machinery has been hijacked to accomplish this targeting strategy. Finally, we discuss why retrotransposons selected Pol III-transcribed genes as a target during evolution and how retrotransposons have shaped genome architecture.

Keywords: Integrase; RNA polymerase III; Retrotransposon; S. cerevisiae; Ty element; tRNA.

Fig. 1

LTR and non-LTR retrotransposons that target to tRNA genes. a. LTR retrotransposons. Ty1, Ty3, DGLT-A and Tj1 elements are depicted in dark green. The boxed black arrows represent the LTRs flanking the two ends of the elements. The first ORF of the Ty1 element encodes Gag and the second ORF encodes a polypeptide (Pol) which is further processed into protease (PR), integrase (IN), and reverse transcriptase (RT)/ ribonuclease H (RH). Ty3 differs in structure from Ty1 by swapping positions of IN and RT/RH. For both Ty1 and Ty3, the Pol polypeptide is generated by a + 1 translational frameshift 38 bp upstream of the 3’end of Gag [–171]. The D. discoideum DGLT-A element contains one ORF that encodes for both Gag and Pol proteins. DGLT-A belongs to the Ty3-gypsy clade, signified by the arrangement of pol with IN after RT/RH [172]. S. japonicas Tj1 has a similar structural arrangement as Ty3 with GAG  and POL as two separate ORFs. The GAG ORF has a stop codon that is thought to be translationally suppressed to allow for translation of the POL ORF which lacks a start codon [121]. The length of each element is depicted by the scale at the bottom in kb. b. non-LTR retrotransposons. D. discoideum TRE5-A and TRE3-A, D. purpureum NLTR-A and P. pallidum NLTR-B are depicted in dark orange and all share a similar structural arrangement. All elements except NLTR-B have two ORFs flanked by untranslated regions (UTR), with TRE5-A and TRE3-A ending with an oligo(A) tail. The 5′ and 3’UTR of TRE5-A are arranged into A- and B-modules, and B- and C-modules respectively. The protein domain arrangement of TRE5-A and TRE3-A ORF2 is the same and encodes a protein containing an apurinic/apyrimidinic endonuclease (APE), RT, and zinc-finger (ZF) domain. Both TRE5-A and TRE3-A require a − 1 frameshift for translation of ORF2 [137, 173]. NLTR-A and NLTR-B have a similar arrangement to the TRE5-A and TRE3-A elements except that an RH domain substitutes for the ZF domain. In addition, NLTR-B has three separate ORFs for APE, RT and RH. It is not yet known if the 5′ and 3’ UTRs of NLTR-A and NLTR-B are arranged into modules. NLTR-A ORF1 overlaps with ORF2 by 13 bp but whether a frameshift occurs or not for translation of ORF2 is not yet known [124]. NLTR-B does not contain overlapping ORFs, however RT does not contain a start codon [124]. The length of each element is depicted by the scale at the bottom in kb

Fig. 2

Pol III structure highlighting subunits that may interact with Ty1-IN. The Pol III surface view is based on the cryoelectron microscopy structure of the initially transcribing Pol III complex (Protein Data Bank code 6f41) [44] with TBP, Brf1 and Bdp1 structures excluded. The arrow points to downstream DNA and the DNA template and non-template strands are coloured in light blue and dark blue, respectively. a The highlighted Pol III subunits are Rpc31 (dark green), Rpc34 (purple), Rpc82 (beige), Rpc1 (light pink), Rpc2 (light green), Rpc40 (magenta), Rpc53 (orange) and Rpc37 (red). The N-terminus of Rpc53 (amino acids 1–270) is not depicted due to lack of structural data. b Same as in (a) except rotated 165^o

Fig. 3

tRNA targeted retrotransposon insertion site profiles. The insertion site preference for S. cerevisiae, Dictyostelium and P. pallidum are shown here upstream and downstream of a tRNA gene. The tRNA gene (gray) contains box A (red) and box B (blue) internal promoters and the external box B (ex B, blue) for social amoeba. LTR-retrotransposons are in green and non-LTR retrotransposons are in orange. Inverted orange or green triangles denote retrotransposon insertion windows ranging from 2 to ~ 1000 bp upstream or 7 to ~ 450 bp downstream of the tRNA gene (not drawn to scale). For the social amoeba, split orange and green inverted triangles denote overlapping insertion footprints for LTR (DGLT-A, Skipper-2) and non-LTR (NLTR-A, NLTR-B, TRE5, TRE3) retrotransposons. For P. pallidum, a specific DLGT-A (DGLT-A.4) is indicated because DGLT-A.1–3 do not target to tRNA genes in this organism [124]. The green triangle with a broader base represents the larger insertion window for S. cerevisiae Ty1 which can insert up to ~ 1 kb upstream of a Pol III-transcribed gene. Nucleosomes are depicted upstream of the S. cerevisiae tRNA gene as Ty1 inserts into nucleosomes