Unformatted text preview: , 75% of
DNA in egg cells versus 50% of DNA in sperm nuclei should be
fully demethylated. The remaining DNA (25% in female
gametes, 50% in male) should be hemimethylated (Fig. 3d). If, as
suggested21, A. thaliana gametes undergo an additional round of
DNA synthesis and fertilize in G2, demethylation ratios should
be higher. We determined the actual frequency of demethylation
and reactivation of the GUS locus in reciprocal backcrosses of F1
hybrids (genotype MET1+/– GUS+/–). The frequency of reactivation was signiﬁcantly higher when met1 and GUS were maternally inherited (74.5% among backcrossed plants; 102 active and
35 silent, hypothesis 3:1; χ2 = 0.004; P > 0.99) than after paternal
transmission (42% among backcrossed plants; 53 active and 72
silent, hypothesis 1:1; χ2 = 2.59; P > 0.10).
nature genetics • volume 34 • may 2003 © 2003 Nature Publishing Group http://www.nature.com/naturegenetics letter
We conﬁrmed this ‘epigenetic segregation’ using Southern blots
for the methylation status of the GUS locus in backcrossed plants
(Fig. 3b). The accurate match between observed segregation frequencies and expected parental effects support the hypothesis that
MET1 is required during gametogenesis. We may also conclude
that the residual hemimethylation provides necessary and sufﬁcient
information for remethylation of the second DNA strand as soon as
MET1 is again provided in the heterozygous zygotes. This remethylation of hemimethylated templates probably occurs before the ﬁrst
S phase, that is, before zygotic DNA replication, as the backcross
progeny included plants without detectable GUS activity in any tissue (indicating complete silencing and methylation of the gene; Fig.
3d). The observed reactivation ratios are consistent with fertilization occurring at G1 stage; however, carryover of MET1 activity
into the ﬁrst postmeiotic mitosis resulting in delayed demethylation cannot be excluded.
Notably, we observed gametophytic effects only after total
depletion of MET1 activity. Previous genetic analysis in F2 populations, using a partial loss-of-function allele of MET1 (met1-1,
met1-2) that retains up to 30–50% of CpG methylation in early
generations of homozygous mutant plants9, suggested a clearly
recessive character of the mutation9,19. In contrast, comparable F2 examination of met1-3 or met1-4 crosses showed clear gametophytic effects (see Supplementary Table 1 online).
Divergent methylation in met1 heterozygotes was not restricted
to the transgenic GUS genes but was also shown for endogenous
loci. Southern-blot analysis of the Rap2.1 promoter on chromosome I (ref. 22) and of the FWA promoter on chromosome IV (ref.
18), both known to be methylated in wild-type A. thaliana, showed
that T2 plants heterozygous with respect to the met1 mutation
inherited either one or no hypomethylated allele of each gene in a
process resembling random segregation (Fig. 3c). These arbitrary
combinations of methylated and unmethylated loci match the phenotypic variation observed among heterozygous met1 plants (Fig.
2) and reﬂect independent epigenetic segregation.
It is well documented that maintenance of DNA methylatio...
View Full Document
- Fall '09
- DNA, DNA methylation, MET1