Circular retrotransposition products generated by a LINE retrotransposon

Abstract

Non-long terminal repeat (non-LTR) retrotransposons are highly abundant elements that are present in chromosomes throughout the eukaryotic domain of life. The long interspersed nuclear element (LINE-1) (L1) clade of non-LTR retrotransposons has been particularly successful in mammals, accounting for 30-40% of human genome sequence. The current model of LINE retrotransposition, target-primed reverse transcription, culminates in a chromosomally integrated end product. Using a budding yeast model of non-LTR retrotransposition, we show that in addition to producing these 'classical', chromosomally integrated products, a fungal L1 clade member (Zorro3) can generate abundant, RNA-derived episomal products. Genetic evidence suggests that these products are likely to be formed via a variation of target-primed reverse transcription. These episomal products are a previously unseen alternative fate of LINE retrotransposition, and may represent an unexpected source for de novo retrotransposition.

Figure 1.

A novel class of LINE retrotransposition products. (A) Structure of L1 clade members and proposed mechanism of non-LTR retrotransposon replication. Top, comparison of the structures of human L1 and Candida albicans Zorro3. Both elements contain two ORFs and endonuclease (endo), reverse transcriptase (RT) and zinc finger (zf) domains. Genomic L1s are typically flanked by target site duplications (flanking arrows). Genomic Zorro3s are flanked by poly(A) tracts, which may include poly(A) target site duplications. Bottom, general mechanism for the initiation of non-LTR retrotransposon replication. (B) Assay to induce Zorro3 retrotransposition. A donor Zorro3 marked with a HIS3AI (47) retrotransposition cassette (Zorro3mHIS3AI) is present at the LYS2 locus of chromosome II. To induce retrotransposition, cells are grown on galactose and then plated on SC-His to select retrotransposition events. H = HaeII. Bar below Zorro3 represents the probe used for Southern blots in all subsequent figures. (C) Example of LM-PCR products. Each lane represents the LM-PCR product from an independent Zorro3 retrotransposition event. Owing to variable EcoRI distances from the Zorro3 3′ end, Class A LM-PCR products are not always amplified. ‘mHIS3AI only’ is an LM-PCR reaction from a strain containing only the mHIS3AI cassette. ‘Donor only’ is an LM-PCR reaction from a strain containing Zorro3mHIS3AI, but not a retrotransposition event. Class A products are noted with an asterisk. (D) Southern analysis of class A and class B events. Genomic DNA from the indicated clones was digested with HaeII and probed for Zorro3 sequence. Arrow indicates the band representing the original Zorro3mHIS3AI donor. * = class A insertions. n = nicked. l = linear. s = supercoiled.

Figure 2.

Instability and transfer of class B products between strains. (A) Representative examples of class A and class B product stability. (B) Quantitation of the experiment described in panel A. (C) Transfer of class B product between strains. Total genomic DNA purified from strains containing a class A or class B insertion was used to transform His⁻ yeast cells without pre-existing Zorro3 sequence. Transformations were plated on SC-His. His⁺ colonies represent successful maintenance of the retrotransposition product under selection. (D) Class B transformants harbour the class B episome. Genomic DNA from His⁺ transformants described in panel C was digested with HaeII and subjected to Southern analysis. The donor (arrow) and class B products are both present in the original class B strains (13–16), but only the class B product is transferred to the transformants (13-1, 13-2, etc.). (E) Class B products transferred to new strains exhibit instability. His⁺ stability testing described in panels A and B was repeated with class B transformants generated in panel C.

Figure 3.

Structural characterization of class B products. (A) Predicted structures and restriction fragment sizes of class B-I and class B-II circles. (B) Verification of predicted class B structures. Genomic DNA from donorless strains containing a class B circle were digested with the indicated enzyme and subjected to Southern analysis with the Zorro3mHIS3 probe shown in Figure 1B.

Figure 4.

Class B products form with high frequency. (A) A cPCR assay to detect class B products. The indicated primers (in red) were used for colony PCR reactions to specifically amplify products in strains containing class B-I or class B-II circles. (B) The cPCR assay is sensitive and specific for class B products. Original class B strains and strains converted to His⁺ by class B genomic DNA transformation (Figure 2C and D) were subjected to the cPCR assay. (C) His⁺ phenotype tracks with the class B circle. A His⁺ stability experiment (described in Figure 2E) was performed on parent strains transformed with a class B product. Colonies that lost the His⁺ phenotype (YPD, His⁻) and rare colonies that retained the phenotype (YPD, His⁺) were tested by the cPCR assay for the presence of a class B episome. (D) Class B products constitute the majority of Zorro3 retrotransposition events. For two strains (containing Zorro3mHIS3AI or Zorro3pA⁻mHIS3AI), 44 independently derived retrotransposition events (all unlabelled lanes) were tested by cPCR for the presence of class B-I or class B-II products. 13 = an original class B-I clone (control). 14 = an original class B-II clone (control).

Figure 5.

Pre-existing homology is not required for class B product formation. (A) Structure of Zorro3 donor and view of Zorro3 regions where most class B junctions occur. Class B-I circles typically form from a joining between the 3′ poly(A) region and the 5′ poly(A) region, whereas class B-II circles typically form from a joining between the 3′ poly(A) region and the interORF poly(A) region. Zorro3pA⁻ sequences are identical to Zorro3 except the stretch of 20 (A)s in the 5′ poly(A) region is deleted. (B–E) Sequences of cloned class B product junctions derived from wild-type Zorro3 or Zorro3pA⁻ donors. Untemplated nucleotides are indicated in green. Yeast genome sequences are indicated in brown.

Figure 6.

Mutations affecting class B frequency and ratio. (A) Zorro3 was induced for retrotransposition in the indicated mutants. In all, 94 independently derived clones were picked for each mutant and assayed by cPCR for the presence of Class B products. **P < 0.001. *P < 0.05. (B) Loss of RNH1 leads to an increase in total Zorro3 retrotransposition. A quantitative retrotransposition assay was performed on the indicated strains. Frequency is presented as the average of five independent cultures. Black bars indicate the range of frequencies obtained.

Figure 7.

A speculative model for the formation of class B products. (A) Genomic target site. (B) Bottom chromosome strand nick and LINE mRNA annealing. (C) Minus strand synthesis. (D–E) Top strand chromosomal nick and template jump to top strand. (F) Recleavage of top and bottom strands (at arrowheads) to release the retrotransposition intermediate. The timing and placement of recleavages do not necessarily need to occur as indicated in the figure. Chromosomal target site DNA between the sites of recleavage have been highlighted in green for emphasis. As Zorro3 integrates into poly(A) tracts, the DNA highlighted in green would likely be poly(A) sequence. For other LINE elements, the DNA in green would likely correspond to part of the target site duplication. (G) Resolution of the excised circle by annealing of complementary strands. Red lines indicate RNA. Blue lines indicate newly synthesized retrotransposon DNA.