a system for insertional mutagenesis and chromosomal...

The Plant Journal (1995) 7(4), 687-701

TECHNICAL ADVANCE

A system for insertional mutagenesis and chromosomal rearrangement using the Ds transposon and Cre-lox

Brian I. Osborne, Uwe Wirtz t and Barbara Baker* Plant Gene Expression Center, University of California, Berkeley and USDA-ARS, 800 Buchanan St, Albany, CA 94710, USA

Summary

A system for insertional mutagenesis and chromosomal rearrangement in Arabidopsis has been developed. The T- DNA vectors are based on the maize trsnspoeen Ds, Iox sites from the Cre-lox site-specific recombination system, and transcriptional fusions expressing Ac transposese or Cre recombinese. The engineered transposon is termed Dslox. Transposed Dslox insertions were created by crossing plants bearing Dslox with plants expressing Ac transposase, then simultaneously selecting for excision and reinsertion in F 2 seedlings using the herbicides chlorsuifuron and phosphonothricin, respectively. 1=2 plants bearing stable Dslox insertions were identified by scoring for the absence of the Ac transposese T-DNA, using a novel, visual marker in that T-DNA. Two independent Dslox insertions were characterized and placed 5.6 and 16.6 ¢M from their T-DNAIox, which mapped close to m506 on chromosome 4. Plants bearing either of the two different transposed Dsloxs and T-DNAIox were crossed to plants expressing Cre recombinese, which catalyzed recombination between the Iox site in transposed Dslox and the Iox site in T- DNAIox. Lox--Iox recombinants were identified selectively amongst progeny of these crosses. Molecular and genetic analysis of the Iox-lox rearrangements indicated that both were inversions. The smaller inversion was germinally transmitted from generation to generation as a simple trait, whereas the larger inversion was not transmitted to progeny of plants bearing the rearrangement.

Introduction

Transposons are ubiquitous genetic elements with the property of mobility. Their existence was deduced by McClintock, whose work was based on observations of mutable phenotypes in maize tissues (McClintock, 1948,

Received 22 August 1994; revised 19 December 1994; accepted 23 December 1994. *For correspondence (fax +510 559 5718). tPresent address: Forschungsinstitut Borstel, Abt. Zellbiochimie, Parkalee 22, D-2061 Borstel, Germany.

1950). She named the master, trans-activating element Activator (Ac), and the defective element under its control was named Dissociation (Ds). Studies in maize have shown that Ac often transposes to sites linked to its origin, in roughly half of the transpositions (Dooner and Belachew, 1989; Greenblatt, 1984). Ac has been cloned (Fedoroff eta/., 1983) and introduced into dicotyledonous plants, where it transposes and remains structurally intact (Baker et al., 1986; Schmidt and Willmitzer, 1989). Both Ac and Ds transpose to linked sites frequently in the transgenic plants bearing these elements (Bancroft and Dean, 1993a; Jones et al., 1990; Osborne et al., 1991).

An effective way of mobilizing these transposons is the 'two element' approach (Hehl and Baker; 1989; Lassner et al., 1989; Masterson et al., 1989), where two transgenes are introduced into independent lines: one which expresses the Ac transposase protein, and a second Ds transposon construct that may contain markers or selectable genes (for example, Fedoroff and Smith, 1993; Long eta/., 1993a). In this way Ds transposes in F1 plants derived from the cross of Ds-bearing plants and plants expressing Ac transposase, and is stabilized in F 2 progeny plants that have lost the transposase gene by segregation.

Our 'two element' system is designed for the efficient generation of insertional mutations. We have chosen selections and marker genes that are effective in the greenhouse, using inexpensive reagents for selection of plants bearing transposed Dslox (tr-Dslox). These readily available herbicides are applied to seedlings in terra, obviating the need for in vitro steps. Resistant plants are inspected visually to assess the presence or absence of the Ac transposase T- DNA, without the need for chromogenic or in vitro assays such as those that detect luciferase (Ow et al., 1986) or ~- glucuronidase (Jefferson et al., 1987) expression. We use only those Dslox plants in our crosses that contain single- copy T-DNAIox-Dslox T-DNAs, making selections for tr- Dslox maximally effective, as discussed below.

We have modified Ds to include sites for site-specific recombination derived from the Cre-lox system, making Dslox. The components of Cre-lox are the Iox site, a palindromic site 34 bp in length with an asymmetric core of 8 bp, and Cre recombinase, which catalyzes recombination between two Iox sites (Abremski et al., 1983; Hoess et al., 1982). The outcome of recombination is dependent on the orientation of the two Iox sites and the contiguity of the substrate DNAs, and may be either deletion, inversion, or translocation. Cre-dependent Iox-lox recombination occurs

687

688 Brian L Osborne et al.

efficiently in dicots, as assessed by recombination between two /ox sites integrated proximally at one chromosomal locus (Bayley et al., 1992; Odell eta/., 1990; Russell et al., 1992). Furthermore, recent work has shown that Cre can catalyze Iox-lox recombination between T-DNA-borne sites on different chromosomes in tobacco (Qin et al., 1994).

We have extended these observations by showing that /ox-lox recombination occurs efficiently between sites that have been dispersed in the Arabidopsis genome by Dslox transposition. We have identified plants bearing /ox-lox rearrangements by making the recombination event selectable, and the ease with which we can identify such recombinant plants suggests that we will be able to use our existing lines to generate a variety of chromosomal rearrangements.

The application of these methods is particularly suitable to Arabidopsis. This popular weed has a small genome, with a haploid size of approximately 80 Mb (Meyerowitz, 1992). Thus, useful sets of rearrangements for mapping purposes may be generated with a reasonable amount of effort. The existence of effective methods for placing cloned sequences on the Arabidopsis genetic map (Chang et aL, 1988; Lister and Dean, 1993) means the elements of this system can be placed with respect to other cloned sequences. Furthermore, the map is becoming increasingly integrated, so that characterized T-DNA/ox-Dslox T-DNAs, Ds/ox insertions, or Iox-lox rearrangements can be placed relative to useful or interesting mutations.

In addition, we are engaged in the mapping of our T- DNA/ox-Ds/ox T-DNAs, which places these loci on the Arabidopsis integrated map. This step is necessary for targeted tagging, where a researcher chooses to mutagenize a particular mapped locus or chromosomal region. The locus in this case may be either a mapped gene that had been identified through a corresponding, mutant allele or a mapped, cloned sequence without a genetic identity, such as an Expressed Sequence Tag (EST, Hofte et al., 1993). We propose that a large collection of such mapped T-DNA/ox-Ds/ox T-DNAs should enable the insertional m utagenesis of most of the loci in the Arabidopsis genome, and that such a collection would be a valuable resource for the community (as in Knapp et al., 1994 and Thomas eta/., 1994).

Results

The T-DNAIox-Dslox vector

We designed a T-DNA vector that would allow us to select for Dslox transposition (Figure la), which may create mutations (Bancroft et al., 1993b; Long et aL, 1993a). The Dslox transposon is carried within a T-DNA bearing a Iox site and other useful elements, named T-DNAIox. The Dslox contains a 35S-bar transcriptional fusion (Botterman and

a.

" ,, [Ppt"] [Gin], , ,

I ~ Iox site

surA B su F Nos-N tll

b.

[Cs] [KnR]

T-ONAIox -Dslox T-ONA

C.

35S2-Ac U 35S~R-Lc 1135S-HygB I [R-Lc] [Hyg.]

Ac transposase T-DNA

35S2-Cre ~135S~-R-Lc I~ 35S-HygB I [R-Lc] [HygR]

Cre recombinase T-DNA

Figure 1. Schematic map of the T-DNA vectors. (a) The transgenes and relevant sites contained within the T-DNAIox-Dslox T-DNA. (b and c) The transgenes contained within the 35S-Ac transposase and Cre recombinase T-DNAs. The phenotype conferred by a transgene is indicated in the brackets. (Cs) 35S-surA,B transcriptional fusion that may confer chlorsulfuron resistance, (Gn) gentamicin-3-N-acetyltransferase ORF that may confer gentamicin resistance, (Kn R) Nos-Npt-II transcriptional fusion that confers kanamycin resistance, (Ppt R) 35S-bar transcriptional fusion that confers phosphonothricin resistance, (R-Lc) 35S-R-Lc transcriptional fusion, (Hyg R) 35S-HygB transcriptional fusion that confers hygromycin resistance.

Leemans, 1988; Thompson et aL, 1987), which confers resistance to the herbicide phosphonothricin (Ppt, or glu- phosinate ammonium). The transposon was inserted into a 35S-surA,B fusion (Lee et aL, 1988; Mazur et aL, 1987), which confers resistance to the herbicide chlorsulfuron (Cs, or sulfonylurea). The insertion of Dslox into T-DNAIox inactivates 35S-surA,B, and Dslox excision confers a Cs R phenotype.

The Iox site in Dslox is adjacent to a promoterless gentamicin-3-N-acetyltransferase AAC(3) ORF (AAC, Hayford et aL, 1988), and the Iox site in T-DNAIox is adjacent to a 35S promoter. Thus, Iox-lox recombination may fuse 35S sequences to the AAC ORF, creating a 35S- AAC transgene that confers resistance to gentamicin (Gn). The Gn a selection was applied to seeds germinating in vitro. We reasoned that such a selection may be necessary for the isolation of Iox-lox recombinants in instances where Iox-lox recombination occurs infrequently. In addition, both T-DNAIox and Dslox contain bacterial markers (supF, Phadnis et aL, 1989) and yeast genes (LEU2, HIS3) for selective cloning in Escherichia coil and Saccharomyces cerevisiae, as well as different 'rare cutter' sites for cloning and mapping purposes (see Experimental procedures).

Ds/Cre-lox system for insertion and rearrangement 689

We introduced the T-DNAIox-Dslox vector into Arabidopsis using Agrobacterium-mediated T-DNA transformation (Valvekens et al., 1988). The selectable marker for the transformation was Nos-Npt-II-Nos 3' (Rogers eta/., 1987, Figure la), which conferred resistance to kanamycin (Kn). Eighty-nine independent T-DNAIox-Dslox transformants were recovered. Plants bearing single locus, single-copy T-DNAIox-Ds/ox T-DNAs were identified by Southern blot analysis of transformants using probes from the left and right ends of the T-DNA (data not shown) and by the segregation of T-DNA marker genes. We reasoned that if multiple-copy, single-locus T-DNAs were used then a single Dslox excision without reinsertion would satisfy the double herbicide selection discussed below, and Ac or Ds excision without insertion can occur frequently in transgenic dicots bearing Ac (Honma eta/., 1993; Jones et al., 1990; Masterson eta/., 1989). In addition, the ideal substrates for the action of Cre recombinase are the single /ox site in one tr-Ds/ox and the single Iox site in a single- copy T-DNA/ox. If the genome contained more than two /ox sites the consequences of recombination could be varied or complex, and the resultant rearrangement might not be useful. Twenty-six of 89 (29%) T-DNAs were shown to be single-copy by Southern analysis and by the 3:1 segregation of the T-DNA/ox Kn R and Dslox Ppt R markers (data not shown).

The Ac transposase and Cre recombinase T-DNA vectors

T-DNA vectors were constructed to carry transgenes that express Ac transposase and Cre recombinase (Figure lb and c). The Ac transposase and Cre recombinase coding regions were fused to the 35S promoter of CaMV or the 'double' 35S promoter (Kay et al., 1987) for strong expression. The 'double' version of the 35S promoter has been previously shown to increase Ds excision frequency in Arabidopsis (Grevelding et al., 1992). In addition, the promoter in the 'double' 35S-Ac transposase fusion contains the 'mutB' point mutation (Qin and Chua, 1991), which also increases 35S expression.

Both Ac transposase and Cre recombinase T-DNAs also carry a useful visual marker gene, a 35S-R-Lc transcriptional fusion (Lloyd et al., 1992). The R-Lc gene product is a transcriptional activator in maize, where it positively regulates anthocyanin biosynthesis (Ludwig et al., 1989). This fusion confers two obvious phenotypes on Arabidopsis plants bearing the transgene: an increase in trichome number, which gives a hirsute appearance, and an increase in anthocyanin concentration which results in more darkly pigmented tissues. Both phenotypes can be detected from the time of leaf emergence through senes- cence, independent of diverse growth conditions. This direct visual assessment saves significant time and labor as compared with a histochemical assay based on lucifer-

ase (Ow et aL, 1986) or ~-glucuronidase (Jefferson et al., 1987) transgenes. The phenotype conferred by the 35S-R- Lc transgene is stable, heritable, and penetrant, and does not appear to affect fertility. The transgene also does not appear to affect Dslox transposition adversely, when compared with trans-activation by lines possessing the same 35S-Ac transposase fusion but lacking the 35S-R-Lc gene (Honma et a/., 1993, and data not shown). Thus, plants bearing either an Ac transposase or Cre recombinase T-DNA were identified by detection of the hirsute, darkly pigmented phenotype (R-Lc+). Conversely, plants without Ac transposase or Cre recombinase T-DNAs were identified by the normal appearance of their leaves (R-Lc-).

The Ac transposase and Cre recombinase vectors were introduced into Arabidopsis using the hygromycin resistance (Hyg R) phenotype conferred by a 35S-hygB fusion (Van den Elzen eta/., 1985, Figure lb and c). Transformants bearing single-locus T-DNAs were identified by the 3:1 segregation of T-DNA-borne Hyg R and R-Lc+ phenotypes (data not shown).

Generation of 1098 independent 35S-Ac × Dslox F1 plants

We selected 13 single-copy T-DNAIox-Dslox lines and 43 single-locus 35S-Ac transposase lines for crossing. The 35S-Ac plants were used as females in these crosses and each 35S-Ac transposase line was crossed to at least three different T-DNA/ox-Ds/ox lines. We reasoned that maximizing the number of different pairwise combinations of 35S-Ac and T-DNAIox-Dslox lines in the Fls would enable us to assess the activities of the lines most accur- ately. The seeds from these crosses were germinated on Kn, and 1098 Kn R, R-Lc+ F1 plants were grown to maturity. The F 2 generation seed harvested from these plants was used for the following studies on germinal transposition of Ds/ox.

Herbicide treatments select for Dslox transposition

Three hundred F 2 seeds from each of the 1098 35S-Ac × DsloxF1 individuals were sown in soil. Seedlings germinating in terra were treated with Cs and Ppt in a solution of fertilizer (see Experimental procedures), and Cs R, Ppt R plants were scored for the R-Lc phenotype. Three hundred of the 1098 F 2 families (27.3%) contained at least one Cs R, Ppt R, R-Lc- F2 plant, which is the phenotype expected of a plant bearing a germinally transmitted Dsloxtransposition. Certain 35S-Ac T-DNAs appeared to be inactive (seven of 43, or 16.3%), as F1 plants containing these T-DNAs bore no Cs R, Ppt R F2 progeny. However, other 35S-Ac T-DNA lines generated Cs R, Ppt R, R-Lc-F 2 plants frequently when used as parents. The data from these more active 35S-Ac lines suggest that one may expect that 1/2 of the 35S-Ac x

690 Brian I. Osborne et al.

,, ..

. , o

IV

10 cM Figure 2. Location of the 522 T-DNA/ox-Dslox T-DNA, tr-Ds/ox 'A', and tr-Dslox 'B' on Arabidopsis chromosome 4.

Ds/ox Fls bearing one of these active 35S-Ac T-DNAs will give rise to at least one Cs R, Ppt R, R-Lc- F 2 progeny plant, when 300 F z seeds are sown and selected (data not shown). Both 35S-Ac and 'double' 35S-Ac transcriptional fusions appeared to act as effective Ac transposase parents equally well (data not shown).

We confirmed that Cs R, Ppt R, R-Lc- F2 plants bore germinally transmitted tr-Ds/ox by Southern analysis. In this experiment we isolated DNA from 118 randomly chosen Cs R, Ppt R, R-Lc- F2 individuals descended from different 35S-Ac × T-DNAIox-Dslox FlS. One hundred and twelve of 118 (95%) clearly contained tr-Dsloxand no Dslox remained 'in-place' in the T-DNAIox (data not shown).

Characterization of the 'A' and "B" tr-Dslox insertions

We have examined two Dslox insertions arising from the B22 T-DNA/ox-Dslox locus, and the resultant /ox- Iox rearrangements. The B22 T-DNAIox-Dslox locus was selected from a group of cloned, mapped loci. The B22 T- DNA/ox-Dslox T-DNA and the genomic region flanking its right border were cloned in ~, using the selection for growth of an amber phage in a sup ° host enabled by the supF marker within Dslox (see Experimental procedures). The B22 locus was placed on the Arabidopsis RFLP map by examining the segregation of RFLPs detected by the B22 genomic flanking region using DNAs extracted from recombinant inbred lines supplied to us by C. Dean and C. Lister (Lister and Dean, 1993). The resultant data were analyzed by the MapMaker 3.0 program (Lander et al., 1987) on a Sun workstation, which placed the B22 locus close to the anonymous marker m506 on chromosome 4, as shown in Figure 2.

Dslox insertions were generated by crossing a plant with the B22 T-DNAIox-Dslox T-DNA to plants bearing a 35S- Ac transposase T-DNA (35S2-Ac, in the scheme shown in Figure 3). An F1 plant containing both T-DNAs was identified by its Kn R, R-Lc+ phenotype. This plant was allowed to self and 700 F 2 seeds were plated on Cs and

Ppt to select for seedlings bearing tr-Dslox. Eighty-three of 700 seedlings were Cs R, Ppt R, R-Lc-. Seventeen randomly selected Cs R, Ppt R, R-Lc- F2 plants from this family were examined by Southern blotting using Dslox sequence as a probe (Figure 3). Each of the 17 plants contained a hybridizing restriction fragment distinct from the fragment characteristic of Ds/ox inserted into T-DNAIox. Two different fragments, corresponding to the putative 'A' and 'B' tr- Ds/oxs, can be seen in Figure 3.

The flanking regions to the putative 'A' and 'B' tr-Dsloxs were cloned to demonstrate that we had generated bona fide transpositions and to aid in our analysis of /ox-lox recombination. The genomic regions flanking the 'A' and 'B' tr-Dsloxs were isolated by IPCR (Triglia et al., 1988) and cloned into plasmids for sequence analysis. This analysis showed that the IPCR products were formed as predicted, with the expected restriction enzyme sites marking the junctions between Ds/ox sequence and unknown, genomic sequence. We also found that the 5' and 3' termini of tr- Dslox'A' had sustained deletions of 5 and 4 bp, respectively. In addition, the 3' terminus of tr-Ds/ox 'B' had sustained a 4 bp deletion. We expect that these deletions would prevent the 'A' and 'B' tr-Dsloxs from excising (Coupland et al., 1989; Hehl and Baker, 1989). Neither of these 2 Dslox insertions were flanked by the 8 bp direct duplication characteristic of Ac and Ds insertion (Muller-Neumann et al., 1984; Pohlman et al., 1984; Sutton et al., 1984). The IPCR products corresponding to the 'A' and 'B' insertions were shown to contain the genomic DNA flanking those insertions by Southern blot analysis of DNA isolated from plants bearing the 'A' or 'B' tr-Dsloxs (data not shown).

Tr-Dslox "A" and "B" are linked to their T-DNA

The genetic distances between the B22 T-DNAIox and the 'A' and 'B' tr-Dsloxs were determined by placing the genomic regions flanking these three elements on the Arabidopsis RFLP map. The genomic region flanking tr- Ds/ox 'A' was used to screen YAC libraries of Arabidopsis

Ds/Cre~ox system for insertion and rearrangement 691

Figure 3. A scheme that generates both Dslox insertions and /ox-lox chromosomal rearrangements using selectable and screen- able marker genes. 35S-Ac transposase and Dslox-bearing plants were crossed, and the resultant Fls were allowed to self. Their F2 seedlings were selected with chlorsulfuron (Cs) and phosphonothricin (Ppt). Plants containing the 35S- Ac transposase T-DNA were identified by their R-Lc+ phenotype and discarded. The South- ern autoradiogram shows an analysis of Cs R, Ppt R, R-Lc- siblings bearing either troDs/ox 'A' or tr-Dslox 'B', as revealed by hybridization with a Dslox-borne AAC probe. The leftmost lane shows the size of hybridizing fragment when Dslox rests 'in place' within the B22 T- DNAIox. F 2 plants were selfed and F3 plants were examined for the presence of mutations. In addition, CS R, Ppt R, R-Lc- F2 plants were crossed to plants bearing a 35S-Cre recombinase T-DNA, and the resultant Fls were allowed to self. Their F2 seedlings were selected with gentamicin (Gn). Plants lacking the 35S-Cre recombinase T-DNA (Gn R, R-Lc-) bore germinally transmitted /ox-lox rearrangements. (Kn R) kanamycin resistant, (Cs R) chlorsulfuron resistant, (Plot R) phosphonothricin resistant, (Gn R) gentamicin resistant, (R-Lc+) hirsute phenotype of plants bearing the 35S-R-Lc transcriptional fusion, which resides in the 35S-Ac transposase and Cre recombinase T-DNAs, (R-Lc-) wild type phenotype of plants lacking the 35S-R-Lctran- scriptional fusion.

genomic DNA and the results of these experiments placed tr-Ds/ox'A' proximal to the RFLP marker g3843 on chromosome 4 (Schmidt, West and Dean, John Innes Centre, personal communication). The genomic region flanking tr- Ds/ox 'B' was placed on the Arabidopsis RFLP map by examining the segregation of RFLPs detected by this probe using DNAs extracted from recombinant inbred lines supplied to us by C. Dean and C. Lister (Lister and Dean, 1993). The resultant map of Arabidopsis chromosome 4 is shown in Figure 2. The distances between the B22 T-DNAIox and the 'A' and 'B' tr-Dsloxs are consistent with the observation of frequent, linked transposition by Ds in Arabidopsis (Bancroft and Dean, 1993). We also determined the recombination frequencies between the Cs R and Ppt R marker genes in the B22 T-DNAIox and the 'A' and 'B' tr-Ds/oxs respectively. The values from these experiments are in agreement with the positions of the elements on the RFLP map (data not shown).

Lox- lox recombination occurs efficiently in planta

The T-DNAIox and Dslox elements were designed to al low a selection for the /ox - lox recombination event. The Dslox contains the promoterless ORF of the E. coil gentamicin- 3-N-acetyltransferase AAC(3) gene adjacent to its single /ox site (Figure la). This gene confers a gentamicin-resist-

ant (Gn R) phenotype on Arabidopsis when properly expressed (Hayford et al., 1988). The single Iox site in the T-DNA/ox is immediately adjacent to the 35S promoter (Figure la). Thus, precise Iox-/ox recombination can fuse the AAC ORF to the 35S promoter, activating it and confer- ring Gn R. Plants bearing the B22 T-DNAIox and either the 'A' or 'B' tr-Dslox were crossed to plants bearing the 35S- Cre recombinase transgene in order to catalyze Iox-lox recombination. F 1 seedlings bearing the B22 T-DNA/ox, the 'A' or 'B' tr-Ds/ox, and the Cre recombinase T-DNA were identified by their Gn R phenotype, which was due to somatic Iox-/ox recombination. As expected, all F1 Gn R seedlings also displayed the R-Lc+ phenotype that should be conferred by the 35S-R-Lc transgene in the Cre recombinase T-DNA (Figure lc). Gn R, R-Lc+ F 1 plants were grown and allowed to self, and their seed was sown on Gn to select for seedlings bearing Iox-lox rearrangements.

Table 1 presents the data on segregation of the Gn and R-Lc phenotypes in F 2 families from 19 F 1 plants derived from crosses of Cre recombinase-expressing plants to plants bearing the B22 T-DNA and either the 'A' or 'B' tr- Dslox. If somatic Iox-loxrecombination occurred efficiently and conferred Gn R we would expect that up to 75% of F 2 seedlings from a given F 1 would be Gn R, that is, any seedling that bore the Cre recombinase T-DNA (and is R- Lc+) could be Gn R. Table 1 shows that 53.7% (tr-Dslox


Table 1. Frequency of Gn a F2 progeny arising from F~ plants bearing the B22 T-DNAIox, a tr-Dslox, and a Cre recombinase T-DNA

F 1 plant Total F 2 Total Gn s F 2 Total Gn R, R-Lc+ F 2 Total Gn R, R-Lc- F 2 % Gn R, R-Lc- % Gn R, R-Lc+

tr-Dslox 'A' X Cre a 2391 1082 1284 25 1.04 53.7 tr-Dslox 'B' X Cre b 3815 1883 1932 0 0 50.6 tr-Dslox 'A'-selff 90 90 0 0 0 0 tr-Dslox 'B'-selP 15 15 0 0 0 0 B100-2 d 22 0 n/a 22 100 n/a

aF 2 seed from 14 independent F1 plants bearing tr-Dslox 'A', the B22 T-DNAIox, and a Cre recombinase T-DNA were analyzed. bF 2 seed from 5 independent F~ plants bearing tr-Dslox 'B', the B22 T-DNAIox, and a Cre recombinase T-DNA were analyzed. CThese lines contained the B22 T-DNA/ox and either tr-Dslox 'A' or 'B' prior to Iox-lox recombination. dB100-2 was homozygous for a 35S-AAC transcriptional fusion borne in a T-DNA (see Experimental procedures).

'A' × Cre) or 50.6% (tr-Dslox'B' × Cre) of F2 seedlings from these crosses were Gn R, R-Lc+, indicating that somatic Iox- /ox recombination occurred efficiently between the B22 T- DNA/ox and both the 'A' or 'B' tr-Ds/oxs. Furthermore, all seedlings from the selfing of plants bearing the B22 T- DNA and either the 'A' or 'B' tr-Ds/ox prior to Iox-lox rearrangement (tr-Ds/ox 'A'--self, tr-Dslox 'B'--self) were Gn s as expected. The line labeled 'B100-2' shows a positive control where we made a 35S-AAC transcriptional fusion in vitro using Cre recombinase (see Experimental procedures) and introduced this transgene into Arabidopsis by Agrobacterium-mediated transformation (Valvekens et al., 1988). This particular line, B100-2, was homozygous for the 35S-AAC ToDNA as all its progeny were Gn R. No Gn R seedlings were found when seeds from No-O or Cre recombinase-expressing plants were grown on Gn (data not shown).

The B22 T-DNAIox-tr-Dslox "A' rearrangement is germinally transmitted

Lox-lox recombination between the B22 T-DNAIox and the 'A' or 'B' tr-Dsloxs may generate either a deletion or an inversion, as all three elements lie on chromosome 4. We would not expect large deletions to be germinally transmitted to progeny, as such deletions could remove genes that are essential for the development or viabil i ty of either gametophyte. However, large inversions which have not caused mutations deleterious to the gametes could be transmitted. A germinally transmitted Iox-lox rearrangement would be detected as a Gn R, R-Lc- seedling.

We recovered germinally transmitted rearrangements only in the progeny of F 1 plants containing the B22 T- DNAIox and the 'A' tr-Dslox. Table 1 shows that Gn R, R- Lc- seedlings were found amongst progeny of Gn R, R-Lc+ F 1 plants bearing the Cre recombinase T-DNA, B22 T- DNAIox and the 'A' tr-Dslox. However, no Gn R, R-Lc- seedlings were found amongst progeny of Gn R, R-Lc+ F1 plants bearing the Cre recombinase T-DNA, B22 T-DNAIox, and the 'B' tr-Dslox (tr-Dslox 'B' × Cre). Seedlings that were Gn R, R-Lc- did not contain the Cre recombinase T-

DNA and any Iox-lox rearrangement that they bore must have been germinally transmitted from the previous F1 generation. The incidence of these Gn R, R-Lc- seedlings is the critical point as our interests are in those rearrangements that can be transmitted, demonstrating their genetic utility.

We also attempted to recover germinally transmitted B22-'B' rearrangements by examining the progeny of a Gn R, R-Lc+ F 2 plant bearing the B22 T-DNAIox and the 'B' tr-Dslox. The data from Table 1 state that we examined 2391 seedlings from B22-'A' Fls and 3815 seedlings from B22-'B' FlS. In addition, we examined 2198 seedlings from the selfing of one Gn R, R-Lc+ F2 plant (bearing the B22-'B' rearrange~ment) in an attempt to obtain a Gn R, R-Lc- F 3 seedling (this plant was heterozygous for the Cre recombinase T-DNA, data not shown). One thousand-and-eighty of the 2198 F 3 seedlings (53.7%) were Gn R, R-Lc+. However, as in the previous generation, we were unable to find Gn R, R-Lc- F3 seedlings in the progeny of this Gn R, R-Lc+ F 2 plant which bore the B22-'B' rearrangement.

The B22-'A" rearrangement is transmitted as a simple Mendelian trait

We crossed and selfed Gn R, R-Lc- F 2 plants bearing the B22-'A' rearrangement in order to assess the heritability of the rearrangement. Table 1 shows that the B22-'A' rearrangement was germinally transmitted to F2 progeny at a frequency of 1.04% from 14 F1 parents. The frequency can also be considered as 4.16% when expressed as the ratio of Gn R, R-Lc- seedlings to total R-Lc- seedlings. However, this result is inaccurate as an indication of the heritability of the Iox-lox recombination product as the recombination events occurred somatically in the F1 parents, and the recombination product may have been present in only some fraction of the F1 parents' tissues. The true heritability of the B22-'A' rearrangement was assessed in two ways: as the frequency of Gn R progeny in the selfed progeny of F 2 plants heterozygous for the B22-'A' rearrangement and as the frequency of Gn R progeny amongst progeny of an outcross of F 2 B22-'A'

Ds/Cre-tox system for insertion and rearrangement 693

Table 2. Germinal transmission of the B22-'A' rearrangement

Male Female Total Gn R Total Gn s GnR: Gn s

F1-2 - 95 31 3.06 F2-2 - 55 22 2.5 F7-1 - 101 20 5.05 F12-1 - 73 20 3.65

Totals 324 93 3.46 - 0.95

F1-2 No-O 100 68 1.47 F1-3 No-O 27 39 0.69 F1-5 No-O 70 65 1.08 F2-1 No-O 3 1 3 F2-2 No-O 8 12 0.67

Totals 208 185 1.12 _+ 0.33

rearrangement heterozygotes to wild type No-O plants. The latter experiment is aimed at detecting a defect specific to transmission through the pollen.

Table 2 shows the results of these experiments. Four plants heterozygous for the B22-'A' rearrangement (lines 'F1-2' to 'F12-1') were allowed to self, and their seed was plated on Gn. Four hundred and seventeen seeds were plated, and 324 of these were Gn R. These data can be expressed as a ratio of Gn R to Gn s of 3.48 _+ 0.95, where 3.0 is the expected ratio if the rearrangement was efficiently transmitted through both gametes.

In addition, five plants heterozygous for the B22-'A' rearrangement (lines 'F1-2' to 'F2-2') were outcrossed to wild-type No-O and the seeds from these crosses were plated on Gn. Table 2 shows that a total of 393 seeds were plated, and 208 of these were Gn R. The resulting ratio of Gn R to Gn s in this experiment was 1.12 _+ 0.33, where 1.0 is the expected ratio if the rearrangement was efficiently transmitted through the pollen. These experiments show that the B22-'A' rearrangement was transmitted efficiently through both gametes. Furthermore, plants bearing the B22-'A' rearrangement in heterozygous and homozygous form appeared as wild-type in growth habit and fertility (data not shown).

Lox-lox recombination creates new resistance phenotypes

Accurate inversion formation by Iox-lox recombination should simultaneously create a 35S-AAC transgene (and the Gn R phenotype) and disrupt the 35S-surA,B transgene contained within T-DNAIox. Thus, the recombinant plant's phenotype would be Cs s, and the parental plant would be Cs R. To test this prediction we examined the expression of the relevant transgenes contained in the B22 T-DNAIox and tr-Dslox 'A' before and after Iox-lox recombination. Table 3 presents the resultant data when the seeds from the selfing of the appropriate plants were plated on the

Table 3. Herbicide resistance phenotypes generated by Iox-/ox recombination

Plant CsR:Cs s GnR:Gn s PptR:Ppt s

G 1- la 0:46 33:0 27:0 D87-2 b 35:5 0:35 28:8 No-O 0:26 0:41 0:24 B100-2 c 0:41 47:0 0:60

aGl-1 was homozygous for the B22-'A' inversion. bD87-2 was heteroxygous for both the B22 T-DNAIox and tr-Dslox 'A'. CB100-2 was homozygous for a 35S-AAC transcriptional fusion (see Experimental procedures).

antibiotics (the diagram in Figure 4 shows these genes as they would lie on the chromosome before recombination and after Iox-lox recombination created a chromosomal inversion). Specifically, the phenotype of plants bearing the B22 T-DNAIox and tr-Dslox'A' before recombination is Cs R, Gn s, Ppt R, whereas expression from these transgenes after an inversion event should confer a Cs s, Gn R, Ppt R phenotype. Seeds from a plant that was homozygous for the B22-'A' inversion ('G1-1') were plated on Cs, Gn, and Ppt, separately, and the expected 100% Cs s, 100% Gn R, and 100% Ppt R phenotypes were seen on these three different selections (Table 3). This result is consistent with the proposal that the B22-'A' rearrangement was an inversion, as a deletion should create a plant with a Cs s, Gn R, Ppt s phenotype. Control seedlings from the selfing of a plant heterozygous for both the B22 T-DNAIox and the 'A' tr-Ds/ox prior to Iox-lox recombination (D87-2, which is one of the parents to G1-1) segregated with the expected 3:1 ratio of CsR:Cs s and PptR:Ppt s. The line labeled 'B100- 2' depicts a positive control showing the transmission of Gn R to progeny by a plant homozygous for a 35S-AAC transcriptional fusion made in vitro using Cre recombinase (see Experimental procedures).


Figure 4. PCR analysis of DNAs from Gn R F2 plants bearing the B22-'A' and B22-'B' inversions. (a) The results of PCR using primers 'A' and 'B', where a product 223 bp in size should be synthesized from DNAs containing the 35S- AAC sequence. (b) The results of PCR using primers 'C' and 'D', where a product 410 bp in size should be synthesized from DNAs containing an inversion generated by Iox-lox recombination. The lane 1 ('No-O') reaction contained DNA f rom untransformed Arabidopsis. The lane 2 and 3 ('D87-7', 'D87- 20') reactions contained DNAs from plants bearing the B22 T-DNAIox and tr-Dslox 'A' and tr-Dslox 'B' respectively, prior to Iox-/ox recombination. The lane 4, 5, 6, and 7 ('E1-1', 'E2-1', 'E7-1', 'E12-1') reactions contained DNA from Gn R, R-Lc- F 2 plants. The lane 8 (B22- 'B') reaction contained DNA from a Gn R, R- Lc+ plant. The lane 9 ('B100-2') reaction contained DNA from a plant bearing a 35S- AAC transcriptional fusion created by Cre recombinase in vitro (see Experimental procedures). A schematic diagram of the B22 T-DNAIox and a tr-Dslox before and after Iox-lox recombination and the PCR primers is shown at the bottom. (Cs R) chlorsulfuron resistant, (Cs s) chlorsulfuron sensitive, (Ppt R) phosphonothricin resistant, (Gn s) gentamicin sensitive, (Gn R) gentamicin resistant.

The B22-'A' and B22-'B" rearrangements are both inversions

We examined the DNA from F 2 Gn R plants to prove that the B22-'A' rearrangement was an inversion, and to determine the nature of the B22-'B' rearrangement. We used PCR and Southern blotting to distinguish inversions from deletions, and to determine whether the expected new junction fragments had been generated by Iox-lox recombination. Figure 4 shows the results of PCR analysis of DNAs taken from F 2 Gn R plants which bore /ox-/ox rearrangements. Lanes 4-7 ('E1-1', 'E2-1', 'E7-1', 'E12-1') contained PCR reactions using DNA from Gn R, R-Lc- plants which bore the germinally transmitted B22-'A' /ox-/ox rearrangement. Lane 8 ('E71-1') shows a reaction using DNA from a Gn R, R-Lc+ plant that bore the somatically generated B22-'B' rearrangement, as no plants were found with a germinally transmitted B22-'B' rearrangement. Gn R plants should bear a new, rearranged gene formed by the precise fusion of 35S promoter and AAC sequences, and

a PCR product of a predicted size of 223 bp should be

formed only from DNA taken from these plants when using

PCR primers A and B. Figure 4(a) shows this expected result, as well as the expected finding that this product

was not found in control plants, which are those plants

bearing the B22 T-DNA/ox and either the 'A' or 'B' tr-Dslox

prior to /ox-lox recombination (lanes 2 and 3, 'D87-7',

bearing tr-Ds/ox 'A', and 'D87-20', bearing tr-Dslox 'B').

Lane 9 shows the result of PCR using DNA taken from a

positive control plant, and contains a product of the

expected size. This plant bore a 35S-AAC transcriptional fusion that was constructed in vitro using Cre recombinase

(see Experimental procedures) and introduced into Arabidopsis

using Agrobacteriurn-mediated transformation. If the Iox-lox rearrangement between B22 T-DNAIox and

the 'A' tr-Dslox (B22-'A') or the 'B' tr-Ds/ox (B22-'B') were an inversion we would expect to see a PCR product of a

predicted size (410 bp) after PCR with primers C and

D. This product corresponds to the second of the two


breakpoints flanking an inversion, and would not be found in Iox-/ox recombinants bearing deletions, which have single, remnant breakpoints. Lanes 4-8 in Figure 4(b) ('E1- 1', 'E2-1', 'E7-1', 'E12-1', 'E71-1') show that this inversion- specific product was found when PCR was performed on DNAs from F 2 Gn R, R-Lc- plants that bore rearrangements between the B22 T-DNA/ox and the 'A' tr-Ds/ox (B22- 'A'), and from DNAs from F 2 Gn R, R-Lc+ plants bearing rearrangements between the B22 T-DNAIox and the 'B' tr- Ds/ox (B22-'B'). Control DNAs from plants that bore these elements prior to /ox-/ox recombination (lanes 2 and 3, 'D87-7', bearing tr-Dslox'A', and 'D87-20', bearing tr-Dslox 'B') did not support synthesis of the inversion-specific PCR product. These analyses are consistent with the proposal that recombination had occurred between the Iox site in the B22 T-DNAIox and the/ox sites in both the 'A' tr-Dslox and 'B' tr-Dslox to form chromosomal inversions in vivo.

We extended these findings by examining the DNAs from F 2 Gn R plants bearing the B22-'A' rearrangement by Southern blotting (Figure 5). In these experiments we used genomic sequences flanking tr-Dslox 'A', T-DNAIox, and Dslox sequences as probes to examine the size differences of genomic restriction fragments adjacent to the/ox sites before and after Iox-lox recombination. Figure 5(a) shows restriction fragment size differences before and after/ox- /ox recombination when a genomic fragment flanking tr- Dslox 'A' at its 5' end was used as a probe. Lane 1 ('No- O') shows the Bcll genomic fragment before Dslox 'A' insertion, in No-O wild type DNA. The size of the Bcll genomic fragment after Ds/ox'A' insertion, and before/ox- /ox recombination, is 4.0 kb (Figure 5(a), 'D87-7'). Lanes 4-7 ('E1-1', 'E2-1', 'E7-1', 'E12-1') contained DNA from independent F 2 Gn R, R-Lc plants, after/ox-/ox recombination. The size of the hybridizing fragment has shifted from 4.0 kb to 4.1 kb, as expected from the restriction map of the Ds/ox. This predicted size would be 4.1 kb regardless of whether an inversion or deletion had occurred due to Iox-lox recombination. It is important to note that the same size fragment is found in these four independent Iox-/ox recombinants, showing that the recombination event is accurate and reproducible. In addition, all four recombinants retain the uninterrupted genomic fragment found in lane 1 (No-O, wild type), indicating that they are all heterozygous for the Iox-lox rearrangement.

Hybridization to a bar probe (Figure 5b) should distinguish a deletion event from an inversion: the bar sequence should be deleted in a Iox-lox mediated deletion and this deletion should be detectable since all four recombinant plants are heterozygous for the rearrangement. However, the bar-hybridizing fragment of 3.2 kb seen before Iox-/ox recombination ('D87-7') should remain unperturbed after an inversion ('E1-1', 'E2-1', 'E7-1', 'E12-1'). Figure 5(b) shows this latter result, again, as in the PCR results, consistent with the B22-'A' rearrangement being an inver-

sion. The lane labeled 'D87-20' contained DNA from a plant bearing the B22 T-DNAIox and tr-Ds/ox 'B' prior to the /ox-lox recombination, and would be expected to contain a hybridizing fragment differing in size from the lane labeled 'D87-7'.

Hybridization to a surA,B probe (Figure 5c) should also distinguish an inversion event from a deletion event. How- ever, in this case the probe was taken from the opposite end of the inversion from the barprobe used in Figure 5(b). Prior to recombination, the surA,B probe hybridized to an expected 3.0 kb Bcll fragment (lane 2, 'D87-7'). If Iox-lox recombination created a deletion, the 3.0 kb Bcll fragment should be lost, whereas an inversion should generate a new junction fragment of 2.9 kb in size. Lanes 4-7 ('E1-1', 'E2-1', 'E7-1', 'E12-1') show the latter result. These three hybridizations unambiguously demonstrate that the B22- 'A' rearrangement was an inversion.

We also used Southern hybridization to further character- ize the B22-'B' rearrangement. We assumed that the Gn R phenotype in a Gn R, R-Lc+ plant bearing the B22-'B' inversion was due solely to somatic Iox-lox recombination that had occurred in some fraction of the Gn R, R-Lc+ plant's cells, since we were never able to detect germinal transmission of the B22-'B' inversion (as evinced by the Gn R, R-Lc- phenotype). However, we were able to detect the B22-'B' inversion junctions by using PCR (Figure 4). The inset on the right in Figure 5(c) shows hybridization by the surA, B probe to DNA from a Gn R, R-Lc+ plant that bore the B22-'B' inversion. This inset shows two hybridization signals: a major fragment of 3.0 kb, corresponding to the surA, B gene prior to Iox-lox recombination, and a predicted minor fragment of 2.9 kb corresponding to the surA, B gene after Iox-/ox recombination. The inset comes from an autoradiogram that had been overexposed to the hybridized filter. We were unable to detect strong hybridization signals corresponding to the products of B22- 'B' Iox-lox recombination that were equivalent to those seen for the B22-'A' inversion in this and other Southern hybridizations. We could only detect the 2.9 kb recombinant Bcll fragment when the filter was greatly overexposed, as in this autoradiogram. This result is consistent with the interpretation that the Gn R, R-Lc+ phenotype of plants that bore the B22-'B' inversion was due solely to somatic Iox- Iox recombination.

In addition, we performed PCR on serial dilutions of genomic DNAs containing the germinally transmitted B22- 'A' inversion and the somatically generated B22-'B' inversion. These experiments were performed to compare the relative concentrations of the B22-'A' and B22-'B' inversion junctions in vivo (data not shown). The primers in these PCR experiments were the same as those used previously (Figure 4). By comparing the intensities of PCR products from reactions using serial dilutions of B22-'A' and B22- 'B'-containing DNAs we estimated that the ratio of the


Figure 5. Southern analysis of DNAs from Gn R F z plants bearing the B22-'A' and B22- 'B' inversions. Lane, 1 ('No-O') contained DNA from untransformed Arabidopsis. Lanes 2 and 3 ('D877', 'D8720') contained DNAs from plants bearing the B22 T-DNAIox and tr-Dslox 'A' and tr- Dslox 'B' respectively, prior to Iox-lox recombination. Lanes 4, 5, 6, and 7 ('E1-1', 'E2-1', 'E7-1', 'E12-1') contained DNA from independent Gn R, R-Lc- F 2 plants bearing the germinally transmitted B22-'A' inversion. (c) An additional lane is inserted on the right. This lane ('E71-1') contained DNA from one Gn R, R-Lc+ F 2 plant bearing the somatically generated B22-'B' inversion. A schematic diagram of the B22 T*DNAIox and a tr-Dslox before and after Iox-lox recombination and the probes for the three hybridizations are shown st the bottom: (a) a genomic fragment flanking tr-Dslox 'A', (b) the bar gene from within Dslox, and (c) a portion of the surA,B gene from within T- DNAIox. This figure also shows the expected BcJI restriction fragments before and after Iox-lox recombination. (Gn R) gentamicin resistant.

relative concentrations of the B22-'A' and B22-'B' inversion junction fragments in vivo was 13.3 _+ 6.0. The actual values of the dilutions of B22-'A'-containing DNAs that yielded junction-specific products of equal intensity to the products from B22-'B'-containing DNAs were 1/8, 1/8, 1/8, 1/8, 3/64, 3/64, 1/16, and 1/16. These results help explain our inability to detect strong hybridizations signals corresponding to the B22-'B' inversion in Southern blots.

Discussion

An efficient system for generating Dslox insertions and chromosomal rearrangements

We have developed and employed a system to insertionally mutagenize and rearrange the Arabidopsis genome. Our strategy is based on three types of vectors: a T-DNAIox


containing a Iox site, a marked Ds/ox transposon, and a Dsloxexcision marker, a T-DNA containing the Ac transposase gene fused to a strong plant promoter, and a T-DNA containing the Cre recombinase gene fused to a strong plant promoter (Figure 1). Both the Ac transposase and Cre recombinase T-DNAs contain a 35S-R-Lc transcriptional fusion (Lloyd et al., 1992), which visually marks the plants containing these T-DNAs. We have used this system to generate 300 independent F 2 families that we have screened for tr-Ds/ox induced insertion mutations. We have also generated two large inversions arising from the B22 T-DNAIoxlocus and two tr-Dsloxs on chromosome 4. These rearrangements arose when Cre recombinase catalyzed recombination between the Iox site in T-DNAIox and the /ox site in a tr-Dslox.

These T-DNA vectors and corresponding transformed Arabidopsis lines have been used for the generation of insertion mutations. This system is particularly well suited for this purpose, for the following reasons. First, Ds/ox transposition is selectable both in vitro and in terra, as shown by the finding that 95% of randomly selected F 2 individuals that are both Ppt R and Cs R, and lack the Ac transposase T-DNA (R-Lc-), bore tr-Dslox. The efficacy of the double selection may partly be due to our utilization of single-copy T-DNAIox-Dslox T-DNAs--if multiple-copy T-DNAs were used then Dslox excision without reinsertion would satisfy the double selection, and Ac excision without insertion can occur frequently in dicots (Jones et al., 1990; Long et al., 1993b).

Second, the 35S-R-Lc gene is very well suited to be a visual marker for the Ac transposase and Cre recombinase T-DNAs. We have found this marker to be heritable, penetrant and easily scored, and it offers a significant savings in effort as compared with luminescent or chromogenic markers such as luciferase (Ow et aL, 1986) or I~-glucuronidase (Jefferson et al., 1987) that could be used for the same purpose.

Both the Dsloxand its starting T-DNA (T-DNAIox) contain Iox sites, which are the substrates for site-specific recombination mediated by Cre recombinase. This reaction occurs efficiently in planta (Bayley et aL, 1992; Odell et al., 1990; Qin et al., 1994; Russell et al., 1992). Thus, Dslox transposition moves a Iox site about the genome, and we have demonstrated that Cre recombinase will act on these sites to generate chromosomal rearrangements. These rearrangements may be inversions, deletions, and translocations.

Lox-lox rearrangements are efficiently generated in planta

Many of the features noted above also make this system a reasonable one for the generation of chromosomal rearrangements. Furthermore, since most of the Cs R, Ppt R,

R-Lc- plants contain only one tr-Dsloxthey may be crossed immediately to plants expressing Cre recombinase. Since the tr-Dslox arise from single-copy T-DNAIox loci we can expect that only one possible Iox-/ox recombination event will ensue through the action of Cre recombinase. The complexity of rearrangements arising from recombination between more than two Iox sites may preclude their usefulness.

Previous experiments have shown that Cre recombinase expressed from within a T-DNA can catalyze Iox-/ox recombination efficiently when the Iox sites are integrated together at a second T-DNA locus (Bayley eta/., 1992; Odell et al., 1990; Russell et al., 1992). These experiments have also demonstrated that the rearrangements can be germinally transmitted, as expected. An important recent finding showed that Iox-loxrecombination can occur between sites on unlinked T-DNAs in tobacco, generating a heritable rearrangement. The results reported here extend this finding to Iox sites on the same chromosome. These findings taken together suggest that /ox-lox recombination may occur between any 2 sites integrated into the tobacco or Arabidopsis genomes. To assess this hypothesis more thoroughly we are crossing a large number of plants bearing T-DNAIoxs and single tr-Dslox to plants with Cre recombinase T-DNAs. In this way we will be able to examine Iox-lox recombination frequency, and generate a collection of chromosomal rearrangements derived from T-DNA/ox- Dslox T-DNAs dispersed in the genome.

However, though Iox-/ox recombination may occur efficiently, the rearrangement itself may not be germinally transmitted, as in the instance of the B22-'B' inversion described above. In this case we examined 6013 progeny seedlings derived from six Gn R, R-Lc+ F1 and F 2 plants containing Cre recombinase, the B22 T-DNAIox, and the 'B' tr-Ds/ox on chromosome 4. No Gn R seedlings were found that lacked the Cre recombinase T-DNA (R-Lc-). The existence of the somatically generated rearrangement was evinced not only by the Gn R phenotype, but also from the existence of predicted rearranged fragments detected by PCR and Southern hybridization analysis. The inability to germinally transmit the B22-'B' inversion was unexpected, whereas the inability to generate germinally transmitted deletions of similar size would be anticipated. One simple explanation for the absence of germinal transmission of this inversion is that the Iox-lox recombination event, but not the initial 'B' Dslox insertion, disrupted a gene and caused gametophytic lethality. In this model the 'B' Ds/ox insertion would not be mutagenic, for example, because of its location within an intron (Wessler, 1989), but subsequent disruption of the target gene by Iox-lox recombination would create a gametophytically lethal mutation. The con- comitant disruption at the B22 T-DNAIox locus would not be expected to be gametophytically lethal as the B22-'A' inversion is efficiently transmitted.


Mapped rearrangements are useful

The ability to induce and recover Iox-lox recombination events in Arabidopsis should prove useful for a number of applications, in the same sense that chromosomal rearrangements have been used in genetically tractable organisms such as Drosophila, Caenorhabditis elegans, and Zea mays. Inversions (or translocations) may be used to map loci (Craymer, 1981), or to suppress recombination, as used in balancer chromosomes. Deletions may be used for mapping purposes as well (Bardoni et al., 1991), and to vary gene dosage, or to heavily mutagenize specific regions of the genome. Some classical mapping methods that use rearrangements have been superseded by more efficient molecular methods, but many of these applications will still be important to Arabidopsis geneticists. One important difference between the Iox-lox rearrangements and those rearrangements induced by radiation or chemical mutagenesis is that the Iox-lox rearrangements have breakpoints that can be precisely mapped relative to genetic markers.

A second difference is that the rearrangements may be induced at specific times in development through control of a Cre recombinase transcriptional fusion. Cre-lox recombination could provide an effective means for analyzing cell lineages in plants or the interaction between genes whose action or perception varies between cells or cell layers (for example, Xu and Rubin, 1993). An inducible recombination system would be a valuable addition when combined with current techniques of clonal analysis in higher plants. A regulable Cre recombinase transcriptional fusion may also be extremely useful in the study of essential genes, where /ox-/ox recombination may be used to introduce a mutation in a specific cell lineage. An attempt to studythe role of DNA polymerase ~ in T cell development in this way has been described (Gu et al., 1994).

An inherent feature of this method to generate chromosomal rearrangements is the tendency of Ds to transpose to linked sites in Arabidopsis (Bancroft and Dean, 1993). This proclivity means that the majority of rearrangements generated by Iox-lox recombination in this system will be either deletions or inversions. We expect that large inversions may be formed and transmitted, as in the case of the B22-'A' inversion described here. Inversions may be useful for mapping purposes as well as for the construction of balancer chromosomes. However, it is possible that large deletions would not be transmitted to subsequent generations, and that they would only be useful when formed and analyzed somatically. Alternatively, deletions which remove small numbers of genes could be transmitted, and may be as informative as point mutations or insertions are for the study of single genes. It may be relevant that linked transposition by Ac in maize often traverses short distances, on the order of kilobases (Moreno et al., 1992).

These T-DNAIox and Ds/ox elements can be used to generate chromosomal rearrangements. A related effort would be to use the variety of rare-cutter sites in both elements to cleave chromosomal DNA in vitro. Indeed, the Iox site itself may be considered to be a rare-cutter site. If the elements in the in vitro reaction are mapped, then cleavage and purification is a means of isolating and subcloning specific chromosomal regions, as well as a means of measuring the physical distances between mapped insertions or genetic loci.

Targeted Dslox tagging in Arabidopsis

Work in our laboratory and other laboratories has estab- lished that Ds usually transposes to linked sites in the Arabidopsis genome (Bancroft and Dean, 1993, Osborne, Wirtz, and Baker, unpublished work). This feature means that the most useful T-DNA/ox-Ds/ox T-DNAs are those that have been placed on the Arabidopsis integrated map with respect to cloned sequences and mutations. The development and dissemination of recombinant inbred lines and integrated maps means that researchers will frequently map mutations of interest to them with respect to these markers, and that all members of the community use the same markers as reference points. If a mutation of interest is not generated by element insertion, then a researcher may choose to mutagenize that gene by targeted tagging, which enables the cloning of the gene.

One meaning of targeted mutagenesis is the effort to insertionally mutagenize a locus that has previously been mutagenized, and that has a map position. An alternative targeted approach is to mutagenize a locus that has a map position, has molecular definition, and yet has no known mutant phenotype. For example, a gene may have been cloned and mapped, have some proposed biochemical role, yet have no genetic identity. One could choose a linked T-DNAIox-Dslox T-DNA, trans-activate Ds/ox to transpose, and screen plants for the presence of a Dslox insertion within that sequence using a PCR-based assay, using primers within Dslox and the targeted sequence. The fact that an insertion need not be in homozygous form to be detected is a salient feature of this approach, as the targeted gene may be essential. This strategy can also be used to mutagenize a gene whose function is redundant, or whose mutant phenotype is subtle or unknown. This aspect of targeted tagging will become increasingly relevant as Expressed Sequence Tags (ESTs, Hofte eta/., 1993) become assigned to the Arabidopsis integrated map.

We are mapping our 25 T-DNA/ox-Ds/ox T-DNAs and will offer them to the community as starting points for targeted mutagenesis. In fact, the tendency of Ds to transpose to linked sites in Arabidopsis argues that these T- DNA/ox-Ds/oxT-DNA lines are only useful if placed on the Arabidopsis map when one is interested in isolating both

Ds/Cre-lox system for insert ion and rear rangement 699

mutations and clones corresponding to specific genes. We

expect that our 25 T-DNA/ox-Dslox T-DNAs constitute a

reasonable collection for the mutagenesis of most of the Arabidopsis genome, as T-DNA insertion appears to be random (Chyi et al., 1986; Knapp et al., 1994; Thomas et al., 1994). In addition, all the chromosomal rearrangements that we may generate using tr-Dslox and T-DNAIox have T-DNA/ox as one breakpoint, and these rearrangements

are only useful when mapped.

Experimental procedures

Construction of the binary vectors

The termini of the Dslox transposon were synthesized by PCR using Ac as template and ligated together, making Dsl,2. These ends correspond to positions 1-254 and 4243-4563 of Ac. The following genes were inserted into Dsl,2: a 35S-bar-g7 3' transcriptional fusion (35S-bar from pGSFR280, De Block et aL, 1987). supF (from Z656 cysQ::TnSsupF, Phadnis et al., 1989), Saccharo- myces HIS3 (from p J J215, Jones and Prakash, 1990) and a promoterless Escherichia coil gentamicin-3-N-acetyltransferases AAC (111) ORF with Nos 3' end (from pMON847, Rogers et aL, 1987). An oligonucleotide containing a Iox site (5'-TCGATAACTTCGTATA- GCATACATTATACGAAG'I-I'ATCTCGA-3') was inserted into a Sail site above the ATG of the AAC ORF.

The T-DNAIox-Dslox T-DNA vector was constructed by modi- fication of MOCA1916 (Honma et aL, 1993). The following genes were inserted into the binary vector: Nos-Npt-II-Nos 3' (from pMON200, Rogers et aL, 1987), Saccharomyces LEU2 (from p J J283, Jones and Prakash, 1990), supF(from ;L656 cysQ::TnSsupF, Phadnis et al., 1989) and a 35S-surA, B transcriptional fusion (from pUC35AD, gift of B. Mazur). The Iox oligonucleotide described above was inserted into a Xhol site at the junction of the 35S promoter and the surA, B gene. The Dslox described above was inserted 3' to this Ioxoligonucleotide. Oligonucleotides containing lacO (5'-AATTGTGAGCGCTCACAA'I-r-3'), Sh], Notl, Ascl, and an XbaI-Xbal synthetic 12-mer (Patel et aL, 1990) were inserted into both Dslox and T-DNAIox.

The 35S-Ac transposase fusion has been described (Honma et al., 1993). The 'double' 35S-Ac transposase fusion was constructed by addition of 35S sequence (5' to the EcoRV site) from pKYLX71-35S 2 (Schardl et al., 1987) to the 35S-Ac transposase fusion. The mutB point mutation at position -77 of 35S (Qin and Chua, 1991) was introduced by PCR mutagenesis. The 35S-Cre recombinase fusion was constructed by inserting the Cre ORF (from pBS7, Sauer, 1987) between the 'double' 35S promoter and the Nos 3' end.

The 35S-Ac transposase and 35S-Cre recombinase-Nos 3' transcriptional fusions were inserted into a binary vector based on MOCA1916 (Honma et al., 1993). This vector contained a 35S- hygB-Nos 3' transcriptional fusion (from pCMVHygNos, gift of R. Hehl) and a 35S-R-Lc-Nos 3' transcriptional fusion. The latter gene contained the R-Lc cDNA (Ludwig etal., 1989) placed between the 'double' 35S promoter and a Nos 3' end.

A detailed construction scheme is available on request.

Analysis of genomic DNA

DNA was isolated from Arabidopsis according to the method of Dellaporta (Dellaporta et al., 1983). Southern blotting was

performed according to method of Church (Church and Gilbert, 1984). PCR was performed using Taq polymerase from Cetus according to the accompanying instructions.

Isolation of regions flanking ToDNAIox and the transposed Dsloxs

The B22 T-DNAIox and its flanking regions were isolated in the bacteriophage Z vector EMBL3a using the supFselection. Genomic DNA from plants carrying the T-DNA was partially digested with Bc/I and ligated into purified, BamHI-cut ~. EMBL3a arms. Ligations were packaged in extracts according to the manufacturer (Stratagene) and plated on the sup ° host MC1061. Comparison of plating efficiencies on MC1061 and the sup+ host LE392 showed that I in 5000 plaques was phenotypically sup+. Southern blotting analysis of the ;~ DNA from all sup+ plaques showed that these DNAs contained internal fragments of the introduced T-DNAIox- Dslox T-DNA.

Genomic DNA flanking both ends of tr-Ds/oxs 'A' and 'B' were isolated by IPCR as described previously (Osborne et al., 1991) using Bc/I-digested DNA. The enzymes used for cleavage after ligation were Pad (3' end) and BamHI (5' end). Primers for the first round of PCR were 5'-GAGATCTCTC'I-I'GCGAGATGA (HIS- UW-1) and 5'-'I-I'ATCCCG'I-I'CG'I-rTTCGTTACCG (B3)for the Pacl- cut end of Dslox and 5'-TGGTACCATTGGGCGAGGT (HIS-UW-4) and 5'-GAAACGGTCGGGAAACTAGCTCT (A5) for the BamHI-cut end of Dslox. Primers for the second round of PCR were 5'- GAAAGCTTTGCAGAGGCTAGCAG (HIS-UW-2) and 5'-GT'I-rT- CGTI-rCCGTCCCGCAAG (B4) for the Pad-cut end of Dslox and 5'- AAGC'I-I'ATGGCAACCGCAAGAGCC (HIS-UW-3) and 5'-TCGGGTT- CGAAATCGATCGGGAT (A4) for the BamHI-cut end of Dslox. The identity of IPCR products corresponding to both 5' and 3' ends of these tr-Dsloxs was confirmed by sequence analysis of these products cloned into the EcoRV site of pKS(+).

Plant transformation and in vitro selection

Arabidopsis of the No-O ecotype was transformed according to the method of Valvekens et al. (1983). The selection for the introduction of the T-DNAIox-Dsloxvector was kanamycin (Sigma, 50 mg 1-1) and the selection for the 35S-Ac transposase and 35S- Cre recombinase T-DNAs was hygromycin (Sigma, 25 mg I-1). Gentamicin sulfate (Sigma) was used at a concentration of 200 mg 1-1. Phosphonothricin (Ignite, Hoechst) was used at a concentration of 5 mg 1-1. Chlorsulfuron (Glean, DuPont) was used at a concentration of 50 ilg 1-1.

In terra selection

Phosphonothricin and chlorsulfuron (60 mg 1-1 and 500 I~g 1-1, respectively) were applied together to seedlings in soil after leaf emergence in a solution of 10-10-10 fertilizer. Approximately 20 ml of this solution was applied superficially to 300 seedlings growing in 4" pots.

Construction of a 35S-AAC transcriptional fusion using Cre recombinase

A vector containing Dslox inserted into the 35S-surA, B gene (as in Figure la) was treated with Cre recombinase in vitro according to the manufacturer's instructions (NEN Products, Boston, MA, USA). The expected recombination product was linearized with


restriction enzyme and cloned into a binary vector based on MOCA1916 (Honma et al., 1993) that contained a Nos-Npt-U-Nos 3' transgene. This binary vector was introduced into Arabidopsis by Agrobacterium-mediated transformation (Valvekens et al., 1983).

Acknowledgments

We would like to thank R. Schmidt, J. West and C. Dean (John Innes Centre) for their efforts that confirmed our placement of the B22 T-DNAIox and that placed the Dslox 'A' genomic flanking region on the chromosome IV physical map. We would also like to thank C. Lister and C. Dean for providing their Arabidopsis recombinant inbred lines to us prior to their description in print. We also thank A. Lloyd, R. Davis, V. Walbot, and S. Wessler for materials and information prior to their description in print. Finally, we would like to thank S. Cho, J. Liao, T. Milos, and K. Nelson for their excellent assistance, and D. Hantz for guidance.

References

Abremski, K., Hoers, R. and Sternberg, N. (1983) Studies on the properties of P1 site-specific recombination: evidence for topologically unlinked products following recombination. Cell, 32, 1301-1311.

Baker, B., Schell, J., Lorz, H. and Fedoroff, N. (1986) Transposition of the maize controlling element 'Activator' in tobacco. Proc. Natl Acad. ScL USA, 83, 4844-4848.

Bancroft, I. and Dean, C. (1993) Transposition pattern of the maize element Ds in Arabidopsis thaliana. Genetics, 134, 1221-1229.

Bancroft, I., Jones, J. D. and Dean, C. (1993). Heterologous transposon tagging of the DRL1 locus in Arabidopsis. Plant Cell, 5, 631-638.

Bardoni, B., Zuffardi, O., Guioli, S., Ballabio, A., Simi, P., Cavalli, P., Grimoldi, M.G., Fraccaro, M. and Camerino, G. (1991) A deletion map of the human Yq11 region: implications for the evolution of the Y chromosome and tentative mapping of a locus involved in spermatogenesis. Genomics, 11,443-451.

Bayley, C.C., Morgan, M., Dale, E.C. and Ow, D.W. (1992) Exchange of gene activity in transgenic plants catalyzed by the Cre-lox site-specific recombination system. Plant MoL BioL 18, 353-361.

Botterman, J. and Leemans, J. (1988) Engineering herbicide resistance in plants. Trends Genet. ScL 4, 219-222.

Chang, C., Bowman, J.L., Dejohn, A.W., Lander, E.S. and Meyerowitz, E.M. (1988) Restriction fragment length polymorphism linkage map for Arabidopsis thaliana. Proc. Nat/ Acad. ScL USA, 85, 6856-6860.

Church, G.M. and Gilbert, W. (1984) Genomic sequencing. Proc. Nat/Acad. ScL USA, 81, 1991-1995.

Chyi, Y.S., Jorgensen, R.A., Goldstein, D., Tanksley, S.D. and Loaiza-Figueroa, F. (1986) Location and stability of Agrobacterium-mediated T-DNA insertions in the Lycopersicon genomes./Viol. Gen. Genet. 204, 64--69.

Coupland, G., Plum, C., Chatterjee, S., Post, A. and Starlinger, P. (1989) Sequences near the termini are required for transposition of the maize transposon Ac in transgenic tobacco plants. Proc. Natl Acad. ScL USA, 86, 9385-9388.

Craymer, L. (1981) Techniques for manipulating chromosomal rearrangements and their application to Drosophila melanogaster. I. Pericentric inversions. Genetics, 99, 75-97.

De Block, M., Botterman, J., Vandeweile, M., Dockx, J., Thoen, C., Gossele, V., Mowa, N.R., Thompson, C., Van Montagu, M. and Leemans, J. (1987) Engineering herbicide resistance in

plants by expression of a detoxifying enzyme. EMBO J. 6, 2513-2518.

Dellaporta, S.L., Wood, J. and Hicks, J.B. (1983) A plant DNA minipreparation: version II. Plant Mol. Biol. Rep. 1, 19-21.

Doonar, H.K. and Belachew, A. (1989) Transposition pattern of the maize element Ac from the bz-m2(Ac) allele. Genetics, 122, 447--458.

Fedoroff, N. V. and Smith, D. L. (1993) A versatile system for detecting transposition in Arabidopsis. Plant J. 3, 273-289.

Fedoroff, N., Weesler, S. and Shure, M. (1983) Isolation of the transposable maize controlling elements Ac and Ds. Ceil, 35, 235-242.

Gresnblatt, I.M. (1984) A chromosome replication pattern deduced from pericarp phenotypes resulting from movements of the transposable element Modulator in maize. Genetics, 108, 471- 486.

Grevelding, C., Backer, D., Kunze, R., yon, M.A., Fantes, V., Schell, J. and Masterson, R. (1992) High rates of Ac/Ds germinal transposition in Arabidopsis suitable for gene isolation by insertional mutagenesis. Proc. Nat/Acad. Sci. USA, 89, 6085- 6089.

Gu, H., Marth, J. D., Orban, P. C., Mossmann, H. and Rajewsky, K. (1994) Deletion of a DNA polymerase ~ gene segment in T cells using cell type-specific gene targeting. Science, 265, 103-106.

Hayford, M.B., Medford, J.l., Hoffman, N.L., Rogers, S.G. and Klee, H.J. (1988) Development of a plant transformation selection system based on expression of genes encoding gentamicin acetyltransferases. Plant Physiol. 86, 1216-1222.

Hehl, R. and Baker, B. (1989) Induced transposition of Ds by a stable Ac in crosses of transgenic tobacco plants. Mol. Gen. Genet. 217, 53-59.

Hoess, R.H., Ziese, M. and Sternberg, N. (1982) P1 site specific recombination: nucleotide sequence of the recombining sites. Proc. Nat/Acad. Sci. USA, 79, 3398-3402.

Hofte, H., Desprez, T., Amselem, J., et al. (1993) An inventory of 1152 expressed sequence tags obtained by partial sequencing of cDNAs from Arabidopsis thaliana. Plant J. 4, 1051-1061.

Honma, M.A., Baker, B.J. and Waddell, C.S. (1993) High-frequency germinal transposition of DsALS in Arabidopsis. Proc. Nat/ Acad. ScL USA, 90, 6242-6246.

Jefferson, R.A., Kavanagh, T.A. and Bevan, M.W. (1987) GUS fusions: beta-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6, 3901-3907.

Jones, J.D.G., Carland, E, Lim,-E., Ralston, E. and Dooner, H.K. (1990) Preferential transposition of the maize element Activator to linked chromosomal locations in tobacco. Plant Cell, 2, 701-707.

Jones, J.S. and Prakesh, L. (1990) Yeast Saccharomyces cerevisiae selectable markers in pUC18 polylinkers. Yeast, 6, 363-366.

Kay, R., Chan, A., Daly, M. and McPherson, J. (1987) Duplication of CaMV 35S promoter sequences creates a strong enhancer for plant genes. Science, 236, 1299-1302.

Knapp, S., Larondelle, Y., Rossberg, M., Furtek, D. and Theres, K. (1994). Transgenic tomato lines containing Ds elements at defined genomic positions as tools for targeted transposon tagging. MoL Gen. Genet. 243, 666-673.

Lander, E.S., Green, P., Abrahamson, J., Barlow, A., Daly, M.J., Lincoln, S.E. and Newburg, L. (1987) Mapmaker: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics, 1, 174-181.

Lassner, M. W., Palys, J. M. and Yoder, J. I. (1989). Genetic transactivation of Dissociation elements in transgenic tomato plants. Mol. Gen. Genet. 218, 25-32.

Ds/Cre-lox system for insert ion and rearrangement 701

Lee, K.¥., Townsend, J., Tepperman, J., Black, M., Chui, C.F., Mezur, B., Dunsmuir, P. and Bedbrook, J. (1988) The molecular basis of sulfonylurea herbicide resistance in tobacco. EMBO J. 7, 1241-1248.

Lister, C. and Dean, C. (1993) Recombinant inbred lines for mapping RFLP and phenotypic markers in Arabidopsisthaliana. Plant J. 4, 745-750.

Lloyd, A.M., Walbot, V. and Davis, R.W. (1992) Arabidopsis and Nicotiana anthocyanin production activated by maize regulators R and C1. Science, 258, 1773-1775.

Long, D., Martin, M., Sundberg, E., Swinburne, J., Puangsomlee, P. and Coupland, G. (1993a). The maize transposable element system Ac/Ds as a mutagen in Arabidopsi~. identification of an albino mutation induced by Ds insertion. Proc. Nat/Acad. Sci. USA, 90, 10 370-10 374.

Long, D., Swinburne, J., Martin, M., Wilson, K., Sundberg, E., Lee, K. and Coupland, G. (1993b). Analysis of the frequency of inheritance of transposed Ds elements in Arabidopsis after activation by a CaMV 35S promoter fusion to the Ac transposase gene. MoL Gen. Genet. 241,627-636.

Ludwig, S.R., Habera, L.F., Dellaporta, S.L. and Wessler, S.R. (1989) Lc, a member of the maize R gene family responsible for tissue- specific anthocyanin production encodes a protein similar to transcriptional activators and contains the myc-homology region. Proc. Nat/Acad. ScL USA, 86, 7092-7096.

Msstercon, R.V., Furtek, D.B., Grevelding, C. and Schell, J. (1989) A maize Ds transposable element containing a dihydrofolate reductase gene transposes in Nicotiana tabacum and Arabidopsis thaliana. MoL Gen. Genet. 219, 461-466.

Mazur, B.J., Chui, C.-F. and Smith, J.K. (1987) Isolation and characterization of plant genes coding for acetolactate synthase, the target enzyme for two classes of herbicides. Plant Physiol. 85, 1110-1117.

McClintock, B. (1948) Mutable loci in maize. Carnegie Inst. Wash. Year Book, 47, 155-169.

McClintock, B. (1950) Mutable loci in maize. Carnegie Inst. Wash. Year Book, 49, 157-167.

Meyerowitz, E.M. (1992) Introduction to the Arabidopsis genome. In Methods in Arabidopsis Research (Koncz, C., Chua, N.-H. and Schell, J., eds). Singapore: World Scientific, pp. 100-118.

Moreno, M.A., Chen, J., Greenblatt, I. and Dellaporta, S.L. (1992) Reconstitutional mutagenesis of the maize P gene by short- range Ac transpositions. Genetics, 131,939-956.

Muller-Neumann, M., Yoder, J.I. and Starlinger, P. (1984) The DNA sequence of the transposable element Ac of Zea mays L. Mol. Gen. Genet. 198, 19-24.

Odell, J., Caimi, R, Sauer, B. and Russell, S. (1990) Site-directed recombination in the genome of transgenic tobacco. Mol. Gen. Genet. 223, 369-378.

Osborne, B.I., Corr, C.A., Prince, J.R, Hehl, R., Tanksley, S.D., McCormick, S. and Baker, B. (1991) Ac transposition from a T- DNA can generate linked and unlinked clusters of insertions in the tomato genome. Genetics, 129, 833-844.

Ow, D.W., Wood, K.V., DeLuca, M., De Wet, J.R., Helinski, D.R. and Howell, S.H. (1986) Transient and stable expression of the firefly luciferase genes in plant cells and transgenic plants. Science, 234, 856-859.

Patel, Y., Van, C.E., Wilson, G.G. and McClelland, M. (1990) Cleavage at the twelve-base-pair sequence 5'-TCTAGATCTAGA-

3' using M.Xbal (TCTAGm6A) methylation and Dpnl (Gm6NTC) cleavage. Nucl. Acids Res. 18, 1603-1607.

Phadnis, S.H., Huang, H.V. and Berg, D.E. (1989) TnSsupF, a 264- base-pair transposon derived from Tn5 for insertion mutagenesis and sequencing DNAs cloned from phage I. Proc, Nat/Acad. Sci. USA, 86, 5908-5912.

Pohlman, R.E, Fedoroff, N.V. and Messing, J. (1984) The nucleotide sequence of the maize controlling element Activator. Cell, 37, 635-643.

Qin, M., Bayley, C., Stockton, T. and Ow, D.W. (1994) Cre recombinase-mediated site-specific recombination between plant chromosomes. Proc. Natl Acad. Sci. USA, 91, 1706-1710.

Qin, X.-F. and Chua, N.-H. (1991) Changing binding affinity of a cis element for a transcription factor alters tissue-specific expression in transgenic plants. J. Cell. Biochem. Suppl. 15A, 19.

Rogers, S.G., Kles, H.J., Horsch, R.B. and Fraley, R.T. (1987) Improved vectors for plant transformation: expression cassette vectors and new selectable markers. Methods Enzymo/. 153, 253-277.

Russell, S.H., Hoopes, J.L. and Odell, J.T. (1992) Directed excision of a transgene from the plant genome. Mol. Gen. GeneL 234, 49-59.

Sauer, B. (1987) Functional expression of the cre-lox site-specific recombination system in the yeast Saccharomyces cerevisiae. Mo/. Cell. Biol. 7, 2087-2096.

Scherdl, C.L., Byrd, A.D., Benzion, G., Altschuler, M.A., Hildebrand, D.E and Hunt, A.G. (1987) Design and construction of a versatile system for the expression of foreign genes in plants. Gene, 61, 1-11.

Schmidt, R. and Willmitzer, L. (1989) The maize autonomous element Activator (Ac) shows a minimal germinal excision frequency of 0.2% - 0.5% in transgenic Arabidopsis thaliana plants. Mol. Gen. Genet. 220, 17-24.

Sutton, W.O., Gerlach, W.L., Schwartz, D. and Peacock, W.J. (1984) Molecular analysis of Ds controlling element mutations at the ADH1 locus of maize. Science, 223, 1265-1268.

Thomas, C.M., Jones, D.A., English, J.J., Carroll, B.J., Bennetzen, J.L., Harrison, K., Burbidge, A., Bishop, G.J. and Jones, J.D. (1994) Analysis of the chromosomal distribution of transposon- carrying T-DNAs in tomato using the inverse polymerase chain reaction. Mol. Gen. Genet. 242, 573-585.

Thompson, C.J., Mowa, N.R., "13zard, R., Crameri, R., Davies, J.E., Lauwereys, M. and Botterman, J. (1987) Characterization of the herbicide-resistance gene bar from Streptomyces hygroscopicus. EMBO J. 6, 2519-2523.

Triglia, T., Peterson, M.G. and Kemp, D.J. (1988) A procedure for in vitro amplification of DNA segments that lie outside the boundaries of known sequences. Nucl. Acids Res. 16, 8186.

Valvekens, D., Van Montagu, M. and Van Lijsebettens, M. (1988) Agrobacterium tumefaciens-mediated transformation of Arabidopsis thaliana root explants by using kanamycin selection. Proc. Nat/Acad. Sci. USA, 85, 6856-6860.

Van den Elzen, RJ.M., Townsend, J., Lee, K.Y. and Bedbrook, J.R. (1985) A chimaeric hygromycin resistance gene as a selectable marker in plant cells. Plant Mol. Biol. 5, 299-302.

Wessler, S.R. (1989) The splicing of maize transposable elements from pre-mRNA-a minireview. Gene, 82, 127-133.

Xu, T. and Rubin, G. (1993) Analysis of genetic mosaics in developing and adult Drosophila tissues. Development, 117, 1223-1237.

a system for insertional mutagenesis and chromosomal...

Documents