全 文 ：Received 31 Mar. 2003 Accepted 15 Jul. 2003
Supported by the National Science Fund for Distinguished Young Scholars of China (30225003) and the Outstanding Youth Research Plans of
the Jilin Provincial Government.
* Author for correspondence. Tel: +86 (0)431 5269367; E-mail: .
http://www.chineseplantscience.com
Alien DNA Introgression into Rice Causes Heritable Alterations in DNA
Methylation Patterns in an Active Retrotransposon Tos17
DONG Yu-Zhu1, LIU Zhen-Lan1, DONG Ying-Shan2, HAN Fang-Pu3, HE Meng-Yuan1, HAO Shui1, LIU Bao1*
(1. Institute of Genetics and Cytology, Northeast Normal University, Changchun 130024, China;
2. National Center of Transgenic Plant Research and Commercialization, Jilin Academy of Agricultural Sciences,
Gongzhuling 136100, China;
3. Eastern Cereal and Oilseed Research Center, Agri-Food Canada, Ontario K1A0C6, Canada)
Abstract: Heritable alteration in DNA methylation patterns was detected in all five rice lines with
introgressed DNA segments from wild rice (Zizania latifolia (Griseb.)) by DNA gel blotting analysis
with an endogenous retrotransposon Tos17 as a probe. The changing patterns include simultaneous loss
of parental fragments and appearance of novel fragments in each of the four methylation-sensitive enzyme
digests. Methylation modifications include cytosines at both symmetrical and asymmetrical sites, as well
as adenine bases. Sequence analysis at two critical regions of Tos17, i.e. the 5-LTR region (region Ⅰ) and
the reverse transcriptase region (region Ⅱ) showed complete conservation for all five introgression lines
compared with the parent. Sequence-specific PCR assay, however, confirmed that methylation changes
occurred in both regions. Moreover, concordance in the collective methylation changes between 5-LTR
and RT regions was observed in two of the introgression lines. The methylation changes are stably
inherited to the next generation. Because earlier studies showed that there had been activation and
mobilization of Tos17 in these introgression lines following alien DNA integration, it appears likely that
DNA methylation may have played some roles in controlling activity of Tos17 in rice, although the exact
relationship between the two phenomena remains to be established.
Key words: alien DNA introgression; DNA methylation; retrotransposon activity; epigenetics
Retrotransposons, also called class Ⅰ mobile elements,
transpose via reverse transcription of RNA intermediates and
are ubiquitous genomic components in all eukaryotes (Kumer
and Bennetzen, 1999). Retrotransposons comprise two main
types, i.e. Non-LTR (long terminal repeat) retrotransposons
and LTR ones. The former includes long or short interspersed
nuclear elements (LINES or SINES), and the later includes
copia-like and gypsy-like elements. While LINES and SINES
are non-autonomous, the LTR retrotransposons usually con-
tain sequences encoding all proteins required for their
retrotransposition and hence are autonomous. It is found,
however, that in plants, even the LTR-retrotransposons are
largely quiescent during normal growth and development but
may be activated by biotical and abiotical stresses, such as
wounding, pathogen attacks, cell culture, environmental cues
(e.g. dryness) and interspecies hybridization (Hirochika, 1993;
Wessler, 1996; Grandibastien, 1998; O’Neill et al., 1998;
Kalendar et al., 2000). Because such stresses are prevalent in
na tur a l p l an t po p ula t io ns , and unco nt ro l l ed
retrotranspositions could potentially lacerate the host’s
genome, it is not surprising that plant genomes have evolved
mechanisms to repress retrotransposon activity. One such
likely mechanism is based on an epigenetic process, i.e. DNA
methylation, whereby retrotransposons are rendered transcrip-
tionally inactively (Liu and Wendel, 2003). A close correlation
between methylation status of a given transposon and its ac-
tivity was first found in a classic maize transposons Mu (class
Ⅱ mobile elements) (Chandler and Walbot, 1986), and has
since been demonstrated in Arabidopsis for both
endogenenous transposons and retrotransposons (Miura et
al., 2001), as well as for an introduced exogenous
retrotransposon (Hirochika et al., 2000).
Tos17 is an endogenous copia-like retrotransposon in
rice (Hirochika et al., 1996). The copy number of Tos17 is
extraordinarily low, ranging from 1 to 4, in various cultivars
growing under normal conditions, but can be significantly
elevated by tissue culture (Hirochika et al., 1996; Hirochika,
1997). Although no target preference at the nucleotide se-
quence level could be generalized, the activated elements
have a propensity to transpose into low-copy, genic regions
of the rice genome (Yamazaki et al., 2001), thus providing an
excellent system for gene-tagging in rice (Hirochika, 2001;
Acta Botanica Sinica
植 物 学 报 2004, 46 (1): 100-109
101DONG Yu-Zhu et al.: DNA Methylation in a Rice Retrotransposon Tos17
Kumar and Hirochika, 2001).
We have reported that Tos17 could also be mobilized by
sexual hybridization-mediated introgression into the rice
genome of genomic DNA from wild rice (Zizania latifolia);
copy number of the element was increased in five represen-
tative introgression lines studied based on genomic South-
ern analysis (Liu and Wendel, 2000). Nevertheless, activity
of Tos17 was ephemeral and completely inactive after a few
generations. We postulated that cytosine methylation was
potentially responsible for rapid element silencing, because
methylation status at the CCGG sites within and/or flank-
ing the reverse transcriptase (RT) regions of the element
was found to have changed in all lines based on fragment
positions in genomic Southern hybridization profiles gen-
erated with a pair of isoschizomers (HpaⅡ /MspⅠ).
However, because only one pair of isoschizomers was used,
it is not clear whether cytosine at positions other than the
symmetrical CCGG of the element was subjected to methy-
lation alterations, nor is known if DNA adenine methyla-
tion is involved. The recent findings in Arabidopsis sug-
gest that asymmetric cytosine methylation might play a
more important role in repressing act ivi ty of
retrotransposons than symmetric cytosine methylations
(Martienssen and Colot, 2001), it is thus of significance to
investigate whether mobilization of Tos17 in the rice intro-
gression lines is also associated with methylation alter-
ations other than the CCGG sites.
The present study was aimed to address the following
questions: (1) Have methylation changes in Tos17 of the
introgression lines occurred to a broader extent than the
CCGG sites? (2) Is there any concordance between methy-
lation status in the promoter (5-LTR) region and down-
stream coding regions (e.g. RT)? (3) Is there a correlation
between DNA methylation status and retrotransposon
activity? By digesting genomic DNAs from five geneti-
cally homogeneous introgression lines and their rice par-
ent with multiple methylation-sensitive enzymes, followed
by DNA gel blotting analysis and sequence-specific PCR
amplifications, we investigated the possible methylation
pattern changes in Tos17 as a result of alien DNA
introgression, and found that: (1) Compared with the rice
parent, all methylation-sensitive enzymes studied revealed
methylation alterations at various sites, including symmetri-
cal and non-symmetrical cytosines as well as adenines in
all the five introgression lines. (2) In some introgression
lines, there exists apparent concordance between methyla-
tion changes in 5-LTR and RT regions. (3) Higher copy
number of Tos17 appears to be accompanied by concomi-
tantly elevated level of DNA methylation within the element.
1 Materials and Methods
1.1 Plant material
Production and characterization of a series of rice (Oryza
sativa L.)- Zizania latifolia Griseb. introgression lines with
integrated genomic DNA from wild rice, Z. latifolia were
described previously (Liu et al., 1999b). The same genomic
DNAs that had been isolated from expanded leaves of indi-
vidual plants from five genetically homogeneous introgres-
sion lines and their rice parent, which had been used in a
previous study (Liu and Wendel, 2000) were used for the
present study. To test heritability of the methylation
changes, progeny plants of each of the five introgression-
lines were also used.
1.2 DNA gel blotting analysis
DNA was isolated by the modified CTAB method
(Kidwell and Osborn, 1992). To ensure complete digestion
by the various methylation-sensitive enzymes, the DNAs
were further eluted by phenol, phenol/chloroform (1:1) and
chloroform purifications, followed by ethanol precipitation.
DNAs (3 mg each sample) were digested by the following
six methylation-sensitive enzymes (New England Biolabs
Inc.), namely, AluⅠ, HaeⅢ, HphI, Sau3AⅠ, HpaⅡ and
MspⅠ, whose recognition sequences as well as the methy-
lation-sensitive base(s) are summarized in Table 1. For each
enzyme, a 3-fold excess of enzyme (1 mg DNA/3 U enzyme)
was used and the reaction was performed overnight at tem-
peratures specified by the supplier. Digested DNAs were
blotted onto the Hybond N+ nylon filter (Amersham
Pharmacia Biotech) according to procedures of the supplier.
A 666 bp fragment amplified by PCR with primers RTP1 and
RTP2 (detailed below) from the RT region of Tos17 were
used as a hybridization probe. The fragment was gel-puri-
fied and labeled with fluorescein-11-dUTP by the Gene Im-
ages random prime- labeling module (Amersham Pharmacia
Biotech). Hybridization signals were detected by the Gene
Images CDP-Star detection module (Amersham Pharmacia
Biotech). As a control for complete digestion by the me-
thylation-sensitive enzymes, a fragment of a unique-copy
genic sequence (T26-2) was used to re-probe the same blots
after stripping off the Tos17 hybridization signals. All ma-
nipulations in the hybridization were exactly following the
supplier’s protocol. Post-hybridization washing stringency
is as reported (Liu and Wendel, 2000).
1.3 Template DNA preparation, sequence-specific primer
design and PCR amplification
Because sensitivity of PCR reactions, additional pre-
cautions were taken for the preparation of genomic DNA
templates. These include: (1) For each enzyme, two doses,
Acta Botanica Sinica 植物学报 Vol.46 No.1 2004102
i.e. a 3-fold excess of enzyme (1 mg DNA/3 U enzyme) and a
10-fold excess (1mg DNA/10 U enzyme) were used for each
DNA sample. (2) For each plant material, both undigested
DNA and DNA digested with a methylation-insensitive
enzyme (FokⅠ), which respectively serves as positive and
negative controls, were included. Enzymes were heat-inac-
tivated as advised by the supplier, and the DNAs were
then extracted by phenol/chloroform (1:1) and chloroform,
followed by ethanol precipitation. The DNAs were dis-
solved in 20 mL H2O and 1 mL (50 ng) was used as PCR
templates.
The following two sets of primers, which respectively
flank the 5-LTR and RT regions of Tos17, were designed.
LTRP1: 5-CTGTATAGTTGGCCCATGTCC-3 (nucleotide
positions 28-48), LTRP2: 5-GGCGGTCAACGACAAATC-
3 (nucleotide positions 323-306), and RTP1: 5-
GCTACCCGTTCTTGGACTAT-3 (nucleotide positions 2
817-2 837), RTP2: 5-CTGAAATCGGAGCACTGACA-3
(nucleotide positions 3 483-3 463). Amplification cycle
numbers were optimized to ensure that amplification is within
the exponential range. Five mL amplification products were
run through 1.5% agrose gels, stained with ethiudium bro-
mide and photographed under UV illumination. The gel
images were also semi-quantified (for relative abundance)
using the Quantity One software for Geldoc (Bio-Rad,
Hercules, California, USA).
2 Results
2.1 DNA methylation changes in Tos17 of the introgres-
sion lines detected by DNA gel blotting analysis
Tos17 is a copia-like retrotransposon that was isolated
by Hirochika and colleagues (1996) from cDNA of rice callus.
Using a cDNA fragment of the RT region of Tos17 (accession
number D85876) as a query, we searched the Genbank da-
tabase by BlastN, and found a 139 kb rice BAC genomic
DNA fragment (nbxb0019M20, accession number
AC087545) located on chromosome 10; a region of this frag-
ment showed perfect match with the RT region of Tos17.
Further analysis identified the two identical LTRs, the tar-
get site direct repeats (five base pairs) flanking the two
LTRs (CTCCT), and all intact components for a typical
copia-like retrotransposon, thus indicates that this BAC
clone encompasses the complete sequence of Tos17
(Fig.1).
Using primers RTP1 and RTP2 (Materials and Methods),
an expected 666 bp fragment was amplified from the paren-
tal line Matsumae. The fragment was sequenced to con-
firm its identity and then used as a probe to hybridize against
blots containing genomic DNAs isolated from the rice par-
ent (cv. Matsumae) and five introgression lines (lines 1 to
5) that were digested with each of the four methylation-
sensitive enzymes (AluⅠ, HaeⅢ, HphⅠ, Sau3AⅠ) that
are sensitive to methylation changes at symmetric or asym-
metric cytosine or adenine bases (Table 1, Fig.2 upper panel).
It was found that compared with the parent (lane 1), all five
introgression lines (lanes 2 through 6) showed conspicu-
ous changes in the hybridization patterns in some or all of
the enzyme digests, as being reflected by simultaneous
loss of parental fragments (denoted by arrows) and/or gain
of novel fragments (denoted by circles). We interpret the
pattern changes in the introgression lines as predominantly,
if not exclusively, due to methylation alterations at the sen-
sitive restriction sites, because sequence analysis at 5-
LTR and RT (the probe region) regions of multiple clones
from each of the five lines showed perfect sequence con-
servation (detailed below), and hence suggesting absence
of structural changes within the element in any of the intro-
gression lines. To rule out the possibility that some of the
pattern changes may result from incomplete digestion of
genomic DNAs by the methylation-sensitive enzymes, we
Table 1 Collective DNA methylation changes at the 5-LTR (region Ⅰ) and RT (region Ⅱ) of Tos17 in five introgressed rice lines
Enzyme Sequence
No. of sites DNA methylation changes
LTR (regionⅠ) RT (region Ⅱ)
LTR RT R L1 L2 L3 L4 L5 R L1 L2 L3 L4 L5
AluⅠ 5-AGCT-3 1 4 1 0 2 0 4 3 0 0 2 0 3 4
HaeⅢ 5-GGCC-3 2 3 0 0 4 0 4 3 2 2 1 1 4 3
HphⅠ 5-GGTGA(N)8-3 1 2 0 0 4 0 4 4 4 4 1 1 1 1
Sau3AI 5-GATC-3 1 3 4 0 0 0 3 3 2 0 1 0 4 1
HpaⅡ 5-CCGG-3 3 1 1 0 3 0 4 2 4 0 4 0 4 3
MspⅠ 5-CCGG-3 3 1 3 0 2 0 4 1 4 0 3 0 2 1
Colective methylation level is defined as: 0, no amplification; 1, 1%-25% of the strongest amplification of the six lanes for a given enzyme
digest (Fig.2); 2, 26%-50% of the strongest amplification of the six lanes for an given enzyme digest; 3, 50%-75% of the strongest
amplification of the six lanes for an given enzyme digest; 4, 76%-100% of the strongest amplification of the six lanes for an given enzyme
digest. The methylation-sensitive base(s) is underlined and in bold. R, the rice parent (cv. Matsume); L1 to L5, the rice— Zizania latifolia
introgression lines 1 to 5.
103DONG Yu-Zhu et al.: DNA Methylation in a Rice Retrotransposon Tos17
hybridized the same set of blots to a fragment from a
unique-copy cellular gene (similar to Arabidopsis permease
1 gene, Genbank accession number AAK59508) that is
known to be unmethylated at all otherwise methylatable
sites within the gene in rice (unpublished data). As was
evident from the lower panel of Fig.2, in all four enzyme-
digests, a monomorphic fragment was detected in the par-
ent and in each of the five introgression lines, therefore
confirming complete digestion of the genomic DNAs with
all the enzymes. The foregoing results, together with data
obtained with a pair of methylation-sensitive isoschizomers,
HpaⅡ /MspⅠ, which showed similar results (Liu and
Wendel, 2000), led to the conclusion that introgression of
alien DNA into rice has caused conspicuous alteration in
DNA methylation patterns not only at symmetric cytosine
sites, but also at non-symmetric cytosines and even in ad-
enine bases.
The DNAs used for this analysis are all from the 9th-
generation introgression lines (Liu and Wendel, 2000). To
test stability/heritabilty of the changed methylation pattern,
we also used progeny of these lines (the 10th generation)
for the same analysis, and we found near identical hybrid-
ization profiles as in Fig.2 (data not shown), thus indicating
stable inheritance of the methylation changes.
2.2 DNA methylation changes within the 5-LTR and RT
regions of Tos17 in the introgression lines detected by
PCR-based assay
Although the above indicates that extensive methyla-
tion pattern alterations occurred in Tos17 in all five intro-
gression lines studied, it is possible that only part of the
alternations detected by DNA gel blotting analysis are at-
tributable to changes in methylation status of the element
per se; that is, some of the changes might result from either
DNA methylation changes or differential background me-
thylation status in genomic regions flanking Tos17. To
investigate methylation changes within Tos17, a method
enabling dissecting crucial domains of the element is
desirable. Because sequence analysis reveled dozens of
restriction sites for each of the six studied enzymes within
Tos17 (data not shown), DNA gel blotting analysis is ap-
parently not the method of choice, because it is incapable
of detecting small fragments (<100 bp) or to distinguish
similar-sized fragments if the element is delineated. We
therefore used a PCR-based methylation assay for the 5-
LTR and RT regions of the element, which has been demon-
strated suitable for this type of analysis (Fu et al., 2000).
Specifically, based on the sequence information, we de-
signed the following two pairs of primers: (1) LTRP1/LTRP2,
Fig.1. Schematic diagram of the general structure and two assayed regions of retrotransposon Tos17. The structure of Tos17 is based
on Hirochika et al. (1996) and on a BAC clone of rice chromosome 10 (nbxb0019M20, accession number AC087545). int, integrase;
LTR, long terminal repeat; PBS, primer binding sites; PPT, polypurine tracts; pr, protease; rh, RNaseH; rt, reverse transcriptase. Region
Ⅰ is 296 bp that encompasses the 5-LTR and part of the PBS region; region Ⅱ is 666 bp that is an internal part of the RT/RNaseH gene.
This fragment (marked by a hatched rectangle) is also used as a probe for Southern analysis. Arrows in region Ⅰ and region Ⅱ
respectively refer to the specific primers, LTRP1/LTRP2 and RTP1/RTP2. All the assayed methylation-sensitive sites and their
positions within the two regions are presented. Al, AluⅠ; Ha, HaeⅢ; Hp, HphⅠ; Sa, Sau3AⅠ; H/M, HpaⅡ/MspⅠ.
Acta Botanica Sinica 植物学报 Vol.46 No.1 2004104
which flanks a 296 bp fragment (designated regionⅠ,
Fig. 1) including the complete 5 -LTR (138 bp) and part of
the PBS region (158 bp); (2) RTP1/RTP2, which delineates a
666 bp fragment (designated region Ⅱ, Fig.1), is part of the
reverse transcriptase/RNaseH gene of Tos17. Using these
two pairs of primers, we performed PCR amplification on
genomic DNAs of the rice parent (cv. Matsumae) and the
five introgression lines. All materials produced amplifica-
tion products of the expected sizes, i.e. 296 and 666 bp
respectively for region Ⅰ and region Ⅱ (Fig.3a). To inves-
tigate if any changes at the sequence level have occurred
within the amplified regions in the introgression lines, we
cloned both PCR products (LTR and RT) into the pGEM-T
vector (Promega, Madison, WI, USA) and sequenced three
arbitrarily selected clones from the rice parent and from
each of the five introgression lines. In all cases, complete
sequence conservation was observed, thus indicating ab-
sence of sequence change at the LTR and RT regions of
Tos17 in the introgression lines.
The two regions (LTR and RT) of Tos17 contain one to
several restriction sites for each of the six methylation-sen-
sitive restriction enzymes, as are depicted in Fig.1. Be-
cause no differential amplification was observed between
the parent and any of the introgression lines when intact
genomic DNAs were used as templates (Fig.3a), and be-
cause sequence data showed no base change within these
two regions in any of the five introgression lines, it is evi-
dent that, on condition that differential digestion owing to
difference in DNA quality could be confidently ruled out,
variations in amplification on templates digested with me-
thylation-sensitive enzymes should reflect differences in
the collective methylation status of the critical sites. Here
we define the methylation status as collective, because on
the one hand, of all the six methylation-sensitive enzymes
used, multiple sites exist within the defined regions (LTR or
RT, Table 1), therefore, for a given Tos17 copy, only a single
unmethylated site is sufficient to render no amplification
after digestion; on the other hand, because more than one
Fig.2. Gel blotting analysis on DNA methylation pattern changes in Tos17 of five rice introgression lines. Disappearance of parental
(lane 1) hybridization fragments (marked by arrows in the upper panel for each enzyme digest) and appearance of novel fragments in the
introgression lines (marked by empty circles in lanes 2 through 6 of the upper panel for each of the enzyme digest) denote methylation
changes at specific sites of the enzymes assayed. The lower panel is re-probing the blots with a unique-copy genic sequence (T26-2)
known to be unmethylated at all sites. A monomorphic hybridization fragment detected in parent and in each of the five introgression
lines for a given enzyme-digest indicate complete restriction by the methylation-sensitive enzymes for all samples. Molecular masses are
given on the right side.
105DONG Yu-Zhu et al.: DNA Methylation in a Rice Retrotransposon Tos17
copy of Tos17 exists in all five introgression lines (Liu and
Wendel, 2000), at least one unmethylated site for every
copy is needed to produce the overall absence of
amplification.
To rule out variations in DNA quality that might affect
completeness of digestion among the various samples, the
following precautions were taken: (1) We used genomic
DNAs known to be of high purity (Liu and Wendel, 2000),
which were further purified by phenol extraction for the
present study. (2) A methylation-insensitive enzyme,
FokⅠ, whose recognition site (5-GGATG(N)9-3) exists in
both assayed regions, was used to digest the DNAs and
hence served as a negative control. As was shown in Fig.
3b, no amplification was detected in any of the DNAs di-
gested with FokⅠ, thus confirming complete digestion of
the genomic DNAs from all the lines by this methylation-
insensitive enzyme. (3) For all six methylation-sensitive
enzymes, more than 3-fold difference in enzyme quantity
(Materials and Methods) produced identical results for each
of the DNA samples (Fig.3, data not shown). Generally, we
are confident that any differential digestibility by methyla-
tion-sensitive enzymes between the rice parent and intro-
gression lines, as well as among the introgression lines are
the results of bona fide differences in methylation status
on the critical sites.
Evidently, methylation differences at all six methylation-
sensitive restriction sites could be detected by the PCR
method at both qualitative level (presence/absence of am-
plification products, Fig.3) and semi-quantitative level
(relative abundance of amplification products after amplifi-
cation within the exponential range, Table 1). Detailed analy-
sis of the data revealed the following observations: (1)
Methylation changes occurred in all five introgression lines
in at least some of the restriction sites of both LTR and RT
regions of Tos17 (Fig.3). Two lines (1 and 3) exhibited only
demethylation changes while three lines (2, 4 and 5) exhib-
ited both demethylation and hypermethylation changes,
depending on the restriction sites (Table 1). (2) Methyla-
tion occurred at both symmetrical and non-symmetrical
cytosine sites, as well as at adenine bases (Fig.3; Table 1).
(3) Although for the rice parent, there is apparent overall
difference in the methylation status between LTR and RT
regions for four of the six enzyme restriction sites, two lines
(1 and 3) exhibited a high level of concordance between the
collective methylation changes in LTR and those in RT for
almost all the restriction sites (Fig.3; Table 1). Nevertheless,
it is not known whether all sites for a given enzyme within
a given region (LTR or RT) underwent concordant changes,
as only demethylation was detected in these two lines, there-
fore lack of methylation in any single site (regardless of the
methylation status of other sites) would result in absence
of amplification products (Fig.3). In contrast, no such con-
cordance between collective methylation changes in LTR
and RT regions is evident in the other three lines.
2.3 Comparison of DNA methylation changes in Tos17
of the introgression lines detected by DNA gel blotting
analysis and PCR-based assay
By comparing the results obtained with the two types
of analysis, all the main findings of this study are corrobo-
rating each other, and hence the conclusions are reinforced.
For example, both methods have unequivocally showed
that, compared with the parent, DNA methylation alterations
occurred in Tos17 of all five introgression lines, and the
changes include both symmetric and asymmetric cytosines
as well as adenines. Nevertheless, the PCR-based assay
apparently revealed more specific and detailed information
in terms of types of methylation changes (demethylation or
hypermethylation) and the extent of changes (semi-
quantitative). We note however that in some cases, incon-
gruent results may emerge from the two types of analysis.
For example, in the DNA gel blotting analysis, appearance
of novel fragments larger than 5 kb (the complete length of
Fig.3. DNA methylation changes in the LTR and RT regions of
Tos17 in five introgressed rice lines as revealed by methylation-
sensitive enzyme restriction coupled with sequence-specific PCR
amplification. Lanes 1 to 6 are respectively the rice parent (cv.
Matsumae), introgression lines 1, 2, 3, 4 and 5; the panels respec-
tively refer to templates of undigested (a), digested with FokⅠ
(b), AluⅠ (c), HaeⅢ (d), HphⅠ (e), Sau3AⅠ (f), HpaⅡ (g)
and MspⅠ (h). M is molecular marker jx174.
Acta Botanica Sinica 植物学报 Vol.46 No.1 2004106
Tos17 is 4.2 kb, Fig.1) was detected in all five lines (Fig.2
upper panel, denoted by circles), thus implicating increase
at methylation levels in all lines. In contrast, in the PCR
assay, two lines (1 and 3) showed only demethylation
changes (Fig.3; Table 1). This incongruity between the
two methods can be reconciled if the high molecular frag-
ments are due to methylation changes and/or differential
methylation levels in the flanking genomic regions, instead
of changes in the element per se.
3 Discussion
Accumulated evidence has implicated that DNA methy-
lation plays dual roles in regulating cellular gene expres-
sion and in genome defense against parasitic or invasive
nucleic acid sequences in eukaryotes (Martienssen and
Colot, 2001). In mammalian animals, disruption of intrinsi-
cally programmed DNA methylation patterns is lethal at
early stages (Li et al., 1992). In Arabidopsis, decreased
DNA methylation, resulting from either mutation in the DDM
gene (decrease in DNA methylation) or experimental knock-
out of a DNA methyltransferase gene, causes abnormal
development and heritable alteration in multiple phenotypic
traits (Finnegan et al., 1996; Ronemus et al., 1996). Al-
though these lines of evidence appears to support the es-
sential role of DNA methylation in development, an emerg-
ing notion regarding the primary function of DNA methyla-
tion in eukaryotes is to serve as a genome defense system
(Yoder et al., 1997). Among the data lending support to
this paradigm shift, is the finding that in almost all studied
eukaryotes, the primary targets for cytosine methylation
are transposons and foreign nucleic acids (DNA or RNA
sequences, e.g. transgenes or viruses) that are often sup-
pressed or silenced as a result of hypermethylation (Bender,
1998). Moreover, recent works on both plants and animals
have demonstrated that methylation status and activity
states of transposons are highly correlated or even caus-
ally linked (O’Neill et al., 1998; Hirochika et al., 2000; Miura
et al., 2001).
The mechanism for DNA methylation changes in Tos17
as a result of DNA introgression from wild rice (Z. latifolia)
remains elusive. Nevertheless, we suspect that it could be
part of the genome-wide methylation changes in these in-
trogression lines whereby both repetitive and low-copy
number sequences could be targeted, as was supported by
circumstantial evidence (Liu et al., 1999a). Alternatively, it
is also possible that mobile elements are preferentially tar-
geted for modification, probably because of their sudden
copy number increase and/or special structural features.
Although reports on host genome methylation changes
due to foreign DNA introgression is scarce in plants, this
phenomenon has been well documented in mammalian cul-
tured cells (Heller et al., 1995). It was proposed that foreign
DNA integration into host mammalian genomes could cause
local chromatin state changes, which on the one hand could
serve as a mark to attract de novo DNA methylation, and on
the other hand could change accessibility of the local ge-
nomic region to DNA methyltransferases (Remus et al.,
1999). This particular site(s) could in turn serve as foci for
spreading of the modified methylation status to flanking
sequences and even to homologous ectopic regions, prob-
ably through transit and aberrant somatic homologous pair-
ing (Meyer et al., 1994). This hypothesis could certainly
suite the rice introgression lines used in the present study:
these lines contain less than 0.1% (unpublished data) inte-
grated genomic DNA from wild rice (Z. latifoila), a related
but sexually incompatible species to rice, which is appar-
ently a rich source for both homologous and divergent
(can be perceived as foreign) DNA sequences.
The paradox between the extremely low-copy number
of Tos17 in the rice genome and its propensity to be acti-
vated by certain conditions, such as tissue culture
(Hirochika et al., 1996) or foreign DNA introgression (Liu
and Wendel, 2000), strongly suggests a tight control on its
activity by epigenetic means. Thus, changes in DNA me-
thylation status of the element (Figs. 2, 3; Table 1) follow-
ing mobilization (Liu and Wendel, 2000) may be related to
its epigenetic control. This possibility is corroborated by
the observation in the PCR assay showing that the collec-
tive methylation levels at both the LTR and RT regions in
lines 2, 4 and 5 are much higher compared with those in
lines 1 and 3 (Fig.3; Table 1), and that the element copy
number in the former three lines are also much higher than
that of the later two lines (Liu and Wendel, 2000). This
indicates that higher copy number of Tos17 is accompa-
nied by concomitantly elevated level of DNA methylation,
suggesting that the newly mobilized copies of the element
have been rapidly and extensively methylated.
Nevertheless, because Tos17 is no longer active in these 9-
generation-old introgression lines and earlier generations
are not available, an exact relationship between methyla-
tion status and element activity remains to be established.
It has been established for decades that cytosine me-
thylation is exclusively restricted to CpG dinucleotide in
mammals and predominately occurs at CpG and CpNpG
sites in plants and fungi (reviewed in Martienssen and Colot,
2001). But recent study demonstrates that in plants and
fungi cytosine methylation can occur at cytosines of virtu-
ally any sequence context (Finnegan et al., 1998), thus raises
107DONG Yu-Zhu et al.: DNA Methylation in a Rice Retrotransposon Tos17
the question concerning possible function and mechanism
for maintenance of the asymmetrical cytosine methylations
in these organisms. Studies in both plants and fungi indi-
cated that asymmetrical cytosine methylation is not only
associated with epigenetic silencing of transgene and
endogens (Selker et al., 1993; Dieguez et al., 1997; Fu et al.,
2000), but also play crucial roles in repressing activity of
retrotransposons (Martienssen and Colot, 2001). For asym-
metric cytosine methylations to have important biological
functions, their stable perpetuation and inheritance is
required. Indeed, it was found that in maize and tobacco at
least some asymmetric cytosine methylations could be sta-
bly inherited by using the nearby faithfully maintained sym-
metrical CpG and/or CpNpG methylated sites as a marker to
attract de novo methylation and then spread the methy-
lated state to the entire local region (Dieguez et al., 1997).
In the present study, no apparent coordination between
CCGG methylation changes (detected by Hpa Ⅱ /MspⅠ)
and cytosine methylations of several other sequence con-
texts in introgression lines could be established (Fig.3; Table
1), although even asymmetrical methylation changes that
occurred in these two lines are stably perpetuated through
selfed generations (between generations 9 and 10; data not
shown).
An interesting observation of the present study is the
apparent concordance between the collective methylation
status at the LTR and the RT regions of Tos17 in lines 1 and
3 (Fig.2; Table 1). In contrast, it was shown in transgenic
plants that the spreading of methylation status is usually
confined to the promoter or coding region of the transgene
but not across the two regions, and hence absence of con-
cordance in methylation status between promoter and the
coding region is expected. It would be interesting to eluci-
date reasons for the difference between a transgene and a
retrotransposon.
Adenine methylation plays essential biological roles in
prokaryotes (Heithoff et al., 1999), but is believed to occur
only infrequently in plants. Nevertheless, it was found
that adenine methylation changes are associated with
somaclonal var iat ions in r ice (Lorz, personal
communication). We showed here that adenines within an
endogenous rice retrotransposon are also subjected to
heritable methylation changes as a result of foreign DNA
introgression, although their possible roles is not
understood.
To conclude, we have demonstrated that introgression
of alien DNA from wild rice has caused extensive and ver-
satile DNA methylation pattern alterations in a quiescent
rice endogenous retrotransposon Tos17. These changes
in epigenetic state are likely associated with activity of the
element. Further work is needed to elucidate the exact na-
ture of relationship between methylation and the element
activity. Our results have apparent implications for breed-
ing programs involving introgression of alien genes from
related wild species. It is likely that the introgression pro-
cess is not only introducing genetic variations, but also
causing an array of heritable epigenetic changes.
Acknowledgements: We are grateful to Prof. J F WENDEL
of the Iowa State University, USA for constructive sugges-
tions to this study, and for critical reading of the manuscript.
References:
Bender J. 1998. Cytosine methylation of repeated sequences in
eukaryotes: the role of DNA pairing. Trends Biochem, 23:
252-256.
Chandler V L, Walbot V. 1986. DNA modification of a maize
transposable element correlates with loss of activity. Proc
Natl Acad Sci USA, 83:1767-1771.
Dieguez M J, Bellotto M, Afsar K, Mittelsten O, Paszkowski J.
1997. Methylation of cytosines in nonconventional methyla-
tion acceptor sites can contribute to reduced gene expression.
Mol Gen Genet, 253:581-588.
Finnegan E, Peacock J, Dennis E. 1996. Reduced DNA methyla-
tion in Arabidopsis results in abnormal plant development.
Proc Natl Acad Sci USA, 93:8449-8454.
Finnegan E J, Genger P K, Peacock W J, Dennis E S. 1998. DNA
methylation in plants. Ann Rev Plant Physiol Plant Mol Biol,
49:223-248.
Fu X, Kohli A, Twyman R M, Christou P. 2000. Alternative
silencing effects involve distinct types of non-spreading cy-
tosine methylation at a three-gene, single-copy transgenic lo-
cus in rice. Mol Gen Genet, 263:106-118.
Grandibastien M A. 1998. Activation of plant retrotransposons
under stress conditions. Trends Plant Sci, 3:181-187.
Heithoff D M, Sinsheimer R L, Low D A, Mahan M J. 1999. An
essential role for DNA adenine methylation in bacterial
virulence. Science, 284:967-970.
Heller H, Kammer C, Wilgenbus P, Doerfler W. 1995. The chro-
mosomal insertion of foreign (adenovirus type 12, plasmid of
bacteriophage) DNA is associated with enhanced methylation
of cellular DNA segments. Proc Natl Acad Sci USA, 92:5515-
5519.
Hirochika H, Okamoto J, Kakutani T. 2000. Silencing of
retrotransposons in Arabidopsis and reactivation by the ddm1
mutation. Plant Cell, 12:357-368.
Hirochika H, Sugimoto K, Otsuki Y, Kanda M. 1996.
Acta Botanica Sinica 植物学报 Vol.46 No.1 2004108
Retrotransposons of rice involved in mutations induced by
tissue culture. Proc Natl Acad Sci USA, 93:7783-7787.
Hirochika H. 1993. Activation of tobacco retrotransposons dur-
ing tissue culture. EMBO J, 12:2521-2528.
Hirochika H. 2001. Contribution of the Tos17 retrotransposon to
rice functional genomics. Curr Opin Plant Biol, 4:118-122.
Hirochika H. 1997. Retrotransposons of rice: their regulation and
use for genome analysis. Plant Mol Biol, 35:231-240.
Kalendar R, Tanskanen J, Immonen S, Nevo E, Schulman A H.
2000. Genome evolution in wild barley (Hordeum spontaneum)
by BARE-1 retrotransposon dynamics in response to sharp
microclimatic divergence. Proc Natl Acad Sci USA, 97:6603-
6607.
Kidwell K K, Osborn T C. 1992. Simple plant DNA isolation
procedures. Beckman J S, Osborn T C. Plant Genomes:
Methods for Genetic and Physical Mapping. Dordrecht, The
Netherlands: Kluwer Academic Publishers. 1-13.
Kumar A, Hirochika H. 2001. Applications of retrotransposons
as genetic tools in plant biology. Trends Plant Sci, 6:127-
134.
Kumer A, Bennetzen J L. 1999. Plant retrotransposons. Annu
Rev Genet, 33:479-532.
Li E, Bestor T H, Jaenisch R. 1992. Targeted mutation of the
DNA methyltransferase gene results in embryonic lethality.
Cell, 69:915-926.
Liu B, Piao H M, Zhao F S, Zhao J H, Liu Z L, Huang B Q. 1999a.
DNA methylation changes in rice induced by Zizania latifolia
Griseb. DNA introgression. Hereditas, 131:75-78.
Liu B, Piao H M, Zhao F S, Zhao J H, Zhao R. 1999b. Production
and molecular characterization of rice lines with introgressed
traits from a wild species of Zizania latifolia Griseb. J Genet
Breeding, 53:279-284.
Liu B, Wendel J F. 2000. Retrotransposon activation followed by
rapid repression in introgressed rice plants. Genome, 43:874-
880.
Liu B, Wendel J F. 2003. Epigenetic phenomena and genome
evolution of plant allopolyploids. Mol Phylogenet Evol, 14.
(in press)
Martienssen R A, Colot V. 2001. DNA methylation and epige-
netic inheritance in plants and filamentous fungi. Science, 293:
1070-1074.
Meyer P, Niedenhof I, Ten Lohuis M. 1994. Evidence for cy-
tosine methylation of non-symmetrical sequences in transgenic
Petunia hybrida. EMBO J , 13:2084-2088.
Miura A, Yonebayashi S, Watanabe K, Toyama T, Shimada H,
Kakutani T. 2001. Mobilization of transposons by a muta-
tion abolishing full DNA methylation in Arabidopsis. Nature,
411:212-214.
O’Neill R J W, O’Neil l M J, Graves J A M. 1998.
Undermethylation associated with retroelement activation and
chromosome remodeling in an interspecific mammalian hybrid.
Nature, 393:68-72.
Remus R, Kammer C, Heller H, Schemitz B, Schell G, Doerfler
W. 1999. Insertion of foreign DNA into an established mam-
malian genome can alter the methylation of cellular DNA
sequences. J Virol, 73:1010-1022.
Ronemus M, Galbiati C, Ticknor J, Chen J, Dellaporta S. 1996.
Demethylation-induced developmental pleiotropy in
Arabidopsis. Science, 273:653-657.
Selker E U, Fritz D Y, Singer M J. 1993. Dense non-symmetrical
DNA methylation resulting from repeat induced point
mutation. Science, 262:1724-1728.
Wessler S R. 1996. Plant retrotransposons: turned on by stress.
Curr Biol, 6:959-961.
Yamazaki M, Tsugawa H, Miyao A, Yano M, Wu J, Yamamoto S,
Matsumoto T, Sasaka T, Hirochika H. 2001. The rice
retrotransposon Tos17 prefers low-copy-number sequences
as integration targets. Mol Genet Genomics, 265:336-344.
Yoder J A, Walsh C P, Bestor T H. 1997. Cytosine methylation
and the ecology of intragenomic parasites. Trends Genet, 13:
335-340.
(Managing editor: ZHAO Li-Hui)