Eva Yi Chou
Mar 10, 2015
Tutorial #7 Summary (Yanofsky et al., 1990)
A) Introduction
Homeotic genes are genes involved in the regulation of the anatomical structures of an
organism throughout its developmental stages; these genes are also involved in the development
of the Arabidopsis flower. The flowers of wild-type (WT) Arabidopsis plants are organized into
four whorls. The outermost whorl is composed of four sepals, the second whorl is made up of
four petals, the third whorl has six stamens and the innermost fourth whorl is composed of two
fused carpels. In the agamous (AG) mutants, the first two whorls are identical to the WT (sepals
then petals) while the six stamens in the third whorl are replaced by six petals and the two fused
carpels becomes a new flower with a repeating sepal, petal, petal pattern. The flower also goes
from having a determinant growth status (ending with the two fused carpel) to becoming
indeterminate. The objective of this paper is to clone the AG gene. By cloning the AG gene,
information such as the gene’s nucleotide sequence, protein sequence and function can be
obtained. This gene can then also be used as a tool for future research (such as using the amino
acid sequence of AG to find other genes with similar binding motifs). The cloning of the AG
gene was also important in helping to clone the other members of the ABC floral development
genes.
B) Result
The well characterized AG mutant (ag-1) was obtained by screening an EMS
mutagenized plant population. Since it was difficult at the time to clone a gene using point
mutation mutants, the authors obtained a previously found T-DNA insertion mutant plant with a
similar phenotype as the AG mutant. To test if this T-DNA insertion mutant was allelic to the
EMS ag-1 mutant, they crossed heterozygous ag-1 plants with heterozygous T-DNA insertion
mutants (the mutations are recessive and heterozygotes were used since homozygous mutants are
sterile). A quarter of the resulting cross had the AG phenotype (three quarters had WT
phenotype) indicating the genes failed to complement and are allelic. The T-DNA insertion
mutant was named ag-2.
Their next step was to ascertain if the T-DNA insertion could be used as a probe to
isolate the AG gene. To do this, the authors demonstrated that the T-DNA co-segregated with the
AG phenotype. The T-DNA insertion in ag-2 confers kanamycin resistance in plants and if
inserted in or near the AG gene, then the kanamycin resistance should co-segregate with the
mutant phenotype. Results were not shown for this test but the kanamycin resistance was found
to co-segregate with the AG mutant phenotype. Together with the previous complementation test
showing ag-1 and ag-2 to be allelic, the authors concluded that the T-DNA insertion could be
used as a molecular probe to isolate sequences of the AG gene.
The authors then used plasmid rescue to obtain sequences flanking the T-DNA insertion
for use to probe a WT Arabidopsis genomic library. One of the plasmids isolated from the
plasmid rescue (named pCIT505) was obtained. To test if pCIT505 contains sequences from the
AG region, RFLP mapping was conducted. Heterozygous ag-1 (in the LE background) was
crossed with WT plants (in the Nd-0 background). The F2 homozygous mutants were then used
for the mapping analysis. No crossovers were observed for all 59 F2 plants tested, indicating
pCIT505 either contains regions of the AG gene or is strongly linked (very close) to the AG gene.
They then used this to probe an Arabidopsis WT cosmid library to screen for the region that
contained the full length AG gene. When one of the isolated clones (pCIT540) was introduced
into the genome of homozygous ag-2 mutant plants (transgene complementation) the mutant
Eva Yi Chou
Mar 10, 2015
phenotype was rescued (flowers displayed WT phenotype). Since pCIT540 was able to rescue
the mutant phenotype, pCIT540 should contain the AG gene.
Next the authors wanted to characterize the AG gene. They first made a cDNA library
from WT Arabidopsis flowers (before stage 12) and used restriction fragments of sequences
flanking the T-DNA insertion site as probes. They were able to isolate one cDNA clone which
spanned the length of the T-DNA insertion site (there were many clones of varying lengths found
but they all overlapped). They sequenced this clone as well as the WT AG region. By comparing
the cDNA sequences against the genomic DNA sequences, the authors were able to find that the
AG gene contained 9 exons and 8 introns. By sequencing the T-DNA flanking sequences the
authors were also able to find that the T-DNA was inserted in the second intron in ag-2 mutants.
The authors also sequenced the AG region in ag-1 mutants and compared this sequence with the
WT sequence. They were able to confirm the ag-1 mutants had a point mutation (G to A) in the
fourth intron and further confirm the gene they characterized was the AG gene (EMS induces
point mutations).
Their next step was to use the cDNA to predict the amino acid sequence of the AG gene.
Once done, the authors realized they did not have a translation initiation codon. Due to this, they
predicted that the cDNA they obtained was not full length and could be due to secondary
structures in the RNA preventing the reverse transcriptase from extending further. From there,
they took the amino acid sequence and ran it through a protein database to look for homologous
sequences and predict protein function. The AG protein shared sequences with three known
transcription factors (MCM1 and AGR80 from yeast and SRF from humans) as well as a then
newly cloned DEF A gene from Antirrhinum. Due to the sequence homology between these
proteins, AG is predicted to function as a transcription factor.
To find the expression pattern of the AG gene, a northern blot was probed with
radiolabelled AG cDNA. Floral buds (stage 9) showed hybridization to the probe while floral
stems and vegetative tissue did not, indicating the AG gene is expressed in floral buds not in the
floral stem or vegetative tissues. To further narrow down expression of the AG gene within the
different organs of the flower, the authors performed in situ hybridization with both radiolabelled
sense (control) and antisense AG RNA. The stamens and carpels were shown to express the AG
gene while the sepals and the carpels did not. Combined with the earlier conclusion of the AG
gene possibly acting as a transcription factor, the authors hypothesize that the AG gene is a
transcription factor which regulates the genes involved in carpel and stamen development.
C) Discussion
The authors successfully cloned the AG gene, characterized the gene and were able to
predict its function based on amino acid sequence homology to other known transcription factors
as well as through RNA expression profiling.
At the time this paper was written, the authors believed they were unable to obtain the
full length cDNA since they lacked the ATG translation initiation codon. Later work by
Reichmann et al. (1999), showed through in vitro protein synthesis that the authors actually did
obtain the entire AG cDNA and the initiation codon was an ACG instead. Now that the AG gene
has been cloned and the amino acid sequence deduced, this could be used as a tool to identify
other genes with similar amino acid homologous regions. The AG gene could also be used as a
probe to clone any other genes with similar motifs (later on known as the MADS box family of
genes). The authors were able to find that the AG gene was expressed in stamens and carpels of
stage 9 floral buds. Though this tells us where expression occurs it does not tell us when
expression begins and ends. To find out, in situ hybridizations can be done for all stages of floral
Eva Yi Chou
Mar 10, 2015
development. The hypothesis that AG is a transcription factor could also be further tested as well
as the role AG plays as a transcription factor in the regulation of floral development (does it bind
DNA, complex with other proteins, inhibit other proteins/genes and/or transcription factors?).
Some future experiments that may help to answer these questions are biochemical assays such as
colorimetric ELISA protein binding assays and the creation of ectopic AG overexpression
mutants. In situ hybridization of AG in different floral mutant backgrounds could also reveal if
AG is being inhibited by other genes (conversely, in situ of other floral genes in AG knockout
mutants could reveal if AG inhibits any floral genes). How AG interacts with other floral
development genes is also largely unknown at the time. Double or triple mutants of the known
floral development genes could help to build a model of how these genes work together (later
known as the ABC model for floral development).
Reference
Riechmann, J. L., Ito, T., & Meyerowitz, E. M. (1999). Non-AUG Initiation of AGAMOUS mRNA
Translation in Arabidopsis thaliana. Molecular and Cellular Biology, 19(12), 8505–8512.