Site-specific DNA double-strand breaks greatly increase stable transformation efficiency in Trypanosoma brucei

Abstract

Genetic manipulation in African trypanosomes typically relies upon electroporation with chromosomal integration of DNA constructs by homologous recombination. Relatively little is known about chromosomal recombination and repair in these organisms however and low transformation efficiency and position effects can limit forward genetic approaches. In yeast and mammalian cells, site-specific DNA double-strand breaks (DSBs) stimulate targeted integration through homologous recombination-based repair where the exogenous DNA serves as the template. We have explored the effect of DSBs on targeted integration in bloodstream-form Trypanosoma brucei, focusing on the ribosomal RNA-spacer target commonly used to integrate recombinant constructs. DSB-repair within the ribosomal RNA tandem gene-repeats is likely dominated by single-strand annealing allowing approximately 80% of cells to survive the break. In the presence of exogenous DNA, transformation efficiency is increased approximately 250-fold by DSB-induction. In the example presented, more than 1% of cells that survive the procedure were transformed generating 80,000 transformants from a typical experiment.

Fig. 1

Response to a DSB at the RRNA-spacer locus. (A) The schematic map illustrates an RFP–PAC fusion gene (R^sP) with an embedded I-SceI site (indicated by DSB) at the RRNA-spacer (RRNA^s) locus. pR^sP_RRNA was assembled as follows: an RRNA promoter (P_RRNA) fragment was amplified from genomic DNA using primers RpF (GATCcggcggTAGCTTTCCACCCAGCGC) and RpR (GATCcggccgggcccACTGggatccTCTGAGAGCGGTCAGTTGC), digested with EagI (relevant restriction sites in lower-case) and ligated to a NotI-digested RRNA-spacer fragment in pBlusescript. An R^sP cassette was then added using the BspI201 and BamHI sites. The R^sP_RRNA construct was then digested with SacI/AgeI and introduced into the 2T1 bloodstream-form T. brucei strain that also contained a tetracycline-inducible I-SceI ORF introduced using the pRPaⁱ construct . These Lister 427, clone 221a cells were grown and manipulated as described . (B) A clonogenic assay to assess recovery from a DSB. Cells in all un-induced wells tested remained puromycin-resistant and cells in every induced well were puromycin-sensitive indicating loss of the R^sP cassette in the latter case. Cell counts were carried out using a haemocytometer and tetracycline (used at 1 μg ml⁻¹) was from Sigma. Data are derived from a pair of independent R^sP_RRNA strains and error bars represent one standard deviation. (C) Physical monitoring of DNA resection adjacent to the lesion was carried out by slot-blot assay as described . Genomic DNA samples were ‘native’, to detect ssDNA or denatured, to detect total DNA. The probes used on each blot are indicated on the right; the control probe is from chromosome 11 (Tb11.01.7240). Phoshorimager analysis was used to quantify the signals and ssRFP values were derived after correction for background, ssDNA versus total DNA and loading. The RFP ssDNA and total DNA plots indicate resection kinetics and DNA loss respectively.

Fig. 2

A genomic DSB increases transformation efficiency. (A) The schematic map illustrates the genomic target (reproduced from Fig. 1A) and the exogenous DNA construct. (B) Transformation assays. A DSB was induced by growth in tetracycline (1 μg ml⁻¹) for 3 h. Nucleofection (Lonza) was carried out as described . Briefly, 2.5 × 10⁷ R^sP_RRNA cells were resuspended in 100 μl of human T-cell Nucleofector solution, mixed with 10 μg of purified linear DNA and subjected to Nucleofection using program X-001 in a 1 mm-gap cuvette. G418 was added <6 h later at 2 μg ml⁻¹. To estimate the number of transformed clones, we initially used serial dilutions in 96-well plates but, due to concerns with loss of accuracy during extensive serial dilution, we used a modified approach to generate the data presented. Briefly, in duplicate experiments, 6 h after Nucleofection and drug addition, we distributed a sample predicted to contain 32 transformants (based on estimates from serial dilutions) over a 96-well plate. This approach yielded 15–40% positive wells per plate and was therefore deemed to have provided accurate scores. Error bars represent one standard deviation.