Somatic and Germinal Recombination of a Direct Repeat in Arabidopsis
Total Page:16
File Type:pdf, Size:1020Kb
Copyright 0 1992 by the Genetics Society of America Somatic and Germinal Recombination of a Direct Repeatin Arabidopsis Farhah F. Assaad' and Ethan R. Signer Department of Biology, Massachusetts Institute of Technology, Cambridge Massachusetts 02139 Manuscript received April 10, 1992 Accepted for publication June 25, 1992 ABSTRACT Homologous recombination between a pair of directly repeated transgenes was studied in Arabi- dopsis. The test construct included two different internal, non-overlapping deletion alleles of npt (neomycin phosphotransferase) flanking an activeHPT (hygromycin phosphotransferase) gene. This construct was introduced into Arabidopsis by agrobacterium-mediated transformationwith selection for resistance to hygromycin, and two independent single-insert lines were analyzed. Selection for active NPT by resistance to kanamycin gave both fully and partly (chimeric) recombinant seedlings. Rates for one transgenic line were estimatedat <2 X 10-5 events per division for germinal and>1 0-6 events per division for somatic recombination, a much smaller difference than between meiotic and mitotic recombination in yeast. Southern analysis showed that recombinants could be formed by either crossing over or gene conversion. A surprisingly high fraction (at least 2/17) of the recombi- nants, however, appeared to result from the concerted action of two or more independent simple events. Some evolutionary implicationsare discussed. ENOMES are not static but rather in constant suggested for other organisms (HILLISet al. 1991), G dynamic flux. That is especially so in plants, can either remove a mutation from a gene copy or which lack a permanent germline and instead produce conversely allow a new mutation to spread through gametes from somatic lineages. Thus changes in the the repeat population. genome of a somatic plant cellmay eventually be Repeated sequence recombinationhas been studied passed, in what amounts to a Lamarckian fashion, extensively in bacteria (PETESand HILL 1988) and through meiosis to gametes and thence to individual yeast (PETESand HILL 1988; KLEIN 1988). For higher progeny. eukaryotes, somatic repeated sequence recombination Plant genomes have a high content of repeated has beenstudied in animal cell culture(BOLLAG, sequences and are unusually variable. For example, WALDMANand LISKAY 1989) as well as in plants DNA content per nucleus varies over three orders of (PETERHANSet al. 1990; GALet al. 1991; TOVARand magnitudeamong known plant species (FLAVELL LICHTENSTEIN1992). In tobacco protoplasts, somatic 1985). Thisseems unlikely to represent a correspond- recombination was observed between two non-over- ing variation in number of expressed genes,especially lappingterminal deletions of the NPT (neomycin as repeated sequences are much more abundant in phosphotransferase) gene in either direct (PETERHANS plants with larger genomes (FLAVELL 1985). Although et al. 1990) or inverted repeat (TOVARand LICHTEN- in general the function of most repeated sequences is STEIN 1992), and shown by sequence analysis to be obscure, such DNA has a strong effect on genome conservative (PETERHANSet al. 1990). In Brassica nu- plasticity. Any genome structure may be altered by pus carrying an integrated multimer of cauliflower mechanisms such as DNA transposition and amplifi- mosaic virus (CaMV), the crossover required to excise cation, but repeated sequences, whether unexpressed the virus genome was scored by virus production in or in gene families, may promote heritable alteration mature leaves, and provided evidence for mismatch by homologous recombination as well. Thus crossing repair in somatic recombination(GAL et al. 1991). over between repeated sequences can either increase Germinal repeated sequence recombination in higher or decrease the numberof repeat copies (SMITH1976) eukaryotes has been reported in Drosophila in a tan- and can fix alterations by genetic drift (SZOSTAKand dem repeat of the alcohol dehydrogenase (ADH)gene WU 1980). Indeed, in at least one plant species, flax, (HIPEAU-JACOTTE,BRUTLAG and BREGEGERE 1989), selection by nutrient stress favors very rapid heritable and in reversion of the bar and white-ivory mutations alterations in copy number of rDNA repeats, presum- (STURTEVANT1928; BOWMAN 1965; BECKER 1975; ably by homologous recombination (SCHNEEBERGER KARESS and RUBIN 1982). Two repeatedsequence and CULLIS1991). Likewise gene conversion in re- recombinantsthat might have beengerminal have peated sequences, perhaps with a directional bias as also been reported in tobacco (TOVARand LICHTEN- STEIN ' Present address: University of Munich, Institute of Genetics and Micro- 1992). biology, Maria-Ward Strasse lA, D-8000 Munich 19, Germany. As a first step toward evaluatingthe role of repeated Genetics 132 553-566 (October, 1992) 554 Assaad F. F. and E. R. Signer sequences in the evolution and plasticity of plant ge- BamHI site between T-DNA left and right borders in broad nomes, we have investigated homologous recombina- host range vector pAC547 (a pRK290 derivative obtained from ABDULCHAUDHURY, CSIRO, Canberra) for Agrobac- tion between a pair of directly repeated genes intro- terium-mediated plant transformation. duced as a transgenic constructin Arabidopsis thaliana. Plant transformation:pDIR was introduced into the recA Arabidopsis is particularly well suited for this as its Agrobacterium strain AGLl (obtained from ROBERTLUD- small size, rapid growth and ease of handling allow WIG,University of California, Santa Cruz) by triparental detection of both somatic and germinal recombinants mating according to VAN HAUTEet al. (1983) with pRK20 1 3 as helper plasmid. Arabidopsis ecotypeBe-0 cotyledon, stem even at low frequency. Here we report somatic and and leaf explants were transformed with AGLl(pD1R) with germinalrecombination frequencies, calculate ap- selection on 20 pg/ml hygromycin (Calbiochem),according proximaterecombination rates, and characterize a to SCHMIDTand WILLMITZER(1988) except that regener- number of recombination products. ated shoots were elongated on hormone-free medium and In yeast, the eukaryotic organism in which recom- transferred to soil for seed setafter 10 days onroot inducing medium. Tobacco leaf disc transformation was according to bination has been studied best, meiotic repeated se- HORSCHet al. (1985). quence recombination rates exceed mitotic rates by Screen for recombinants: We anticipated that recombi- several hundredfold(KLEIN and PETES198 1 ;JACKSON nation rates might be as low as 10-5, so that a large number and FINK 1981). In a tobacco inverted repeat, somatic of seed would have to be screened. In order to have homo- and meiotic frequencies have been reported as com- geneous pools of seeds for individual experiments, we am- plified single T1 seeds over two generations as follows. T1 parable (TOVARand LICHTENSTEIN1992), but assay seed from primary transformants were surface sterilizedand conditions differed (regenerating protoplasts vs. ger- plated on hygromycin. Hyg' seedlings were transferred to minating seedlings, respectively), and the meiotic fre- soil individually and their T2 progeny seed harvested. T2 quencies were based on only two recombinants that seed from a single T1 seedling were planted in a large flat. might instead have been somatic (in the parent spo- Cross fertilization was prevented by shielding of the flats with Saran wrap, with small holes for aeration. Plants were rophyte). grown at 22" under constant light. Pools of roughly lo5T3 In the Arabidopsis direct repeat studied here, we seeds were harvested from each flat. Each pool of T3 seed have been ableto estimate both somatic and germinal from a singleflat was surface sterilized individually and rates, albeit only roughly, in the same conditions of plated on MS medium containing0.5% agarose, 1% sucrose germinating seedlings. We findthat in contrast to and 75 pg/ml kanamycin at a density of 1500 seeds/Petri plate. Plates were incubated at 20" with 16 hr day/8 hr yeast, Arabidopsis germinal and somatic rates differ night cycles. Both fullygreen and green-sectored (chimeric) by at most an order of magnitude. We have charac- recombinants were found (see text). Fully green recombi- terized recombinants produced by crossing over as nants were scored after 6 to 7 days and transferred to soil. well as gene conversion. A relatively high fraction of Sectored recombinants were scoredafter 2 weeks and trans- the recombinants appear to be the result of the con- ferred to kanamycin-free MSO or callus/regeneration (HORSCHet al. 1985) medium. Frequenciesof recombinants certedaction of two or more independentsimple were expressed as the number of recombinants divided by events. the total number of seeds plated, with total number calcu- lated as grams platedX 35,000 seeds/gram (for this ecotype, MATERIALS AND METHODS allowing for seeds lost during surface sterilization) X ger- mination efficiency. DNA construct: All genes in the construct are expressed Analysis of recombinants: T4 seeds were harvested from from the cauliflower mosaic virus P35S promoter and no- recombinants that had been transferred to kanamycin-free paline synthaseno&' terminator. The direct repeat plasmid medium. Those seeds were surface-sterilized, and plated pDIR was constructed in six steps, as follows. First, the 3'A on MSO containing 50 pg/ml kanamycinor 20 pg/ml hygro- 284-bp npt deletion was obtained by a partial NaeI digestof mycin. Resistant and sensitive progeny were scored after pFA39 (ASSAADand SIGNER1990), removing bp 665 to