Methods Construction of AGL5 knockout vector.The AGL5 KO construct was created using pZM104A14, which carries the kanamycin-resistance cassette flanked by several cloning sites. In addition, pZM104A has the screenable marker gene encoding beta-glucuronidase (GUS), which should be present in non-homologous insertions but absent from homologous integrations. The GUS marker can be useful to identify non-homologous events, but was not used in the initial screen for targeted insertions since many primary transformants contain multiple independent insertions. The 3 kb region representing the 5' portion of AGL5 was obtained by PCR amplification using the oligos A (5'- CGG ATA GCT CGA ATA TCG - 3') and B (5'- AAC CAT TGC GTC GTT TGC - 3') and the resulting fragment cloned into the pCRII vector (Invitrogen). This fragment was excised with EcoRI and inserted into the EcoRI site of pZM104A. The 3' portion of AGL5 was excised as an XbaI fragment from an AGL5 genomic clone in the vector pCIT3016 and inserted into the XbaI site of pZM104A. The resulting plasmid, designated AGL5 KO construct (Fig. 1), was used in Agrobacterium-mediated infiltration of wild-type Arabidopsis plants of the Columbia ecotype. The knockout construct was derived from Landsberg erecta genomic DNA. Identification of knockout plant. Approximately 750 primary (T1) kanamycin-resistant transformants were selected, and DNA was extracted from individual leaves17 in pools representing ten plants. To identify a pool that contained a candidate targeted disruption, isolated DNAs were subjected to PCR amplification using oligo #1 (5'- GTAATTACCAGGCAAGGACTCTCC - 3'), which represents AGL5 genomic sequence that is not contained within the AGL5 KO construct, and oligo #2 (5'- GTCATCGGCGGGGGTCATAACGTG - 3') which is specific for the kanamycin-resistance cassette. Amplified products were size fractionated on agarose gels, and used for standard DNA blotting assays with the 6.0 kb probe (Fig.3).. One pool of ten plants revealed the anticipated hybridizing band of the correct size, and this pool was subsequently broken down into individual plants. A single (T1) plant was identified that appeared to contain the desired event, and this plant was allowed to self-pollinate for analyses in subsequent generations. This T1 plant was shown to contain the GUS-reporter gene, indicating that in addition to the putative homologous integration event, there were independent non-homologous events. Segregation in the subsequent generations allowed the identification of plants that no longer contained the GUS-reporter gene, and it was these lines that were used for subsequent analyses. Plants homozygous for the disruption were identified by PCR amplification using primers #4 (5'- GAGGATAGAGAACACTACGAATCG- 3') and #5 (5'- CAGGTCAAGTCAATAGATTC - 3'), which yielded a single 1.5 kb product in wild type plants, and a single 2.6 kb product in the mutant. Further confirmation that these plants contained the desired disruption was obtained by PCR amplification with oligos #3 (5'- CAGAATTTAGTGAATAATATTG - 3') and #2, which gave the expected amplified product in the mutant but no product in wild-type plants.
Confirmation of targeted disruption. Confirmation of targeted disruption. To confirm that the desired disruption had occurred, a series of genomic DNA blots representing wild-type and homozygous mutant (T4 generation) plants were analyzed (Fig. 3). Probe 1 hybridized to the expected 3.9 kb XbaI fragment in wild-type and mutant plants, whereas the 1.3 kb XbaI fragment was present only in wild-type. This same probe hybridized to a 6 kb EcoRI fragment in wild-type and to the expected 4.1 and 2.8 kb EcoRI fragments in the mutant. Additional digests with BglII and with HindIII confirmed that the mutant plants contained the desired targeted event (data not shown). To confirm that there were no detectable deletions and/or rearrangements outside the targeted region, genomic DNA blots of wild type and homozygous mutant plants were further analyzed. Probe 2 hybridized in wild-type and mutant DNAs to the expected 2.9 kb XmnI fragment, the 1.5 kb and 0.4 kb HincII fragments, and the 0.6 kb HindIII fragment (Fig 4). Probe 3 hybridized in wild-type and mutant DNAs to the 9 kb ScaI fragment, the 3.9 kb XbaI fragment, and the 1.8 kb NdeI fragments (Fig 5). The faintly-hybridizing bands in the ScaI digests represent fragments that span the insertion site, and are, as expected, different sizes in wild-type and agl5 mutant plants. The targeted allele described here is designated agl5-1.
RNA blotting analyses. RNA blotting analyses. Approximately 6 ug of polyA+ RNA was purified using Dynabeads (Dynal) from wild-type and agl5 mutant inflorescences, size fractionated and hybridized using standard procedures18 with a gel-purified 450 bp HindIII-EcoRI fragment from pCIT22429 specific for the 3' end of the AGL5 cDNA. The expected AGL5 transcript was detected in wild-type but not in the mutant plants (Fig. 6). This same filter was subsequently stripped and re-hybridized with a tubulin-specific probe to verify that approximately equal amounts of RNA were present in each lane.