Transgene expression in rice is strongly dependent on the integrity and
organization of the transgenic locus. It is therefore a matter of ongoing
importance to study and elucidate the mechanisms of transgene integration,
and how this affects transgene integrity and organization, in order to
design more efficient transformation strategies. Many studies in this
area have focussed on transgene structure at the molecular level, but
we have found that the higher order structure of the transgene locus may
play an equally important role in transgene expression and stability.
This article focuses on higher order aspects of transgene organization,
reflecting the combined results of molecular, cytogenetic and genetic
analyses of transgene behavior in rice and wheat.
Using a combination of molecular analysis techniques (Southern blot hybridization,
PCR and DNA sequencing), fluorescence in situ hybridization to entire
chromosomes, and genetic segregation analysis of transgene-conferred phenotypes,
we have discovered that transgenes are organized in a three-tier hierarchy
(Fig. 1). At the first level, individual transgenes may join together
in contiguous transgene arrays (Kohli et al. 1999). These may be
direct repeats or a mixture of transgenes in different orientations, and
may contain from two to ten copies. In cotransformation strategies, the
arrays usually contain mixtures of different transgenes with no apparent
bias as to which are included, suggesting that inclusion is dependent
on non-specific free DNA ends rather than any particular sequence. Some
of the transgenes in each array may be truncated or otherwise rearranged.
The defining feature of the arrays is the absence of any genomic DNA between
the transgenes. Occasionally, there are short sequences of so-called 'filler
DNA' which are thought to represent the consequences of DNA repair synthesis
across gaps that remain after concatemerization (Gorbonova and Levy 1997,
Salamon and Puchta 1998), but the absence of significant stretches of
genomic DNA indicates that the arrays form prior to integration
(Kohli et al. 1999). We suggest that the nucleases and DNA repair
enzymes induced following damage to the plant cell by particle bombardment
both the degradation of exogenous DNA and its concatemerization,
in a parody of the processes that occur during the repair of damaged genomic
The second level of organization is called a transgene cluster (Kohli
et al. 1998). Here, individual transgenes and transgene arrays
integrate into the genome in close proximity. The individual copies and
arrays may be separated by stretches of genomic DNA, but these range from
1-10 kbp and can be isolated by standard PCR or long PCR strategies in
which primers face outwards from the extremities of the integrated transgenes.
The existence of such intervening genomic regions can be demonstrated
on Southern blots, when the genomic DNA has been cut with restriction
enzymes that do not cut within the transgene. In the absence of
intervening genomic DNA, all the transgenes would resolve to a single
restriction fragment defined by the nearest restriction sites for the
'noncutter' enzyme on either side of the transgenic locus. Instead, it
is often the case that two or more hybridizing bands are revealed when
such noncutters are used, allowing regions of intervening genomic DNA
to be mapped. The existence of genomic DNA between individual transgenes
and arrays indicates that this level of organization arises during
or following integration (Kohli et al. 1998). Why should several
transgene arrays integrate in the same region? The analysis of transgene
integration sites at the genomic level suggests integration occurs randomly,
perhaps at naturally-occurring DNA breaks. It would therefore seem more
logical for the genome to become peppered with dispersed transgene arrays
following transformation. The reason for clustering may reflect a two-phase
integration mechanism, in which the initial integration event acts as
a nucleation point and stimulates further integration events nearby
(Kohli et al. 1999). Local phenomena of this nature have been reported
previously, such as the tendency for excised transposable elements to
reintegrate close to their site of origin, and this may involve the activity
of proteins that move along the DNA duplex (Moreno et al. 1992).
In the case of transgene integration, a simple explanation is the recruitment
of DNA repair complexes to sites of DNA damage, and their tendency to
introduce nicks into the DNA as they slide along it. Alternatively, the
exogenous DNA may interact with a group of local replication forks. In
each case, this would result in a dense cluster of free genomic DNA ends,
combined with an excess of free exogenous DNA. There would also be a tendency
for genomic DNA in this region to be eliminated or inverted as part of
the repair process and such deletions and rearrangements have indeed been
observed with great frequency at transgenic loci in rice.
FISH analysis revealed an unexpected third level of transgene organization
(Abranches et al. in press). The analysis was carried out as part
of a routine study to map the distribution of transgene integration events
in the genome. Wheat was used for this study because the structure of
wheat chromosomes is well characterized and many chromosome-specific probes
are available to enable unambiguous identification of integration sites.
However, analogous experiments are in progress using rice, and preliminary
data suggest that the results will be similar. We have found that, in
almost all cases, the transgenic phenotype segregates as a single locus.
We therefore expected a single FISH signal on the appropriate chromosome,
but instead, we observed multiple separable FISH signals clustered in
a particular chromosome region (Abranches et al. in press). Single
transgene copies are below the resolution of the FISH procedure in cereals,
so the individual signals must represent transgene arrays/clusters. At
cytogenetic resolution, the gaps between the signals correspond to genomic
DNA in the order of hundreds of kilobase pairs. A different mechanism
must lie behind this third level of organization, and an intriguing result
of our further FISH analysis has suggested what that mechanism may be.
Although the fluorescent signals are separable when FISH is carried out
on metaphase chromosomes, if the same technique is carried out at interphase,
the signals resolve to a single spot. The ramification of this result
is that the signals distributed along the metaphase chromosome occupy
the same location in the interphase chromatin. We propose several models
to explain these results (Abranches et al. in press). First, homologous
transgenes may be associating in trans during interphase. This
has been put forward as a possible silencing mechanism in dicots, and
we are currently investigating the possibility of links between transgene
silencing and the coincidence of FISH signals, which would provide physical
evidence for this silencing process. Second, transgenes with the same
promoter may be recruited to a common transcription factory in the nucleus.
An important question concerning both these theories is whether such interactions
would only occur among transgenes linked in the same chromosome region,
or would also occur between, for example, transgenic loci on different
chromosomes. Experiments are in progress to address these issues. Finally,
it is possible that the distribution of FISH signals at metaphase may
simply reflect differences between metaphase and interphase DNA organization.
It can be presumed that most transformation events occur during interphase,
when the chromatin is arranged in a specific three-dimensional configuration
in the nucleus. Following particle bombardment, the metal particle lodges
in the nucleus providing a local source of DNA. Integration events may
therefore be favored in the genomic DNA immediately adjacent to the particle,
and such DNA may even be damaged when the particle arrives. However, there
is no reason to believe that DNA sites close together in trans
during interphase should be adjacent in the metaphase chromosome. Indeed,
the arrangement of interphase chromatin into loops attached to the nuclear
matrix may bring together sites which, when the DNA is stretched out,
lie hundreds of kilobase pairs apart (Fig. 1).
These exciting discoveries suggest that the integrity and hierarchical
organization of a transgenic locus may arise from a combination of factors,
ranging from the molecular mechanisms of DNA joining during transgene
integration to the three-dimensional organization of chromatin in the
interphase nucleus. Improved vector design is likely to influence only
the primary level of transgene organization, thus it is of paramount importance
to investigate factors influencing higher order structure in order to
optimize transgene expression.
Abranches, R., A.P. Santos, S. Williams, A. Castilho, P. Christou, P.
Shaw, and E. Stoger. Multiple transgene integration sites in wheat metaphase
chromosomes are brought together at interphase. Plant J. (in press).
Gorbunova, V., and A.A. Levy, 1997. Non-homologous DNA end joining in
plant cells is associated with deletions and filler DNA insertions. Nucleic
Acids Res. 25: 4650-4657.
Kohli A.,S. Griffiths, N. Palacios, R.M. Twyman, P Vain, D.A. Laurie,
and P. Christou, 1999. Molecular characterization of transforming plasmid
rearrangements in transgenic rice reveals a recombination hotspot in the
CaMV 35S promoter and confirms the predominance of microhomology-mediated
recombination. Plant J. 17: 591-601.
Kohli, A., M. Leech, P. Vain, D.A. Laurie, and P. Christou, 1998. Transgene
organization in rice engineered through direct DNA transfer supports a
two-phase integration mechanism mediated by the establishment of integration
hot-spots. Proc. Natl. Acad. Sci. USA 95: 7203-7208.
Moreno M.A., J. Chen, I. Greenblatt, and S.L. Dellaporta, 1992. Reconstitutional
mutagenesis of the maize P gene by short-range Ac transpositions.
Genetics 131: 939-956.
Salomon, S., and H. Puchta, 1998. Capture of genomic and T-DNA sequences
during double-strand break repair in somatic plant cells. EMBO J. 17: