T-DNA Integration in Arabidopsis Chromosomes.
Presence and Origin of Filler DNA Sequences1[w]
Pieter Windels, Sylvie De Buck, Erik Van Bockstaele, Marc De Loose, and Ann Depicker*
Department of Plant Genetics and Breeding, Center of Agricultural Research-Gent, Caritasstraat 21, B–9090
Melle, Belgium (P.W., E.V.B., M.D.L.); Department of Plant Systems Biology, Flanders Interuniversity
Institute for Biotechnology, Ghent University, Technologiepark 927, B–9052 Gent, Belgium (S.D.B., A.D.); and
Department for Plant Production, Ghent University, Coupure Links 653, B–9000 Gent, Belgium (E.V.B.)
To investigate the relationship between T-DNA integration and double-stranded break (DSB) repair in Arabidopsis, we
studied 67 T-DNA/plant DNA junctions and 13 T-DNA/T-DNA junctions derived from transgenic plants. Three different
types of T-DNA-associated joining could be distinguished. A minority of T-DNA/plant DNA junctions were joined by a
simple ligation-like mechanism, resulting in a junction without microhomology or filler DNA insertions. For about one-half
of all analyzed junctions, joining of the two ends occurred without insertion of filler sequences. For these junctions,
microhomology was strikingly combined with deletions of the T-DNA ends. For the remaining plant DNA/T-DNA
junctions, up to 51-bp-long filler sequences were present between plant DNA and T-DNA contiguous sequences. These filler
segments are built from several short sequence motifs, identical to sequence blocks that occur in the T-DNA ends and/or
the plant DNA close to the integration site. Mutual microhomologies among the sequence motifs that constitute a filler
segment were frequently observed. When T-DNA integration and DSB repair were compared, the most conspicuous
difference was the frequency and the structural organization of the filler insertions. In Arabidopsis, no filler insertions were
found at DSB repair junctions. In maize (Zea mays) and tobacco (Nicotiana tabacum), DSB repair-associated filler was normally
composed of simple, uninterrupted sequence blocks. Thus, although DSB repair and T-DNA integration are probably closely
related, both mechanisms have some exclusive and specific characteristics.
The bacterial phytopathogen Agrobacterium tumefaciens provides a natural system to transfer and introduce genetic information into susceptible plant cells.
The T-DNA region, a portion of the tumor-inducing
plasmid delimited by two 24-bp border repeats, is
transferred to the plant cell as a single-stranded DNA
after vir (virulence) gene induction (Gheysen et al.,
1998; Gelvin, 2003). Upon arrival in the plant cell, the
T-strand is targeted to the cell nucleus, and the
T-DNA becomes integrated into the host genome by
the plant’s illegitimate recombination apparatus
(Matsumoto et al., 1990; Gheysen et al., 1991; Mayerhofer et al., 1991; Tinland, 1996). Recently, T-DNA
integration in the Arabidopsis genome has been
shown to depend on the sequence of the pre-insertion
site (Brunaud et al., 2002).
That T-DNA strands can be captured at doublestranded breaks (DSBs; Salomon and Puchta, 1998) in
tobacco (Nicotiana tabacum) and also that T-DNA integration in yeast (Saccharomyces cerevisiae) depends
on nonhomologous end-joining (NHEJ) enzymes
1
This work was supported by the European Community (grant
no. 18687–2001–11), by the European Union BIOTECH program
(grant no. QLRT–2000 – 00078), and by the Fund for Scientific
Research Flanders (G.0118.01).
[w]
The online version of this article contains Web-only data.
* Corresponding author; e-mail [email protected]; fax
32–9 –3313809.
Article, publication date, and citation information can be found
at http://www.plantphysiol.org/cgi/doi/10.1104/pp.103.027532.
(van Attikum et al., 2001) suggest that doublestranded break repair via NHEJ is involved as a
pathway in T-DNA integration. More recently, the
T-DNA integration efficiency has been found to decrease in Ku80- and DNA ligase IV-deficient plants
(Friesner and Britt, 2003). The first molecular data on
NHEJ in plants came from the footprint analysis of
excised transposons. Repair at transposon-induced
DSBs seems to generate only minor changes, such as
small deletions, insertion of filler, or inversions (Gorbunova and Levy, 1999; Yan et al., 1999). Also,
T-DNA integration has been studied as an example
of NHEJ. Short regions of microhomology, insertions
of filler DNA, and small deletions have been reported
as features of T-DNA/plant DNA junctions (Matsumoto et al., 1990; Gheysen et al., 1991; Mayerhofer et
al., 1991; Tinland, 1996; Fladung, 1999; Kumar and
Fladung, 2002; Meza et al., 2002). More direct NHEJ
analysis in plants involved the characterization of
naturally occurring deletions in plant genomes (Ralston et al., 1988; Wessler et al., 1990) and the sequencing of newly formed junctions at repaired DSBs (Gorbunova and Levy, 1997; Salomon and Puchta, 1998).
For five spontaneous deletions of the waxy locus in
maize (Zea mays; Wessler et al., 1990), for a spontaneous deletion in the maize bz-R allele (Ralston et al.,
1988), and for NHEJ-repaired junctions in tobacco
cells (Gorbunova and Levy, 1997; Salomon and Puchta, 1998), the majority of repaired junctions contained uninterrupted filler insertions (Gorbunova
Downloaded
on Junewww.plantphysiol.org
16, 2017 - Published by©www.plantphysiol.org
Plant Physiology, December 2003, Vol.
133, pp. from
2061–2068,
2003 American Society of Plant Biologists
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
2061
Windels et al.
and Levy, 1999). Recently, species-specific differences in plant DSB repair have been reported by
Kirik et al. (2000), who showed that DSB repair in
tobacco cells is accompanied by insertions but not in
Arabidopsis in which large deletions of the joined
plant DNA ends are frequent. These findings have
been further substantiated by Orel and Puchta (2003).
These authors suggest that free DNA break ends in
Arabidopsis are less stable and more prone to end
degradation than free DNA ends in tobacco cells.
To investigate the relationship between DSB repair
by NHEJ and T-DNA integration in Arabidopsis, we
sequenced and analyzed 67 T-DNA/plant DNA junctions and 13 junctions between linked T-DNAs. We
compared the outcome of the T-DNA/plant DNA
and T-DNA/T-DNA joining reaction with the characteristics of NHEJ-repaired DSBs reported in Arabidopsis (Kirik et al., 2000; Orel and Puchta, 2003) and
other plant species (Gorbunova and Levy, 1997;
Salomon and Puchta, 1998). This comparison mainly
focused on the frequency, the presence, and the origin of filler DNA sequences.
plant DNA for T-DNA/plant DNA junctions could
be determined. Similarly, for linked T-DNAs, the end
points of the two adjacent T-DNA borders were
known. Filler sequences longer than 17 bp were
screened for complete colinear sequence identity
with the Arabidopsis genome to evaluate whether
the observed filler sequences could be uninterrupted
sequence blocks that had been copied from the plant
genome. In addition, for 21 filler DNAs larger than 6
bp, we analyzed the origin of the filler DNA by
looking for identical sequence motifs with: (a) the
first 100 bp of the T-DNA sequence situated immediately adjacent to the transition point of the T-DNA
junction, (b) a 200-bp plant DNA segment surrounding the T-DNA integration point, and (c) the complete T-DNA plasmid sequence (see “Materials and
Methods” and Supplemental Data). An experimental,
statistical analysis was performed to determine
thresholds for reporting identities, allowing us to
discriminate mechanistic relevance from by chance
occurrence (see “Materials and Methods” and Supplemental Data).
RESULTS
Outcome of the Joining Reaction between T-DNA and
Plant DNA Ends
Experimental Setup
T-DNA junctions, derived from a population of
Arabidopsis transformants (see “Materials and Methods”), were amplified and sequenced. We characterized 44 left T-DNA/plant DNA junctions, 23 right
T-DNA/plant DNA junctions, eight junctions between tandemly linked T-DNAs, and five junctions
connecting two T-DNAs in inverted orientation, for
which one T-DNA was truncated for the T-DNA
border region (see supplemental data, available in
the online version of this article at http://www.
plantphysiol.org). To determine the end point of the
T-DNA borders, the T-DNA region was aligned
against the T-DNA plasmid sequence. We will refer
throughout to a correct nicking process of the
T-strand when the T-DNA is cleaved between the
third and fourth base of the 24-bp border repeat (Van
Haaren et al., 1988).
Further sequence analysis revealed three distinct
junction types. A first class included T-DNA junctions that were joined by a simple ligation-like mechanism, resulting in a junction without microhomology or filler DNA insertions. A second class had
microhomology at the transition point between two
adjacent DNA segments. Here, microhomology is defined as small DNA sequences, ranging from 1 to 7
bp and found at the junction point between the
T-DNA and the plant DNA for plant DNA/T-DNA
junctions and between two T-DNAs, when linked
T-DNAs are studied (Roth et al., 1985; Roth and
Wilson, 1986). A third class grouped junctions with
insertion of filler DNA. These new, intermediate filler
sequences were detected easily because the end point
of the T-DNA border, and the starting point of the
2062
We wanted to address the question whether T-DNA
junction characteristics and the mode of T-DNA joining are related. Therefore, we characterized the outcome of the joining reaction of 67 plant DNA/T-DNA
junctions, taking into account the T-DNA border type
involved and the degree of T-DNA end processing
(Table I). Of 44 left border T-DNA/plant DNA junctions, four had the T-DNA end joined to the plant
DNA without the presence of microhomology or the
formation of filler DNA sequences, 23 displayed microhomology of 1 bp up to a stretch of 6 bp, and 17 had
filler DNA insertions ranging from 1 up to 48 bp
(Table I; Supplemental Data). Similar results were obtained for the right border junctions: three of 23 junctions were joined by means of a simple ligation-like
joining reaction, 10 had microhomology of 2 up to 7
bp, and 10 had filler DNA ranging from 1 to 51 bp
(Table I; Supplemental Data). These results clearly
indicate that the frequency by which simple ligation,
microhomology, or filler DNA is observed at left and
right border junctions is comparable.
We also analyzed whether T-DNA joining characteristics and the degree of T-DNA end processing
Table I. Features of 67 T-DNA/plant DNA junctions
T-DNA end
Simple
Ligation
Microhomology
Filler
Sequences
Left T-DNA border
Right T-DNA border
Intact T-DNA ends
Processed T-DNA ends
4/44
3/23
2/13
5/54
23/44
10/23
3/13
30/54
17/44
10/23
8/13
19/54
7/67
33/67
27/67
Total
Downloaded from on June 16, 2017 - Published by www.plantphysiol.org
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
Plant Physiol. Vol. 133, 2003
Characterization of Filler DNA
were correlated (Table I; Supplemental Data). Regarding the T-DNA end processing, our results confirm recent observations (Brunaud et al., 2002;
Krysan et al., 2002; Meza et al., 2002) that, although
the deletion size at the left T-DNA ends is larger than
that at the right borders, processing and trimming is
a ubiquitous feature of T-DNA ends. Interestingly,
when the degree of T-DNA end processing is correlated with the outcome of the joining reaction, three
of 13 intact T-DNA ends were joined to the plant
DNA with microhomology (Table I), in contrast to 30
of 54 processed T-DNA ends (Table I). Also striking
is that eight of 13 intact and only 19 of 54 processed
T-DNA ends were joined with insertion of filler sequences. In conclusion, microhomology was on average 2.4-fold more associated with deleted than with
intact T-DNA borders (Table I). Intact T-DNA ends
were on average 1.7-fold more associated with filler
DNA than were T-DNA junctions with deleted
T-DNA ends (Table I). A Fisher exact test was performed that verified that the correlation between the
intactness or the processing of T-DNA ends, and the
presence of filler DNA or microhomology, respectively, was significant (P ⬍ 0.05; see “Materials and
Methods” and Supplemental Data). However, these
conclusions should be considered with caution because of the small number of analyzed junctions.
Origin of Filler DNA at T-DNA/Plant DNA Junctions
Analysis of DSB lesions repaired via NHEJ in Arabidopsis has shown that none of 40 analyzed junctions had filler insertions (Kirik et al., 2000). Because
of the striking difference with our observation that in
Arabidopsis 27 of 67 T-DNA junctions (approximately 40%) harbor filler insertions (Table I), we
investigated the origin and structure of the filler
sequences found at the T-DNA junctions.
The origin of the filler DNA sequences was determined by comparative sequence analysis. The results
of this analysis are presented in Figure 1A. For a full
overview, containing all filler sequences and their
origin analysis, see Supplemental Data. None of the
filler sequences larger than 17 bp had a complete
colinear sequence identity with the Arabidopsis genome sequence, indicating that the observed filler
insertions did not consist of uninterrupted sequence
blocks duplicated from the Arabidopsis genome. For
15 of the 17 T-DNA/plant DNA junctions with intermediate filler DNA of 6 bp or more, the filler DNA
was made up of short duplicated sequences, identical
to sequence motifs occurring in either the plant DNA
surrounding the T-DNA insertion point and/or the
T-DNA plasmid sequence (Fig. 1; Supplemental
Data; plant DNA and T-DNA plasmid-derived repeats are indicated above or below the junction sequence, respectively). The remaining two filler insertions of 8 and 11 bp did not contain a sequence motif
longer than the determined thresholds for reporting
Plant Physiol. Vol. 133, 2003
sequence identities. Therefore, in these two cases, we
were unable to discern mechanistic relevance and by
chance occurrence (see “Materials and Methods”). Of
the 15 analyzed filler DNAs for which the origin
could be attributed, four contained sequence motifs,
identical to the plant DNA target or the T-DNA end,
four had sequence blocks identical to the plant DNA
target only, and seven consisted of sequence identities with the T-DNA end only (Fig. 1; Supplemental
Data).
The first class of filler DNAs, with both plant DNAand T-DNA-derived identical sequence blocks, is
clearly illustrated by junction kg353LB-21 (Fig. 1A-3).
This T-DNA/plant DNA junction had a 13-bp filler
sequence, made up of five identified, different, and
overlapping sequence motifs. Two sequence blocks
of 9 bp (Fig. 1A-3, motifs 9 and 10) were identical to
sequence motifs present in the plant target, whereas
the other motifs, one of 7 bp (Fig. 1A-3, motif 13) and
two of 6 bp (Fig. 1A-3, motifs 11 and 12), were
identical to sequence blocks occurring in the T-DNA
border sequence. The sequence identities with the
plant genome corresponded either with the plant
DNA immediately adjacent to the left border end of
the T-DNA (Fig. 1A-3, motif 9) or with the plant DNA
adjacent to the opposite T-DNA end of the same
T-DNA insert (Fig. 1A-3, motif 10). The different
sequence motifs, which make up the filler, shared
small stretches of microhomology of up to a few
bases, and microhomology was usually restricted to
the ends of adjacent motifs. Moreover, microhomology was also observed between the filler DNA and
the T-DNA end and/or the plant DNA end (Fig.
1A-3, motif 13). On the other hand, sequence blocks
with a full overlap were also detected. For junction
kg353LB-21, motif 12 lies entirely within motif 9 (Fig.
1A-3). Similar results were obtained for the other
filler DNA sequences (see Supplemental Data), and
mutual microhomology seemed to be a general characteristic of all sequence identities that make up a
filler sequence at T-DNA/plant DNA junctions. Furthermore, junction kg353LB-21 also nicely illustrated
that sequence identities that make up a filler can be in
direct or inverted orientation relative to their template DNAs (Fig. 1, solid or arrowed line, respectively). One 9-bp sequence block, identical to a sequence motif present in the plant target (Fig. 1A-3,
motif 9), was in inverted orientation, whereas the
other four sequence identities were in direct orientation relative to their template DNAs. In general, 20%
of all sequence identities in the filler had an inverted
orientation, but the majority was in direct orientation
relative to their template DNA.
The second class of filler sequences, containing
sequences identical with the plant target only, is
illustrated by junction kg353RB-8 (Fig. 1A-1). This
right border T-DNA/plant DNA junction harbored a
51-bp filler segment composed of three sequence motifs, identical to sequence blocks occurring in the
Downloaded from on June 16, 2017 - Published by www.plantphysiol.org
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
2063
Windels et al.
Figure 1.
2064
(Figure legend continues on following page)
Downloaded from on June 16, 2017 - Published by www.plantphysiol.org
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
Plant Physiol. Vol. 133, 2003
Characterization of Filler DNA
plant target. First, one 20-bp sequence block (Fig.
1A-1, motif 4) was identical to the plant DNA that
was immediately adjacent to the filler sequence itself.
Then, a 30-bp sequence (Fig. 1A-1, motif 1) was identical to the plant DNA adjacent to the opposite
T-DNA end of the same T-DNA insert. The third, a
21-bp sequence block (Fig. 1A-1, motif A), was identical to part of the target site deletion.
The final class of filler sequences consisted solely of
sequence motifs identical to sequence blocks occurring in the T-DNA. Junction kg150LB-8 (Fig. 1A-4)
contained an 8-bp filler DNA harboring identical
sequences in inverted orientation to two left-end
T-DNA-derived sequence motifs (Fig. 1A-4, motifs 16
and 17). Also, a 16-bp sequence motif enclosing the
8-bp filler DNA was completely identical with a
16-bp sequence block present in the T-DNA 1,360 bp
distant from the left border T-DNA end. This observation shows that sequence motifs observed in filler
DNAs did not always originate from sequences proximal to the filler segment itself. Similarly, for the filler
insertions at junctions kg44LB-2, CK2L102RB-7, and
kg135RB-4, sequence motifs identical to sequence
blocks positioned 293, 1,693, and 638 bp, respectively,
from the corresponding T-DNA end were found (see
Supplemental Data, A1, B1, and B4).
In addition, chimeric filler sequences (Fig. 1, dotted
lines) were identified, meaning that the observed
sequence motifs in these filler segments were duplicated within the filler DNA itself. Chimeric filler
sequences were present in junctions kg353LB-20 (Fig.
1A-3, motifs 14 and 15), kg135RB-4 (Supplemental
Data, B4, motif 43), and kg165-12 (Fig. 1C-1, motifs 28
and 29).
Origin of Filler DNA at T-DNA/T-DNA Junctions
Because the frequency of filler insertions differed in
T-DNA/plant DNA junctions and DSB-repaired
plant DNA/plant DNA joints, we also analyzed the
frequency and origin of filler insertions at T-DNA/
T-DNA junctions to investigate the influence of the
plant DNA end in a T-DNA/plant DNA joining re-
action. Of eight junctions between tandemly linked
T-DNAs, three contained filler DNA sequences of 1,
21, and 38 bp; of three junctions between T-DNAinverted repeats over the left border, one junction
harbored an intermediate 21-bp filler DNA, and of
two junctions between two right border T-DNA ends,
both contained filler sequences of 4 and 40 bp. Taken
together, six of 13 linked T-DNAs harbor filler insertions at their junction (see Supplemental Data).
For the two junctions between tandemly linked
T-DNAs with filler DNA of 21 and 38 bp, the origin
of the filler sequences was determined (Fig. 1B). For
these two junctions, the origin of the filler was attributed to T-DNA sequences. The filler DNA at junction
kd73-4 (Fig. 1B-1) was identical to a sequence motif
found in the T-DNA sequence 1,181 bp distant from
the T-DNA border end. It is the only filler reported in
this study that consisted of an uninterrupted sequence motif. In contrast, the 38-bp filler sequence at
junction kg150-7 (Fig. 1B-2) is a nice example of how
alternating sequence motifs identical to different
T-DNA sequences make up a filler DNA. Similar
results were observed for the filler DNAs of two
inverted repeat junctions (Fig. 1, C-1 and C-2).
In addition, we analyzed whether the observed
filler DNAs at linked T-DNAs could be attributed to
Arabidopsis genomic sequences. In none of the filler
DNAs could an identical sequence longer than 15 bp
be detected with 100% identity with the plant genome. Only for junction kd12-1, a 16-bp plant DNA
sequence was found that overlapped with two
T-DNA-located sequence identities (Fig. 1C-2, italicized sequence). Nevertheless, the statistical significance of this 16-bp sequence motif, identical to a
sequence block occurring in the plant genome, is
difficult to evaluate because the surrounding plant
DNA for the kd12 T-DNA locus had not been characterized. In conclusion, our findings with regard to
the origin of the filler sequences at T-DNA/T-DNA
junctions resemble those for T-DNA/plant DNA
junctions.
Figure 1. (Continued from previous page)
Origin of filler DNA sequences for T-DNA/plant DNA junctions and for junctions between linked T-DNAs. The sequence
of each junction for which the origin of the filler DNA (represented by different types of lines) was determined is presented
and so are the border characteristics for each junction. The T-DNA end sequences are given in uppercase, plant DNA
sequences in lowercase, and intermediate filler DNA in bold lowercase. All sequence motifs, identical to sequence blocks
occurring in the plant DNA surrounding the T-DNA integration point or the T-DNA plasmid sequence, are positioned above
and below the junction sequence, respectively. All sequence motifs for which the origin could be indicated on the figure
are given consecutive numbers, and corresponding motifs have the same number. Motifs designated by uppercase letters are
identical to sequence blocks found in the plant DNA target, for which the originating sequence are not represented on the
figure. Motif A is identical to a sequence block occurring in the target site deletion, whereas motif B is either identical to
a sequence block from the T-DNA target site deletion or from the plant DNA adjacent to the opposite T-DNA end of the same
T-DNA insert. For sequence motifs that are identical to T-DNA plasmid sequence blocks that are not present on the figure
(internal T-DNA plasmid sequences), the position of the original sequence was determined, and its position is given between
parentheses. Inverted, direct, and chimeric sequence motifs are indicated with arrowed, solid, and dotted lines, respectively.
LB and RB, Left and right border T-DNA ends for linked T-DNAs, respectively. For junction kd12-1, the italicized sequences
correspond to the Arabidopsis genomic DNA. Asterisk, Base mismatch with the original sequence; double slash, Both
junction sequences correspond to one and the same T-DNA insert (see “Experimental Setup”).
Plant Physiol. Vol. 133, 2003
Downloaded from on June 16, 2017 - Published by www.plantphysiol.org
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
2065
Windels et al.
DISCUSSION
That T-DNAs can be captured at DSB sites in tobacco (Salomon and Puchta, 1998), that NHEJ enzymes in yeast are involved in T-DNA integration
(van Attikum et al., 2001), and that Ku80- and DNA
ligase IV-deficient plants exhibit a reduced T-DNA
integration efficiency (Friesner and Britt, 2003) have
been taken as evidence for the role played by NHEJ
mechanisms in a pathway for T-DNA integration.
However, when we compare the outcome of the joining reaction at the T-DNA junctions with junction
characteristics of repaired DSBs in plant cells, some
remarkable differences appear. First, NHEJ-repaired
junctions in Arabidopsis are characterized by the
occurrence of large deletions ranging up to 2,207 bp
(Kirik et al., 2000). T-DNA insertions in Arabidopsis
are normally associated with much smaller target site
deletions. We characterized target site deletions for
eight single-copy CK2L-transgenic lines, and we
found that on average, 48 bp of the genomic preinsertion target site are deleted (data not shown),
confirming earlier findings that on average the target
site deletion in the host genome is small after T-DNA
integration (Krysan et al., 2002; Meza et al., 2002).
However, the most conspicuous difference is that
approximately 40% of all T-DNA junctions harbor
filler, whereas filler insertions are absent at NHEJrepaired junctions in Arabidopsis (Kirik et al., 2000).
Moreover, not only the frequency of filler insertions
seems different, but also the composition of T-DNAassociated filler can be discriminated from filler insertions found at repaired DSBs in plant species other
than Arabidopsis (Ralston et al., 1988; Wessler et al.,
1990; Gorbunova and Levy, 1997; Salomon and Puchta, 1998; Yan et al., 1999). Filler insertions at TDNA/plant DNA junctions consist of several, predominantly short, sequence motifs that are identical
to sequence blocks occurring in: (a) the immediate
surrounding of the plant T-DNA target site, (b) the
T-DNA sequence, or (c) both. The T-DNA target sitederived motifs were identical to the pre-insertion site
deletion or to the plant DNA adjacent to either site of
the T-DNA, whereas T-DNA-derived sequence identities were scattered along the whole T-DNA sequence. Alternatively, chimeric sequence motifs that
form internal repeat structures in the filler DNA
sequence are observed. The majority of the sequence
motifs are in direct orientation relative to their template DNA (Fig. 1; Supplemental Data). In contrast,
the majority of the filler insertions found at DSBrepaired junctions in tobacco and maize are made up
of simple, uninterrupted sequence blocks identical to
sequences of the plant genome. Analogously, in
yeast, in which illegitimate recombination is normally not accompanied by insertion of filler DNA,
filler sequences are observed at the junctions upon
T-DNA integration into the yeast genome (Bundock
and Hooykaas, 1996). Moreover, DNA ligase IV and
Ku80, which are required for repair of DNA damage
2066
in Arabidopsis, seem not to be involved in the integration of T-DNA into plants (Gallego et al., 2003;
van Attikum et al., 2003). Taken together, although
DSB repair and T-DNA integration might be associated, both mechanisms probably have some exclusive
and specific characteristics. The observed differences
are seemingly linked to the T-DNA integration process, meaning that the T-DNA may follow a pathway
of its own during T-DNA integration rather than an
entirely plant cell-determined route.
The differences between T-DNA integration and
DSB repair might be explained in several ways. First,
the T-DNA is transferred to the plant nucleus as a
nucleoprotein complex, and T-DNA-associated factors or cotransferred virulence proteins might influence the outcome of the joining reaction. After transposon excision, transposition-specific proteins
probably play a role in protecting and repairing the
broken sites (Rinehart et al., 1997; Gorbunova and
Levy, 1999). We show that right and left T-DNA
borders are joined to the plant DNA in a similar
manner (Table I), implying that the right borderassociated VirD2 does not play a major role in it,
confirming the findings of Ziemienowicz et al. (2000).
However, other—yet unidentified—virulence factors
might take part in the T-DNA joining process. Second, subtle differences between T-DNA integration
and DSB repair may affect the NHEJ process. In our
opinion, complex filler at T-DNA junctions could
result from unstable interactions between the plant
DNA target site and the invading T-DNA end. DNA
repair via NHEJ has been shown to involve DNA
end-stabilizing factors, such as DNA end-binding Ku
proteins, a DNA-dependent protein kinase, the
XRCC4/DNA ligase IV complex, and the Mre11/
Rad50/Xrs2 complex (for review, see Tsukamoto and
Ikeda, 1998; Pfeiffer et al., 2000), for which analogs in
Arabidopsis have been identified (Hartung and Puchta, 1999; West et al., 2000; Gallego et al., 2001;
Daoudal-Cotterell et al., 2002; Tamura et al., 2002).
This DNA end-joining protein complex might protect
the DNA ends against nuclease-mediated degradation and also stabilize the intramolecular association
of the interacting DNA ends (Tsukamoto and Ikeda,
1998). Our results suggest that when filler insertions
are observed at T-DNA/plant DNA junctions, the
initial interaction between the invading T-DNA end
and the plant target is not stabilized immediately by
this kind of protein complexes. Before stabilization, a
free 3⬘ protruding T-DNA end either from the left or
right T-DNA end lands and screens for microcomplementarity. Upon landing, the free 3⬘ end is taken as
primer for simultaneous template-based DNA synthesis. Because at this stage the initial interaction is
not stabilized by host-dependent NHEJ-associated
protein complexes, the T-DNA takes off and lands in
the neighborhood. Landing and taking off is repeated
regularly. Different regions, such as the surroundings of the T-DNA target site and the T-DNA plasmid
Downloaded from on June 16, 2017 - Published by www.plantphysiol.org
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
Plant Physiol. Vol. 133, 2003
Characterization of Filler DNA
sequence, can be invaded, resulting in a patchworklike filler sequence. It is tempting to speculate that at
least during this initial interaction, DNA endstabilizing protein complexes are absent, whereas a
repair-polymerase complex might be involved. The
presence of polymerase activity and the absence of
exonuclease activity during this initial interaction
might explain why filler DNA has been associated
more frequently with intact T-DNA ends than with
T-DNA ends that are joined via a microhomologybased mechanism. In contrast, DSB repair sites might
not recruit DNA replication enzyme activities in
combination with the plant’s NHEJ machinery,
which might result in a faster and stable interaction
between both DSB ends.
MATERIALS AND METHODS
Transgenic Lines
Two populations of transgenic Arabidopsis C24 lines were used for
amplification of T-DNA junction fragments. A first population included 21
transgenic lines that were cotransformed with two Agrobacterium tumefaciens
strains containing each a different T-DNA, of which six transgenic lines (the
kd lines) were cotransformed with the T-DNAs from plasmids pAK1202 and
pAD1201 (De Neve et al., 1997), 14 lines (the kg lines) with those from
plasmids pAK1202 and pAG1201 (De Neve et al., 1997), and one (Hsb-K2L
67/1) with those from plasmids pHsb and pK2L610 (De Buck et al., 1999). A
second population included 12 CK2L-transgenic lines that were transformed
with the A. tumefaciens strain C58C1RifR (pGV2260 and pK2L610; De Buck et
al., 2000) and subsequently screened for single-copy inserts by T-DNA
fingerprinting (Theuns et al., 2002).
Amplification of T-DNA Junctions
All T-DNA junctions were amplified with T-DNA fingerprinting (Theuns
et al., 2002). From each transgenic line, a T-DNA fingerprint was generated
either with MseI- or with BfaI-digested DNA as template DNA. Subsequently, the sequence of these fragments was determined after amplification
upon elution (Windels et al., 2001). All transgenic lines were fingerprinted
by using a right and left border-specific anchor primer, situated 129 and 277
bp internally to the right and left border repeats, respectively.
deletion and the plant DNA adjacent to the opposite T-DNA junction. For
the filler in T-DNA junctions for which the corresponding, opposite T-DNA
junction had not been amplified, sequence identities with the target site
deletion or the plant DNA adjacent to the opposite T-DNA border could not
be distinguished. We chose to evaluate the statistical relevance of the
reported sequence identities by means of an experimental statistical test. To
perform a numeral, statistical analysis for each of the observed identical
sequence motifs is extremely difficult because the statistical relevance is
influenced by several parameters, such as length of the sequence identity,
distance between the reported identity and the originating sequence, number of identities observed per filler insertion, and length of the filler sequence itself. Therefore, we compared the actual, observed filler sequences
with the 200-bp plant target regions of other, characterized T-DNA inserts.
Because this analysis revealed that sequence identities of 8 bp and shorter
could occur simply by chance when a 200-bp plant region is screened (see
Supplemental Data), only sequence identities of at least 9 bp with the plant
target site were reported. Second, we compared the filler DNA with the first
100 bp of the T-DNA sequence that are situated immediately adjacent to the
filler DNA. To determine a threshold for statistically relevant sequence
identities when this 100-bp T-DNA border region was taken into account,
the actual filler sequences were shuffled and compared with a 100-bp
T-DNA border region. In that case, identities of at least 6 bp in length were
found to be relevant. To detect identical sequences in the filler with T-DNA
plasmid sequences that were not situated immediately adjacent to the filler
DNA, we performed a pair-wise comparison between the T-DNA plasmid
and filler sequences. Here, only sequence identities of at least 11 bp are
reported. The probability of finding an 11-bp repeat in a 12-kb sequence
(which is the average length of the T-DNA plasmids used in this study) is
less than 0.01.
Fisher Exact Test for Statistical Significance
To test whether the differences observed with regard to the degree of
T-DNA processing and the frequency of either filler DNA or microhomology at the T-DNA junction were significant, a Fisher exact test was performed. A two-by-two contingency table was constructed with the following values: A1, 3; A2, 30; B1, 8; and B2, 19 (see Supplemental Data). A Fisher
one-sided exact P value of 0.043 was obtained.
ACKNOWLEDGMENTS
The authors thank Isabel Roldán-Ruı́z, Rosalinde van Lipzig, and Sergei
Kushnir for critical reading of the manuscript and helpful comments, and
Martine De Cock for layout.
Received May 26, 2003; returned for revision July 20, 2003; accepted August
21, 2003.
Characterization of the T-DNA Junctions and
Filler Sequences
LITERATURE CITED
Pair-wise sequence alignment was used to screen for sequence identity
between the amplified junctions and the T-DNA plasmid. Sequence identity
between the plant region of the T-DNA junction fragment and the Arabidopsis genome was determined with the BLAST algorithm (Altschul et al.,
1990) against the Arabidopsis database (http://www.arabidopsis.org/
BLAST; Huala et al., 2001).
We first screened the complete Arabidopsis genome sequence with the
BLAST algorithm for the filler sequences larger than 17 bp. In this manner,
we could determine whether the observed filler DNAs were uninterrupted
sequence blocks that had been copied from the Arabidopsis genome. A
threshold was set at 17 bp because the probability of finding a 17-bp
identical sequence motif in a 125-Mb sequence, which is the length of the
complete Arabidopsis genome sequence (The Arabidopsis Genome Initiative, 2000), is less than 0.01.
Subsequently, for those junctions for which a filler DNA segment of 6 bp
or more was detected, their origin was analyzed as follows. First, we
screened the filler DNA at plant DNA/T-DNA junctions for sequence
motifs, identical either in direct or inverted orientation to sequence blocks
occurring in the 200-bp plant DNA region surrounding the T-DNA integration point. For those T-DNA inserts for which both T-DNA junctions were
amplified, the filler sequence was easily compared with the target site
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local
alignment search tool. J Mol Biol 215: 403–410
The Arabidopsis Genome Initiative (2000) Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408: 796–815
Brunaud V, Balzergue S, Dubreucq B, Aubourg S, Samson F, Chauvin S,
Bechtold N, Cruaud C, DeRose R, Pelletier G et al. (2002) T-DNA
integration into the Arabidopsis genome depends on sequences of preinsertion sites. EMBO Rep 3: 1152–1157
Bundock P, Hooykaas PJJ (1996) Integration of Agrobacterium tumefaciens
T-DNA in the Saccharomyces cerevisiae genome by illegitimate recombination. Proc Natl Acad Sci USA 93: 15272–15275
Daoudal-Cotterell S, Gallego ME, White CI (2002) The plant Rad50-Mre11
protein complex. FEBS Lett 516: 164–166
De Buck S, De Wilde C, Van Montagu M, Depicker A (2000) Determination of the T-DNA transfer and the T-DNA integration frequencies upon
cocultivation of Arabidopsis thaliana root explants. Mol Plant-Microbe
Interact 13: 658–665
De Buck S, Jacobs A, Van Montagu M, Depicker A (1999) The DNA
sequences of T-DNA junctions suggest that complex T-DNA loci are
formed by a recombination process resembling T-DNA integration. Plant
J 20: 295–304
Plant Physiol. Vol. 133, 2003
Downloaded from on June 16, 2017 - Published by www.plantphysiol.org
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
2067
Windels et al.
De Neve M, De Buck S, Jacobs A, Van Montagu M, Depicker A (1997)
T-DNA integration patterns in co-transformed plant cells suggest that
T-DNA repeats originate from ligation of separate T-DNAs. Plant J 11:
15–29
Fladung M (1999) Gene stability in transgenic aspen (Populus): I. Flanking
DNA sequences and T-DNA structure. Mol Gen Genet 260: 574–581
Friesner J, Britt AB (2003) Ku80- and DNA ligase IV-deficient plants are
sensitive to ionizing radiation and defective in T-DNA integration. Plant
J 34: 427–440
Gallego ME, Bleuyard J-Y, Daoudal-Cotterell S, Jallut N, White CI (2003)
Ku80 plays a role in non-homologous recombination but is not required
for T-DNA integration in Arabidopsis. Plant J 35: 557–565
Gallego ME, Jeanneau M, Granier F, Bouchez D, Bechtold N, White CI
(2001) Disruption of the Arabidopsis RAD50 gene leads to plant sterility
and MMS sensitivity. Plant J 25: 31–41
Gelvin SB (2003) Agrobacterium-mediated plant transformation: the biology
behind the “gene-jockeying” tool. Microbiol Mol Biol Rev 67: 16–37
Gheysen G, Angenon G, Van Montagu M (1998) Agrobacterium-mediated
plant transformation: a scientifically intriguing story with significant
applications. In K Lindsey, ed, Transgenic Plant Research, Harwood
Academic Publishers, Amsterdam, pp 1–33
Gheysen G, Villarroel R, Van Montagu M (1991) Illegitimate recombination in plants: a model for T-DNA integration. Genes Dev 5: 287–297
Gorbunova V, Levy AA (1997) Non-homologous DNA end joining in plant
cells is associated with deletions and filler DNA insertions. Nucleic Acids
Res 25: 4650–4657
Gorbunova V, Levy AA (1999) How plants make ends meet: DNA doublestrand break repair. Trends Plant Sci 4: 263–269
Hartung F, Puchta H (1999) Isolation of the complete cDNA of the Mre11
homologue in Arabidopsis (Accession No. AJ243822) indicates conservation of DNA recombination mechanisms between plants and other eukaryotes (PGR 99-132). Plant Physiol 121: 311
Huala E, Dickerman AW, Garcia-Hernandez M, Weems D, Reiser L,
LaFond F, Hanley D, Kiphart D, Zhuang M, Huang W et al. (2001) The
Arabidopsis Information Resource (TAIR): a comprehensive database
and web-based information retrieval, analysis, and visualization system
for a model plant. Nucleic Acids Res 29: 102–105
Kirik A, Salomon S, Puchta H (2000) Species-specific double-strand break
repair and genome evolution in plants. EMBO J 19: 5562–5566
Kumar S, Fladung M (2002) Transgene integration in aspen: structures of
integration sites and mechanism of T-DNA integration. Plant J 31:
543–551
Krysan PJ, Young JC, Jester PJ, Monson S, Copenhaver G, Preuss D,
Sussman MR (2002) Characterization of T-DNA insertion sites in Arabidopsis thaliana and the implications for saturation mutagenesis. OMICS 6:
163–174
Matsumoto S, Ito Y, Hosoi T, Takahashi Y, Machida Y (1990) Integration of
Agrobacterium T-DNA into a tobacco chromosome: possible involvement
of DNA homology between T-DNA and plant DNA. Mol Gen Genet 224:
309–316
Mayerhofer R, Koncz-Kalman Z, Nawrath C, Bakkeren G, Crameri A,
Angelis K, Redei GP, Schell J, Hohn B, Koncz C (1991) T-DNA integration: a mode of illegitimate recombination in plants. EMBO J 10: 697–704
Meza TJ, Stangeland B, Mercy IS, Skårn M, Nymoen DA, Berg A, Butenko
MA, Håkelien A-M, Haslekås C, Meza-Zepeda LA et al. (2002) Analyses
of single-copy Arabidopsis T-DNA transformed lines show that the presence of vector backbone sequences, short inverted repeats and DNA
methylation is not sufficient or necessary for the induction of transgene
silencing. Nucleic Acids Res 30: 4556–4566
2068
Orel N, Puchta H (2003) Differences in the processing of DNA ends in
Arabidopsis thaliana and tobacco: possible implications for genome evolution. Plant Mol Biol 51: 523–531
Pfeiffer P, Goedecke W, Obe G (2000) Mechanisms of DNA double-strand
break repair and their potential to induce chromosomal aberrations.
Mutagenesis 15: 289–302
Ralston EJ, English JJ, Dooner HK (1988) Sequence of three bronze alleles of
maize and correlation with the genetic fine structure. Genetics 119:
185–197
Rinehart TA, Dean C, Weill CF (1997) Comparative analysis of non-random
repair following Ac transposon excision in maize and Arabidopsis. Plant J
12: 1419–1427
Roth DB, Porter TN, Wilson JH (1985) Mechanisms of nonhomologous
recombination in mammalian cells. Mol Cell Biol 5: 2599–2607
Roth DB, Wilson JH (1986) Nonhomologous recombination in mammalian
cells: role for short sequence homologies in the joining reaction. Mol Cell
Biol 6: 4295–4304
Salomon S, Puchta H (1998) Capture of genomic and T-DNA sequences
during double-strand break repair in somatic plant cells. EMBO J 17:
6086–6095
Tamura K, Adachi Y, Chiba K, Aguchi K, Takahashi H (2002) Identification
of Ku70 and Ku80 homologues in Arabidopsis thaliana: evidence for a role
in the repair of DNA double-stranded breaks. Plant J 29: 771–781
Theuns I, Windels P, De Buck S, Depicker A, Van Bockstaele E, De Loose
M (2002) Identification and characterization of T-DNA inserts by T-DNA
fingerprinting. Euphytica 123: 75–84
Tinland B (1996) The integration of T-DNA into plant genomes. Trends
Plant Sci 1: 178–184
Tsukamoto Y, Ikeda H (1998) Double-strand break repair mediated by DNA
end-joining. Genes Cells 3: 135–144
van Attikum H, Bundock P, Hooykaas PJJ (2001) Non-homologous endjoining proteins are required for Agrobacterium T-DNA integration.
EMBO J 20: 6550–6558
van Attikum H, Bundock P, Overmeer RM, Lee L-Y, Gelvin SB, Hooykaas
PJJ (2003) The Arabidopsis AtLIG4 gene is required for the repair of DNA
damage, but not for the integration of Agrobacterium T-DNA. Nucleic
Acids Res 31: 4247–4255
Van Haaren MJJ, Sedee NJA, Krul M, Schilperoort RA, Hooykaas PJJ
(1988) Function of heterologous and pseudo border repeats in T region
transfer via the octopine virulence system of Agrobacterium tumefaciens.
Plant Mol Biol 11: 773–781
Wessler S, Tarpley A, Purugganan M, Spell M, Okagaki R (1990) Filler
DNA is associated with spontaneous deletions in maize. Proc Natl Acad
Sci USA 87: 8731–8735
West CE, Waterworth WM, Jiang Q, Bray CM (2000) Arabidopsis DNA
ligase IV is induced by ␥-irradiation and interacts with an Arabidopsis
homologue of the double strand break repair protein XRCC4. Plant J 24:
67–78
Windels P, Taverniers I, Depicker A, Van Bockstaele E, De Loose M (2001)
Characterisation of the Roundup Ready soybean insert. Eur Food Res
Technol 213: 107–112
Yan Y, Martinez-Ferez IM, Kavchok S, Dooner HK (1999) Origination of Ds
elements from Ac elements in maize: evidence for rare repair synthesis at
the site of Ac excision. Genetics 152: 1733–1740
Ziemienowicz A, Tinland B, Bryant J, Gloeckler V, Hohn B (2000) Plant
enzymes but not Agrobacterium VirD2 mediate T-DNA ligation in vitro.
Mol Cell Biol 20: 6317–6322
Downloaded from on June 16, 2017 - Published by www.plantphysiol.org
Copyright © 2003 American Society of Plant Biologists. All rights reserved.
Plant Physiol. Vol. 133, 2003