Download Structural Domains and Matrix Attachment Regions

The Plant Cell, Vol. 12, 249–264, February 2000, www.plantcell.org © 2000 American Society of Plant Physiologists
Structural Domains and Matrix Attachment Regions along
Colinear Chromosomal Segments of Maize and Sorghum
Alexander P. Tikhonov, Jeffrey L. Bennetzen, and Zoya V. Avramova1
Department of Biological Sciences, Purdue University, West Lafayette, Indiana 47907
Although a gene’s location can greatly influence its expression, genome sequencing has shown that orthologous genes
may exist in very different environments in the genomes of closely related species. Four genes in the maize alcohol dehydrogenase (adh1) region represent solitary genes dispersed among large repetitive blocks, whereas the orthologous
genes in sorghum are located in a different setting surrounded by low-copy-number DNAs. A specific class of DNA sequences, matrix attachment regions (MARs), was found to be in comparable positions in the two species, often flanking
individual genes. If these MARs define structural domains, then the orthologous genes in maize and sorghum should
experience similar chromatin environments. In addition, MARs were divided into two groups, based on the competitive
affinity of their association with the matrix. The “durable” MARs retained matrix associations at the highest concentrations of competitor DNA. Most of the durable MARs mapped outside genes, defining the borders of putative chromatin
loops. The “unstable” MARs lost their association with the matrix under similar competitor conditions and mapped
mainly within introns. These results suggest that MARs possess both domain-defining and regulatory roles. Miniature
inverted repeat transposable elements (MITEs) often were found on the same fragments as the MARs. Our studies
showed that many MITEs can bind to isolated nuclear matrices, suggesting that MITEs may function as MARs in vivo.
INTRODUCTION
Our recent results have illustrated that genes residing in the
maize alcohol dehydrogenase (adh1) region display two distribution patterns: individual genes amid a sea of repetitive
DNA and clustered genes occupying a space uninterrupted
by highly repetitive DNAs (Tikhonov et al., 1999). Sorghum, a
close relative of maize with an ⵑ3.5-fold smaller genome
(Laurie and Bennett, 1985), has orthologous genes located
in an environment that lacks highly repeated DNAs
(Tikhonov et al., 1999). Similarly, the maize shrunken2/a1
genes are located ⵑ140 kb apart (Civardi et al., 1994), separated by blocks of repeats of retroelements (P. SanMiguel,
unpublished observations), whereas no large repeats were
found in the orthologous regions of sorghum and rice (Chen
and Bennetzen, 1996; Chen et al., 1997). Given the established importance of the genomic context for the function of
a gene, best illustrated by the phenomenon known as position-effect variegation, these findings point to an apparent
paradox: orthologous genes in related species may show
similar functions in very different chromosomal settings.
However, all results describing the effects of the surrounding DNA on a gene’s behavior have been obtained either in transgenic systems or with translocation events in
which tested genes are placed in an environment that is dif-
1 To
whom correspondence should be addressed. E-mail zavramov
@bilbo.bio.purdue.edu; fax 765-496-1496.
ferent from their wild-type location. Therefore, the chromosomal context and its putative effect on a gene’s function in
its natural environment have not been assessed. The existence of functional genes among blocks of retroelement repeats indicates that natural systems must have evolved
mechanisms and strategies that would allow orthologous
genes to function at their respective, evolutionarily established locations, despite the very different natures of the
neighboring sequences.
A current model accounting for this apparent paradox
suggests that genes and blocks of repetitive DNAs might
exist in different, structurally separated nuclear compartments (Henikoff and Comai, 1998; Lamond and Earnshaw,
1998; Cockell and Gasser, 1999) and that mislocation alters
expression. Each gene in the nucleus may have only one
“address” at which it functions correctly, and during evolution,
natural genes may have acquired “anchors” to stably position
them in the spatial architecture of the nucleus (Flavell, 1994).
A specific class of DNA sequences, matrix attachment regions (MARs), may be involved in this anchoring function.
MARs are operationally defined as DNA sequences that
preferentially bind to the proteins of the nuclear matrix, potentially defining and delimiting individual structural units of
chromatin (reviewed in Bode et al., 1996). One step in testing
such a model therefore would be to screen for the distribution
of MARs along large chromosomal continuums. MAR location would suggest a pattern for a putative folding of the
250
The Plant Cell
region into individual structural domains. Defining how biologically active units are positioned within these domains
would be an obvious next step.
In an earlier study, we reported that the long stretches of
mixed classes of repetitive DNAs in the maize adh1 region were
segregated into topologically sequestered units (Avramova
et al., 1995). The lack of DNA sequence data, however, precluded more detailed comparisons between the possible
structural organization of the region and gene/genome function. Subsequent studies of the region (Avramova et al.,
1996; SanMiguel et al., 1996; Tikhonov et al., 1999) provided
a molecular basis for renewing our pursuit of a possible relationship between gene structure and function. Although some
limitations of our analysis and screening approach, at that
time, have been acknowledged and discussed (Avramova et
al., 1995), most of them have now been overcome, as described below.
Recently, we were able to demonstrate that microcolinearity of gene composition between maize and rice was paralleled by genomic folding into colinear structural units.
More importantly, the pattern of potential structural organization was followed by similar placement of the genes within
the putative loops, suggesting that the positioning of the
genes in their respective structural domains might be conserved in evolution (Avramova et al., 1998). However, the
colinear shrunken2/a1–homologous regions of both rice and
sorghum lack large amounts of repetitive DNA and hence
are in gene-rich environments (one gene per 8 to 10 kb). The
colinear maize and sorghum adh regions differ greatly in the
amount of intervening repetitive DNAs and resulting gene
richness (one gene per 25 kb in maize versus one gene per 6
kb in sorghum; Tikhonov et al., 1999).
Here, we studied a possible structural folding of large
colinear segments carrying orthologous genes in different
genomic environments by using the distribution of MARs
along the regions as genomic elements that may define
structural units. The placement of candidate genes within
the putative structural domains supports a model in which
plant genes have evolved within structural domains separated from neighboring nongenic or genic sequences.
RESULTS
Comprehensive Analysis of MARs in the Maize
adh1 Region
The screening for MARs was conducted by using an approach that tests the capacity of isolated nuclear matrices to
interact with end-labeled tested DNA fragments in the presence of excess unlabeled competitor DNA. In our previous
study, due to the absence of sequence information, MARs
were positioned on the basis of the available map created
by a set of three restriction endonucleases (Springer et al.,
1994). As a result, MARs could have been destroyed by the
digestion (Avramova et al., 1995) or missed because of too
short or too long fragment lengths (Izaurralde et al., 1988;
Mielke et al., 1990). Recent data on the sequence composition of this region (Tikhonov et al., 1999) allowed MAR
screening after digestions with sets of carefully chosen restriction endonucleases (see Methods). This approach, however, generated a very large number of fragments, and we
have chosen to reexamine only the low-copy-number regions of the maize contig under new binding conditions.
Multiple overlapping fragments from the low-copy-number regions were generated and studied. The shortest fragments still displaying binding activities were defined as
MARs. This approach made it possible to identify individual
attachment points within larger matrix binding regions. Each
identified MAR was characterized under several different
competitive concentrations and on several overlapping fragments. These multiple reexaminations allowed independent
confirmation of the presence of a MAR and its characteristics. Moreover, these experiments also permitted a more
precise definition of MAR location and complexity. Knowing
that the largest low-copy-number domain from the maize
adh1 contig was segregated in a putative loop of ⵑ45 kb
(Avramova et al., 1995), we wished to test the possible structural organization within this domain, particularly with regard
to several candidate genes (Tikhonov et al., 1999). A goal of
this study was to compare the possible chromatin organization of orthologous genes in colinear genomic regions of two
species differing in the amount and nature of intervening
DNAs. Hence, we have focused on the possible structural
organization of the putative gene-containing regions.
In our earlier study, MAR screening was conducted under
stringent binding conditions (Avramova et al., 1995). High
concentrations of competitor DNA and nonsaturating quantities of matrices in the assays allowed only fragments with
the highest matrix affinity to be identified. Here, binding assays were performed with saturating amounts of matrix proteins. These have been established empirically by testing
increasing amounts of isolated matrices, until a point was
reached at which all of a labeled MAR fragment was matrix
bound. The concentration of matrix proteins at which all of
the labeled MAR fragments were recovered from the pellet,
with none being found in the supernatant, was defined as
saturating. It corresponded, approximately, to the residual
protein obtained after high-salt extraction of 1 A260 OD of
starting nuclei (see Methods). A higher concentration of matrix proteins provides more binding sites in the system, allowing for the discovery of weaker but specific MARcontaining fragments (Mielke et al., 1990).
Matrices isolated from the nuclei of sorghum seedlings
were used. Matrix proteins are evolutionarily conserved
such that they recognize and bind MARs across species in
test assays conducted in vitro (Cockerill and Garrard,
1986b; Avramova and Bennetzen, 1993). Here, matrix preparations from sorghum seedlings were found to bind maize
MAR-containing fragments with the same specificity shown
earlier for matrices isolated from maize leaves and roots
MARs and Putative Genomic Structures
(Figures 1A and 1B). Also, a MAR flanking a tobacco root–
specific gene, RB7 (Hall et al., 1991), bound with very high
specificity to the matrix isolated from nuclei of sorghum
seedlings (Figure 1C), supporting the idea that matrices from
one species can be used to discover MARs in the genome
of another.
Finally, three different concentrations of competitor DNA
were tested. The results allowed us to evaluate both the relative binding strength of individual MARs and to follow the
profile of their binding to the matrix under competition (Figures 2 and 3). Two classes of MARs have been tentatively
discriminated: “durable” MARs and “unstable” MARs.
Two Classes of MARs
Under the binding conditions described, many fragments
from the tested clones bound specifically to the nuclear matrices. This created ambiguities for defining possible loops.
251
We looked for criteria that would allow us to distinguish between different types of matrix binding DNAs. Usually,
MARs are arbitrarily defined as strong and weak, based on
their different affinities for the nuclear matrix in the presence
of an excess of nonspecific DNA.
Following this criterion, we quantified the amount of remaining matrix-bound DNA after competition with various
concentrations of nonspecific DNA. On the basis of the density-profile peaks of a labeled fragment recovered from the
pellet (MAR-containing fragment), its amount was calculated
as a percentage of the density of each fragment present in
the input fraction, taken as 100% (Figure 2; see Methods).
Restriction fragments recovered from the pellet were quantitatively analyzed for their matrix binding affinity in the
presence of 100 to 400 ␮g/mL competitor DNA. These correspond to 5000- to 20,000-fold excesses of Escherichia
coli DNA over the substrate genomic DNA. Approximately
270 DNA fragments were analyzed and 36 MARs were identified along the two contigs.
Figure 1. In Vitro Matrix Binding Assays near the Orthologous adh Genes.
(A) The sorghum adh-containing clone p13MC (a 13-kb MluI-ClaI fragment in the pGEM7⫹ vector) was digested with two combinations of enzymes: MluI, SacII, Bsu36I, SnaBI, MfeI, and NcoI (left and middle sections) or MluI, SacII, ApaLI, SmaI, MfeI, SpeI, and XbaI (right section). The
amount of competitor DNA is shown above each lane in ␮g/mL.
(B) Maize subclones p5zMAR and p3zMAR from the 5⬘ and 3⬘ regions near adh1-F, respectively. p5zMAR (left two lanes) was digested with XbaI
and XhoI, whereas p3zMAR (right two lanes) was digested with BamHI, BglII, and EcoRI. The amount of competitor DNA is shown on top of each
lane in ␮g/mL.
(C) The tobacco RB7 MAR (Hall et al., 1991) as a heterologous control. Clone pGHNC5, containing the matrix attachment fragment as a dimer,
was digested with NotI and Bsp120I. The amount of competitor DNA is shown on top of each lane in ␮g/mL.
In, the input DNA before binding (one-twentieth of the amount used for binding was loaded); M, a 1-kb ladder DNA marker; Pellet, the DNA
bound to the matrix proteins; Supernatant, the unbound DNA (one-tenth of the supernatant was loaded); v, the fragments of the cloning vector
used as a negative control; v⫹, vector fragment with an attached cloned fragment. Black arrows indicate matrix-bound fragments; open-headed
arrows indicate fragments resulting from partial star activity.
252
The Plant Cell
The matrix binding curves drop steeply under increasing
concentrations of competitor DNA (Figure 3B). Although
some of the MARs in this class could qualify as strong at
lower excesses of E. coli DNA (e.g., Figure 3B, curves 1 to 3),
their binding to the matrix was blocked by higher concentrations of competitor DNA. These less competitive, unstable
MARs mapped to regions inside genes and, predominantly,
within putative introns (Tikhonov et al., 1999).
A notable exception is the dissociation curve of a MAR located internal to maize gene 334B7.4; this MAR displays a
durable profile (Figure 3C). It is discussed in more detail later
in the article.
Orthologous adh Genes of Maize and Sorghum and
Their Structural Domains
Figure 2. Matrix Binding Analysis.
(A) X-ray film image of a representative matrix binding assay. The input lane shows a complete set of fragments, from the maize clone
pL347, tested for binding activity. The pellet lanes show the DNA
fragments recovered from the pellet, that is, bound to the matrix, under competition with 100, 200, or 400 ␮g/mL of E. coli DNA. v, mobility of vector fragments; v⫹, vector fragments with adjacent
genomic sequences.
(B) Density-profile plots of the gel in (A) representing the average
gray value of the pixels across the gel lane. The scale is different for
each plot.
After plotting the amount of recovered matrix-bound fraction over the concentration of the competitor DNA, we were
able to follow the changes in binding of the different MARs.
Analyzing the profiles of the curves, we observed that MARs
could be divided into two classes. A common feature of the
first class, as illustrated by the representative profiles of several MARs (Figure 3A), is that these MARs continue to bind
the matrix in the presence of the highest tested concentrations of competitor DNA. Some of these MARs, if considered at only one competitor DNA concentration (even the
lowest, i.e., a 5000-fold excess), could have been classified
as weak and possibly dismissed from further analysis. However, these MARs still could bind the matrix (15 to 30% recovered from the pellet) under higher (up to 20,000-fold)
excesses of competitor DNA (Figure 3A, curve 5). This class
represents the durable MARs, and 16 MAR fragments were
found to belong to this category. All but one of them
mapped in the space outside genes.
The MARs in the second class exhibit different profiles.
The maize adh1 gene is an example of a lone gene amid
large blocks of highly repetitive elements. Previous studies
have localized a MAR 5 ⬘ to the gene (Avramova and
Bennetzen, 1993; Paul and Ferl, 1993) and two MARs at its
3⬘ end (Avramova et al., 1995). The precise location of the
weaker MAR closer to the 3 ⬘ end, however, has remained
unclear. Thus, the size of the putative adh1-containing loop
could vary from 3 to 9 kb, depending on the position of the
MAR inside a 7-kb restriction fragment. Also, a small MARcontaining fragment (⬍500 bp) from the region was not positioned on the previously available map (Avramova et al.,
1995).
Here, the 5⬘ and 3⬘ regions flanking maize adh1 have been
examined in detail after being subcloned and digested with
a combination of restriction endonucleases. The recovery of
the tested fragments from the pellet (MARs) and the supernatant (nonbound DNA) is shown (Figure 1B). The gradual
increase in the percentage of the fragment in the supernatant, as a result of increasing concentration of competitor, illustrates the different affinities for the matrix proteins
displayed by the various MAR-containing and non-MAR
fragments.
Analysis of the matrix-bound fragments indicated that
they were adjacent on the contig, probably constituting multiple attachment points. A model for the structure of a putative (ⵑ4 kb) adh1 domain, bordered by the two flanking
MARs, is shown (Figure 4). Thus, both MARs surrounding
the adh1 gene appear to be composite and durable (Figure
3A, curves 1 and 5) and to be tightly bordering the long terminal repeats of the flanking retroelements ( Cinful-2 and
Milt). These complex MARs are positioned to precisely separate the gene from the neighboring 70-kb (5 ⬘) and 14-kb
(3⬘) blocks of highly repetitive DNAs.
It was of interest to determine the structure of the region
around the orthologous sorghum adh gene because the regions occupied by MARs in maize did not exhibit obvious
sequence homology with sorghum and because there are
no large repetitive DNA blocks around the sorghum adh
gene (Tikhonov et al., 1999). A 13-kb region containing the
MARs and Putative Genomic Structures
253
sorghum adh gene was subcloned and digested with two
different sets of restriction endonucleases, and the overlapping fragments were tested for binding activity. The results
from the binding of two different sets of fragments are
shown (Figure 1A). Both 5 ⬘ and 3⬘ MARs flanking the sorghum adh1 homolog were found to be durable (data not
shown).
Analysis of the matrix-bound fragments suggested a potential structural organization of the sorghum adh gene–containing domain (Figure 4). The similarity in the sizes of the
putative loops, defined by the positioning of the matrix binding regions, is quite remarkable, given the absence of MAR
sequence homology. In the different chromosomal contexts
in which the orthologous genes exist in the two species, the
preservation of the structural organization of the gene may
be relevant to its expression.
Putative Genes and Putative Domains along the
Maize Contig
Figure 3. Competition Profiles for Various MARs.
(A) Matrix associations of durable MAR fragments under three different concentrations of E. coli competitor DNA: 100, 200, 300, or 400
␮g/mL. Representative graphs are shown for MARs from both maize
and sorghum: curve 1, a 1.2-kb MAR fragment 5⬘ of the maize adh1
gene; curve 2, a 245-bp MAR fragment between the sorghum
110K5.10 and 110K5.11 genes; curve 3, a 2.7-kb MAR fragment between the maize 334B7.7 and 334B7.8 genes; curve 4, a 160-bp MAR
between the sorghum 110K5.8 and 110K5.9 genes; curve 5, a 2.5-kb
MAR 3⬘ of the adh1 gene; and curve 6v, a control vector fragment.
(B) Matrix associations of intronic MARs under similar binding conditions. Curves 1 and 2 are 2.3- and 2.2-kb MAR fragments, respectively, located in introns of the sorghum 110K5.4 gene; curve 3, a
440-bp MAR fragment inside the sorghum 110K5.11 gene; curve 4,
a 2.3-kb MAR fragment in the last intron of the maize 334B7.7 gene;
and curve 5v, a control vector fragment.
The possible chromatin-loop arrangements of the gene candidates recently identified on the maize contig (Tikhonov et
al., 1999) were examined. Four genes, 334B7.2, 334B7.3,
334B7.4, and 334B7.9, are individual loci, surrounded by
blocks of retroelement DNAs. Screening for MARs has identified MARs around the genes, suggesting structural organization similar to the 334B7.3 ( adh1) gene. Putative gene
334B7.2, discovered at one end of the contig, is separated
from its nearest neighbor, adh1, by ⵑ70 kb of retroelements
(Figure 5). A MAR marks one end of its putative structural
loop, whereas the other end was not defined because it was
outside the contig. Downstream of the adh1-containing
loop, after an ⵑ14-kb block of highly repeated retroelements, another solitary putative gene, 334B7.4, was discovered amid repetitive DNAs (Tikhonov et al., 1999). In this
region, both a newly found and a previously discovered
MAR (fragment 47 in Avramova et al., 1995) mapped inside
the gene (Figure 5). One of the internal MARs, overlapping
putative exon sequences of the gene, displayed a matrixaffinity profile relating it to the durable, possibly loop-defining MARs (Figure 3C). The curve of the other was somewhat
intermediate (see Discussion). The closest MARs located
outside the 334B7.4 gene sequence were, on the one side,
the MAR from the 3⬘ end region of adh1 and, on the other,
two MARs on adjacent fragments (see below). These would
define a putative loop of ⵑ60 kb. This predicted loop is unusual because it is composed of various highly repetitive
(C) Matrix affinity of a MAR located inside the maize 334B7.4 gene.
Curve 1, a 1.4-kb MAR fragment found inside a putative exon sequence (this fragment also contains a MITE); and curve 2v, a control
vector fragment.
254
The Plant Cell
Figure 4. Comparison of the Orthologous adh Domains in Maize
and Sorghum.
The maize yeast artificial chromosome (YAC) bar (top) represents a
225-kb region of maize chromosome 1 containing adh1. The empty
numbered boxes correspond to the nested retrotransposon blocks.
The open boxes with the arrows inside and numbers on top indicate
gene-containing regions. The arrows designate putative directions
of transcription of identified candidate genes. The sorghum bacterial
artificial chromosome (BAC) bar (bottom) represents the orthologous
78-kb region of sorghum. The figures in the middle illustrate the putative adh1 chromatin domains. The potential chromatin loops, defined by MARs (empty boxes), are shown as dotted loops. The
composite structures of the MARs are shown by open boxes, indicating individual matrix binding fragments. The asterisks designate
identified MITEs in the MAR-containing regions. The boxes containing triangles, flanking the maize adh1 gene domain, correspond to
the long terminal repeats (LTR) of the neighboring Cinful-2 and Milt
retrotransposons.
retrotransposons and a gene. The maize 334B7.4 gene had
been considered, hitherto, as an example of a plant gene
“broken” by the insertion of three large blocks of highly repetitive elements, scattering the genic sequence over 42 kb
of genomic space (Tikhonov et al., 1999). Here, we show
that 334B7.4 may have an unusual domain organization as
well.
The last gene from the maize contig surrounded by repetitive DNAs is putative gene 334B7.9. However, its apparent
loop also includes remnants of two members of retroelement families, Cinful-1 and Kake-1 (Figure 5). The possible
loop-defining MARs of this (ⵑ10 kb) domain have been established previously (Avramova et al., 1995) and have not
been retested in this study.
The remaining four genes, 334B7.5, 334B7.6, 334B7.7,
and 334B7.8, are organized in a cluster, uninterrupted by repetitive retroelement DNA (Tikhonov et al., 1999). Multiple
overlapping fragments from the region were generated and
tested for matrix binding. The pattern of MAR distribution
was rather complex, and the durable MARs were considered
domain defining.
The first gene of the cluster, 334B7.5, is positioned in a
putative 5-kb loop. The upstream binding region seems to
contain at least two attachment points that, together, sepa-
rate the low-copy-number region from the highly repetitive
DNA block upstream. Each of these two attachment fragments behaved as weak MARs, but, nonetheless, they
bound the matrix under the stringent assay conditions of the
earlier study (Avramova et al., 1995) and displayed durable
MAR profiles under the highest competition conditions of
this assay (not shown).
The MAR closing the loop containing the 334B7.5 gene
and defining the beginning of the adjacent loop was found
positioned very near the 3⬘ end of the neighboring 334B7.6
gene. The other end was located ⵑ20 kb downstream,
where a durable MAR was found between the 334B7.7 and
334B7.8 genes (Figure 3A, curve 3). An unusual feature of
this putative structural loop is that it encompasses two
genes, because no MAR was found in the space separating
the 334B7.6 and 334B7.7 loci. These two putative genes,
apparently transcribed in opposite directions, were found as
an unusual example of closely positioned maize genes that
are not gene duplication products (Tikhonov et al., 1999).
They are separated by only 200 bp of DNA and apparently
share the same structural domain. Gene 334B7.7 contains
several unusually large introns and two internal MARs (Figure 3B, curve 4). The last gene from the cluster, 334B7.8, is
positioned in a putative loop defined by two durable MARs.
An unstable MAR was found inside a putative intron (data
not shown).
Putative Folding of the Colinear Region of Sorghum
As a result of the absence of retrotransposons from the homologous region of sorghum, the genes are more densely
arranged, with an average density of one gene per 6 kb. An
unexpected observation was that three putative genes, in
the otherwise colinear sorghum region, were missing in
maize (Tikhonov et al., 1999). The differences in the genomic
structures, at a sequence level, prompted us to compare the
possible domain-folding patterns of the two regions. The
distribution of the MARs along the 78-kb sorghum continuum colinear with the 225-kb maize contig was examined.
Each of seven subclones was digested with combinations of
restriction nucleases (see Methods), and overlapping fragments were tested for binding capacity under the same conditions as described for maize (Figure 6).
The distribution of the intergenic MARs along the sorghum contig outlines a pattern, placing most of the genes in
apparently well-defined putative loops. Notable exceptions
are two genes, 110K5.9 and 110K5.10, that appear to share
the same loop. The ends of this putative structural domain
are defined by durable MARs positioned between 110K5.8
and 110K5.9 (Figure 3A, curve 4) and between 110K5.10
and 110K5.11 at the other end (Figure 3A, curve 2). The region separating putative gene 110K5.9 from 110K5.10 displayed no matrix binding potential. Putative intronic MARs
were identified in both 110K5.10 (not shown) and 110K5.11
(Figure 3B, curve 3).
MARs and Putative Genomic Structures
255
Figure 5. MARs and Predicted Loops in the Maize adh1 Region.
Maize YAC 334B7 is represented by a bar with open boxes (genic DNA). Putative exons are shown as vertical boxes. Diamonds and filled arrowheads designate MITES and long interspersed nuclear element–like retroelements, respectively. The thin lines correspond to regions occupied
by repetitive retrotransposon DNAs. Putative chromatin loops are shown by curves above the boxes. The anchors and the asterisks indicate the
positions of domain-defining and intragenic MARs, respectively. The boxes below the YAC designate the restriction enzyme fragments tested in
the binding assays: multiple overlapping fragments were tested, and the MAR signs were placed above the shortest restriction fragments recovered from the matrix-bound fraction. The scale is in kilobases. The numbers above the boxes correspond to the putative genes as defined in
Tikhonov et al. (1999) and as referred to in the text.
MARs and Miniature Inverted Repeat
Transposable Elements
Based on the molecular analysis of two different orthologous chromosome segments, the shrunken2/a1 regions of
rice and sorghum and the adh1 regions of maize and sorghum, ⵑ80 miniature inverted repeat transposable elements
(MITEs; reviewed in Wessler et al., 1995) were identified
along the four genomic continuums (Chen and Bennetzen,
1996; Chen et al., 1997; Tikhonov et al., 1999). Screening
the regions for MARs revealed ⵑ40 restriction fragments
with matrix affinity that also contained a MITE, raising the
question of whether MITEs carry the MAR function and/or
whether MITEs have inserted preferentially around MARs.
To study this question, we tested the capacity of isolated
MITEs to bind to the matrix. Approximately 20 elements
from the four genomic regions were examined under the
binding conditions described in this study. The MITEs, in all
tested cases, displayed matrix binding activity (Figure 7).
The MITEs were obtained either by excision with restriction
endonucleases or by polymerase chain reaction (PCR) amplification, using devised primers (shown in Methods).
In approximately a dozen cases, the presence of convenient restriction sites allowed the isolation of MITEs as separate recognizable fragments from the surrounding genomic
sequences. The binding profiles of a few examples are
shown in Figures 7A to 7E.
Alternative digestions with restriction nucleases of the region containing two nested Tourist elements 3⬘ to the sorghum 110K5.3 gene (Figure 6) released either the whole set
(Figure 7A) or an individual element (Figure 7B; for detailed
nucleotide information on the structure and location of the
elements and the position of the restriction sites, see the
legend to Figure 7 as well as GenBank accession number
AF124945 and Tikhonov et al. [1999]). The matrix binding
tests showed that both nested and individual Tourist elements can bind the matrix (Figures 7A and 7B). In accordance, an individual Tourist near the 5⬘ end of the 110K5.5
gene from the same contig (Figure 6) binds also to the matrix (Figure 7C, marked by the arrow). The higher molecular
weight bands belong to adjacent genomic sequences not
discussed here.
The matrix binding of two MITEs of maize origin are
shown as well: the MITE from the putative exon of maize gene
334B7.4 (Figure 7D) and the Tourist-Zm1 element (Bureau
and Wessler, 1992; Figure 7E). The latter element is very
short, ⵑ130 bp, which may be a reason for its apparent
weaker binding. It is possible that it is a component of a
larger binding region at its genomic location.
Conveniently located restriction sites, however, often
would leave flanking genomic sequences or eliminate sequences from the tested elements, which could impact the
MAR activity of the tested fragments. To avoid this possibility, we examined several MITEs after PCR amplification of
their sequences. The 894-bp fragment containing three complex MITEs, between maize genes 334B7.7 and 334B7.8
(Figure 5), has been PCR amplified using primers tightly
flanking the whole set. The primers and their genomic positions are shown in Methods and provided under GenBank
accession number AF123535. The matrix binding capacity
of the whole fragment is shown in Figure 7G (arrow 1).
The potential involvement of individual elements in the
256
The Plant Cell
Figure 6. MARs and Predicted Loops in the Sorghum adh Region.
A region of 78 kb from sorghum BAC 110K5 is shown as a bar containing the exon–intron structures of candidate genes. All designations, including MITEs, long interspersed nuclear element–like elements, intragenic MARs, MARs outside genes, and putative chromatin loops, are as
given in Figure 5. The boxes below the graph designate the tested restriction fragments. The seven BAC subclones (p11CM, p13MC, p5.4, p17,
p2, p16, and p18) are shown; their cloning sites are marked by vertical dotted lines.
matrix binding interaction is illustrated for two adjacent
Tourist elements revealed in the sorghum shrunken2/a1
region, close to the 3⬘ end of the shrunken2 gene homolog
(Chen et al., 1997; GenBank accession number AF010283).
Their structure, the location of the various primers used, and
the amplified regions are illustrated in Figures 8A and 8B.
The binding capacity of the individually amplified fragments
are shown in Figures 7F and 7G. Bands 2 and 4, corresponding to the individual MITEs, can bind the matrix; an
amplified fragment containing only the terminal repeats at
the junction of the two elements (band 5) does not bind to
the matrix. The whole fragment (band 1) binds as well, although with an apparent lower affinity. The reason for this is
not clear at present, although one possibility is the lower
overall outcome of this particular labeling (as determined by
the intensity of the input fragment). Nonetheless, this apparently weaker fragment retained persistent matrix binding
capacity when tested under three different competition concentrations. Its matrix-affinity curve as well as the binding
curves of all examined MITEs with MAR activity place them
in the class of the durable MARs (data not shown).
DISCUSSION
9), and two genes that apparently cohabit with retroelements inside their structural domains.
The genic sequence of the maize 334B7.4 gene has been
scattered over 42 kb of genomic space as a result of the insertion of three large blocks of retroelements (Tikhonov et
al., 1999). The closest external MARs would delimit a putative structural domain of ⵑ60 kb (Figure 5). However, a MAR
located inside the gene displayed matrix-affinity characteristics of the durable MARs (Figure 3C). This could indicate
that structural attachment point(s) might exist inside the
genic sequence. Thus, this region might be involved in a
structural role, particularly if the gene is not functional anymore, as suggested by the location of a MAR and a MITE inside an exon (Tikhonov et al., 1999). In the absence of
expression data regarding the functional capacity of the
334B7.4 gene, however, it is not possible to decide whether
it is located within a large structural domain, flanked by and
interrupted by retroelements, or whether it serves as an attachment structure itself.
The only other gene in the studied region sharing a loop
with retrotransposable elements is the putative 334B7.9
gene (Figure 9, loop l⬘). There are no data regarding whether
this is a functional gene, but both 334B7.4 and 334B7.9
genes provide attractive systems to further pursue a relationship between genes’ structural organization and their
expression.
Domain Organization along the Maize Contig
Domain Structures in Sorghum
Comprehensive analysis of MAR distribution in the lowcopy-number regions of the maize contig provided an insight into the potential of these regions to fold into putative
structural loops. The general arrangement along the maize
contig shows genes occupying individual loops that are
separated from neighboring retroelement repeats. Three exceptions have been observed: the 334B7.6 and 334B7.7
genes, which are positioned in the same loop (Figures 5 and
In the 78-kb colinear region of sorghum, 18 MARs were
identified. Most of the genes, with the apparent exception of
genes 110K5.9 and 110K5.10, seem to be positioned in individual loops (Figures 6 and 9). Seven of the MARs colocalize
with fragments containing MITEs. It is interesting that a fragment in the space between genes 110K5.9 and 110K5.10
did not bind to the matrix, despite the fact that it contained
MARs and Putative Genomic Structures
257
a MITE; thus, the two genes apparently share the same
domain.
Comparison of the Predicted Domain Structures in
Maize and Sorghum
The degree of similarity between the orthologous adh-containing regions in maize and sorghum has been reported at
the DNA sequence level (Tikhonov et al., 1999). The possible
folding into structural domains (loops) of these colinear continuums has been assessed here. Despite the presence of
more genes (nine and 14 putative genes on the maize and
sorghum contigs, respectively), the sorghum contig contains
approximately three times less DNA than does the orthologous maize region. Of the 225 kb surrounding maize adh1,
Figure 7. In Vitro Nuclear Matrix Binding Assays of MITEs.
Fragments corresponding to MITEs are shown by arrows in (A) to
(G). Open arrowheads denote fragments from the vector; m, I, and P
denote lanes containing size-marker fragments, input fragments,
and fragments recovered from the pellet, respectively.
(A) Clone p13MC (Table 1), digested with PmlI, SnaBI, and ScaI, released nested Tourist elements located 3⬘ to the sorghum 110K5.3
gene.
(B) Alternative digestion of the same clone with StuI alone, showing
the binding activity of one of the Tourist elements.
(C) Clone p17 (Table 1) after digestion with NotI and StuI, generating
a solo Tourist at the 5⬘ of the sorghum 110K5.5 gene.
(D) Maize clone p95d (Table 1) after digestion with AflII and StuI, releasing the MITE inside the putative exon of maize gene 334B7.4.
(E) Clone pB2wx carries the 128-bp Tourist-Zm1 element (Bureau
and Wessler, 1992; GenBank accession number S48688). Digestion
with EcoRI released a 184-bp fragment; its binding is shown by the
arrow.
(F) and (G) PCR-amplified MITEs. Fragment 1 contains three MITEs
(two nested and a closely adjacent one) from the spacer region between maize genes 334B7.7 and 334B7.8 (see Figure 8B). The primers are shown in Table 3. Fragments numbered 2 to 4 correspond to
amplified elements, as shown in Figure 8A. Fragment 5 is a short region containing the terminal inverted repeats of two adjacent Tourist
elements and does not display binding activity.
Figure 8. Schematic Representation of the MITEs Shown in Figure 7.
(A) The region containing two closely positioned Tourist elements at
the 3⬘ end of the sorghum shrunken2 homolog (Chen et al., 1997;
GenBank accession number AF010283).
(B) The region containing three MITEs (two nested and a closely adjacent one) from the spacer region between maize genes 334B7.7
and 334B7.8 (Tikhonov et al., 1999).
The individual MITEs are shown by the boxes, and the terminal inverted repeats as blank arrows. The position of the primers is shown
by the black arrowheads. The numbers at the primers correspond to
the genomic location of the respective sequence. The PCR products
are shown by the numbered lines under the boxes, corresponding to
the numeration of the binding fragments in Figures 7F and 7G. The
cross-hatched box inside the empty box illustrates a MITE nested
within another MITE.
160 kb is occupied by highly repetitive DNAs represented by
23 long terminal repeat retrotransposons, organized in six
blocks of various sizes (Tikhonov et al., 1999). To be able to
compare the gene-containing structures of the two regions,
we have graphically “subtracted” the retrotransposon blocks
from the maize contig (Figure 9).
It is shown that the gene-containing domains and the
sizes of the putative loops containing orthologous genes in
the two species are comparable. This fact suggests that the
structural folding of the homologous genomic regions, excluding the blocks of highly repetitive DNA, might be very
similar. Thus, despite an enormous difference in the sizes of the
loops occupied by the two orthologous genes, 110K5.4 and
334B7.4 (7 and 60 kb, respectively), if the retroelement blocks
in the maize gene were excluded, the genic sequence would
be located in comparable loops (Figure 9, d and d⬘).
A notable difference is that in sorghum, a gene-containing
loop (Figure 9, loop e) immediately adjacent to the 110K5.4
gene has been deleted from what would have been its location in maize (Figure 9). A relevant question is whether the
perturbation of the maize 334B7.4 domain structure by the
inserted large blocks of retroelements might be related to
the deletion of its immediate neighbor.
Another major difference is that two adjacent genes in
sorghum, 110K5.8 and 110K5.9, are also missing from the
258
The Plant Cell
Figure 9. Comparison of Putative Chromatin Domain Structures of the Orthologous adh Regions in Maize and Sorghum.
The maize YAC represents 225 kb of the maize adh1-F region, with 22 identified long terminal repeat retroelements and eight genes. The
sorghum BAC bar represents 78 kb of the sorghum adh region and 14 candidate genes. The black arrows in the boxes designate the putative direction of both gene and retroelement transcription. The genes in sorghum are designated with letters, whereas the homologous
maize genes are denoted by the same letters with the prime signs. Nonhomologous genes are designated with open arrows. Dashed lines
connecting the bars indicate the nonconserved (deleted) regions between maize and sorghum. The anchors in the boxes indicate durable
domain-defining MARs.
maize region. These regions differ in their domain organization as well: 110K5.8 is predicted to be in an individual small
loop (Figure 9, loop h) on the sorghum contig, whereas
110K5.9 shares a putative loop with 110K5.10 (Figure 9,
loop i to j). The orthologous maize gene, 334B7.7, also is
positioned in a loop as a pair. However, because the two
corresponding neighboring genes 5⬘ to 334B7.7 are missing
in maize, its partner in the loop is its current 5⬘ neighbor,
gene 334B7.6 (Figure 9, loop g⬘ to j⬘). Therefore, sharing a
structural domain with a closely positioned 5⬘ neighboring
gene seems to be a preserved feature for the two orthologous genes, 334B7.7 and 110K5.10. The possible significance of such an arrangement is unclear, but it points again
to a possible relationship between the positioning of a gene
within a structural unit and its functional requirements.
The apparent tendency of high-copy-number retroelements to insert into other retroelements, rather than into
genic sequences, was reported by SanMiguel and co-workers (SanMiguel et al., 1996). Here, we noted that the insertion sites in the maize genome of the initial retroelements
(those at the base of the stacks [SanMiguel et al., 1996])
map right at, or in a very close proximity to, MARs that flank
genes. This observation suggests that MARs may act both
as potential target sites and as barriers for the genes against
invasion by transposable elements.
Intronic MARs
It has been reported that under conditions of nonsaturating
amounts of matrix proteins (i.e., binding sites), only DNAs
with the highest affinity would exhibit binding. Equilibrium
binding studies, at saturating levels of matrix proteins, have
revealed that the process is cooperative and reversible
(Mielke et al., 1990). Competitor DNA, added to prevent
nonspecific binding, is used to measure the matrix affinity of
a MAR-containing fragment. On this basis, MARs have been
arbitrarily defined as strong and weak, and it has been suggested that they may have different functions (Allen et al.,
1993; Bode et al., 1996). However, the criteria for defining
strong and weak MARs have not been unified.
The detection of MARs in the introns of several genes
(Cockerill and Garrard, 1986a; Mielke et al., 1990; Avramova
and Paneva, 1992; Romig et al., 1994) pointed to yet another feature that divides MARs into different categories.
Intronic MARs, by definition, cannot block progression by polymerase II, implying either that they do not function in vivo
or that their binding to the matrix is regulated. In any case,
intronic MARs are not considered to be defining structural domains (Bode et al., 1996). An important question,
therefore, is what makes the two types of MARs different.
It has been shown that a simple classification of MARs as
strong and weak would not discriminate between intronic
and external MAR elements (Romig et al., 1994). We therefore chose to examine the various MARs for their affinity to
the nuclear matrix, under increasing concentrations of competitor DNA, instead of making a one-point strong/weak
comparison. Tentatively, the tested MARs were divided into
two classes on the basis of their binding patterns. One
group contains elements that could bind the matrix in the
presence of the highest competitor concentrations. These
MARs were termed durable, and they were usually found
MARs and Putative Genomic Structures
outside the genes. The second group, unstable MARs, were
competed off their matrix binding sites under similar conditions. Most of the intronic MARs were found in this class. It
should be emphasized that the difference in the competition
curves is not regarded as an absolute criterion separating the
two classes of MARs. Some of the tested MAR fragments,
particularly longer restriction fragments and fragments encompassing internal and external genic sequences, exhibited curves intermediate to the two profiles, indicating a
continuum of MAR affinities. The continuous series of MAR
binding behavior suggest that various components collaborate and contribute to the interactions. Therefore, the difference in the curves suggests that intronic and external MARs
might differ in the nature of their association contacts with
the nuclear matrix. Persistent binding activity might be more
relevant for MARs with a structural function, whereas a capacity for an easier dissociation of even strong MARs might
be a prerequisite for a controlled association with the matrix.
Intronic MARs were found in several sorghum and maize
genes. It is interesting that orthologous loci in maize/sorghum (334B7.4/110K5.4, 334B7.7/110K5.10, and 3334B7.8/
110K5.11) have preserved MARs in their respective introns.
These MAR-containing introns also display extensive sequence conservation (Tikhonov et al., 1999), distinguishing
them from the external MARs.
Little is known about the role, if any, of intronic MARs. It
has been suggested that MARs, particularly those in introns,
may act as a sink for absorbing the positive supercoils generated in front of a progressing polymerase transcription
complex, increasing the efficiency of transcript elongation
by removing a geometric impediment to progress (Benham
et al., 1997). Other data show an involvement of the intronic
␬-immunoglobulin MAR in the regulated demethylation of
the gene upon activation (Kirillov et al., 1996) and in somatic
hypermutation (Betz et al., 1994). An intronic MAR from a
potato gene, when placed between the gene encoding human interferon ␤ and neo, caused activation of transcription
in transfected mouse L cells (Mielke et al., 1990). The findings that intronic MARs are preserved in homologous genes
(Avramova and Paneva, 1992) and in orthologous introns in
maize and sorghum (see earlier text) support the idea of a
conserved function, possibly in gene regulation. Specific
MAR binding soluble factors might be involved in mediating
a regulated binding of intronic MARs to the nuclear matrix
(Boulikas and Kong, 1993; Bode et al., 1996).
MARs and MITEs
Inverted repeats are cruciform-forming sequences (Schroth
and Ho, 1995), and MARs are often enriched in inverted repeats (Boulikas and Kong, 1993). Plant genomes harbor
short interspersed repeats, MITEs, which have terminal inverted repeats and have a potential to form secondary
stem–loop structures (Bureau and Wessler et al., 1994;
Charrier et al., 1999). Nineteen of the 36 MARs discovered in
259
the adh-homologous genomic regions of the two grasses
were found to colocalize on the same fragments with MITEs.
Frequent close associations between MITEs and MARs in
the shrunken2/a1–homologous regions of rice and sorghum
have been noticed as well (Avramova et al., 1998), suggesting a possible correlation between the potential secondary
DNA structure of these elements and MAR function.
Hence, the capacity of isolated MITEs to bind to the matrix was tested. Approximately 20 elements contained within
MAR fragments from the four genomic regions were examined. The MITEs were obtained either by excision with restriction nucleases or by PCR amplification, and in all tested
cases, the MITEs displayed matrix binding activity. In several cases, MITEs were found in a nested arrangement or
immediately adjacent to each other, and our results showed
that the set as a whole and its individual components display a capacity for matrix binding. This had been observed
earlier for the maize 5⬘ adh1 MAR, before recognizing that this
region was made of two nested Tourist elements. Each of
these elements bound to the matrix, although with reduced
activity compared with the nested structure (Avramova and
Bennetzen, 1993).
The apparent involvement of these mobile elements in a
genome structural function concerns the basic question of
whether transposable elements and repetitive DNAs may
play any role in host genomes. As is also true of the MARs
associated with some members of retroelement families, it
does not seem likely that mobile DNAs would be key determinants of chromosome structure. However, once present at
a new genomic location, a mobile DNA element might be
selected for new local functions, such as gene regulation, recombinational initiation (Galliano et al., 1995), telomere structure
(Danilevskaya et al., 1994; Eickbush, 1997), or MAR function.
Loop Organization of Plant Genes
A major consequence of the model suggesting that chromatin in the nucleus is folded into discrete structural domains
(loops) is the attractive possibility that each domain constitutes an independent unit of genetic function. Several
classes of structural elements have been suggested to play
this role (reviewed in Geyer, 1997), but there is controversy
regarding the chromatin-boundary function for some of
them (Avramova and Tikhonov, 1999).
MARs are good candidates for playing a structural, domain-defining role, and their distribution in eukaryotic genomes typically places them at the borders of gene domains.
Adjacent MARs commonly outline potential loops of 3 to
400 kb (Boulikas and Kong, 1993), exhibiting a general inverse relationship between domain size and gene activity
(reviewed in Bode et al., 1996). In plants, relatively small (3
to 10 kb) loops seem to represent a common pattern for
the organization of plant genes (Breyne et al., 1992; van
der Geest et al., 1994; Chinn and Comai, 1996; Avramova
et al., 1998). In a 16-kb region of the small Arabidopsis
260
The Plant Cell
genome, for example, loops averaging ⵑ5 kb were found to
harbor two genes each (van Drunen et al., 1997).
Recently, it was reported that the maize adh1 gene is located within a predicted 90-kb loop, based on the size of a
genomic fragment obtained by preferential cleavage at superhypersensitive sites (Paul and Ferl, 1998). Although it is
not possible to interpret this result in terms of a structural
loop, it localizes the closest nuclease-susceptible sites to
the adh1 gene. Our sequence data indicate that ⵑ90 kb is
the approximate distance between the promoter of adh1
gene and the closest flanking promoter of maize gene
334B7.5, suggesting that the nuclease hypersensitive sites
map at potentially active promoter regions that often colocalize with MARs (reviewed in Bode et al., 1996).
A difficult unresolved matter is the relationship between
genome structure and regulated gene expression. An approach providing information on a possible link between
chromatin folding into structural units and the organization
of the genes inside has been reported here. From this perspective, MARs have been analyzed and considered as
structural elements only outlining putative domains. The
comparative analysis enabled us to observe preservation of
common structural features of gene organization, despite
the very different nature of the surrounding sequences. An
important corollary is that if MARs define structural do-
mains, then the orthologous genes in maize and sorghum
should experience similar chromatin environments. The
preservation of MAR function around orthologous genes, including cases of genes sharing a domain (334B7.7 and
110K5.10), points to functional significance.
Although numerous MARs have been found involved in
a variety of regulated phenomena (Allen et al., 1993;
Mlynarova et al., 1995; Iglesias et al., 1997; Holmes-Davis
and Comai, 1998; Matzke and Matzke, 1998; Ulker et al.,
1999; Vain et al., 1999), others have failed to play the role
expected of them in some transgenic studies. This suggests
that the role of MARs may be more elusive and complicated
than initially postulated. To understand how a genome functions, it is necessary to know how it is constructed, and a
description of the organization of complex native systems is
a step in that direction.
METHODS
DNA Clones
The yeast artificial chromosome (YAC) 334B7, contains an ⵑ225-kb
insert from the maize adh1 F region (Edwards et al., 1992). The bac-
Table 1. Subclones Analyzed in the Matrix Binding Assays
Clone
Sorghum
p11CM
P13MC
p5.4
p17
p2
p16
p18
p5sMARda
p3sMARa
Maize
pL243d
p5zMARa
p3zMAR1a
p3zMAR2a
pL95d
pL201
pL347
pL21
p3.3BamHI
pB2wxb
Tobacco
pGHNC5a,c
a Subclones
Position (bp)
Size (bp)
Cloning Sites
Vector
Source
1–10,408
10,403–22,564
22,559–27,930
27,923–45,076
45,069–47,184
47,177–64,422
64,417–78,195
8,096–10,362
14,433–16,851
10,408
12,161
5,371
17,153
2,115
17,245
13,778
2,266
2,418
ClaI-MluI
MluI-ClaI
ClaI-NotI
NotI-NotI
NotI-NotI
NotI-ClaI
ClaI-NotI
XbaI-XbaI
SacII-MunI
pGEM7⫹
pGEM7⫹
pBSIIKS⫹
pMOB
pBSIIKS⫹
pMOB
pMOB
pBSIIKS⫹
pGEM-TEazy
This study
This study
This study
This study
This study
This study
This study
This study
This study
1–6,028
65,831–67,056
71,014–73,511
71,014–73,333
81,429–93,499
144,388–159,323
154,148–172,938
169,902–186,579
186,298–189,670
N/A
7,000
1,225
2,496
2,319
12,070
14,935
18,790
16,676
3,272
184
BsiWI-BsiWI
XhoI-XbaI
PstI-BamHI
PstI-SacI
NotI-XhoI
NotI-NotI
NotI-NotI
NotI-NotI
BamHI-BamHI
EcoRI-EcoR
pGEM7⫹
pBSIIKS⫹
pBSIIKS⫹
pBSIIKS⫹
pBSIIKS⫹
pMOB
pMOB
pMOB
pBSIIKS⫹
pCRII
This study
This study
This study
This study
This study
This study
This study
This study
This study
Bureau and Wessler (1992)
N/A
2.2 kb
NotI-Bsp120I
pBSIISK⫹
Hall et al. (1991)
with minimal MAR fragments.
b Tourist-Zm1 element of wx-B2 locus of maize (GenBank accession number S48688; Bureau and Wessler, 1992).
c Dimer RB7 scaffold attachment regions (Hall et al., 1991).
MARs and Putative Genomic Structures
261
Table 2. Restriction Enzyme Digestions for Matrix Binding Assays
Clone
Assay 1
Assay 2
Assay 3
p11CM
XbaI, EcoRI, ApaLI, ClaI, MluI
MluI, MfeI, BamHI, NruI
MfeI, EcoRI, XbaI, BsiWI, NcoI, NruI
P13MC
MluI, XbaI, Bsu36I
MluI, ClaI, ApaLI, SpeIa
p5.4
p17
p2
p16
p18
pL243d
ClaI, SacI
BsiWI, PmeI, ApaLI, XbaI, NotI
NotI
HindIII, Ecl136, StuI, ClaI
Bsp120I, HindIII, MfeI, NdeI, NotI
NcoI, NotI, XhoI, BamHI
p5zMAR
p3zMAR1
pL95d
XhoI, XbaI
BamHI, BglII, EcoRI
Bsu36I, HindIII, SalI, XbaI, XhoI
pL201
pL347
pL21
p3.3BamHI
PB2wx
pGHNC5
Bsu36I, XhoI, XbaI, NotI
Bsu36I, XbaI, XhoI, NotI
Bsu36I, XhoI, XbaI, NotI
AatII, SnaBI, EcoRI
EcoRI
NotI, Bsp120I
ClaI, BstXI
Bsp120I, SacII, XhoI
NotI, StuI
ApaLI, NdeI, NotI, ClaI
ApaLI, NotI, StuI
Bsu36I, PstI, NotI, XhoI,
BamHI
BamHI, XhoI, XbaI, BsrGI
BamHI, EcoRI
SpeI,a SphI, StuI, BsiWI,
XhoI
BsrGI, BsiWI, EcoRI, NotI
ApaLI, AvrII, NotI
ApaLI, AvrII, NotI
BglII, NcoI, StuI, BamHI
a Enzyme
Assay 4
BsiWI, ClaI, MluI, NcoI,
PacI, SpeIa
MluI, SacII, Bsu36I, SnaBI, MfeI, NcoI MluI, SacII, ApaLI, SmaI,
MfeI, SpeI, XbaI
StuI, NotI
BsiG, EcoRI, NarI, NotI, ClaI
BsrGI, SalI, XbaI, NotI
EcoRI, BspHI, ClaI
StuI, AscI
AflIII, StuI
BspHI, SpeI, PstI, NotI
MfeI, NcoI, SpeI, NotI
BglII, BspHI, NotI
BglII, StuI, AatII
that produced partial digestion.
terial artificial chromosome (BAC) 110K5 contains ⵑ165 kb of the
colinear sorghum region (Woo et al., 1994). YAC 334B7 and BAC
110K5 were subcloned, and the clones analyzed here are shown in
Table 1. All 225 kb of the maize YAC and the colinear 78 kb from the
sorghum BAC have been screened for matrix attachment region
(MAR) distribution, after their complete sequence analysis (Tikhonov
et al., 1999). The sizes of the cloned fragments, the cloning sites, and
the vectors they were cloned in are shown in Table 1.
The clones were subjected to digestions with various combinations of restriction endonucleases, as shown in Table 2. The purpose
was to create multiple overlapping fragments, all of which were
tested for matrix binding activity. Their positions on the respective
contigs are shown in Figures 5 and 6. Restriction endonucleases and
T4 polynucleotide kinase were obtained from New England Biolabs
(Beverly, MA) or Promega (Madison, WI). Calf intestinal alkaline
phosphatase used for the dephosphorylation of DNA fragments was
obtained from Pharmacia (Piscataway, NJ).
Isolation of Miniature Inverted Repeat Transposable Elements
The location of the genomic clones containing colocalized miniature
inverted repeat transposable elements (MITEs) and MARs is shown
in Table 1. Restriction digestion of the clones was conducted as indicated in the legend to Figure 7.
Oligonucleotide primers (summarized in Table 3) were synthesized
by Integrated DNA Technologies, Inc. (Coralville, IA). Polymerase
chain reactions (PCRs) were performed as follows: one cycle at 95⬚C
for 3 min; 35 cycles at 55⬚C for 30 sec, 72⬚C for 2 min, 94⬚C for 45
sec; and one cycle at 75⬚C for 5 min. Samples were stored at 4⬚C.
Isolation of Nuclei and Nuclear Matrices
Nuclei were prepared following the procedure described in Avramova
and Bennetzen (1993). Sorghum nuclei were isolated from 15-dayold etiolated seedlings. Ten grams of seedling tissue was frozen in
liquid nitrogen and ground in three portions to a fine powder in a
mortar. All subsequent steps were performed at 0 to 4⬚C. Intactness
and purity of nuclei were monitored under a light or fluorescent microscope after staining with 4⬘,6-diamidino-2-phenylindole. The nuclear preparation from the final pellet was adjusted to 70% glycerol,
aliquoted at 3 A260 ODs, and kept at ⫺70⬚C.
Nuclear matrices were isolated by the 2 M NaCl extraction protocol, as described in Cockerill and Garrard (1986a), with minor modifications. Three A260 units of nuclear preparation were thawed on
ice and spun down for 3 min at maximum speed in a microcentrifuge, and the pellet was washed in 1 mL of 10 mM NaCl, 3 mM
MgCl2, 10 mM Tris-HCl, pH 7.4, 0.5 mM phenylmethylsulfonyl fluoride, with the addition of leupeptin and aprotinin (1 ␮g/mL each),
pelleted at maximim speed, and resuspended in 250 ␮L of the
same buffer. DNase I and RNase A were added (5 ␮g of each), and
the nuclei were incubated for 1 hr at 37⬚C. An equal volume of icecold 4 M NaCl was added, and the tube was incubated on ice for
10 to 20 min with two to three inversions of the tube. The nuclear
proteins that were insoluble in 2 M NaCl were pelleted by centrifugation at maximum speed for 3 min at room temperature. The
supernatant was removed carefully, and the pellet (nuclear matrix)
was washed twice with 1 mL of 2 M NaCl as described previously.
The nuclear matrices were washed three times, as described previously, in matrix binding buffer (MBB; 10 mM Tris-HCl, pH 7.4, 50
mM NaCl, 20 mM EDTA, 250 mM sucrose, and 0.25 mg/mL BSA).
262
The Plant Cell
Table 3. Clones and DNA Primers Used for Amplification of MITEs
PCR
Clones
Primers
Sequencesa
1
pL21
Z180369, Z181263
2
pSh2-1b
Sb7106, Sb7475
3
pSh2-1
Sh7384, Sh7784
4
5
pSh2-1
pSh2-1
Sb7106, Sh7784
Sh7384, Sb7475
cctacgtaTGCATTACTCATCTGTCAATGAC
cctgatcagCTGAGCCCGCAGAAAATCGCA
gagctcgaGTACTCAAACCTTAGACCCTG
cacgtgtcGAcAGGACTACAGTCTAAGGCCC
ccgaaTTCATGCATGTGTCCCAAAGATTCG
cctcgcgaATTCAACCTCCCACTGTTGG
See above
See above
a Capital
letters designate genomic sequence. Underlined nucleotides designate sequences from terminal inverted repeats. Double-underlined
nucleotides are direct duplications at the insertion site.
b Clone pSh2-1 containing a 13,945-bp BamHI fragment from the sorghum Shrunken2-homologous region (GenBank accession number
AF010283; Chen and Bennetzen, 1996).
Finally, matrices from 3 A260 units of the nuclear preparation were
dispersed in 100 ␮L of MBB.
Matrix Binding Assay
An aliquot (30 ␮L) of the matrix preparation described above was
used in each binding assay, in a total volume of 50 ␮L. Sheared (0.3
to 2 kb) competitor Escherichia coli DNA was added to the suspension of nuclear matrices to a final concentration of 50 to 400 ␮g/mL.
Depending on the size of the tested plasmid, 20 to 100 ng of DNA
was dephosphorylated and end-labeled with ␥-32P-ATP by T4 polynucleotide kinase. Five percent of the labeling reaction was added to
the binding assay, and the volume was adjusted to 50 ␮L with MBB.
Microcentrifuge tubes were placed in a horizontal position on a tube
rotary shaker and incubated at room temperature. After 5 to 12 hr,
the matrices were collected by centrifugation at maximum speed for
3 min, and the supernatants were saved. The pellets were washed by
resuspending once in 1 mL of MBB, followed by centrifugation as described earlier, and finally washed once with 1 mL of Tris-EDTA, pH
8.0. Final pellets were dispersed in 25 ␮L of 0.5% SDS and 2 mg/mL
proteinase K in Tris-EDTA. After incubation at 37⬚C for 5 to 12 hr, the
reactions were incubated at 70⬚C for 15 min. One-half of each reaction was loaded onto 0.7 to 1.2% agarose gels in 0.5 ⫻ Tris-borateEDTA buffer. After electrophoresis, agarose gels were dried onto nylon membranes in a vacuum gel drier at 80⬚C. Dried gels were quantitatively analyzed either after exposing x-ray film to them and
scanning by densitometer or by exposing them directly to a PhosphorImager plate (Molecular Dynamics, Sunnyvale, CA). Both methods yielded comparable results.
Image Analysis
Exposed x-ray films were analyzed by ImageQuant version 3.3 (Molecular Dynamics, Sunnyvale, CA) and/or Scion Image version 3, a
Windows version of NIH Image (Scion Corp., Frederick, MD). The
mean gray value of each fragment recovered from the pellet was
computed, and the values for the background were subtracted. The
mean gray values for the background were determined by measuring
a band of the same size from a region adjacent to the respective
fragment in the pellet. The mean gray value of each respective frag-
ment from the “input” lanes was used as a reference value, taken as
100%. MAR activity was quantified as a percentage of each matrixbound fragment relative to its value in the input fraction. These values were then plotted against the step-gradient concentration of the
competitor DNA.
ACKNOWLEDGMENTS
The authors are grateful to Phillip SanMiguel (Purdue University) for
subcloning the maize YAC and to Steve Spiker (North Carolina State
University, Raleigh) and Susan Wessler (University of Georgia,
Athens) for providing the clones containing the MAR from the tobacco root–specific gene RB7 and the clone pB2wx carrying the
Tourist-Zm1 element, respectively. This research was supported by
the U.S. Department of Agriculture Cooperative State Research Education and Extension Service grants No. 98-35300-6167 to Z.V.A.
and No. 97-35300-4594 to J.L.B.
Received September 14, 1999; accepted December 5, 1999.
REFERENCES
Allen, G.C., Hall, G.E., Childs, L.C., Weissinger, A.K., Spiker, S.,
and Thompson, W.F. (1993). Scaffold attached regions increase
reporter gene expression in stably transformed plant cells. Plant
Cell 5, 603–613.
Avramova, Z., and Bennetzen, J.L. (1993). Isolation of nuclear
matrices from maize leaves: Identification of a matrix binding site
adjacent to the maize Adh1 gene. Plant Mol. Biol. 22, 1135–1143.
Avramova, Z., and Paneva, E. (1992). Matrix attachment sites in the
murine ␣-globin gene. Biochem. Biophys. Res. Commun. 182,
78–85.
Avramova, Z., and Tikhonov, A.P. (1999). Are scs and scs⬘ ‘neutral’
chromatin domain boundaries of the 87A7 locus in vivo? Trends
Genet. 15, 138–139.
MARs and Putative Genomic Structures
Avramova, Z., SanMiguel, P., Georgieva, E., and Bennetzen, J.L.
(1995). Matrix attachment regions and transcribed sequences
within a long chromosomal continuum containing maize adh1.
Plant Cell 7, 1667–1680.
Avramova, Z., Tikhonov, A., SanMiguel, P., Jin, Y.-K., Liu, C.,
Woo, S.-S., Wing, R., and Bennetzen, J.L. (1996). Gene identification in a complex chromosomal continuum containing maize
adh1. Plant J. 10, 1163–1168.
Avramova, Z., Tikhonov, A., Chen, M., and Bennetzen, J.L.
(1998). Matrix-attachment regions in colinear segments of the sorghum and rice genomes. Nucleic Acids Res. 26, 761–767.
Benham, C., Kohwi-Shigematsu, T., and Bode, J. (1997). Stressinduced duplex DNA destabilization in scaffold/matrix attachment
regions. J. Mol. Biol. 274, 181–196.
Betz, A., Milstein, C., Gonzalez-Fernandez, A., Panneli, R., Larson,
T., and Neuberger, M. (1994). Elements regulating somatic
hypermutation of the immunoglobulin ␬ gene: Critical role for the
intron enhancer/matrix attachment region. Cell 77, 239–248.
Bode, J., Iber, M.C., Kay, V., Schlake, T., and Dietz-Pfeilstetter,
A. (1996). Scaffold/matrix–attached regions: Topological switches
with multiple regulatory functions. Crit. Rev. Eukaryotic Gene
Expression 6, 115–138.
Boulikas, T., and Kong, C. (1993). Multitude of inverted repeats
characterize a class of anchorage sites of chromatin loops to the
nuclear matrix. J. Cell. Biochem. 53, 1–12.
Breyne, P., van Montague, M., Depicker, A., and Gheysen, G.
(1992). Characterization of a plant scaffold-attachment region in
DNA fragment that normalizes transgene expression in tobacco.
Plant Cell 4, 463–471.
Bureau, T.E., and Wessler, S. (1992). Tourist: A large family of small
inverted repeat elements frequently associated with maize genes.
Plant Cell 4, 1283–1294.
Bureau, T.E., and Wessler, S. (1994). Stowaway: A new family of
inverted repeat elements associated with the genes of both monocotyledonous and dicotyledonous plants. Plant Cell 6, 907–916.
Charrier, B., Foucher, F., Kondorsi, E., d’Aubenton-Carafa, Y.,
Thermes, C., Kondorsi, A., and Rater, P. (1999). Bigfoot: A new
family of MITE elements characterized from the Medicago genus.
Plant J. 18, 431–441.
Chen, M., and Bennetzen, J. (1996). Sequence composition and
organization in the sh2/a1-homologous region in rice. Plant Mol.
Biol. 32, 999–1001.
Chen, M., SanMiguel, P., and Bennetzen, J. (1997). Sequence
organization and conservation in sh2/a1-homologous regions of
the sorghum and rice genomes. Genetics 148, 435–443.
Chinn, A.M., and Comai, L. (1996). The heat shock cognate 80
gene of tomato is flanked by matrix attachment regions. Plant
Mol. Biol. 32, 959–968.
263
enhancer in a region containing topoisomerase II sites. Cell 44,
273–282.
Cockerill, P.N., and Garrard, W.T. (1986b). Chromosomal loop
anchorage sites appear to be evolutionarily conserved. FEBS Lett.
204, 5–7.
Danilevskaya, O., Slot, F., Pavlova, M., and Pardue, M.-L. (1994).
Structure of the Drosophila HeT-A transposon: A retrotransposonlike element forming telomeres. Chromosoma 103, 215–224.
Edwards, K.J., Thompson, H., Edwards, D., de Saizieu, A.,
Sparks, C., Thompson, J.I., Greenland, A.J., Eyers, M., and
Schuch, W. (1992). Construction and characterization of a yeast
artificial chromosomal library containing three haploid maize
genome equivalents. Plant Mol. Biol. 19, 299–308.
Eickbush, T.H. (1997). Telomerase and retrotransposons: Which
came first? Science 277, 911–912.
Flavell, R.B. (1994). Inactivation of gene expression in plants as a
consequence of specific sequence duplication. Proc. Natl. Acad.
Sci. USA 91, 3490–3496.
Galliano, H., Muller, A.E., Lucht, J.M., and Meyer, P. (1995). The
transformation booster sequence from Petunia hybrida is a retrotransposon derivative that binds to the nuclear scaffold. Mol.
Gen. Genet. 247, 614–622.
Geyer, P.K. (1997). The role of insulator elements in defining
domains of gene expression. Curr. Opin. Genet. Dev. 7, 242–248.
Hall, G.E., Allen, G.C., Loer, D.S., Spiker, S., and Thompson, W.F.
(1991). Nuclear scaffolds and scaffold-attachment regions in
higher plants. Proc. Natl. Acad. Sci. USA 88, 9320–9324.
Henikoff, S., and Comai, L. (1998). Trans-sensing effects: The ups
and downs of being together. Cell 93, 329–332.
Holmes-Davis, R., and Comai, L. (1998). Nuclear matrix attachment regions and plant gene expression. Trends Plant Sci. 3, 91–97.
Iglesias, V.A., Moscone, E.A., Papp, I., Neuhuber, F., Michalowski,
S., Phelan, T., Spiker, S., Matzke, M.A., and Matzke, A.J.M.
(1997). Molecular and cytogenetic analysis of stably and unstably
expressed transgene loci in tobacco. Plant Cell 9, 1251–1264.
Izaurralde, E., Mirkovitch, J., and Laemmli, U.K. (1988). Interaction
of DNA with nuclear scaffolds in vitro. J. Mol. Biol. 200, 111–125.
Kirillov, A., Kistler, B., Mostoslavsky, R., Cedar, H., Wirth, T., and
Bergman, Y. (1996). A role for nuclear NF-␬B in B-cell-specific
demethylation of the Igk locus. Nat. Genet. 13, 435–441.
Lamond, A.I., and Earnshaw, W.C. (1998). Structure and function
in the nucleus. Science 280, 547–553.
Laurie, D.A., and Bennet, M.D. (1985). Nuclear DNA content in the
genra Zea and Sorghum. Intergenic, interspecific and intraspecific
variation. Heredity 55, 307–313.
Matzke, J.M., and Matzke, M.A. (1998). Position effects and epigenetic
silencing of plant transgenes. Curr. Opin. Plant Biol. 1, 142–148.
Civardi, L., Xia, Y., Edwards, K.J., Schnable, P., and Nikolau, B.J.
(1994). The relationship between genetic and physical distances
in the cloned a1-sh2 interval of the Zea mays L. genome. Proc.
Natl. Acad. Sci. USA 91, 8268–8272.
Mielke, C., Kohwi, Y., Kohwi-Shigematsu, T., and Bode, J. (1990).
Hierarchical binding of DNA fragments derived from scaffoldassociated regions: Correlation of properties in vitro and function
in vivo. Biochemistry 29, 7475–7485.
Cockell, M., and Gasser, S.M. (1999). Nuclear compartments and
gene regulation. Curr. Opin. Genet. Dev. 9, 199–205.
Mlynarova, L., Jansen, R.C., Conner, A.J., Stiekema, W.J., and
Nap, J.-P. (1995). The MAR-mediated reduction in position effect
can be uncoupled from copy-number dependent expression in
transgenic plants. Plant Cell 7, 599–609.
Cockerill, P.N., and Garrard, W.T. (1986a). Chromosomal loop
anchorage of the kappa immunoglobulin gene occurs next to the
264
The Plant Cell
Paul, A.-L., and Ferl, R.J. (1993). Osmium tetroxide footprinting of a
scaffold attachment region in the maize adh1 promoter. Plant Mol.
Biol. 22, 1145–1151.
Paul, A.-L., and Ferl, R.J. (1998). Higher order chromatin structures
in maize and Arabidopsis. Plant Cell 10, 1349–1359.
Romig, H., Ruff, J., Fackelmayer, F., Patil, M., and Richter, A.
(1994). Characterization of two intronic nuclear matrix-attachment
regions in the human DNA topoisomerase I gene. Eur. J. Biochem.
221, 411–419.
SanMiguel, P., Tikhonov, A., Jin, Y.-K., Motchoulskaya, N.,
Zakharov, D., Berhan, A., Springer, P., Edwards, K., Lee, M.,
Avramova, Z., and Bennetzen, J. (1996). Nested retrotransposons in the intergenic regions of the maize genome. Science
274, 765–768.
Schroth, G.P., and Ho, P.S. (1995). Occurrence of potential cruciform and H-DNA forming sequences in genomic DNA. Nucleic
Acids Res. 23, 1977–1983.
Ulker, B., Allen, G.C., Thompson, W.F., Spiker, S., and Weissinger,
A.K. (1999). A tobacco matrix attachment regions increase transgene expression levels and stability in transgenic tobacco plants.
Plant J. 18, 253–263.
Vain, P., Worland, B., Kohl, A., Snape, J., Allen, G.C., and
Thompson, W.F. (1999). Matrix attachment regions reduces the
loss of transgene expression in the progeny of transgenic tobacco
plants. Plant J. 18, 233–242.
van der Geest, A.H.M., Hall, G.E., Spiker, S., and Hall, T.C. (1994).
The ␤-phaseolin gene is flanked by matrix-attachment regions.
Plant J. 6, 413–423.
van Drunen, C., Oosterling, R., Keultjes, G., Weisbeek, P., van
Driel, R., and Smeekens, C.M. (1997). Analysis of the chromatin
domain organization around the plastocyanin gene reveals a
MAR-specific sequence in Arabidopsis thaliana. Nucleic Acids
Res. 25, 3904–3911.
Springer, P., Edwards, K.S., and Bennetzen, J.L. (1994). DNA
class organization on maize Adh1 yeast artificial chromosomes.
Proc. Natl. Acad. Sci. USA 91, 863–867.
Wessler, S.R., Bureau, T.E., and White, S.E. (1995). LTR-retrotransposons and MITEs: Important players in the evolution of
plant genomes. Curr. Opin. Genet. Dev. 5, 814–821.
Tikhonov, A.P., SanMiguel, P.J., Nakajima, Y., Gorenstein, N.,
Bennetzen, J.L., and Avramova, Z.V. (1999). Colinearity and its
exceptions in orthologous adh regions of maize and sorghum.
Proc. Natl. Acad. Sci. USA 96, 7409–7414.
Woo, S.-S., Jiang, J., Gill, B.S., Paterson, A.H., and Wing, R.A.
(1994). Construction and characterization of a bacterial artificial
chromosome library for Sorghum bicolor. Nucleic Acids Res. 22,
4922–4931.
Structural Domains and Matrix Attachment Regions along Colinear Chromosomal Segments of
Maize and Sorghum
Alexander P. Tikhonov, Jeffrey L. Bennetzen and Zoya V. Avramova
Plant Cell 2000;12;249-264
DOI 10.1105/tpc.12.2.249
This information is current as of August 3, 2017
References
This article cites 52 articles, 17 of which can be accessed free at:
/content/12/2/249.full.html#ref-list-1
Permissions
https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X
eTOCs
Sign up for eTOCs at:
http://www.plantcell.org/cgi/alerts/ctmain
CiteTrack Alerts
Sign up for CiteTrack Alerts at:
http://www.plantcell.org/cgi/alerts/ctmain
Subscription Information
Subscription Information for The Plant Cell and Plant Physiology is available at:
http://www.aspb.org/publications/subscriptions.cfm
© American Society of Plant Biologists
ADVANCING THE SCIENCE OF PLANT BIOLOGY