The Plant Cell, Vol. 4, 119-128, February 1992 O 1992 American Society of Plant Physiologists

RESEARCH ARTICLE

Cloning the Arabidopsis GA7 Locus by Genomic Subtraction

Tai-ping Sun,’ Howard M. Goodman, and Frederick M. Ausube12 Department of Genetics, Harvard Medical School, and Department of Molecular Biology, Massachusetts General Hospital, Boston, Massachusetts 02114

Arabidopsis thaliana ga l mutants are gibberellin-responsive dwarfs. We used the genomic subtraction technique to clone DNA sequences that are present in wild-type Arabidopsis (ecotype Landsberg erecta, Ler) but are missing in a presump- tive g a l deletion mutant, gal-3. The cloned sequences correspond to a 5.0-kb deletion in the gal-3 genome. Three lines of evidence indicated that the 5.0-kb deletion in the gal-3 mutant is located at the GAl locus. First, restriction fragment length polymorphism mapping showed that DNA sequences within the 5.0-kb deletion map to the GAl locus. Second, cosmid clones that contain wild-type DNA inserts spanning the deletion in gal-3 complemented the dwarf phenotype when integrated into the gal-3 genome by Agrobacterium tumefaciens-mediated transformation. Third, we identified molecular lesions in four additional ga l alleles within the 5.0-kb region deleted in mutant gal-3. One of these lesions is a large insertion or inversion located within the most distal intron encoded by the GAl locus. The three other lesions are all single base changes located within the two most distal exons. RNA gel blot analysis indicated that the GAl locus encodes a 2.8-kb mRNA. We calculated a recombination rate of 10-5 cM per nucleotide for the GAl region of the Arabidopsis genome.

INTRODUCTION

Cloning a plant gene that is defined solely by a mutant pheno- type can be an arduous task. Chromosome walking, which is used to isolate a gene that is linked to a restriction fragment length polymorphism (RFLP) marker, is currently laborious and can be impeded by unclonable sequences or stretches of repetitive DNA. Gene tagging systems that use engineered heterologous transposable elements to generate random mu- tations over an entire genome are being developed but are not yet available for most plant species (Jones et al., 1989; Martin et al., 1989; Masson and Fedoroff, 1989). In the case of Arabidopsis, Agrobacterium tumefaciens T-DNA has been used successfully to tag a variety of genes (Feldman et al., 1989; Feldman, 1991). However, individual transformed plants have to be generated for each T-DNA insertion. In the case of Arabidopsis with a genome size of 108 bp (B. Hauge and H.M. Goodman, unpublished data), over 100,000 independent transformants (Clarke and Carbon, 1979) are required to achieve 95% probability of having an insert every 2 kb in the genome, assuming an average of 1.4 random inserts per plant (Feldman, 1991).

Recently, our laboratory developed a technique called “genomic subtraction” that provides a useful addition to the chromosome walking and gene tagging approaches for clon- ing genes corresponding to a mutant phenotype (Straus and

’ Current address: Department of Botany, Duke University, Durham, NC 27706.

To whom correspondence should be addressed.

Ausubel, 1990). The procedure identifies DNA sequences that are present in a wild-type organism (target sequences) but are missing in a homozygous deletion mutant. Genomic subtrac- tion involves multiple cycles of hybridization between wild-type DNA and an excess of biotinylated deletion mutant DNA. In each cycle, avidin-coated beads are added to remove com- mon sequences present in both wild-type and mutant DNA samples. The remaining DNA after several cycles of subtrac- tion should be enriched for sequences that are absent in the deletion mutant. Following several cycles of subtraction, adap- tors are ligated to the ends of the remaining DNA fragments that are not bound to avidin beads and the DNA fragments are amplified by the polymerase chain reaction (PCR) for further analyses.

Genomic subtraction was developed using a defined 5.0-kb deletion of Saccharomyces cerevísiae, which has a genome size of .u2 x 107 bp. When contemplating the extension of genomic subtraction to clone plant genes, it was necessary to take into account the fact that genomic subtraction requires self-annealing of the target DNA sequences. Because the con- centration of target DNA sequences decreases as the genome size increases, we initially sought to determine whether the technique could be successfully used with Arabidopsis, the plant with the smallest known genome (Meyerowitz and Pruitt, 1985). When we initiated this project, there were no known deletion mutants of Arabidopsis. However, one presumptive deletion had been described at the Arabidopsis GAf locus (Koornneef et al., 1983a).

120 The Plant Cell

gal-4i

gal-1gal-9gal-10

i

gai-2gal-6gal-8 gal

1.5 5.5 6.6 7.1

ga1-3Figure 1. Genetic Map of the Arabidopsis GA1 locus in cM x 10~2

(Koornneef et al., 1983a).

The alleles have been renamed as described in Methods. The extentof the presumptive deletion in ga1-3 is indicated by the heavy horizontalline.

Arabidopsis gal mutants are nongerminating, gibberellin(GA)-responsive, male-sterile dwarfs whose phenotype canbe converted to wild type by repeated application of GA(Koornneef and van der Veen, 1980). GAs, especially GA, andGA*, are diterpenoid plant growth hormones required for seedgermination, leaf expansion, stem elongation, flowering, andfruit set in Arabidopsis (Talon et al., 1990). The biosynthesisof GAs in the gal mutants is blocked prior to the synthesisof enf-kaurene (Barendse et al., 1986; Zeevaart, 1986), a keyintermediate in the pathway (Graebe, 1987). As diagrammedin Figure 1, Koornneef et al. (1983a) constructed a fine-structuregenetic map of the Arabidopsis GA1 locus using nine indepen-dently isolated gal alleles. Three of these gal alleles (ga1-2,ga1-3, gal-4) were generated by fast neutron bombardmentand six (ga1-1, ga1-6, gal-7, ga1-8, ga1-9, ga1-10) were gener-ated by ethyl methane sulfonate mutagenesis. One of the fast-neutron-generated mutants, ga1-3, failed to recombine with thesix alleles indicated in Figure 1 and was classified as an intra-genic deletion (Koornneef et al., 1983a).

Based on the genetic analysis at the GA1 locus, we usedgenomic subtraction to clone the Arabidopsis GA1 locus usingsheared biotinylated DMA from mutant gal-3 and Sau3A-digested nonbiotinylated DMA from the wild-type parent,ecotype Ler. To determine the efficiency of the genomic sub-traction technique for genomes the size of Arabidopsis, wefirst carried out a reconstruction experiment in which bio-tinylated Arabidopsis DNA was used to subtract the same DNApreparation to which a small amount of adenovirus (Ad-2) DNAhad been added.

RESULTS

Extension of Genomic Subtraction to Arabidopsis:Reconstruction Experiment Using Arabidopsis DNAContaining Ad-2 DNA

To determine the efficiency of the genomic subtraction tech-nique for Arabidopsis, we carried out a reconstruction

experiment using the protocol outlined in Methods in which12.5 ug biotinylated Arabidopsis (ecotype Ler) DNA was usedto subtract the same DNA preparation (0.5 ug) to which an ap-proximately single-copy level (0.5 ng) of Ad-2 DNA had beenadded. After four cycles of subtraction, the starting materialsand the end products of each cycle were ligated with Sau3Aadaptors, amplified by PCR, and analyzed on a 1.5% agarosegel. Figure 2 shows that Ad-2-specific bands could be detectedin PCR-amplified DNA starting at the end of the third cycleof subtraction.

The amplified unbound DNA from each cycle of subtractionwas ligated into the Smal site of pUC13. Figure 3 shows thedegree of enrichment of Ad-2 DNA after each of three cyclesof subtraction. Approximately 4000 colonies containing clonedSau3A fragments from each cycle were plated onto nitrocellu-lose filters (Hanahan and Meselson, I983) and hybridized with32P-labeled Ad-2 DNA. In a separate experiment, 200 to 400colonies containing cloned Sau3A fragments from each cyclewere plated per filter to facilitate counting the number of clonesthat contained Ad-2 sequences (data not shown). Before sub-traction, 0.1% of the cloned Sau3A fragments hybridized tothe Ad-2 DNA probe. On average, there was a 10-fold enrich-ment of the Ad-2 DNA per cycle in each of the first three cy-cles of subtraction, and, after four cycles of subtraction, morethan 95% of the colonies contained Ad-2 sequences.

1 2 3 4 5 6kb1.6 -

1.0 -

0.50.40.3

Figure 2. PCR-Amplified Unbound DNA from the Genomic Subtrac-tion Reconstruction Experiment.

PCR-amplified products were size fractionated on a 1.5% agarose gelcontaining 0.5 iig/mL ethidium bromide. Lane 1, starting material be-fore subtraction; lanes 2 to 5, unbound DNA after first to fourth cyclesof subtraction; lane 6, Sau3A-digested Ad-2 DNA.

Cloning the Arabidopsis GA1 Locus 121

BEFORESUBTRACTION FIRST CYCLE

SECOND CYCLE THIRD CYCLE

Figure 3. Enrichment of Ad-2 DMA in the Reconstruction Experiment.Autoradiogram ot colony hybridization experiment is shown for filterscontaining ~4000 recombinant colonies after each cycle of subtrac-tion probed with 32P-labeled Ad-2 DNA.

To determine the extent of the deletion in ga1-3 DNA identi-fied by pGAM, pGAM DNA was used as a hybridization probeto identify larger genomic fragments of wild-type ArabidopsisDNA. Two such clones, XGA1-3, identified in a XFIX library ofLer genomic DNA, and pGA1-4, identified in a pOCA18 cos-mid library of Columbia genomic DNA, are shown in Figure5. A subclone of XGA1-3, pGA1-2 (Figure 5), was used to probea DNA gel blot containing Hindlll-digested DNA from wild-typeLer and from several gal mutants. As shown in Figure 4B anddiagrammed in Figure 5, pGA1-2 hybridized to four Hindlll frag-ments (1.0 kb, 1.2 kb, 1.4 kb, and 5.6 kb) in wild-type DNA thatwere absent in ga1-3 DNA. The deletion mutation results inan extra Hindlll fragment (4.2 kb) in ga1-3 DNA. These resultsand additional DNA gel blot analyses (data not shown) showedthat the deletion in ga1-3 is 5 kb, corresponding to 0.005% ofthe Arabidopsis genome.

RFLP Mapping and Complementation Analysis

Several lines of evidence indicated that the 5.0-kb deletion inmutant ga1-3 identified by genomic subtraction correspondsto the GA1 locus. First, RFLP mapping analysis carried outby the procedure detailed in Nam et al. (1989) using XGA1-3(Figure 5) as a hybridization probe showed that XGA1-3 maps

1 2 3 4B

1 2 3 4

5.6 kb-

Cloning the Arabidopsis GA1 Locus by GenomicSubtraction

Following the successful completion of the Ad-2 reconstruc-tion experiment, we used genomic subtraction, as describedin Methods, to isolate the Arabidopsis G/47 locus using bi-otinylated ga1-3 DNA and Sau3A-digested DNA from the wild-type parent, ecotype Ler. A control experiment was carried outin parallel in which biotinylated gal-3 DNA was used to sub-tract Sau3A-digested Ler DNA that contained an approximatelysingle-copy level (0.5 ng) of Ad-2 DNA. In accordance with theprevious reconstruction experiment, Ad-2-specific bands couldbe detected in the PCR-amplified subtraction products afterthe third cycle of subtraction (data not shown). In the actualsubtraction experiment, DNA fragments remaining after fivecycles of subtractive hybridization were amplified by PCR andcloned in pUC13. Six individual clones were radiolabeled andused as hybridization probes for DNA gel blot analyses. Asshown in Figure 4A, one of these six clones, pGA1-1, containeda 250-bp SauSA fragment that hybridized to a 1.4-kb Hindlllfragment in DNA isolated from wild-type Ler DNA and fromthe ga1-2 and ga1-4 mutants but did not hybridize to ga1-3 DNA.

— 4.2 kb

-« 3.3 kb

( — 1.4kb 1.4kb-1.2kb-1.0kb-

:•*— 1.3kb

Figure 4. Detection of Deletions and Insertions (or Inversion) in ga1-3and gaJ-2 DNA, Respectively.

(A) Autoradiogram of DNA gel blot probed with the 250-bp Sau3A frag-ment from pGA1-1. The blot contains 1 jig Hindlll-digested DNA iso-lated from Landsberg erecfa, lane 1, and from three gar mutants, ga1-3,lane 2; ga1-4, lane 3; and ga7-2, lane 4.(B) The same blot shown in (A) probed with the 6-kb fragment frompGA1-2 that covers the entire deleted region in ga7-3 (Figure 5). Thearrows indicate altered Hindlll fragments in ga1-3 (4.2 kb) and ga1-2(1.3 and 3.3 kb).

122 The Plant Cell

11 11 ' ' gal-3

I A Px I5.6 / 1.21.<M/

, ' gal-6r< n i —— L- 1

gai-8(G— A)

-|A.GA1-q

I - II I.0 8.0

x

ga1-£ N

gal -7(G— A)

Figure 5. Physical Map of the Arabidopsis GA1 Locus.

The heavy horizontal line is a Hindlll restriction map. The location ofa 5-kb deletion in ga7-3 is indicated by the hatched box. The horizon-tal lines above the restriction map indicate the extent of the sequencescontained in the X clone XGA1-3, the plasmid pGA1-2, and the cosmidclone pGA1-4. The diagram below the horizontal lines depicts the lo-cation of introns (lines) and exons (open boxes) of the GA1 gene andthe locations of the insertion or inversion mutation in ga1-2 and thepoint mutations in ga7-6, ga1-7, and gal-8 alleles. The gal locus of ga1-2was interrupted by unknown DMA between nucleotides 159 and 161(with a nucleotide C missing) in the intron. The single-nucleotidechanges in ga1-6, ga1-8, and gal-7 are at nucleotides 200 and 205 ofthe penultimate exon and at nucleotide 72 of the last exon, respectively.

to the telomere proximal region at the top of chromosome 4(data not shown), consistent with the location to which the GA1locus had been mapped previously by Koornneef et al. (1983b).

Second, the cosmid clone pGA1-4 (Figure 5), which con-tains a 20-kb insert of wild-type (ecotype Columbia) DNA span-ning the deletion in gal-3, complemented the dwarf phenotypeof the ga1-3 mutant. pOCA18, the parental vector of pGAl-4,is a cosmid cloning vector that contains the right and leftborders and the "overdrive" sequences of Agrobacterium T-DNArequired for efficient transfer of cloned DNA into plant genomes.It also carries a chimeric neomycin phosphotransferase genethat confers kanamycin resistance in transgenic plants(Olszewski et al., 1988). Agrobacterium strain LBA4404 con-taining pGA1-4 was used to infect root explants of mutant gal-3,and kanamycin-resistant (Kmr) transgenic plants wereselected as described (Valvekens et al., 1988). One hundredand thirty Kmr plants (T1 generation) were regenerated thatset seeds in the absence of exogenous GA. Fifty to 300 seedsfrom each of four different T1 plants showed 100% linkage ofthe GA1 + and Kmr phenotypes, which segregated approxi-mately 3:1 in relation to the GA1~ kanamycin-sensitive pheno-type (T2 generation). As shown in Figure 6, all T2 generationGA1+/Kmr transgenic plants tested did not require exogenousGA for normal growth. Germination, stem elongation, and seedset were the same in the transgenic plant as in the wild-typeplant without exogenous GA treatment.

As shown in lanes 4 and 5 of Figure 7A, DNA gel blot analy-sis, using the 6-kb fragment from pGA1-2 as a probe, indicatedthat both the endogenous mutant gal-3 locus (4.2-kb Hindlllfragment) and the transgene wild-type (Columbia) GA1 locus

(5.0-, 1.4-, 1.2-, and 1.0-kb Hindlll fragments) are present in twoindependent T3 transgenic lines. Further DNA gel blot analy-sis, shown in lanes 3 and 4 of Figure 7B, using pOCA18, whichcontains the T-DNA border sequences as a probe, showed thatonly two border fragments were present in the genomes of bothtransgenic plants. These results indicated that a copy of thewild-type GA1 transgene was integrated at a single locus inthe genomes of both transgenic plants.

Identification of Four Other gal Alleles within theRegion Deleted in gal-3

To obtain unequivocal evidence that the 5.0-kb deletion in ga1-3corresponds to the GA1 locus, we showed that four additional

Figure 6. Phenotypes of 6-Week-Old Arabidopsis Landsberg erecfaPlants.

From left to right, a ga1-3 mutant, a transgenic ga1-3 plant containingthe 20-kb insert from pGA1-4 (Figure 5), and a wild-type Landsbergerecta plant. Seeds of transgenic gaJ-3 and wild-type plants were ger-minated on agarose plates containing 1 x Murashige and Skoog (1962)salts and 2% sucrose (MS plates) with or without kanamycin. Seedsof the mutant ga1-3 were soaked in 100 nM GA3 for 4 days before be-ing germinated on MS plates. Seven-day-old seedlings were trans-ferred to soil. To show the dwarf phenotype, no additional GA3 wasgiven to the mutant gal-3 after germination.

Cloning the Arabidopsis GA1 Locus 123

1 2 3 4 5B

1 2 3 4

4.2 kbkb12 -

Q __

6 -

kb5.65.0

1.41.21.0

Figure 7. DMA Gel Blots of Transgenic Plants Containing the ClonedGA1 Gene.(A) Autoradiogram of DMA gel blot probed with the 6-kb fragment frompGA1-2. Blot contains Hindlll-digested DMA from Landsberg erecta,lane 1; Columbia, lane 2; ga1-3, lane 3; and two T3 generation trans-genic ga1-3 plants transformed with pGA1-4, lanes 4 and 5. The insertin pGA1-4 was isolated from the Columbia ecotype. As seen in lanes1 and 2, pGA1-2 detected an RFLP between the Landsberg (5.6 kb)and Columbia (5.0 kb) DMAs. The DMA in lanes 1, 2, and 3 was puri-fied by CsCI density gradient centrifugation, whereas the DMA in lanes4 and 5 was purified by a minipreparation procedure (Dellaporta etal., 1983). This accounts for the differences in mobilities of the hybridiz-ing bands in lanes 1, 2, and 3 compared to lanes 4 and 5.(B) Autoradiogram of DNA gel blot probed with pOCA18 DMA, whichis the vector for pGA1-4 (Olszewski et al., 1988). Blot contains Hindlll-digested DNA from Columbia, lane 1; ga1-3, lane 2; and the same twoT3 generation transgenic ga1-3 plants transformed with pGA1-4 shownin (A), lanes 3 and 4.

of PCR products from ga1-2 DNA (data not shown) indicatedthat the gal gene of ga1-2 mutant is interrupted by at least3.4-kb DNA of unidentified origin within the last intron definedby the partial cDNA clone (Figure 5). The mutation could bedue to either an insertion or an inversion event. RFLP map-ping of the unknown "insertion" DNA would be needed to dis-tinguish between these two possibilities. DNA gel blot analyses,using pGA1-2 (Figure 4B) and pGA1-4 (data not shown) asprobes, showed that there are no detectable deletions orinsertions in ga1-4 DNA.

The ethyl methane sulfonate-induced ga1-6, ga1-7, and ga1-8alleles are located at or near the gal-2 allele on the geneticmap (Figure 1). Direct sequencing of PCR products amplifiedfrom ga1-6, ga1-7, and ga1-8 mutant DNA templates revealedsingle nucleotide changes in the 1.2-kb Hindlll fragment in eachof the three mutants (Figure 5). Mutant gal-7, which definesone side of the genetic map, contained a single nucleotidechange in the most distal exon of the GA1 gene. Mutants ga1-6and gal-8 contain single base changes in the penultimate exon.The single nucleotide changes in mutants gal-6, gal-7, andgal-8 result in missense mutations, which is consistent withtheir leaky phenotypes (Koornneef et al., 1983a). It is unlikelythat the base changes observed in mutants ga1-6, ga1-7, andgal-8 are PCR artifacts or are due to the highly polymorphicnature of the GA1 locus because the 1.2-kb Hindlll fragmentsamplified and sequenced from ga1-1 and ga1-9 both had thewild-type sequence. Moreover, the PCR products were se-quenced directly and the sequence analysis was carried outtwice, using the products of two independent amplificationsfor each allele examined.

gal alleles contain alterations from the wild-type sequencewithin the region deleted in ga1-3 in the order predicted bythe genetic map. To aid in this analysis, a partial GA1 cDNAclone (0.9 kb), containing poly(A) and corresponding to the1.2-kb Hindlll fragment (Figure 5), was isolated from a cDNAlibrary constructed from RNA isolated from siliques (seedpods) of Arabidopsis ecotype Columbia (J. Giraudat and H.M.Goodman, unpublished data). Four exons and three intronsin the 1.2-kb Hindlll fragment in the GA1 locus (Figure 5) werededuced by comparison of the cDNA and genomic DNA se-quences (data not shown). The identification of this cDNA cloneshowed that the 1.2-kb Hindlll fragment is located at the 3'endof the GA1 gene and suggested that the mutations in the galalleles 2, 3, 6, 7, and 8 should also be located at the 3' endof the GA1 gene (Figures 1 and 5).

In addition to ga1-3, two other mutants, gal-2 and gal-4, weregenerated by fast neutron mutagenesis (Koornneef et al.,1983a). As shown in Figure 4B, the 1.2-kb Hindlll fragmentin ga1-2 DNA was replaced by two new fragments of 1.3 and3.3 kb without alteration of the adjacent 1.4- and 5.6-kb frag-ments. Further DNA gel blot analysis and DNA sequencing

The GA1 Locus Encodes a 2.8-kb mRNA

Figure 8 shows that the GA1 locus encodes a 2.8-kb mRNAdetected by RNA gel blot analysis using the 0.9-kb partial GA1cDNA as a hybridization probe. As expected, this RNA couldnot be detected in the deletion mutant ga1-3. The amount ofRNA in each lane is the same as shown in lanes 1, 2, and4 using the CAB gene as a probe. The decreased hybridiza-tion of the CAB probe in lane 3 is due to low expression ofthe CAB gene in the siliques. The GA1 locus was expressedat an extremely low level as reflected by the length of the ex-posure time needed for the blots: more than 3 weeks for GA1mRNA versus 2 min for the CAB mRNA. The low level of ex-pression of the GA1 locus was also reflected in the fact thatthe frequency of GA1 clones in the silique cDNA library was.5 x 10~5. The level of expression of the GA1 gene in imma-ture siliques (lane 3) seemed to be slightly higher than thatin whole plants (lanes 1 and 2). This is consistent with the ob-servation that the maximum enf-kaurene synthesizing activitywas found in immature siliques (Zeevaart, 1986). However, amarker gene with constitutive expression in different tissuesneeds to be used for quantitative analysis in future studies.

124 The Plant Cell

1 2 3 4

GA1 2.8kb

One of the seven clones contained a highly repetitive sequence,three appeared to be single-copy Arabidopsis sequences thathybridized to both Ler and ga1-3 DNA, and three containedinserts that only hybridized to the preparation of Ler DNA usedin the subtraction experiment but did not hybridize to otherpreparations of Ler DNA.

CAB

Figure 8. Detection of a 2.8-kb mRNA Using a GA1 cDNA Probe.

Autoradiogram of an RNA blot probed with a 32P-labeled 0.9-kb GA1cDNA or CAB DNA (chlorophyll a/b-binding protein gene). RNA wasfrom 4-week-old wild-type plants, lane 1; 5-week-old wild-type plants,lane 2; immature wild-type siliques, lane 3; and 4-week-old ga1-3 plants,lane 4.

Several attempts to isolate the full-length GA1 cDNA by am-plifying the 5' region of the cDNA using the PCR technique(Frohman et al., 1988) have not been successful. This is prob-ably due to the secondary structure and the extremely lowabundance of the GA1 mRNA.

Efficiency of Enrichment of GA1 Sequences byGenomic Subtraction

In contrast to the genomic subtraction experiment carried outin yeast (Straus and Ausubel, 1990), we found that the ampli-fied Sau3A fragments after five cycles of subtraction were notsufficiently enriched for screening an Arabidopsis genomiclibrary directly. To quantitate the degree of enrichment of GA1sequences, 70 clones from the fifth cycle of subtraction wereprobed with pGA1-2 and three positive clones were identified.A similar experiment was carried out with 180 clones from twoadditional cycles of subtraction. Interestingly, no further en-richment was achieved in these two cycles. On average, onein 15 to one in 20 clones contained Sau3A fragments locatedin the 5-kb deleted region. Three classes of sequences werefound among 13 positive clones examined by DNA gel blotanalyses. Four clones were identical to pGA1-1, eight clonescontained a 220-bp Sau3A fragment that is located within the1.0-kb Hindlll fragment, and one clone contained a 150-bpSau3A fragment located within the 1.4-kb Hindlll fragment (datanot shown). The biased enrichment of certain Sau3A fragmentsis most likely due to preferential amplification of particular frag-ments during the PCR reaction (Wieland et al., 1990).

We also examined seven individual clones that did not hy-bridize to pGA1-2 to determine the types of sequences thatwere not removed by five cycles of subtractive hybridization.

DISCUSSION

Genomic Subtraction for Isolating Genes fromArabidopsis

In cloning the GA1 locus by genomic subtraction, only one in20 cloned SauSA fragments following five cycles of subtrac-tion corresponded to the 5-kb deletion in the ga1-3 mutant.This level of enrichment was insufficient to use the amplifiedproduct directly as a hybridization probe for screening anArabidopsis genomic library. Instead, individual clones wereanalyzed by DNA gel blot hybridization. Because no furtherenrichment was achieved after the fifth cycle, it is possiblethat the contaminating Arabidopsis fragments had sequence-specific features that prevented them from being subtracted,despite the fact that they hybridized to Arabidopsis DNA ona DNA gel blot at normal stringency. The contaminating se-quences that only hybridized to the particular preparation ofLer DNA used in the subtraction experiment most likely origi-nated from contaminating microorganisms because theArabidopsis plants used in these experiments were not grownunder sterile conditions. To eliminate foreign DNA, it is clearlydesirable to isolate wild-type DNA from plants grown understerile conditions and to use disposable tubes and pipettesfor DNA purification.

In carrying out genomic subtraction, we found that it wasconvenient to carry out a parallel reconstruction experimentusing biotinylated deletion mutant DNA to subtract Sau3A-digested wild-type DNA containing a small amount of Sau3A-digested adenovirus DNA (as described in Results). Specificadenovirus DNA fragments can be directly visualized on anethidium bromide-stained gel in the amplified DNA productsafter the third cycle of subtraction. This offers a quick indica-tion of the efficiency of the subtraction.

Genomic subtraction is not labor intensive, and the resultsreported here indicate that genomic subtraction could be rou-tinely used to clone other nonessential Arabidopsis genes ifa method were available for generating deletions at reason-able frequency. In addition to the fast-neutron-induced dele-tion in mutant ga1-3 (Dellaert, 1980; Koornneef et al., 1982,1983a), x-ray and y-ray irradiation have also been shown tocause short viable deletions in Arabidopsis at the ch!3(Wilkinson and Crawford, 1991), ff3 (B. Shirley and H.M.Goodman, unpublished data), and g/7 (Oppenheimer et al.,1991) loci. Experiments are in progress to systematically test

Cloning the Arabidopsis GA7 Locus 125

severa1 mutagens to determine the optimal mutagenesis con- ditions for generating deletions in the Arabidopsis genome at high frequency.

Interestingly, the ga7-3 mutant in which we identified a 5.0-kb deletion at the GA7 locus was only backcrossed once to wild type before being used for genetic analysis (M. Koornneef, per- sonal communication). Because genomic subtraction only identified DNA sequences that map to the GA7 locus, it seems likely that the fast neutron mutagenic treatment used to gener- ate gal-3 did not result in the generation of relatively large de- letions at many unlinked loci.

The GA7 Locus of Arabidopsis

The linkage map of Arabidopsis is -600 cM (S. Hanley and H.M. Goodman, unpublished data), and the genome size is -108 bp. Therefore, on average, the recombination frequency per nucleotide is -6 x 10-6 cM. We used the recombination frequency between different ga7 alleles reported by Koornneef et al. (1983a) (assuming only reciproca1 cross-overs, i.e., no gene conversions) to calculate the recombination frequency within the gal locus. The reported recombination frequency between ga7-6 or gal-8 alleles and gal-7 is 0.5 x 10-* cM (Figure 1). Our DNA sequence analysis showed that the mu- tations in ga7-6 and gal-7 and in gal-8 and gal-7 are sepa- rated by 432 and 427 bp, respectively (data not shown). This calculation suggests that the recombination frequency in the GA7 locus is w ~ O - ~ cM per nucleotide, in good agreement with the average recombination frequency of the Arabidopsis genome. The calculation also suggests that the extent of the entire GA7 locus defined by mutants gal-4 and gal-7 is ~7 kb, more than sufficient to accommodate the detected 2.8-kb GA7 mRNA.

The biosynthesis of GAs in the Arabidopsis ga l mutants is blocked at the conversion of geranylgeranyl pyrophosphate (GGPP) to enf-kaurene (Barendse et al., 1986; Zeevaart, 1986). GGPP is a branch point metabolite that also serves as a precur- sor of other diterpenes and tetraterpenes, such as the phytol chain of chlorophyll and carotenoids. Because enf-kaurene is the first committed intermediate in GA biosynthesis, it is likely that the GA7 gene, required for the formation of ent-kaurene, is a point of regulation for GA biosynthesis (Graebe, 1987). In- deed, the biosynthesis of ent-kaurene has been shown to oc- cur preferentially in rapidly developing tissues, such as immature seeds, shoot tips, petioles, and stipules near the young elongating internodes in some species (Chung and Coolbaugh, 1986; Moore, 1989; Moore and Coolbaugh, 1991).

lsolation of the Arabidopsis GA7 gene will allow us to inves- tigate the regulation of GA biosynthesis in response to environ- mental and developmental signals. Future studies on the localization and pattern of expression of the GAl gene should provide information on the site and the regulation of GA bio- synthesis at an early stage in the pathway.

Extension of Genomic Subtraction to Plants with Complex Genomes

To extend genomic subtraction to organisms with complex ge- nomes such as maize, it is desirable to improve the efficiency of the hybridization steps by increasing the rate of reassocia- tion. It would also be desirable to eliminate the need to reas- sociate small amounts of the final product. Attempts are currently in progress to modify the protocol by carrying out the hybridization reactions in the presence of dextran sulfate to increase the reassociation kinetics (D. Straus, personal com- munication) and by ligating adapters onto the Sau3A-digested wild-type DNA before subtraction rather than afterward to per- mit the amplification of single-stranded sequences (D. Straus, personal communication; M. Mindrinos, 6. Shirley, and F.M. Ausubel, unpublished data). The genomic subtraction protocol used successfully to clone the Arabidopsis GA7 locus does not contain these modifications and is not likely to be very ef- ficient for isolating DNA sequences corresponding to small deletions from organisms with complex genomes. Other groups have also published subtractive hybridization procedures that employ multiple rounds of subtraction (Welcher et al., 1986; Bjourson and Cooper, 1988; Wieland et al., 1990). However, the reported levels of enrichment obtained with these proce- dures were less than we obtained with genomic subtraction. It is likely that these procedures will also require modification before they can be used to isolate DNA sequences correspond- ing to small deletions from organisms with complex genomes.

METHODS

Plant Materials

Arabidopsis ga7 mutants were obtained from M. Koornneef. In agree- ment with Dr. Koornneef (Agricultura1 University, Wageningen, The Netherlands), we have renamed gal alleles based on recommenda- tions promulgated at the Third lnternational Arabidopsis Meeting (East Lansing, MI, 1987). ga7-7 to ga7-70 correspond to the alleles NG5, 6.59, 31.89,29.9, 29.423, d352, 6027, A428, NG4, and d69, respectively, de- scribed previously in Koornneef et al. (1983a). Growth conditions of gal mutants for DNA isolation and seed harvesting were the same as previously described (Koornneef et al., 1983a).

Preparatlon of Plant DNA for Genomic Subtraction Experiments

Ler DNA and gal-3 mutant DNA were isolated from 4- to 6-week-old plants and purified by CsCl gradient centrifugation as described else- where (Ausubel et al., 1990). The concentration of the purified DNA was determined by measuring the UV absorbance at 260 nm. To con- firm the DNA concentration, we compared both Ler and gal-3 DNAs with bacteriophage I DNA of known concentration on an agarose gel.

126 The Plant Cell

Sonication and Biotinylation of Presumptive Deletion Mutant DNA

Procedures for sonication and photobiotinylation were as described elsewhere (Straus and Ausubel, 1990) with slight modifications. Two hundred micrograms of DNA isolated from the presumptive deletion mutant was resuspended in 2 mL of 10 mM N-(2-hydroxyethy1)- piperazine-N’-3-propanesulfonic acid (EPPS), pH 8.0,l mM EDTA, pH 8.0 (1 x EE buffer) and sonicated in a 15-mL conical plastic tube until the average size of the DNA was ~3 kb as determined by agarose gel electrophoresis. We used a sonicator (model No. W-375; Ultrasonics, Farmingdale, NY) set at continuous output mode setting No. 5 for 5 to 10 sec. We do not recommend using DNA fragments of average size less than 1 kb. It is important that the presumptive deletion DNA be uniformly biotinylated. The procedure described below only results in the addition of approximately one biotin moiety per 100 to 400 bp (Forster et al., 1985). The sonicated DNA was aliquoted to four 2-mL microcentrifuge tubes, boiled for 2 min, and precipitated with ethanol. It is important to aliquot the DNA solution in several flat-bottomed 2-mL microcentrifuge tubes to ensure even and maximum illumination during the next step. Each DNA sample was resuspended in 50 pL dHpO, and 50 pL of 2 pglpL photoactivatable biotin (model No. K1030; Clontech, Palo Alto, CA) was added to each tube. The tubes with tops open were placed in a styrofoam float in an ice-water bath and placed 10 cm from a sunlamp for 15 min. The solution was dark brown follow- ing this treatment. The aliquots of biotinylated DNA were pooled into one tube, and 20 pL 1 M Tris-HCI (pH 9.0) was added. DNA was ex- tracted several times with equal volumes of water-saturated I-butanol until the butanol phase was colorless. Biotinylated DNA was precipi- tated with ethanol and resuspended in 72 pL 1 x EE (2.5 pglpL, as- suming 90% recovery). The DNA pellet should be dark brown.

Sau3A Restriction Digestion of Wild-Type DNA for Genomic Subtraction

Five micrograms of Ler DNA and Ad-2 DNA (Bethesda Research Laboratories) were digested with 20 units of Sau3A (New England Bio- labs, Beverly, MA), respectively, for 3 to 6 hr at 37°C to ensure com- plete restriction digestion.

Genomic Subtraction

Genomic subtraction experiments were performed similarly to those described previously (Straus and Ausubel, 1990). In the first cycle of subtraction, 0.5 pg of Sau3A-digested wild-type Arabidopsis DNA was mixed with 12.5 pg of biotinylated deletion mutant DNA in a 0.5-mL microcentrifuge tube. In the case of the reconstruction experiment, 0.5 pg of Sau3A-digested Ler DNA plus 0.5 ng of Ad-2 DNA (Bethesda Research Laboratories) were hybridized with 12.5 pg of the same prep- aration of Ler DNA that had been sheared and photobiotinylated. An end-labeled synthetic oligonucleotide (84-mer) with 50,000 cpm (Cerenkov counts) was added as a tracer to determine the recovery after each cycle of subtraction. The DNA mixture was boiled in 10 to 15 pL of 2 x EE for 1 min, lyophilized, resuspended in 3 pL dHpO, and mixed with 1 pL 4 M NaCI, 0.16 M (Na)EPPS (pH 8.0), 20 mM EDTA (pH 8.0). Hybridization was carried out for at least 20 hr in a 65OC oven. Two microliters of yeast tRNA (20 pgIpL) and 100 pL EEN buffer (0.5 M NaCI, 1 x EE) was added by pipeting up and down to thoroughly mix. One milliliter of Fluoricon avidin polystyrene assay particles (IDEXX

Corp., Portland, ME) was washed twice in 1 x EE to remove azide and free avidin, resuspended in 0.9 mL EEN, and stored at 4%. One hundred microliters of washed avidin-coated particles was added to the DNA solution, and the sample was incubated at room tempera- ture for 30 min. The mixture was then transferred to an Ultrafree-MC filter unit (model No. UFC3 OGV 00; Millipore Corp., Bedford, MA) and spun 10 sec in a microcentrifuge. The supernatant was transferred to a fresh 1.5-mL microcentrifuge tube. The particles on the filter were washed with 0.2 mL EEN, and the supernatants were pooled, precipi- tated, and counted using Cerenkov counts.

Recovery should be greater than 70% except for this first cycle. In the first cycle, recovery may be lower due to the presence of unincor- porated label used to phosphorylate the tracer DNA. The DNA pellet from the first cycle was resuspended in 6 pL 1 x EE and transferred to a 0.5-mL microcentrifuge tube. Five percent (0.3 pL) of this DNA was saved and diluted into 5 pL 1 x EE for further analysis. For the second and subsequent cycles of subtraction, 10 pg of biotinylated deletion mutant DNA was added to the remaining DNA sample from the previous cycle. Subtractive hybridization was carried out as de- scribed for the first cycle except that no additional tRNA was added. In the case of the reconstruction experiment, 5% of the starting materials was saved and diluted in 5 pL 1 x EE. In the actual subtraction ex- periment for the GA7 locus, 10% of the unbound DNA was reserved and diluted into 5 pL 1 x EE at the end of the fifth and sixth cycles.

Addition of Sau3A Adaptors

Before analyzing unbound DNA fragments further, a fraction of this DNA was treated at 80°C, ethanol precipitated, and ligated to Sau3A adaptors, as described previously (Straus and Ausubel, 1990), with indicated modifications. After several cycles of subtraction, the unbound DNAfrom the last cycle was resuspended in 5 pL 1 x EE. In the case of the reconstruction experiment, 10 ng Sau3A-digested Ad-2 DNA (as control), 2 pL of the reserved samples from the starting material and the first three cycles of subtraction, and 0.5 pL (10%) of the unbound DNA from the fourth cycle were added to 50 pL of 1 M NaCI, 1 x EE, 400 pglmL yeast tRNA in 1.5-mL microcentrifuge tubes, respectively, and incubated at 80% for 30 min. In the actual subtraction experi- ment for the GAI locus, 2 pL of the reserved samples from the first four cycles, 5 pL of the reserved samples from the fifth and sixth cy- cles, and O5 pL of the DNA from the seventh cycles were treated at 80OC.

Polymerase Chain Reaction Amplification of Unbound DNA after Subtractive Hybridization

DNA with Sau3A adaptors was amplified as described (Straus and Ausubel, 1990) except that 0.2 pg of phosphorylated primer was used for each 50 pL polymerase chain reaction (PCR). Five microliters of each PCR was analyzed on 1.5% agarose gels. For cloning individual unbound fragments, PCR-amplified DNA fragments were incubated with Klenow to create blunt ends and ligated into the Smal site of pUC13.

lsolation of Corresponding Genomic and GAl cDNA Clones

XGA1-3 was isolated from a Landsberg erecta genomic library con- structed in XFlX (Voytas et al., 1990) using J2P-labeled pGA1-1 as probe. pGA1-2 was obtained by ligating a 6-kb Sall-EcoRI fragment from M A 1 9 into Sal1 plus EcoRlcleaved pBluescriptll SK (Stratagene).

Cloning the Arabidopsis GA1 Locus 127

pGA1-4 was isolated from a genomic library of Arabidopsis ecotype Columbia DNA constructed in the binary vector pOCA18 (Olszewski et al., 1988). The 0.9-kb GA1 cDNA clone was isolated by screening a cDNA library, constructed from RNA isolated from siliques (seed pods) of Arabidopsis ecotype Columbia (J. Giraudat and H.M. Goodman, unpublished data), using 32P-labeled 6-kb Sall-EcoRI DNA from pGA1-2 as a hybridization probe.

DNA Gel Blot Analyses

One microgram of Hindlll-digested DNA from Arabidopsis was frac- tionated on 1% agarose gels, transferred to GeneScreen membrane (Du Pont-New England Nuclear), and hybridized with gel-purified, 3zP-labeled DNA fragments, as described (Church and Gilbert, 1984).

Agrobacterium tumefaciens-Mediated Transformation of Arabidopsis Root Explants

The transformation procedure was as described previously (Valvekens et al., 1988) with slight modifications as indicated. To prepare roots of gal-3 for transformation, gal-3seeds were germinated in liquid me- dia containing l x Murashige and Skoog (1962) salts and 2% sucrose (MS media) with 100 (IM GA,. Media were changed weekly. Two weeks after germination, 1 x MS without gibberellin (GA) was used. pGA1-4 (Figure 5) and three other clones (data not shown) isolated from the pOCA18 library were introduced into Agrobacterium LBA4404 by electroporation (Ausubel et al., 1990). The stability of the inserts in these cosmid clones in LBA4404 was tested by hybridizing DNA gel blots containing Hindlll-digested DNA from LBA4404 carrying differ- ent CIO-es with 32P-labeled plasmid DNA. Among four clones isolated from the pOCA18 library, two clones were very unstable in LBA4404, indicating that it is important to monitor the stability of plasmids used in Agrobacterium before transformation.

A fresh overnight culture of LBA4404 carrying pGA1-4 was used to infect root explants of 5-week-old gal-3 plants. Kanamycin-resistant (Kmr) transgenic plants were regenerated as described (Valvekens et al., 1988). To score for GA+/Kmr and GA- kanamycin-sensitive segre- gation, seeds of transgenic plants were placed on MS agar plates con- taining kanamycin (50 pg/mL). GA+/Kmr seeds germinated within 3 days and developed into green seedlings. Nongerminating seeds af- ter 8 days were transferred onto MS plates containing 100 pM GA3 and 50 pg/mL kanamycin. GA- kanamycin-sensitive seeds ger- minated in the presence of GA3 but became white seedlings.

RNA lsolation and RNA Gel Blot Analysis

DNA Sequence Analyses

DNA sequences of GAl genomic DNA and cDNAwere obtained using the dideoxy method (Ausubel et al., 1990) with Sequenase (U.S. Bio- chemical Corp.) and both single- and double-stranded DNA templates. The gal-2 DNA was restricted with Hindlll and was ligated under con- ditions that favor the intramolecular ligation (Innis et al., 1990). The inverse PCR (Ausubel et al., 1990; lnnis et al., 1990) was carried out to amplify the junctions of the insertion or inversion in gal-2 DNA using 20 to 30 ng of Hindlll-digested and religated gal-2 DNA. The inverse PCR amplified fragments were cloned into the pSK vector, and DNA sequences were obtained by using double-stranded DNA templates. The 1.2-kb Hindlll fragments were amplified from 10 ng of genomic DNA isolated from Ler and gal-7, gal-6, gal-i: gal-8, and gal-9 by PCR (Ausubel et al., 1990). For DNA sequence analyses, single-stranded DNA templates corresponding to each of the 1.2-kb Hindlll fragments were obtained by single-primer reamplification (Ausubel et al., 1990) and purified by selective ethanol precipitation with 2 M ammonium acetate (Innis et al., 1990).

ACKNOWLEDGMENTS

We thank Maarten Koornneef for Arabidopsis gal mutants, Donald Straus for help with the genomic subtraction technique, Susan Hanley for RFLP analysis, Gabor Lazar and Ursula Hanfstingl for help with Agrobacterium-mediated transformation, and Brian Seed for oligonu- cleotide synthesis. We also thank Dan Voytas, Neil Olszewski, Jerome Giraudat, and Elliot Meyerowitz for providing Arabidopsis genomic and cDNA libraries. We are grateful to Jean Greenberg for helpful com- ments on the manuscript and to Rache1 Hyde for help in preparing the manuscript. This work was supported by USDA-CRGO Grant No. 91-37300-6384 and by a grant from Hoechst AG to Massachusetts General Hospital.

Received November 26, 1991; accepted December 23, 1991.

REFERENCES

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A., and Struhl, K. (1990). Current Protocols in Mo- lecular Biology. (New York: Greene Publishing Associates/ Wiley-lnterscience).

Barendse, G.W.M., Kepczynski, J., Karssen, C.M., and Koornneef, M. (1986). The role of endogenous gibberellins during fruit and seed development: Studies on gibberellin-deficient genotypes of Arabidop- sis tbaliana. Physiol. Plant. 67, 315-319.

Poly(A)+ RNA of 4-week-old and 5-week-old plants was prepared from the entire plant except the roots, and silique RNA was prepared from

scribed (Ausubel et al., 1990). Approximately 2 pg RNAfrom each sam- ple was treated with glyoxal, size-fractionated on a 1% agarose gel (Maniatis et al., 1982), transferred to GeneScreen membranes, and hybridized with a 3Wabeled 0.9-kb EcoRl DNA fragment from the GAl cDNA and a 3ZP-labeled 1.65-kb EcoRl fragment containing the Arabidopsis CAB gene (AB165) (Leutwiler et al., 1986).

Bjourson, A.J., and Cooper, J.E. (1988). lsolation of Rbizobium loti

immature siliques plus flower buds and stems as previously de- Strain-sPecific DNA sequences bY subtractive hybridization. APPl. Environ. Microbiol. 54, 2852-2855.

Chung, C.H., and CoolbaWh, R.C. (1986). efW"?ne biosynthesis in cell-free extracts of excised parts of tal1 and dwarf pea seedlings. Plant Physiol. 80, 544-548.

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

128 The Plant Cell

Clarke, L., and Carbon, J. (1979). Selection of specific clones from colony banks by suppression or complementation tests. Methods Enzymol. 68, 396-408.

Dellaert, L.W.M. (1980). X-ray- and Fast Neutron-lnduced Mutations in Arabidopsis thaliana, and the Effect of Dithiothreitol upon the Mu- tant Spectrum. Ph.D. Thesis (Wageningen, The Netherlands: Agricul- tural University).

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

Feldman, K.A. (1991). T-DNA insertion mutagenesis in Arabidopsis: Mutational spectrum. Plant J. 1, 71-82.

Feldman, K.A., Marks, M.D., Christianson, M.L., and Quatrano, R.S. (1989). A dwarf mutant of Arabidopsis generated by T-DNA in- sertion mutagenesis. Science 243, 1351-1354.

Forster, A.C., Mclnnes, J.L., Skingle, D.C., andSymons, R.H. (1985). Non-radioactive hybridization probes prepared by the chemical label- ling of DNA and RNA with a nove1 reagent, photobiotin. Nucl. Acids Res. 13, 745-761.

Frohman, M.A., Dash, M.K., and Martin, G.R. (1988). Rapid produc- tion of full-length cDNAs from rare transcripts: Amplification using a single gene-specific oligonucleotide primer. Proc. Natl. Acad. Sci.

Graebe, J.E. (1987). Gibberellin biosynthesis and control. Annu. Rev. Plant Physiol. 38, 419-465.

Hanahan, D., and Meselson, M. (1983). Plasmid screening at high colony density. Methods Enzymol. 100, 333-342.

Innis, M.A., Gelfand, D.H., Sninsky, J.J., and White, T.J. (eds) (1990). PCR Protocols: A Guide to Methods and Applications. (San Diego: Academic Press).

Jones, J.D.G., Carland, F., Maliga, P., and Dooner, H.K. (1989). Vi- sual detection of transposition of the maize element activator (Ac) in tobacco seedlings. Science 244, 204-207.

Koornneet, M., and van der Veen, J.H. (1980). lnduction and analy- sis of gibberellin-sensitive mutants in Arabidopsis thaliana (L.) Heynh. Theor. Appl. Genet. 58, 257-263.

Koornneet, M., Dellaert, L.W.M., and van der Veen, J.H. (1982). EMS- and radiation-induced mutation frequencies at individual loci in Arabidopsis thaliana (L.) Heynh. Mutation Res. 93, 109-123.

Koornneet, M., van Eden, J., Hanhart, C.J., and de Jongh, A.M.M. (1983a). Genetic fine-structure of the GA-1 locus in the higher plant Arabidopsis thaliana (L.) Heynh. Genet. Res., Camb. 41, 57-68.

Koornneef, M., van Eden, J., Hanhart, C.J., Stam, P., Braaksma, F.J., and Feenstra, W.J. (1983b). Linkage map of Arabidopsis thaliana. J. Heredity 74, 265-272.

Leutwiler, L.S., Meyerowitz, E.M., and Tobin, E.M. (1986). Struc- ture and expression of three light-harvesting chlorophyll a/b-binding protein genes in Arabidopsis thaliana. Nucl. Acids Res. 14,

Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982). Molecular Clon- ing: A Laboratory Manual. (Cold Spring Harbor, NY Cold Spring Harbor Laboratory), pp. 197-201.

Martin, C., Prescott, A., Lister, C., and Mackay, S. (1989). Activity of the transposon Tam 3 in Antirrhinum and tobacco: Possible role of DNA methylation. EMBO J. 8, 997-1004.

USA 85, 8998-9002.

4051-4064.

Masson, P., and Fedoroff, N.V. (1989). Mobilityof the maize suppressor- mutator element in transgenic tobacco cells. Proc. Natl. Acad. Sci. USA 86, 2219-2223.

Meyerowitz, E.M., and Pruitt, R.E. (1985). Arabidopsis thaliana and plant molecular genetics. Science 229, 1214-1218.

Moore, T.C. (1989). Biochemistry and Physiology of Plant Hormones. (New York: Springer-Verlag), pp. 113-135.

Moore, T.C., and Coolbaugh, R.C. (1991). Correlation between ap- parent rates of ent-kaurene biosynthesis and parameters of growth and development in Pisum sativum. In Gibberellins, N. Takahashi, B.O. Phinney, and J. MacMillan, eds (New York: Springer-Verlag), pp. 188-198.

Murashige, T., and Skoog, F. (1962). A revised medium for rapid growth and bioassays with tobacco tissue culture. Physiol. Plant. 15,473-497.

Nam, H.-G., Giraudat, J., den Boer, B., Moonan, F., Loos, W.D.B., Hauge, B.M., and Goodman, H.M. (1989). Restriction fragment length polymorphism linkage map of Arabidopsis thaliana. Plant Cell

Olszewski, N.E., Martin, F.B., and Ausubel, F.M. (1988). Special- ized binary vector for plant transformation: Expression of the Arabidopsis thaliana AHAS gene in Nicotiana tabacum. Nucl. Acids Res. 16, 10765-10782.

Oppenheimer, D.G., Herman, P.L., Sivakumaran, S., Esch, J., and Marks, M.D. (1991). Amyb gene required for leaf trichome differen- tiation in Arabidopsis is expressed in stipules. Cell 67, 483-493.

Straus, D., and Ausubel, F.M. (1990). Genomic subtraction for clon- ing DNA corresponding to deletion mutations. Proc. Natl. Acad. Sci.

Talon, M., Koornneet, M., and Zeevaart, J.A.D. (1990). Endogenous gibberellins in Arabidopsis thaliana and possible steps blocked in the biosynthetic pathways of the semidwarf ga4 and ga5 mutants. Proc. Natl. Acad. Sci. USA 87, 7983-7987.

Valvekens, D., van Montagu, M., and van Lijsebettens, M. (1988). Agmbacterium tumefaciens-mediated transformation of Arabidop- sis thaliana root explants by using kanamycin selection. Proc. Natl. Acad. Sci. USA 85, 5536-5540.

Voytas, D.F., Konieczny, A., Cummings, M.P., and Ausubel, F.M. (1990). The structure, distribution and evolution of the Tal retro- transposable element family of Arabidopsis thaliana. Genetics 126,

Welcher, A.A., Torres, AR., and Ward, D.C. (1986). Selective enrich- ment of specific DNA, cDNA and RNA sequences using biotinylated probes, avidin and copper-chelate agarose. Nucl. Acids Res. 14,

Wieland, I., Bolger, G., Asouline, G., and Wigler, M. (1990). A method for difference cloning: Gene amplification following subtractive hy- bridization. Proc. Natl. Acad. Sci. USA 87, 2720-2724.

Wilkinson, J.Q., and Crawford, N.M. (1991). ldentification of the Arabidopsis CHL3 gene as the nitrate reductase structural gene NlA2. Plant Cell 3, 461-471.

Zeevaart, J.A.D. (1986). Plant Research’86, Annual Report of the MSU- DOE Plant Research Laboratory (East Lansing, MI), pp. 130-131.

1, 699-705.

USA 87, 1889-1893.

71 3-721.

10027-1 0044.

DOI 10.1105/tpc.4.2.119 1992;4;119-128Plant Cell

Tp. Sun, H. M. Goodman and F. M. AusubelCloning the Arabidopsis GA1 Locus by Genomic Subtraction.

 This information is current as of September 18, 2020

 

Permissions X

https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists