Role of RNA polymerase III transcription factors in the selection of integration sites by the dictyostelium non-long terminal repeat retrotransposon TRE5-A

Abstract

In the compact Dictyostelium discoideum genome, non-long terminal repeat (non-LTR) retrotransposons known as TREs avoid accidental integration-mediated gene disruption by targeting the vicinity of tRNA genes. In this study we provide the first evidence that proteins of a non-LTR retrotransposon interact with a target-specific transcription factor to direct its integration. We applied an in vivo selection system that allows for the isolation of natural TRE5-A integrations into a known genomic location upstream of tRNA genes. TRE5-A frequently modified the integration site in a way characteristic of other non-LTR retrotransposons by adding nontemplated extra nucleotides and generating small and extended target site deletions. Mutations within the B-box promoter of the targeted tRNA genes interfered with both the in vitro binding of RNA polymerase III transcription factor TFIIIC and the ability of TRE5-A to target these genes. An isolated B box was sufficient to enhance TRE5-A integration in the absence of a surrounding tRNA gene. The RNA polymerase III-transcribed ribosomal 5S gene recruits TFIIIC in a B-box-independent manner, yet it was readily targeted by TRE5-A in our assay. These results suggest a direct role of an RNA polymerase III transcription factor in the targeting process.

FIG. 1.

Setup of the TRE trap. The D. discoideum gene encoding UMP synthase (pyr5-6) is equipped with an intron derived from the D. discoideum cbfA gene (2). The intron sequence is indicated as a dashed line. The white arrow indicates the transcription orientation of the pol III gene inserted into the intron. All tested pol III genes were inserted into the trap as EcoRI fragments, schematically exemplified by D. discoideum Val^UAC. TRE5-A integrations into the trap were isolated by PCR using pyr5-6 exon-specific primers as indicated. The tRNA gene-internal A-box and B-box promoter elements are indicated. Mutations introduced into the consensus GTCnnnnG⁵³TTC⁵⁶RANYC⁶¹ B-box motif of the D. discoideum Val^UAC tRNA gene are indicated.

FIG. 2.

Results from TRE trap assays. (A) In this experiment 4 × 10⁶ D. discoideum cells carrying the empty trap (−), a D. discoideum Glu^sup gene, or a D. discoideum Val^UAC gene were subjected to selection in 5-FOA and uracil. (B) Comparison of targeting frequencies at a human Met^CAU gene and a D. discoideum Val^UAC gene. D. discoideum cells (5 × 10⁶ of each strain) were used for 5-FOA selection. Mean clone numbers from 10 petri dishes are shown ± SD.

FIG. 3.

5′ junctions of de novo integrations of TRE5-A into the TRE trap. (A) Targeting of the D. discoideum Glu^sup tRNA gene and the human Met^CAU gene. The target sequence is written in the first line in lowercase letters. Vertical arrows point to the integration sites of TRE5-A elements; the numbers indicate the distance of the TRE5-A to the first nucleotide of the targeted tRNA gene. The numbers in parentheses indicate the first nucleotides of the inserted TRE5-As, which are written in bold uppercase letters. Extra nucleotides are boxed, and target site deletions are indicated as “Δ.” (B) Integrations of TRE5-A upstream of an isolated B box. The target sequence is shown in lowercase letters. The underlined sequence of 26 bp represents duplication of the intron sequence located 120 bp downstream of the integration site. Note that all integrations shown were isolated from different selection plates and represent independent integration events by definition, even if the integrated elements and integration sites look very similar. Note that the 5′ end of a full-length TRE5-A consists of a 271-bp A module which is composed of a 199-bp core sequence and 72 bp identical with the 5′ end of the core. Thus, elements whose 5′ ends are designated “(1)” have full-length A modules (199 + 72 bp), whereas elements labeled “(200)” contain only the 72-bp repeat.

FIG. 4.

Frequencies of TRE integrations upstream of an isolated B box. The TRE trap assay was performed with the Val^UAC gene (wt), the mutant Val^UAC(C56G) gene, and the empty trap (−). A B box was inserted 34 bp downstream of the wild-type and mutant tRNA gene (referred to as the exB box) and at the corresponding position in the intron of the empty trap. Clone numbers from 10 petri dishes of two independent clones were counted after 5-FOA selection and normalized for the wild-type Val^UAC gene without the exB box and presented ± SDs.

FIG. 5.

In vitro binding of TFIIIC to mutant tRNA genes. The figure shows results of EMSAs with the D. discoideum Glu^sup tDNA derivatives as radiolabeled probe. Radiolabeled probes were wild-type Glu^sup (lanes 1 to 3), Glu^sup(G53T) (lanes 4 to 6), Glu^sup(C56G) (lanes 7 to 9), and Glu^sup(C61A) (lanes 10 to 12). One microgram of plasmid carrying the empty trap was used as competitor. DNA-TFIIIC complexes are indicated by the white arrowhead. Three increasing amounts (0.1 μg, 0.5 μg, and 1 μg) of NE600 fraction were used as a source of TFIIIC.

FIG. 6.

TRE integration frequencies at mutant tRNA genes. The figure shows the results of the TRE trap assay with Val^UAC (white bars) and Glu^sup (black bars) as bait tRNA genes. Mutant tRNA genes (as indicated), empty TRE trap (−), and wild-type tRNA genes were tested. Clone numbers from 10 petri dishes of two independent clones of each strain are normalized for the wild-type (wt) tRNA genes and expressed ± SDs.

FIG. 7.

Examples of TRE5-A integrations upstream of an inactivated tRNA gene equipped with a downstream extra B box. DNA sequences of representative TRE5-A insertions at positions 10 bp (A) and 50 bp (B) upstream of an inactive Val^UAC(C56G) are shown. The intron sequence is shown in lowercase letters. The exB box is presented in uppercase letters inside the black box; the mutated B box of the tRNA gene is boxed with a dashed line. The tRNA gene sequence is shown in uppercase letters. The insertion site is indicated by the vertical arrow. The DNA sequences of the inserted TRE5-As are shown in bold uppercase letters. The first nucleotides of the 5′-deleted TRE5-As are shown in parentheses and TSDs in gray boxes.

FIG. 8.

Results from a TRE trap assay using the D. discoideum ribosomal 5S gene as bait. The targeting frequency at the r5S gene is compared to the empty trap (−) and the trap loaded with a D. discoideum Val^UAC gene. Mean clone numbers from 10 petri dishes are shown ± SDs.

FIG. 9.

5′ junctions of new TRE5-A integrations. (A) Targeting of the D. discoideum r5S gene. The target sequence is written in the first line in lowercase letters. The 5′ end of the r5S gene (reverse orientation) is shown in uppercase letters. The vertical arrow points at the integration site of a TRE5-A element; the numbers indicate the distance of the TRE5-A to the first nucleotide of the targeted r5S gene. The numbers in parentheses indicate the first nucleotides of the inserted TRE5-A, which are written in bold uppercase letters. Target site deletions are indicated as “Δ.” (B) Integrations of TRE5-A upstream of an r5S tandem. The first line indicates the target sequence, which consists of two r5S genes (uppercase) separated by an EcoRI restriction site and the transcription terminator sequence of the first r5S gene (lowercase). The first nucleotides of the inserted TRE5-As (bold uppercase) are written in parentheses; extra nucleotides are boxed.

FIG. 10.

Model of tRNA gene recognition by TRE5-A proteins. A tRNA gene is indicated with an A box and a B box and a downstream exB box. D. discoideum TFIIIB likely consists of three subunits: TATA binding protein, Brf1, and Bdp1 (T. Winckler, unpublished observation). The exact subunit composition of D. discoideum TFIIIC is unknown. The TRE5-A preintegration complex, consisting of ORF1 and/or ORF2 proteins and TRE5-A RNA, is indicated as a single sphere. (A) TFIIIC binds to the tRNA gene-internal B box and recruits TFIIIB to the 5′ end of the tRNA gene. Integration of TRE5-A occurs via interaction with TFIIIB, which leaves the −50 position unprotected. (B) If TFIIIC slides to the exB box during transcription of the tRNA gene, TFIIIB stays at its position, still supporting integration of TRE5-A in the −50 position, while the −10 position is blocked by DNA-bound TFIIIB. (C) If TFIIIC dissociates from the tRNA gene, TFIIIB may stay and further support TRE5-A integration in the −50 position.