Recombination between T-DNA insertions to cause chromosomal deletions in Arabidopsis is a rare phenomenon

Plant material and growth conditions

Arabidopsis lines carrying T-DNA insertions flanking FUS2 (At4g10180), FUS6 (At3g61140), Fus7 (At4g14110), and Fus11 (At5g14250) were obtained from the Arabidopsis Biological Resource Center at The Ohio State University (abrc.osu.edu). FUS2: SALK_046021 and SALK_004560. FUS6: WiscDsLox453-456F5, WiscDsLox402F03, SALK_203944, and SALK_062171. FUS7: SALK_073580 and SALK_150313. FUS11: SALK_150318 and SALK_150319. Seedlings were germinated on 1% agar (w/v) plates containing 0.5× Murashige and Skoog (MS) basal salt mixture under continuous light at 20 °C–23 °C. Seedlings were transferred to a 2:1 mixture (vol:vol) of potting mix:pearlite 5–7 days after germination and grown under constant light at 20 °C–23 °C.

Genotyping

PCR-based genotyping was used to determine the presence of the desired T-DNA insertion elements. Genomic DNA was extracted from 5 mm square pieces of leaf tissue as previously described (Kasajima et al., 2004). PCR was performed using the gene-specific PCR primers indicated in Table 1 in conjunction with the T-DNA left border primer p745: 5′-AACGTCCGCAATGTGTTATTAAGTTGTC-3′. The thermal cycling conditions used were as follows: 96 °C for 2 min, followed by 40 cycles of 94 °C for 10 s, 58 °C for 30 s, and 72 °C for 30 s. PCR products were visualized by agarose gel electrophoresis followed by ethidium bromide staining.

Phenotypic screening

F1 plants were produced by crossing two single-mutant T-DNA lines. PCR genotyping was used to confirm that the F1 plants carried the desired T-DNA insertions. F2 seed was then collected from individual F1 plants. F2 seedlings were germinated on petri dishes containing agar growth media and transplanted to potting mix at a density of 9 seedlings per 3.5 inch square pot. The F2 plants were allowed to self-pollinate, and the resulting seed was collected in bulk from each pot of 9 plants. Ca. 1,200 F3 seeds from each pool of nine F2 plants were placed onto water-saturated 125 mm diameter circular pieces of Whatman filter paper (Whatman cat. #1004-125) inside 140 mm diameter plastic petri dishes. The plates were then incubated in the dark at 4 °C for 2–3 days to stratify the seeds. The plates were then placed under light at 20 °C–23 °C for four hours to stimulate germination, and then transferred to black box with no light at 20 °C–23 °C for four days to allow for etiolated growth. Plates were then visually inspected to search for seedlings displaying the distinctive fus phenotype.

Design of the genetic screen

We use the term “deletion caused by T-DNA recombination” to refer to the situation where meiotic crossing over between T-DNA insertions located nearby to each other leads to deletion of the intervening genomic DNA (Fig. 1). In our previous study describing the identification of a deletion caused by T-DNA recombination in the MEKK1/2/3 gene family, we used a PCR-based method to identify a single deletion mutant after screening ca. 2,000 seedlings (Su et al., 2013). In the present study, our objective was to screen a larger population of individuals in order to better ascertain the frequency with which deletions caused by T-DNA recombination occur. In order to do this in a cost-effective manner, we chose to develop a genetic screen that would let us visually screen for evidence of deletions caused by T-DNA recombination. The rationale for the screen was to choose loci in the Arabidopsis genome for which null mutants produced a striking visual phenotype that could be easily scored at the early seedling stage. We chose to use FUSCA (FUS) genes for this purpose (Misera et al., 1994). The FUS genes are negative regulators of photomorphogenesis such that null mutants cause dark-grown seedlings to display characteristics normally seen in light-grown seedlings. When Arabidopsis seeds are germinated and grown in the dark for four days, the hypocotyls elongate, the apical hook remains closed, and the seedlings are pale white in appearance. By contrast, fus mutant seedlings grown in the dark have short hypocotyls, the apical hook is open, and they are dark purple due to anthocyanin accumulation (Misera et al., 1994). A fus mutant can therefore be easily distinguished on a plate of thousands of wild-type dark grown seedlings as a short purple seedling with open cotyledons against a background of long, skinny, white seedlings. This made the fus mutant phenotype a good candidate for our genetic screen. There are at least fourteen FUS loci (Misera et al., 1994) in the genome, and these genes have also been identified as CONSTITUTIVE PHOTOMORPHOGENIC (COP) and DE-ETIOALTED (DET) (Huang, Ouyang & Deng, 2014).

To identify specific FUS loci for our screen, we searched the available T-DNA mutant collections for FUS genes that had T-DNA insertions in close proximity upstream and downstream of the FUS gene. Suitable T-DNA lines were found for FUS2, FUS6, FUS7, and FUS11 (Fig. 2). In the case of FUS6, we chose two different sets of T-DNA lines flanking FUS6, one set from the SALK T-DNA line collection (Alonso et al., 2003) and the other from the WiscDsLox collection (Woody et al., 2007). The T-DNA lines for FUS2, FUS7, and FUS11 were from the SALK collection.

In our previous study involving deletion of the MEKK1/2/3 locus, we determined that T-DNA left border sequences were present on both sides of the T-DNA inserts for MEKK1 and MEKK3. This observation suggested that the at least two copies of the T-DNA monomer were inserted at each locus in an inverted repeat orientation. Since inverted repeats have the potential to form hairpin structures that may have the ability to stimulate recombination, we were interested in determining if some of the T-DNA inserts used in the present study had left border sequences on both sides of the insert. PCR reactions using left border-specific primers and gene specific primers were used to map the left border sequences present at each locus. As shown in Fig. 2, at least one of the T-DNA inserts in each pair had left border sequence on both sides of the insert. In the case of FUS7 and FUS11, both the upstream and downstream T-DNA inserts had left border sequences on both sides.

For a given FUS gene, the upstream and downstream T-DNA lines were crossed together to produce F1 plants carrying two T-DNA insertions: one upstream of the FUS gene and one downstream (Fig. 1). If recombination occurs between those two T-DNA insertions during meiosis, one would expect one of the meiotic products to carry a deletion of the FUS gene. Unless both the pollen and egg carried deletion events, one would expect T-DNA directed deletions produced by an F1 plant to be heterozygous in the F2 progeny (Fig. 1). Since the fus phenotype is a recessive condition, it is necessary to perform F3 progeny testing to identify the presence of a T-DNA directed deletion carried by an F2 plant. To streamline this process, we grew nine F2 plants together in a single pot and collected their F3 progeny in bulk. Ca. 1,200 F3 seeds from each pot were then germinated in the dark on moist filter paper and visually scored after four days to screen for the fus phenotype. This represents an average of 133 progeny for each F2 parent. If any of the F2 parents were carrying a heterozygous deletion of the FUS gene, then we would expect 25% of those progeny, or ca. 33 seedlings, to display the fus phenotype. In dark growth conditions, fus seedlings are short and purple with open cotyledons, which would stand in stark contrast to the long, white, etiolated form of wild-type seedlings.

Results of genetic screen

F3 progeny collected from a total of 14,157 F2 parent plants were screened for the fus phenotype as described above in order to find seedlings potentially carrying deletions caused by T-DNA recombination of the FUS gene (Table 2). Since the recombination event that would lead to a deletion could occur during meiosis in either the male of female gametophyte, the population of 14,157 F2 plants that we screened represented 28,314 meioses. The distribution of these F2 plants between the different FUS loci used in the study are shown in Table 2. No fus mutant seedlings were observed in any of the F3 progeny tested.