seed-specific gene activation mediated by the cre//ox site-specif ic recombination system · lhe...

Plant Physiol. (1994) 106: 447-458

Seed-Specific Gene Activation Mediated by the Cre//ox Site-Specif ic Recombination System

Joan T. Odell*, Joyce 1. Hoopes, and Wilfred Vermerris'

Agricultural Products, DuPont, Experimental Station, P.O. Box 80402, Wilmington, Delaware 19880-0402

~~ ~ ~

l h e Cre//ox site-specific recombination system was used to activate a transgene in a tissue-specific manner. Cre-mediated activation of a 8-glucuronidase marker gene, by removal of a lox- bounded blocking fragment, allowed the visualization of the activation process. By using seed-specific promoters, the timing and efficiency of gene activation could be followed within the developing tobacco (Nicotiana tabacum) embryo. To serve as a basis for analyzing gene expression after Cre-mediated activation, the timing and patterns of expression of the promoters of the genes encoding French bean (Phaseolus vulgaris) &phaseolin and the a' subunit of soybean (G/ycine max) &conglycinin, as well as the cauliflower mosaic virus 35s promoter, were studied in developing transgenic tobacco embryos using the same visual marker. These seed-specific promoters were expressed earlier than anticipated. l h e 35s promoter was expressed earlier than the seed-specific promoters, but not in globular-stage embryos. Cre-mediated gene activation occurred approximately 1 d after promoter activation, based on developmental staging, and spread progressively throughout the embryo. l h e timing of gene activation was varied by altering Cre expression. Efficient Cre expression ultimately directed gene activation throughout the model tissue, whereas inefficient Cre expression resulted in mosaic tissue. limited gene activation provides a system for cell lineage and developmental analyses.

Site-specific recombination systems derived from yeast plasmids and bacteriophage have been shown to function in plant cells. In each case the recombinase locates and interacts with its target sites, resulting in excision of DNA sequences located between directly oriented target sites or inversion of sequences located between inverted target sites. Integration between target sites on separate DNA molecules can also occur. The site-specific recombination systems expressed in plants thus far include FLP/FRT (recombinase/recognition site) of the Saccharomyces cerevisiae 2p plasmid (Lyznik et al., 1993; Lloyd and Davis, 1994), R/RS of the Zygosaccharomyces rouxii R1 plasmid (Onouchi et al., 1991), Gin/& of bacteriophage Mu (Maeser and Kahmann, 1991), and Crellox of bacteriophage P1 (Dale and Ow, 1990; Odell et al., 1990); all systems are reviewed by Odell and Russell (1994).

The two elements of the Crellox system, Cre Kauses - recombination) protein and loxP @cus of crossing over, 5, - P1; also lox) sites, function well as stable components of the plant genome. loxP sites integrated in the plant genome are

' Present address: Department of Genetics, North Carolina State

* Corresponding author; fax 1-302-695-4296. University, Raleigh, NC 27695-7614.

447

recognized and recombined by Cre. Test genes have been designed such that Cre-mediated excision or inversion of lox- bounded DNA sequences lead to their activation or inactivation. A lox-bounded coding region placed in inverted orientation with respect to the regulating promoter was expressed after inversion (Dale and Ow, 19.90). A lox- bounded fragment placed between a promoter and coding region, disrupting gene expression, was excised to restore activity (Odell et al., 1990; Bayley et al., 1992). Translocation between lox sites located on separate chromosomes also brought together a promoter and coding region, resulting in gene activation (Qin et al., 1994). Excision of an entire gene located between lox sites resulted in loss of expression (Dale and Ow, 1991; Russell et al., 1992). These alterations in the function of a test transgene occurred after the expression of a cre gene, which was introduced either by retransformation of tissue containing the lox construction or by genetic crossing of lox- and cre-containing plants. In all previous examples Cre was expressed from a relatively constitutive promoter, the 35s promoter of CaMV. Alteration in test gene expression was assayed in roots, stems, leaves, and/or germinating seedlings.

Cre-mediated gene activation in a specified tissue may provide a useful tool for developmental analyses. For example, Cre may be used to specifically activate a cell-lethal gene to determine the effects of the loss of the affected cells on surrounding tissues. Similar tissue interactions may be studied by controlling the activation of a gene whose product restores function in a mutant host. Here we examine tissue- specific regulation of gene expression mediated by Crellox. A silent GUS gene is activated by Cre-mediated excision of a lox-bounded fragment located between the promoter and coding region. The seed-specific promoters from the CY' subunit of soybean (Glycine max) P-conglycinin and from French bean (Phaseolus vulgaris) P-phaseolin were used to regulate GUS and Cre expression. These developmentally controlled systems allow an analysis of the timing and efficiency of Cre- mediated gene activation. Comparisons are made to pattems of activation induced by Cre expressed from the 35s promoter and by Cre protein that is targeted to the nucleus.

Abbreviations: Cab, Chl alb-binding protein; CaMV, cauliflower mosaic virus; dap, days after pollination; GUS, @-glucuronidase; Nos, nopaline synthase; npt I and npt 11, neomycin phosphotransferase I and 11; Ocs, octopine synthase; ppb, parts per billion; SV40, simian virus 40.

https://plantphysiol.orgDownloaded on February 2, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

https://plantphysiol.org

448 Odell et ai. Plant Physioi. Val. 106, 1994

MATERIALS AND METHODS

Chimeric Cene Constructions

The French bean (Phaseolus uulgaris) @-phaseolin promoter (Ph/P) and the soybean (Glycine max) a’ subunit of 8- conglycinin promoter (Cg/P; Doyle et al., 1986) were provided by J. Slightom. NcoI sites had been added at each translation initiation site while preparing the promoter fragments using the PCR technique (described by Slightom and Chee, 1991). Plasmids containing the promoter fragments also contained a 1.17-kb fragment including the phaseolin 3’ polyadenylation signal region (Ph 3’), downstream of the NcoI site. A linker was added to create the following sites between the promoters and 3‘ regions: NcoI, SmaI, Asp718, and XbaI. The cre coding region was then added as an NcoI- XbaI fragment to prepare the chimeric genes, Ph/P-cre and Cg/P-cre (Fig. 1A). Each gene was transferred as a HindIII fragment into the binary vector pZS94. This plasmid contains

A +Ph/P( c r e IPh 3+-

+ CR/PI c r e IPh Y+-

+ 3 5 ~ ~ I c r e IOCS 3’+

4 Ph/PI Gus IPh 3’1- C

--I 3 5 s c I

Figure 1. Constructions of chimeric cre and CUS genes. A, Cenes to express the cre coding region from t h e P-phaseolin promoter (Ph/P), a‘ @-conglycinin promoter (Cg/P), or 35s promoter with the Cab leader (35SC). T h e NTcre coding region includes t h e SV40 T- antigen nuclear localization signal. T h e 3’ end is either from the 8- phaseolin gene (Ph 3’) or the Ocs gene (Ocs 3’). B, lnterrupted /ox2-CUS genes designed to be inactive. /ox sites (filled arrowheads) surround a fragment including a polyadenylation signal sequence (polyA). C, Control genes that express CUS from the promoters in A. One has the 3’ end from the Nos gene (Nos 3’). D, Cre-mediated activation of the /ox2-CUS gene. Recombination between directly repeated lox sites results in excision of the segment between the two sites, leaving one /ox site in the untranslated leader of the CUS mRNA.

the origin of replication and ampicillin resistance gene from pBR322!, the replication and stability regions of t he Pseudom- onas ae~~uginosa plasmid pVSl (Itoh et al., 1984), and T-DNA borders described by van den Elzen et al. (1985). To provide a plant transformation selection marker, a PstI-llsp718 fragment containing a mutant 35s-acetolactate sy nthase gene that confers sulfonylurea resistance (described by Russell et al., 1992) was added.

A gene encoding Cre with a nuclear localization signal added at the N terminus was prepared. At the NcoI site in Cg/P-cre an oligonucleotide encoding amino acids 118 through 123 and 126 through 132 of the SV40 laige T antigen (Rihs and Peters, 1989; Fig. 1A) was inserted. ‘rhe oligonucleotide was designed so that the 3’ NcoI site was not regen- erated. The resulting Cg/P-NTcre gene was transferred as a HindIII fragment into pZS96. pZS96 is the vector pZS94 with a Nos/P-nptII-Ocs 3’ gene added to serve as a kanamycin- resistant plant selection marker.

A gene containing the 35s promoter of CaMV and the cre coding region was constructed (Fig. 1A). A promclter fragment containing the 355 promoter and leader from the Chl a/b- binding protein gene 22L (35SC; Harpster et al., 1988) was joined to the cre coding region at the NcoI site located at the translaiion-initiation ATG. The cre coding regon was followed by a 720-bp fragment containing the polyadenylation signal sequence from the Ocs gene (Ocs 3’). An EcoRI-Asp718 fragment containing the entire 35SC-cre gene was transferred to the binary vector pZS198. pZS198 is pZS96 with the bacterial ampicillin-resistance selection marker replaced by a bacterial kanamycin-resistance selection marker (nptI; kindly provided by M. Locke).

Inactive marker GUS genes with a lox-bounded blocking fragment (Iox’) between the promoter and coding; region were constructed as follows (Fig. 1B). Before adding ihe Iox2 fragment to each promoter, the NcoI site at the 3’ end of each promoter was deleted so that there would be no translation- initiatirig ATG 5’ to the first lox site. Plasmids containing each promoter were digested with NcoI, treated with S1 nuclease to remove the overhanging ends, and then religated (Ph/P) or cut at the SmaI site downstream of the NcoI site and religated (Cg/P). The 3’ end of each promcter fragment was sequenced to verify that the ATG from the VcoI site was removed. A linker containing XhoI and NcoI sitw was added between the Asp718 and XbaI sites located 3’ to the Cg/P. The lox-polyA-lox fragment described by Odell et al. (1990) was added as a XhoI-Ncol fragment, after an NcoI linker was added at the HindIII site at the 3’ end. This fragnient contains the polyadenylation signal sequence region frcm a tobacco Rubisco small subunit gene (Mazur and Chui, 15185) between directly oriented lox sites. Both lox sites are orirtnted so that the two ATGs present in the site are in reverse orientation with respect to the promoter. The GUS coding r2gion (Jeffer- son et d., 1987) was joined, via an NcoI site at its translation start site, 3’ to lox’. An Asp718-Asp718 fragment from this clone containing lox-polyA-lox-GUS was added i o the Asp718 site between the Ph/P and Ph 3’, and clones with the desired fragment orientation were identified. The entire Ph/P or Cg/ P-lox-poIyA-lox-GUS-Ph 3 ’ gene was transferred as a HindIII fragment into the binary vector pZS96.

Control Ph/P-GUS and Cg/P-GUS genes wlzre also con-



Crellox-Mediated Seed-Specific Gene Activation 449

structed (Fig. 1C). An NcoI-Asp718 fragment containing the GUS coding region was added between the NcoI and Asp718 sites between the Cg/P or Ph/P and Ph 3' in plasmids described above for the cre constructions. A HindIII fragment containing each gene was transferred into the pZS96 vector.

Plant Transformation

All binary vectors containing cre or lox constructions were transferred into Agrobacterium tumefaciens strain LBA4404 by three-way matings (Ditta et al., 1980), direct DNA uptake (An et al., 1988), or electroporation (Mersereau et al., 1990). Selected transformed A. tumefaciens were used to obtain Nicotiana tabacum cv Xanthi transformants as described by Odell et al. (1990) using either 300 pg/mL kanamycin or 50 ppb chlorsulfuron for selection, depending on the selection marker present.

Seed Germination Assay

Seed was sterilized and assayed for seedling resistance as described previously (Odell et al., 1990) using 200 mg/L kanamycin or 50 ppb chlorsulfuron.

Cross-pollinations

Hand pollinations were performed as described previously (Odell et al., 1990). Initially, single-locus homozygous cre plants, identified through seed-germination assays, were crossed with primary /ox2-GUS transformants that contained one or more (two or three) kanamycin-resistance loci. In this case, all seed inherit the cre gene, but 50% or less do not inherit the lox2-GUS gene. Therefore, some F1 embryos would never show GUS activation, having not inherited the GUS gene. In cases in which GUS was expressed in some F1 embryos, there were also embryos from the same cross that did not show GUS activity as expected; these were not included in figures. In cases in which no GUS activity was detected, such as in very early-stage embryos, the embryo sample size was large enough to assure that embryos inheriting /ox2-GUS would have been included in the assay.

Single-locus homozygous /ox2-GUS transformants were then identified through seed-germination assays. These were crossed with primary transformants bearing 35SC-cre or Cg/ P-NTcre genes. The same inheritance situation described above occurs in the F1 seed but is reversed for the cre and lox*-GUS genes.

Seed Dissection and Analysis

Pods were harvested at various days after selfing (in the case of plants containing Ph/P-GUS, Cg/P-GUS, and 35S/ P-GUS transgenes) or cross-pollination (dap). Seed was removed from the pod, dissected to separate the embryo from the endosperm (which often remained partially attached to the seed coat), placed directly into GUS enzyme assay buffer (Jefferson, 1987), and incubated at 37OC. Samples were examined for GUS staining at various times, and photomicro- graphs were taken using a Zeiss Stemi SR dissecting microscope, or a Wild M8 dissecting microscope for the higher magnifications.

Southern Blot Analysis

Preparation of DNA was based on the procedure of Shure et al. (1983), with modifications by J. Chen. Fresh tissue was ground in a mortar with liquid nitrogen. Buffer containing 7 M urea, 0.31 M NaCl, 50 IT~M Tris, pH 8.0, 20 mM EDTA, and 1% sarkosyl was added and the sample was shaken vigor- ously then incubated at room temperature for 10 min. The sample was then extracted with an equal volume of pheno1:chloroform:isoamyl alcohol at 25:24:1. The supema- tant was passed through Miracloth, then 0.1 volume of 4.4 M NH4Ac and 0.5 volume of isopropanol were added. The DNA was spooled, washed with 80% ethanol, dried, dis- solved in water, and precipitated a final time. Approximately 10 pg of DNA was digested with Asp718 or HindIII, and the resulting fragments were separated by gel electrophoresis then transferred to a Hybond-N (Amersham) filter. Hybridi- zation was to a digoxigenin-labeled probe, which was detected by chemiluminescence using Lumi-Phos 530 (Genius System, Boehringer Mannheim).

RESULTS

Experimental Design for Seed-Specific Gene Activation

To assess the timing and efficiency of Cre-mediated recombination in plant cells, we used the Crellox system to activate a marker GUS (Jefferson et al., 1987) gene in developing seed. An inactive GUS gene was constructed by placing a /ox-bounded DNA fragment (represented by /ox2), containing a polyadenylation signal sequence, between the promoter and coding region (Fig. 1B). The directly repeated lox sites provide the substrate for Cre-mediated excision of the blocking fragment, leading to expression of GUS (Fig. 1D). Cre- mediated recombination of the lox2-GUS gene in vitro did give the expected products. This activatable marker and cre were regulated by two seed-specific promoters, derived from genes encoding French bean ,&phaseolin (referred to as 0- phaseolin promoter and Ph/P) and the soybean a' subunit of P-conglycinin (referred to as a' /3-conglycinin promoter and Cg/P). When this work was initiated, the patterns of expression of these two promoters in regions of the embryo and in the endosperm of transgenic tobacco seed were not well established, and we expected different patterns for their expression. Control GUS genes were constructed without the intervening lox-bounded fragment (Fig. 1 C). For comparison, cre was also expressed from the CaMV 35s promoter. The timing of marker gene expression after activation and the extent to which it is activated throughout the embryo were determined. These analyses characterize the Crellox recombination process in plant cells.

Preparation of Transgenic Plants

Each lox-inactive chimeric GUS gene and control GUS gene was stably integrated into the tobacco genome through A. tumefaciens-mediated transformation. A linked nptII gene, conferring kanamycin resistance, was used to select the transformants and to identify lines carrying a single integrated locus. Lines were screened for GUS activity in mature seed. Of 12 kanamycin-resistant transformants with Ph/P-GUS or



450 Odell et al. Plant Physiol. Vol. 106, 1994

Cg/P-GUS in the introduced T-DNA, only 2 lacked GUSactivity in mature seed. In these 2 lines an intact GUS genemay not have been integrated. Of 15 kanamycin-resistantlines transformed with a GUS gene interrupted by the /ox-bounded fragment, 8 showed no GUS activity, whereas 7expressed a very low level of GUS activity in mature seedrelative to the Ph/P-GUS and Cg/P-GUS levels. Overnightincubation in GUS assay buffer was required to detect lightblue staining, whereas staining began to appear within 15min in seed with no blocking fragment. In the 8 negativelines the intervening /ox-bounded fragment appeared to beeffective in blocking expression of the GUS coding region,although it is possible that a few of them did not integratean intact GUS gene. Seven of these lines were tested for Creactivation of GUS as described below. In all but 1, GUSexpression could be activated, verifying the presence of theintact gene. The blocking fragment was not completely effec-tive, as evidenced by the 7 lines with low GUS activity. Thesewere not examined to determine whether the "inactive" GUSgene construction was transferred in an intact form.

Separate tobacco lines carrying chimeric seed promoter-cregenes were selected using a sulfonylurea-resistant acetolac-tate synthase marker. By seed-germination assays on chlor-sulfuron, four single-locus lines were identified, then homo-zygotes in these lines were identified. Of these four lines,three were able to activate the inactive GUS gene as describedbelow. In the fourth line an active ere transgene may nothave been integrated.

In the course of these experiments, two different geneconstructions containing the 35S promoter and ere codingregion were used. Homozygous tobacco plants with a 35S-cre transgene were produced and assayed in previous studies(cre/Hpt-A, Odell et al., 1990; Russell et al., 1992). Thischimeric gene has an untranslated leader composed of se-quence from the leader of the 35S RNA of CaMV (21 bp),and sequence upstream of the translation start site of thebacteriophage PI ere gene (about 50 bp). In a second gene,called 35SC-cre, the leader of the petunia Cab gene followsthe 35S promoter, and is joined to the start of the ere codingregion by an Ncol site (see "Materials and Methods"). Tobaccotransformants containing the 35SC-cre transgene were se-lected using a linked kanamycin-resistance selection marker.Other control plants bearing a GUS gene expressed from thesame 35S promoter linked to the Cab leader (35SC-GUS; Fig.1C) were produced and described previously (Russell et al.,1992).

Expression of Ph/P, Cg/P, and 35SC/P All Initiate at theHeart Stage

Previously, expression of the 0-phaseolin, a' /3-conglyci-nin, and 35S promoters in developing transgenic tobaccoseed had been determined only by bulk seed analyses. Sincewe planned to visualize GUS staining of individual embryosto analyze Cre-mediated gene activation, it was critical todetermine the expression patterns of these promoters usingthe same procedure. Seed bearing Ph/P-GUS, Cg/P-GUS, or35SC-GUS were dissected at different dap and the embryosand endosperm were stained separately for GUS enzymeactivity. Variation occurred in the timing of embryo devel-

opment, although plants were grown under the same condi-tions. For example, the heart to small torpedo-stage embryosin Figure 2C were collected at 9 dap and those of similarstages in Figure 2A were collected at 10 dap. More advancedtorpedo-stage embryos were also on occasion found at 10dap, such as those in Figure 3F. There was also substantialvariation in development among seed within the same pod.For example, all of the embryos in Figure 3A, which rangefrom torpedo to maturing cotyledon stages, were dissectedfrom seed within the same pod collected at 14 dap. Thus,when assaying individual embryos it is more informative to

*•**Figure 2. Expression of GUS activity from /3-phaseolin, a' /3-congly-cinin, and 35S promoters in tobacco embryos and endosperm. Allembryos are shown at the same magnification except for those inF, bottom row, and G, as indicated below. All samples were incu-bated in GUS assay buffer overnight at 37°C unless otherwisespecified. A, Embryos containing the Ph/P-GUS gene dissected 10dap. B, Endosperm of seed containing Cg/P-GUS dissected 10 dap.C, Embryos containing Cg/P-GUS dissected 9 dap. D, Embryoscontaining Cg/P-GUS dissected 11 dap and incubated for 2 h inGUS assay buffer. E, The same embryos as in D incubated overnight.F, Embryos containing 35SC-GUS dissected at 10 dap. The upperset is shown at the same magnification as those in A-E. The lowerset contains the same embryos shown at 1.6x higher magnificationfor better visualization. G, Embryos containing 35SC-GUS dissectedat 10 dap and incubated in GUS assay buffer for 2 h, shown at 1.6xhigher magnification than those in A to E.



Cre//ox-Mediated Seed-Specific Gene Activation 451

B

A

Figure 3. GUS activity following Cre-mediated activation of /ox2-GUS genes. All samples are from Fi seed unlessotherwise specified. A, Embryos inheriting Cg/P-/ox2-GUS and Cg/P-cre dissected 14 dap. Inset, Embryos inheriting Ph/P-/ox2-GUS and Ph/P-cre dissected 12 dap. B, Endosperm of seed inheriting Ph/P-/ox2-GUS and Ph/P-cre dissected 11dap. C, Embryos inheriting Ph/P-/ox2-GUS and 35S-cre dissected at 13 dap. Inset, Embryos inheriting Ph/P-/ox2-GUS and35S-cre dissected at 12 dap. Note that these embryos are inverted. D, Endosperm of seed inheriting Ph/P-/ox2-GUS and35S-cre dissected at 16 dap. E, Embryos inheriting Cg/P-/ox2-GUS and 35S-cre dissected at 12 dap. F, F2 embryos fromFI plants grown from F, seed that inherited Ph/P-/ox2-GUS and Ph/P-cre dissected 10 dap.

describe stages of embryo development than dap. In Figures2 and 3, embryos are shown at the same magnification (exceptin Fig. 2, F and G) so that the stages of development can becompared by the size and shape of the embryos. Those inFigure 2A are entering the heart stage or are hearts or youngtorpedos. Figure 3A includes embryos of torpedo and earlyto maturing cotyledon stages. Embryos in the bottom row of

Figure 2, F and G, and Figure 4 are shown at higher magni-fication for better visualization.

The expression patterns of the /3-phaseolin and a' fi-conglycinin promoters were similar throughout seed devel-opment. No GUS activity was observed in globular-stageembryos. As embryos entered the heart stage, spots of GUSactivity were observed just below the cotyledonary lobes (see




Cg/P ioxpH——W-

loxP Ph 3'

.5 . 1.6H -s A

- 5.0 kb-3.3 kb

X + Cre

Cg/P ioxp GUS Ph 3'+H •=> A

- 3.5 kb-1.8 kb 1 2 3 4

Figure 4. Molecular analysis of Cre-mediated recombination leading to activation of CUS gene expression. A, Diagramof the inactive Cg/P-/ox2-GUS gene (above) and Cre-activated Cg/P-CUS gene (below). Shown below each diagram arethe DNA fragments produced by H/ndlll (H) or Asp718 (A) digestion from the original construction, and those predictedfollowing recombination. The hatched bar indicates the CUS coding region probe used to detect these fragments. Sizesshown are approximate. Abbreviations are as described in Figure 1. B, Southern blot showing bands from the inactiveand recombined genes. Lanes 1 and 3, Wild-type tobacco DNA spiked with inactive Cg/P-/ox2-CUS gene-containingplasmid. Lanes 2 and 4, DNA from an f, ere x /ox plant. DNA samples in lanes 1 and 2 were digested with Asp718, andthose in lanes 3 and 4 were digested with H/ndlll. Left and right upper arrows mark the bands representing nonrecombinedfragments in lanes 1 and 3, respectively. Left and right lower arrows mark the expected recombination products in lanes2 and 4, respectively. The upper band in lane 4 is likely to result from a rearrangement of integrated DNA. The positionsof size markers are shown to the left.

Fig. 2A). More spots accumulated in this area, creating a ringof GUS activity in the transition zone between the developingcotyledons and embryo axis of heart-stage embryos. This ringthen spread up into the cotyledons and down into the axis,more rapidly on the outside edges (Fig. 2, A and C). Embryosentering the torpedo stage lacked GUS activity only in thetips of the cotyledons and the bottom of the root axis (Fig. 2,A and C). Torpedo-stage embryos stained entirely blue. Withboth promoters, GUS activity was absent in the endospermof very early-stage embryos but then spread throughout thistissue (Fig. 2B). The endosperm expression was unexpectedfor the a' /3-conglycinin promoter, since no activity was seenin soybean endosperm using in situ analysis (Perez-Grau andGoldberg, 1989).

The embryo expression of both promoters was much earlierthan is generally described for seed storage proteins. There-fore, eight independent transformants containing Ph/P-GUSand four containing Cg/P-GUS were assayed. All but oneshowed strong endosperm expression and similar patterns ofembryo expression. In several lines the timing varied slightly,being either a bit more rapid or delayed than that describedabove. One exceptional Ph/P-GUS line showed much re-duced expression; spots of GUS activity appeared first intorpedo stages, which stained entirely blue in all of the otherlines (data not shown). Mature embryos of this line still didnot stain solidly blue, but were blotchy, indicating reducedGUS expression. This pattern could be explained by greatlyreduced functioning of the phaseolin promoter, due to itsposition of integration in the tobacco genome in this partic-ular line.

Incubation of embryos in GUS assay buffer for a shortperiod of time allows the localization of regions with strongestGUS expression. The cotyledon-axis junction region is alwaysfirst to stain for GUS activity, even in torpedo- and later-

stage embryos that stain entirely blue during longer incuba-tions (Fig. 2, D and E). Thus, this region, which is the first toshow phaseolin and conglycinin promoter expression at theheart stage, maintains the highest level of expressionthroughout later stages of development.

The 35SC promoter showed a different pattern of expres-sion, observed in the three independent transformants bear-ing the 35SC-GUS gene that were analyzed. Surprisingly,GUS activity was also absent in globular embryos. Theseunstained globular embryos are not shown due to the diffi-culty of photographing them under the conditions we em-ployed. GUS activity appeared and spread more rapidly thanwhen expressed by the phaseolin and conglycinin promoters.Embryos entering the heart stage showed a patch of GUSactivity in the top central region (Fig. 2F). The expressionspread rapidly; mature heart-stage embryos stained entirelyblue. This timing varied slightly between independent trans-formants, but was always substantially earlier than observedfor the phaseolin and conglycinin promoters. GUS activityalso developed in the endosperm. After a short incubation inGUS assay buffer, the tips of the cotyledons and the lowerpart of the axis remained unstained, indicating that these arethe regions with the weakest promoter expression (Fig. 2G).The short incubation also allows the detection of a spot ofstrong GUS activity at the tip of the root axis, a pattern notobserved in the /3-phaseolin and a' /3-conglycinin promoters.

Cre-Mediated Gene Activation

Once we determined the timing and pattern of GUS expres-sion directed by the /3-phaseolin and «' /3-conglycinin pro-moters in individual embryos, we could compare it to theGUS expression seen after Cre-mediated activation of the/o*2-GUS gene. Cre and lox elements were joined in FI seed



Cre//ox-Mediated Seed-Specific Gene Activation 453

by crossing a plant containing a ere gene and a plant contain-ing a /OA:2-GUS gene. Seeds of varying developmental stageswere dissected and the embryos and endosperm were stainedfor GUS enzyme activity. GUS activity was observed inembryos and endosperm of seed derived from crosses be-tween lox and ere plants, indicating that the /<w2-GUS genewas activated, whereas seed from crosses between lox plantsand control wild-type plants never showed any GUS staining.Activation of the /o;t2-GUS gene by Cre-mediated recombi-nation, diagrammed in Figure ID, was later verified in Ftplants. DNA of ere X lox F] plants analyzed on Southernblots showed GUS probe-hybridizing bands of the sizes thatwould be expected after excision of the DNA segment be-tween the lox sites (Fig. 4).

Similar timing and patterns of GUS activation were ob-served in seed derived from crossing four independent Ph/P-/ox2-GUS transformants with a Ph/P-cre plant, and twoindependent Cg/P-/ox2-GUS transformants with two inde-pendent Cg/P-cre transformants (all four combinations). Ineach case no GUS activity was observed in heart-stage oryoung torpedo-stage embryos. As described above, we un-expectedly found that the /3-phaseolin and a' /3-conglycininpromoters are active during these stages. It was critical tovisualize this early expression of these promoters to deter-mine that a delay in expression occurred when recombinationbetween lox sites was required to activate the GUS gene. Thefirst few spots of GUS activity appeared in torpedo-stageembryos (Fig. 3A). This represents a delay of about 1 d in thedevelopmental time course. Spots of GUS activity occurredfirst at the cotyledon-axis junction, which is the region ofearliest and highest /3-phaseoIin and a' /3-conglycinin pro-moter expression (see Fig. 2, A-D). The GUS-expressingregion then spread up into the cotyledons and down into theaxis, until the whole maturing cotyledon-stage embryostained blue. Activation occurred coordinately in the endo-sperm, initially in spots of activity that then spread through-out the endosperm (Fig. 3B). No difference in the timing ofGUS gene activation was observed in FI embryos derivedfrom reciprocal crosses (data not shown).

Eliminating the Delay in Gene Activation

When the same developmentally regulated promoters di-rect expression of Cre and GUS, the time that is required forCre expression and the recombination process itself is seenas a delay in GUS expression. Earlier expression of Cre couldeliminate this delay. We initially thought that the 35S pro-moter would be expressed throughout embryo developmentin a more constant manner than are the seed-specific pro-moters. If this were true, Cre protein would be present andable to create the activated GUS gene prior to the time whenthe seed-specific promoters are expressed. Therefore, GUSactivity would appear with the same timing as the expressionof Ph/P-GUS. Surprisingly, this was not the case in initialexperiments.

When plants containing Ph/P-/ox2-GUS or Cg/P-/ox2-GUSwere crossed with several different plants containing the35S-cre gene (not 35SC-cre, see "Preparation of TransgenicPlants'), two different patterns of GUS expression were ob-served in the F! seed (Fig. 3, C and E; embryos at same

magnification). The 35S-cre plants were from lines analyzedin previous experiments in which the transformation selectionmarker, bounded by lox sites, was excised from the tobaccogenome in ere by lox F, progeny (Russell et al., 1992). LineCre3 was the best at directing recombination in these previousexperiments, and line Cre2 was the least effective. Crosses of/ox2-GUS plants (both Ph/P-/ox2-GUS and Cg/P-/ox2-GUSplants used) with two plants from line Cre3 produced anactivation pattern in F] seed similar to that described abovefor Ph/P-cre and Cg/P-cre activation (Fig. 3A): a delay ofabout 1 d following initial expression of phaseolin and con-glycinin promoters (Fig. 3E, and earlier stages shown in Fig.5B at higher magnification). This result indicates that theGUS gene was not preactivated. In crosses of /ox2-GUS plantswith plants from line Cre2, FI seed showed a pattern of GUSexpression that indicated even later excision of the /ox-bounded blocking fragment. No GUS activity was observeduntil embryos reached a late torpedo stage. Spots of GUS-expressing cells were observed first in the cotyledon-axis

B

Figure 5. Comparison of CUS gene activation by Cre expressedfrom different ere genes. These F, embryos are shown at 1.6X themagnification of those in Figures 2 and 3. A, Embryos inheriting Cg/P-/ox2-GUS and 35SC-cre dissected at 10 dap. B, Embryos inheritingCg/P-/ox2-GUS and 35S-cre dissected at 10 dap. C, Embryos inher-iting Cg/P-/ox2-GUS and Cg/P-NTcre dissected at 11 dap.




junction region, and then in the cotyledons. In older embryos there were large patches as well as smaller spots of GUS- expressing cells. Endosperm from maturing seed also had localized spots of GUS activity. Surprisingly, one plant from the Cre3 line caused this same delayed and limited pattem of GUS activation (Fig. 3, C and D). This plant appeared to have reduced Cre expression. Its Cre3 siblings activated GUS expression no earlier than when Cre was regulated by the conglycinin and phaseolin promoters.

The gene construction that includes the 35s promoter/Cab leader and cre coding region, 35SC-cre, did direct the earlier activation of the GUS gene that was originally predicted. In Fl embryos from crosses of one Cg/P-lox2-GUS plant with two independent 35SC-cre transformants, GUS was expressed in embryos entering the torpedo stage in all but the tips of the cotyledons and bottom of the axis. Shown in Figure 5A, these embryos are at higher magnification, but are of similar stages as some in Figure 2, A and C. The tip of the root axis did show GUS activity, following the expression pattern specific for the 355 promoter. Young torpedo-stage embryos stained entirely blue. This mimics the expression pattem for Cg/P-GUS (no activation required), with the addition of the blue axis tip. Thus, the blocked GUS gene appears to have been preactivated by Cre present in heart- stage embryos, as predicted from the expression pattem of this same promoter visualized using the GUS marker, 35SC- GUS, described above.

From the patterns of GUS gene activation, we infer that the initial 35s-cre construction directed later and lower levels of expression than 35SC-cre, which probably expressed in a manner similar to 35SC-GUS. The better expression of the 35SC-cre gene, as compared to the 35s-cre gene, may be attributed to the different 5’ untranslated leader sequences. In particular, the sequence surrounding the translation initiation ATG of the 35SC-cre gene is more similar to the consensus sequence desired for maximal translatability (Fig. 6; Kozak, 1986). It is not known if translational efficiency accounts for the difference in expression levels, or whether increased mRNA stability is possibly conferred by the Cab leader as well as other unknown factors.

Nuclear Targeting of Cre

The efficiency of Cre-mediated recombination could poten- tially be increased by targeting Cre protein to the nucleus.

T TAAAT GT CC I I

AAC AAT G C T con sensus I l l I I AACCATGGCC CabL-cre

c r e L-c re

Figure 6. Comparison of sequences surrounding the translation initiation ATG in two different cre genes expressed from the 35s promoter. The top sequence is from t h e 35s-cre gene, which has the natural sequence 5’ to cre in the P1 genome upstream of the ATC (creL-cre). The bottom sequence is from the 35SC-cre gene, which has the leader from the Cab gene 22L (CabL-cre). The middle sequence is the Kozak consensus for maximal translation, with the most important nucleotides underlined. Vertical dashes indicate the identities between t h e sequences surrounding t h e ATGs.

Cre is a small (38 kD) prokaryotic protein that obviously is able to enter the eukaryotic nucleus. However, by facilitating entry into the nucleus, the delay in timing between Cre expression and lox gene activation might be decreased. This hypothesis was tested by placing the SV40 large T-antigen nuclear-localization signal sequence at the N temunus of Cre expressed from an NTcre gene (see “Materials and Methods”). Three independent Cg/P-NTcre transformants were crossed to a Cg,lP-lox2-GUS line, and GUS expression was followed in F1 einbryos to determine whether activation was more rapid than in embryos inheriting Cg/P-cre. The same I-d delay in GUS expression was observed with boih cre genes (Fig. 3A, and Fig. 5C at higher magnification). Thus, the additiori of this targeting signal did not increase the in vivo efficiency of the Crellox recombination system.

Loxz-CIJS I s Preactivated in Fz Seed

To determine whether the blocked GUS gene w as activated in genetically effective cells in the F1 generation, plants were grown from Fl seed derived from crosses between Ph/P-lox2- GUS plants and Ph/P-cre or 35s-cre plants. GUS expression was observed in developing F2 seed produced by selfing each F1 plant. Preactivation of the GUS gene in the Fl generation, prior to F2 seed formation, can be determined in two ways. The GUS expression pattern in F2 embryos should not be delayed, as it is in the F1 embryos, if preactivation occurred. Also, genetic segregation predicts different ratios of GUS- expressing embryos for preactivation versus actixration in the FZ seed.

Each of the original cre parents crossed to produce F1 seed was homozygous for the cre gene at a single locus. Each of the orignal lox parents was hemizygous (primary transformants) at a single locus for the interrupted GUS (lox2-GUS) gene. Therefore, each cross produces 50% of seeds inheriting both crie and lox, and 50% of seeds inheriting only cre. Seedlings that inherited both cre and lox were selected by their resistance to kanamycin, the marker linked to the lox2- GUS gene. Each of these F1 plants is heterozygous for both cre and lox. Selfing of the F1 plant produces F2 seed in the following ratios of genotypes: 9 lox + cre:3 lox alone:3 cre a1one:l neither lox nor cre. If preactivation occurred in the F1 generation, cre is not necessary for activation in F2 seed, so all Fz seed inheriting lox would express GUS. A ratio of 3 b1ue:l white embryos is expected in this situation. If activation is necessary for GUS expression, then cre must be inherited along with lox2-GUS, so a ratio of 9 blue:7 wbite embryos is expected. Ratios of 3 b1ue:l white were obsemed in the F2 embryos of each selfed cre-lox F1 plant in samples of over 40 (for example, 31 blue embryos:12 white embryos). Thus, the genetic analysis indicates that the lox blocking fragment is removed in genetically effective cells prior to F2 seed development.

Results of the second assay for preactivation were consist- ent with the genetic data in that the timing of GU!; expression in F2 embryos was not delayed. Similar express’on patterns were observed in F2 embryos of two individml F1 plants derived from each of the following crosses: two independent Ph/P-lox2-GUS lines with one Ph/P-cre line and one Ph/P- lox2-GUS line with two 35s-cre lines. Spots of GUS activity




appeared first in heart-stage embryos (not shown) and a thick blue ring was seen in mature hearts (Fig. 3F), clearly earlier expression than seen in the F, embryos. The expression level appeared to be slightly reduced, since in young torpedo-stage embryos the axis stained a lighter blue and the axis tip remained white, even after the overnight incubation in GUS assay buffer. Reasons for this reduced expression are unknown.

DISCUSSION

To characterize the timing of Cre-mediated gene activation in seed, it was first necessary to establish the expression patterns of the two seed-specific promoters used in our system. By staining intact embryos we were able to visualize the timing and pattem of expression of the marker GUS gene directed by the promoter of either the French bean P-phaseolin gene or the soybean a‘ subunit of the P-conglycinin gene. In each case the expression was earlier in the development of transgenic tobacco embryos than expected from previous reports. Soybean P-conglycinin genes (a’, a, and P subunits) are highly expressed in soybean embryos in mid- to late-maturation stages (Meinke et al., 1981; Walling et al., 1986). In situ hybridization analysis of P-conglycinin expression (both a’ and /3 subunit mRNAs hybridized to the probe) showed no expression in globular-, heart-, or cotyledon-stage soybean embryos (Perez-Grau and Goldberg, 1989). Mid- to late-maturation-stage embryos expressed ,8-conglycinin in both cotyledons and the axis. In this analysis, there also was no expression in the endosperm.

Although seed development in tobacco and petunia is much more rapid than in soybean, previous results indicated that transgenic tobacco and petunia plants expressed the a’ P-conglycinin promoter in the corresponding mid to late stages of embryo development (Beachy et al., 1985; Chen et al., 1988; Naito et al., 1988; Chamberland et al., 1992). Since promoter sequences to -257 bp upstream of the transcription start site were shown to be sufficient for temporal control, as well as for maximal expression (Chen et al., 1986, 1988), our 550-bp promoter fragment was sufficient to contain all of the regulatory sequences. Using GUS staining of individual embryos, we observed expression in heart-stage embryos. Since the same heart-stage expression was observed in embryos of many independent transformants, this pattem is not related to a position effect. The staining of isolated embryos, which allows the visualization of limited localized expression, may allow detection of expression that is not detected above background in analyses of protein or total RNA prepared from bulk seed samples, the assays used in the other studies. Therefore, the early timing and endosperm expression that we observe, which differ from the endogenous gene expression in soybean seed as determined by in situ hybridization, may result from expression of the a’ P-conglycinin promoter in a heterologous host or from a more sensitive assay. The expression we see in the cotyledons and axis of older embryos is in agreement with the expression pattern in maturation- stage soybean embryos. Chamberland et al. (1992) also observed uniform staining of the cotyledons and axis of mature embryos bearing a GUS gene expressed from the a’ P- conglycinin promoter. They did not determine GUS localiza-

tion in earlier-stage embryos. They describe a light staining of the endosperm in mature seed, which indicates lower expression than we observe in that tissue.

The promoter of the French bean P-phaseolin gene has also been reported to be expressed in mid- to late-maturation stages (Sun et al., 1978; Murray and Kennard, 1984). In transgenic tobacco seeds analyzed as bulk samples, phaseolin protein (Sengupta-Gopalan et al., 1985) and P-phaseolin promoter-expressed mRNA (Riggs et al., 1989) or marker gene activity (Bustos et al., 1989) were first detected at a mid- maturation stage, Deletion analysis of this promoter in transgenic tobacco revealed a complex set of regulatory regions within the sequence between +20 and -795 (Bustos et al., 1991). Since our promoter fragment extends only to -410, it retains only a subset of these domains. It does include upstream activating sequence 1 and lack negative regulatory sequence 2, a combination that results in expression 3 d earlier than that directed by the full-length promoter region. Thus, the shorter length of our &phaseolin promoter may explain the early expression that we observe in heart-stage embryos. However, we do not lose expression in the hypo- cotyl region of mature embryos, as was reported by Bustos et al. (1991) for a -418 promoter. In their report, longer promoters were expressed throughout the embryo except in the root meristem and cap. All promoter fragments analyzed directed expression in the endosperm as well, in agreement with our observations of endosperm expression.

The pattern of expression directed by the a‘ 8-conglycinin and @-phaseolin promoters in tobacco seed was very consist- ent among independent transformants in this study. Others have also observed a relative insensitivity to position effects of seed-specific promoters (Chen et al., 1986). Therefore, we assume that the pattern of Cre expression directed by the a’ B-conglycinin and P-phaseolin promoters is similar to the pattem of GUS expression from the same promoters, initiating at a heart stage and extending throughout the embryo in a young torpedo stage. Cre expressed in this manner is able to activate the lox2-GUS gene throughout the entire embryo and endosperm. The pattem of activation in the embryo follows the pattern of promoter activity. Spots of activated GUS activity are first detected at the junction between the cotyledons and embryo axis, the site of highest seed promoter expression. A direct correlation between amount of Cre protein and frequency of recombination was previously established (Sauer and Henderson, 1988). Blue spots are first seen in torpedo-stage embryos, about 1 d in development after the a’ P-conglycinin and P-phaseolin promoters are first active, at the heart stage. This then is the period of time required for Cre protein to be synthesized, enter the nucleus, identify lox sites within the plant genome, and recombine these sites, producing a functional gene.

Entering the nucleus does not seem to be a limiting factor, since the addition of a nuclear-targeting signal to Cre did not improve the timing of lox2-GUS activation. Cre is a small, 38-kD protein that should be able to pass through the nuclear pore complex (Garcia-Bustos et al., 1991). However, small nuclear-localized proteins do contain localization signals to facilitate their entry (Raikhel, 1992). The SV40 large T- antigen nuclear localization signal did effectively direct the localization of T7 polymerase to the plant cell nucleus (Lass-



456 Odell et al. Plant Physiol. Vcll. 106, 1994

ner et al., 1991). Since we did not directly assay the effect of this signal on nuclear localization of Cre or on Cre enzymatic activity, it is possible that reduction in recombination rate and enhancement from nuclear targeting could have been balanced, opposing effects. In any case, our attempt to reduce the time period required for gene activation by addition of the SV40 large T-antigen nuclear localization signal to Cre was not successful.

Rapid expression of the lox2-GUS gene was achieved by directing recombination prior to the time when the seed- specific promoter was initially active. The GUS gene was then already in the active state when expression of its seed- specific promoter initiated, eliminating the delay in produc- tion of GUS enzyme activity. Cre expressed from the early, highly active 355 promoter/Cab leader directs this preactivation. Earlier expression of the 35s promoter than the a‘ p- conglycinin promoter was inferred previously in a study on seed-specific inhibition of GUS expression, since GUS regulated by a 35s promoter was expressed, then shut off by an antisense RNA expressed from the a’ 8-conglycinin promoter (Fujiwara et al., 1992). Expression of the 35s promoter by 10 dap was also discussed by Scofield et al. (1993). By staining for GUS activity in dissected embryos we were able to determine that the 35SC regulatory sequence is active in heart- stage embryos but not in globular embryos. Since the 35s promoter is active in many tissues and cell types (Nagy et al., 1985; Odell et al., 1985; Jefferson et al., 1987), its inactivity in globular embryos should be noted. Attempts to comple- ment a globular-embryo-stage mutation by expressing the wild-type coding region under control of the 35s promoter would not be effective. Inactivity in the globular embryo might be explained by the modular nature of the 35s promoter. It is composed of domains and subdomains that confer different tissue-specific and developmental pattems of expression (Benfey et al., 1989, 1990; Lam et al., 1989). One identified domain confers expression in radicle and root tissue, whereas another confers expression in cotyledons and most seedling tissues. These tissue-specific cis elements may not be recognized by cells of the undifferentiated globular embryo, but by the heart stage, cells of the embryo may have already acquired enough identity with radicle, axis, or cotyledon tissue to express factors that interact with the domains of the 35S/P, allowing its expression.

The presence of the activated GUS gene in F2 heart-stage embryos, even when the cre gene is not inherited, indicates that recombination occurred prior to FZ embryo formation. Activation could have occurred in the F1 embryonic shoot meristem or in any tissue of the F1 plant from which genetically effective cells arise. By GUS staining, both the 35s promoter and seed-specific promoters appear to be active in the shoot meristem region of the embryo. Activity of the p- phaseolin promoter in the embryonic shoot meristem was seen previously by GUS staining of mature transgenic tobacco seed (Bustos et al., 1991). Seed-specific promoters are not generally known to be active other than in the embryo, in concordance with Cre-mediated GUS gene activation occur- ring in the embryonic shoot meristem. Recently, however, expression of GUS directed by the p-phaseolin promoter was observed in apical meristems of transgenic tobacco plants (Sen et al., 1993).

Comparison of the efficiencies of recombinaíion directed by our 35SC-cre and 35s-cre genes supports the importance of the rnRNA sequence in contributing to the lek el of protein expression. Infemng from the timing of lox2-GUS gene activation, the 35SC-cre gene is expressed much more highly than the 35s-cre gene. These genes do have different 3‘ ends: Ocs 3’ versus Nos 3’, which could contribute to mRNA processing and stability. However, there is a striking difference in the sequence surrounding the translati on-initiating ATG, 35SC-cre having an almost perfect consenijus sequence (Kozak, 1986) and 35s-cre having little homology. Thus, mRNA translatability may be the predominant basis for the discrepancy in Cre expression.

During these experiments we also saw differences in levels of Cre expression, inferred from the timing of lo.rz-GUS gene activation, among plants belonging to the same 35s-cre line. Several different 35s-cre plants from the Cre3 line were used in cros!jes to activate lox2-GUS. They were all grown from seed produced by selfing an RI plant that was homozygous at a single locus for the linked selection marker Thus, these plants were expected to be genetically identical. ‘Two of them directed the same pattem of lox2-GUS-activated GUS activity in F1 seed as expected, whereas one directed much less activation. Reduced Cre activity in this plant, as compared to its siblings, could be due to methylation of the 35s-cre transgene. Variable loss of transgene expressiori among siblings has been correlated with methylation in some studies (Lin et al., 1990; Kilby et al., 1992; Asaad et al., 1993). We have not compared the methylation of the 3551-cre gene in this plant and its siblings.

The reduced recombination directed by this 35s-cre plant, and also by the plants of the Cre2 line, pr0duc.s a spotting pattem on maturing cotyledon-stage embryos. Small spots indicate sites of recent lox2-GUS activation, whereas larger spots indicate an earlier recombination event and include cells descending from the one in which loxZ-GIJS was activated. 1,imited recombination provides a system in which cell lineages may be followed. Limited recombinabon could be directed by low expression of Cre in any tissue, producing a mosaic pattern of expression of a target gene. This system could be used to restore function of a wild-type gene in the corresponding mutant background, in a cell-localized manner. Effects of cell-autonomous mutant restorohon on the tissue would indicate whether a cell’s response to the gene product is transmitted among the cells of the tissue or remains limited to the cell in which it is expressed. This study on tissue-specific Cre-mediated activation of a GUS marker gene, whose expression is initially blocked by a lox-bounded fragment, provides a basis for applying the Cr ./lox recombination system to these types of developmenta I studies.

ACKNOWLEDGMENTS

We wish to thank C. Patterson for generating transformed tobacco plants, :H. Hershey and E. Heppard for preparing the cre coding region with the 5’ NcoI site, C. Wandelt for the modified plasmids containing the seed promoters and for the nuclear localization signal, M. Locke for the binary vector pZS198, K. Reiter for help with chemiluminescent Southem blots, and J. Slightom for the 0-phaseolin and a’ j3-conglycinin promoters. We also thank S . Russell and J.




Shen for valuable discussions and E. Krebbers for critical reading of the manuscript.

Received April 14, 1994; accepted June 28, 1994. Copyright Clearance Center: 0032-0889/94/106/0447/12.

LITERATURE CITED

An G, Ebert PR, Mitra A, Ha SB (1988) Binary vectors. In SB Gelvin, RA Schilperoort, eds, Plant Molecular Biology Manual. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp A31-19

Assaad FF, Tucker KL, Signer ER (1993) Epigenetic repeat-induced gene silencing (RIGS) in Arabidopsis. Plant Mol Biol22 1067-1085

Bayley CC, Morgan M, Dale EC, Ow DW (1992) Exchange of gene activity in transgenic plants catalyzed by the Cre-lox site-specific recombination system. Plant Mol Biol18: 353-361

Beachy RN, Chen Z-L, Horsch RB, Rogers SG, Hoffman NJ, Fraley RT (1985) Accumulation of soybean B-conglycinin in seeds of transformed petunia plants. EMBO J 4 3047-3053

Benfey PN, Ren L, Chua N-H (1989) The CaMV 35s enhancer contains at least two domains which can confer different developmental and tissue-specific expression pattems. EMBO J 8:

Benfey PN, Ren L, Chua N-H (1990) Tissue-specific expression from CaMV 35s enhancer subdomains in early stages of plant development. EMBO J 9 1677-1684

Bustos MM, Begum D, Kalkan FA, Battraw MJ, Hall TC (1991) Positive and negative cis-acting DNA domains are required for spatial and temporal regulation of gene expression by a seed storage protein promoter. EMBO J 10: 1469-1479

Bustos MM, Guiltinan MJ, Jordan0 J, Begum D, Kalkan FA, Hall TC (1989) Regulation of 0-glucuronidase expression in transgenic tobacco plants by an A/T-rich, cis-acting sequence found upstream of a French bean @-phaseolin gene. Plant Cell 1: 839-853

Chamberland S, Daigle N, Bernier F (1992) The legumin boxes and the 3’ part of the soybean P-conglycinin promoter are involved in seed gene expression in transgenic tobacco plants. Plant Mol Biol 1 9 937-949

Chen Z-L, Pan N-S, Beachy RN (1988) A DNA sequence element that confers seed-specific enhancement to a constitutive promoter.

Chen Z-L, Schuler MA, Beachy RN (1986) Functional analysis of regulator elements in a plant embryo-specific gene. Proc Natl Acad Sci USA 8 3 8560-8564

Dale EC, Ow DW (1990) Intra- and intermolecular site-specific recombination in plant cells mediated by bacteriophage P1 recombinase. Gene 91: 79-85

Dale EC, Ow DW (1991) Gene transfer and subsequent removal of the selection gene from the host genome. Proc Natl Acad Sci USA 88: 10558-10562

Ditta G, Stanfield S, Corbin D, Helinski D (1980) Broad host range DNA cloning system for Gram-negative bacteria: construction of a gene bank of Rhizobium meliloti. Proc Natl Acad Sci USA 77:

Doyle JJ, Schuler MA, Godette WD, Zenger V, Beachy RN, Sligh- tom JL (1986) The glycosylated seed storage proteins of GZycine max and Phaseolus vulgaris. J Biol Chem 261: 9228-9238

Fujiwara T, Lessard PA, Beachy RN (1992) Seed-specific repression of GUS activity in tobacco plants by antisense RNA. Plant Mol

Garcia-Bustos J, Heitman J, Hall MN (1991) Nuclear protein localization. Biochim Biophys Acta 1071: 83-101

Harpster MH, Townsend JA, Jones JDG, Bedbrook J, Dunsmuir P (1988) Relative strengths of the 35s cauliflower mosaic virus, l’, 2’, and nopaline synthase promoters in transformed tobacco, sugarbeet, and oilseed rape callus tissue. Mol Gen Genet 212

Itoh Y, Watson JM, Haas D, Leisinger T (1984) Genetic and molecular characterization of the Pseudomonas plasmid-PVS1. Plasmid

Jefferson RA (1987) Assaying chimeric genes in plants: the GUS

2195-2202

EMBO 7: 297-302

7347-7351

Bi0120 1059-1069

182-190

11: 2287-2300

gene fusion system. Plant Mol Biol Rep 5: 387-405

Jefferson RA, Kavanagh TA, Bevan MW (1987) GUS fusions: 0- glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J 6: 3901-3907

Kilby NJ, Ottoline Leyser HM, Furner IJ (1992) Promoter methylation and progressive transgene inactivation in Arabidopsis. Plant Mol Biol20 103-112

Kozak M (1986) Point mutations define a sequence flanking the AUG initiator codon that modulates translation by eukaryotic ribosomes. Cell 4 4 283-292

Lam E, Benfey PN, Gilmartin PM, Fang R-X, Chua N-H (1989) Site-specific mutations alter in-vitro factor binding and change promoter expression pattem in transgenic plants. Proc Natl Acad Sci USA 8 6 7890-7894

Lassner MW, Jones A, Daubert S, Comai L (1991) Targeting of T7 RNA polymerase to tobacco nuclei mediated by an SV40 nuclear location signal. Plant Mol Biol 17: 229-234

Lin F, Heidmann I, Saedler H, Meyer P (1990) Epigenetic changes in the expression of the maize A1 gene in Petunia hybrida: role of numbers of integrated gene copies and state of methylation. Mol Gen Genet 222: 329-336

Lloyd AM, Davis RW (1994) Functional expression of the yeast FLP-FRT site-specific recombination system in Nicotiana tabacum. Mol Gen Genet 242 653-657

Lyznik LA, Mitchell JC, Hirayama L, Hodges TK (1993) Activity of yeast FLP recombinase in maize and rice protoplasts. Nucleic Acids Res 21: 969-975

Maeser S, Kahmann R (1991) The Gin recombinase of phage Mu can catalyze site-specific recombination in plant protoplasts. Mol Gen Genet 230 170-1 76

Mazur BJ, Chui C-F (1985) Sequence of a genomic DNA clone for the small subunit of ribulose bis-phosphate carboxylase-oxygenase from tobacco. Nature 313 810-812

Meinke DW, Chen J, Beachy RN (1981) Expression of storage- protein genes during soybean seed development. Planta 153

Mersereau M, Pazour GJ, Das A (1990) Efficient transformation of Agrobacterium tumefaciens by electroporation. Gene 9 0 149-151

Murray MG, Kennard WC (1984) Altered chromatin conformation of the higher plant gene phaseolin. Biochemistry 2 3 4225-4232

Nagy F, Odell JT, Morelli G, Chua N-H (1985) Properties of expression of the 355 promoter from CaMV in transgenic tobacco plants. In M Zaitlin, P Day, A Hollaender, eds, Biotechnology in Plant Science: Relevance to Agriculture in the Eighties. Academic Press, New York, pp 227-236

Naito S, Dube PH, Beachy RN (1988) Differential expression of conglycinin CY’ and 0 subunit genes in transgenic plants. Plant Mol Biol 11: 109-123

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

Odell JT, Nagy F, Chua N-H (1985) Identification of DNA sequences required for activity of the cauliflower mosaic virus 35s promoter. Nature 313: 810-812

Odell JT, Russell SH (1994) Use of site-specific recombination systems in plants. In J Paszkowski, ed, Homologous Recombination in Plants. Kluwer Academic Publishers, Dordrecht, The Nether- lands, pp 219-270

Onouchi H, Yokoi K, Machida C, Matsuzaki H, Oshima Y, Mat- suoka K, Nakamura K, Machida Y (1991) Operation of an efficient site-specific recombination system of Zygosaccharomyces rouxii in tobacco cells. Nucleic Acids Res 1 9 6373-6378

Perez-Grau L, Goldberg RB (1989) Soybean seed protein genes are regulated spatially during embryogenesis. Plant Cell 1: 1095-1 109

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

Raikhel N (1992) Nuclear targeting in plants. Plant Physiol 100

Riggs CD, Hunt DC, Lin J, Chrispeels MJ (1989) Utilization of luciferase fusion genes to monitor differential regulation of phy- tohemagglutinin and phaseolin promoters in transgenic tobacco. Plant Sci 6 3 47-57

130-139

369-378

1627-1632




Rihs H-P, Peters R (1989) Nuclear transport kinetics depend on phosphorylation-site-containing sequences flanking the kary- ophilic signal of the simian virus 40 T-antigen. EMBO J 8:

Russell SH, Hoopes JL, Odell JT (1992) Directed excision of a transgene from the plant genome. Mol Gen Genet 234 49-59

Sauer B, Henderson N (1988) Site-specific DNA recombination in mammalian cells by the Cre recombinase of bacteriophage P1. Proc Natl Acad Sci USA 8 5 5 166-5 170

Scofield SR, English JJ, Jones JDG (1993) High level expression of the Activator transposase gene inhibits the excision of Dissociation in tobacco cotyledons. Cell 75: 507-517

Sen P, Chlan CA, Burow MD, Lee WS, Murai N (1993) Apical and lateral shoot apex-specific expression is conferred by a promoter of the seed storage protein @-phaseolin gene. Transgenic Res 2

1479-1484

21-28

Sengupta-Gopalan C, Reichert NA, Barker RF, Hall TC, Kemp JD (1985) Developmentally regulated expression of the bean @-phaseolin gene in tobacco seed. Proc Natl Acad Sci USA 3 2 3320-3324

Shure h4, Wessler S, Fedoroff N (1983) Molecular ide itification and isolation of the waxy locus in maize. Cell 35: 225-233

Slightoim JL, Chee PP (1991) Protein expression via seed specific regulatory sequences. Patent application No. WO 9 113993

Sun SM, Mutschler MA, Bliss FA, Hall TC (1978) Protein synthesis and accumulation in bean cotyledons during growth. Plant Physiol

van den Elzen PJM, Townsend J, Lee KY, Bedbrook JR (1985) A chimeric hygromycin resistance gene as a selectable marker in plant cells. Plant Mol Biol 5 299-302

Walling; L, Drews GN, Goldberg RB (1986) Transcriptional and post-iranscriptional regulation of soybean seed protein mRNA levels. Proc Natl Acad Sci USA 83 2123-2127

61: 918-923



seed-specific gene activation mediated by the cre//ox site-specif ic recombination system · lhe...

Documents