B RIEFINGS IN FUNC TIONAL GENOMICS AND P ROTEOMICS . VOL 5. NO 1. 24 ^31
doi:10.1093/bfgp/ell003
The positive aspects of stress: strain
initiates domain decondensation (SIDD)
Silke Winkelmann, Martin Klar, Craig Benham, Prashanth AK, Sandra Goetze, Angela Gluch and Juergen Bode
Advance Access publication date 23 February 2006
Abstract
The conventional string-based bioinformatic methods of genomic sequence analysis are often insufficient to identify
DNA regulatory elements, since many of these do not have a recognizable motif. Even in case a sequence pattern is
known to be associated with an element it may only partially mediate its function. This suggests that properties not
correlated with the details of base sequence contribute to regulation. One of these attributes is the DNA strandseparation potential, known as SIDD (stress-induced duplex destabilization) which facilitates the access of tracking
proteins and the formation of local secondary structures. Using the type 1 interferon gene cluster as a paradigm,
we demonstrate that the imprints in a SIDD profile coincide with chromatin domain borders and with DNAse I
hypersensitive sites to which regulatory potential could be assigned. The approach permits the computer-guided
identification of yet unknown, mostly remote sites and the design of artificial elements with predictable properties
for multiple applications.
Keywords: chromatin domains; interferon gene cluster; remote control elements; non-viral episomes; SIDD; DNAse I
hypersensitive sites
Corresponding author. Juergen Bode, GBF German Research Center for Biotechnology/Epigenetic Regulation Mascheroder Weg 1,
D-38124 Braunschweig, Germany. Tel: þ49 (531) 6181 251; Fax: þ49 (531) 6181 262. E-mail: [email protected]
SilkeWinkelmann is a PhD student in the group of JBO where she develops novel methods to study transcriptional properties as they
relate to nuclear structure.
Martin Klar received his PhD and a subsequent prize in 2005 for his work in the same group dealing with the remote control of the
human and mouse interferon-beta genes. He could verify the major predictions of SIDD profiles. http://opus.tu-bs.de/opus/volltexte/
2005/713/pdf/DissMK.pdf
Craig Benham is the senior biomathematical counterpart of the group. He has an AB degree from Swartmore College and a PhD
in Mathematics from Princeton University. His studies of biopolymer structure started when he worked with John Kozak at the
University structure (proteins) and during his postdoctoral years with Max Delbrueck at Cal Tech. After leading the Department of
Biomathematics at Mount Sinai, NY, he became a founding Associate Director of the UC Davis Genome Center. Contacts to the
Braunschweig group date back to 1994, the first joint publication to 1997.
Prashanth AK joined Craig Benham at Mount Sinai and now he shares his research programmes at Davis. Nowadays he is the active
cooperation partner of the Braunschweig people.
Sandra Goetze did her PhD in the lab of Prof. Bode (1998–2001) working on the identification and biochemical analysis of boundary
elements under consideration of biomathematical models. Currently she is investigating the three-dimensional chromatin organization
in the human interphase nucleus in the context of an FP6 European Program in the lab of Prof. van Driel in Amsterdam. http://
opus.tu-bs.de/opus/volltexte/2004/587/pdf/Dissertation.pdf
Angela Gluch (born Knopp) initiated participation in the human genome project together with JBO. The major concepts realized in
the present contribution go back to her initiative. She received PhD in 2001. She may return to the group after maternity leave in
2006. http://opus.tu-bs.de/opus/volltexte/2001/200/pdf/Dissertati.pdf
Juergen Bode studied Organic Chemistry at the Technical University of Braunschweig. His switch to Biochemistry occurred during
postdoctoral years at the California Institute of Technology (CalTech; enzyme and neuroreceptor structures with Michael A. Raftery)
and University of Oregon, Eugene (UofO; enzyme kinetics with Sidney A. Bernhard). The final direction of research initiated in the
year the nucleosome was discovered. It widened to aspects of nuclear structure as these relate to gene expression. He is a group leader
(Epigenetics) at GBF (German Research Center of Biotechnology at Braunschweig) and a Professor of Biochemistry at the Technical
University (http://juergenbode.de.vu; http://cvjbo.de.vu).
ß Oxford University Press, 2006, All rights reserved. For permissions, please email: [email protected]
Positive aspects of stress: SIDD
INTRODUCTION
The effects of DNA superhelicity have first been
studied in circular plasmids, which naturally occur
in an underwound (i.e. negatively supercoiled) rather
than in a relaxed state. Using pBR322 DNA as a
model, Mung bean nuclease digestion experiments
have shown strand separation at two regulatory
sites, the promoter region and the 30 terminus of the
ampicillin resistance gene [1]. Starting with this
model the SIDD-algorithm has been developed
and continuously extended to provide a universal
tool to predict the location of ‘base-unpairing
regions’ (BURs) where strand separation occurs
under superhelical strain and in a wider sequence
context (Figure 1B0 ). The algorithm is based on
a statistical mechanical procedure that has been
outlined previously [2, 3].
While for a bacterial plasmid, superhelicity is the
direct consequence of gyrase action, it also arises
during the progression of tracking proteins such as
DNA and RNA polymerases and—in eukaryotes—
also due to the loss of nucleosomes (facilitated by
histone hyperacetylation) [4]. Strand separation
initiates at a nucleation centre (core-unpairing
element, CUE) within the BUR, in case the
underwound state is stabilized, either in the context
of a circular DNA or of a chromatin domain that
owes its existence to the bordering elements. These
bordering elements, otherwise known as S/MARs
(scaffold- or matrix attachment regions) are an
essential attribute of higher order chromatin structures as they dissect the eukaryotic genome into
independently regulated units (chromatin domains
[5]). Each domain contains one gene together
with its regulatory elements (upstream regulatory
elements, enhancers, LCR), and accordingly the
number of domains has been estimated (and partially
verified) to be in the order of 30 000 with sizes that
vary between about 4 kb (transcriptionally active)
and 200 kb (rather inactive). Exceptions to this rule
are clusters of co-regulated genes that may, with
advantage, populate a single domain and larger arrays
of transcriptionally inactive genes, which are not in
contact with the transcription and replication
machineries that are components of the nuclear
matrix (i.e. the protein-/hnRNA-backbone of the
nucleus). By necessity, these refinements implicate
a discrimination of S/MARs into either constitutive
or facultative variants. For the constitutive group,
DNA–nuclear matrix contacts exist irrespective of
cell type and transcriptional activity. Members of this
25
family are mostly associated with a constitutive
DNAse I hypersensitive site [6] and these are also
the very sites at which domain-size fragments are
generated early in apoptosis [7]. Facultative S/MARs
on the other hand are elements, which need active
transcription (presence of transcription factors, superhelicity or both) in order to associate with the matrix
[8]. These contacts are usually marked by an activitydependent or even cell-type specific DNAse I
hypersensitive site.
In vivo DNA superhelicity is closely regulated,
and it can induce the formation of locally unpaired
regions that are involved in the initiation of
replication and transcription [9]. The BURs not
only form barriers between independently regulated
domains [5, 10], but also support transcriptional
initiation in a process that has been termed
augmentation [11]. Proteins involved in constitutive
contacts and their cooperation with regulatory
factors that require locally denatured DNA for
their binding have been described [3, 12, 13].
Definition of chromatin domain:
type I interferon paradigm
The human type 1 interferon (IFN) gene cluster
consists of 26 IFN genes and pseudogenes, distributed over 400 kb at Giemsa band 9p22. While
IFNB1 forms an extended domain with many sites of
regulatory potential, which together may constitute
an LCR (locus-control region), the other interferon
genes occur in three subclusters (I–III; Figure 1A).
As a whole, the locus has gained attention because
deletions that initiate here and include the adjacent
tumour suppressor gene(s) are related to some of the
most common genetic abnormalities that occur in
numerous forms of cancer [7]. The S/MARs have
been implicated in the mechanism of these deletions
as they demarcate fragile sites in the genome (see the
site marked ‘A172’ and ‘A1235’ in Figure 1A).
The first studies performed in this laboratory
examined the organization of the 14 kb region
containing the human IFN-b (IFNB1) gene
domain at the telomeric end of the 400 kb locus.
The IFNB1 transcription unit is bounded by two
extended, constitutive S/MARs of 7 kb (upstream)
and 5 kb (downstream; Figure 1B0 ). An additional
facultative S/MAR-element has been demonstrated
between the gene and the downstream domain
border. Owing to its favourable properties, the high
affinity core of the upstream element (Figure 1B00 )
has found multiple uses for the design of novel vector
26
Winkelmann et al.
A
9-Centromere
9-Telomere
400
200
300
Sub-cluster I
100
Sub-cluster II
0
Sub-cluster III
kb
IFNB1
A172
A1235
breakpoints
S/MARs
“E“
B′ 10
8
6
IFNWP18
IFNA10
IFNA7
4
2
0
−2
132.000
130.000
128.000
126.000
124.000
122.000
SAR-E
6
4
2
0
−2
8000
4000
WP18A10A7 [bp]
F20(90%)
F75(96%) F100(61%)
F180(57%)
F230(83%) F275(86%)
B′′
SAR-E
10
8
6
4
2
0
−2
0
40.000
35.000
30.000
4000
8000
12000 16000 20000
IFNB1 domain [bp]
G(x)[kcal/mol]
G(x) [kcal/mol]
D 10
SAR-I
P T
8
G(x) [kcal/mol]
G(x) [kcal/mol]
C 10
HS-1
HS-2
HS-3
B1-Promotor
B1-Terminator
DNaseI-HS
TopoII-sites
25.000
20.000
Intergenic Region [bp]
15.000
AmpT
AmpP
8
6
4
2
0
−2
5000
pTZE20 [bp]
Figure 1: Structural features along the human type I IFN gene cluster. (A) The cluster on the short arm of
human chromosome 9. Solid arrowheads mark functional genes while open arrowheads indicate pseudogenes.
Vertical arrows stand for the position of UEs/DNAse hypersensitive sites. Black boxes below the line symbolize
strong S/MARs while open boxes stand for tested regions without major S/MAR character. A172 and A1235 are fragile
sites (deletion endpoints) adjacent to S/MAR elements. (B0 ) SIDD-profile for the IFNB1 gene domain. A SIDD profile
shows the propensity of a base pair to separate under negative superhelical tension (s ¼ 0.05) and in a given sequence
context. Sites preferred for strand separation have a G < 0 kcal/mol.The slightly destabilized B1promoter and the pronounced B1 terminator are indicated upstream and downstream from the coding-sequence (horizontal arrow). The
strongly destabilized main peaks HS1, HS2 and HS3 represent upstream control elements [9, 10]. (B00 ) SIDD structure
of the IFNB1-upstream S/MAR, analysed as part of a pUVC-type vector. Since all UEs are in competition with one
another, the over-all destabilization of the S/MAR can be estimated from the size of the AmpTand AmpP sites in the
vector backbone. (C) Functional genes (IFNA10 and IFNA7) and a pseudogene (IFNWP18) within subcluster III. Note
that the pseudogene has totally lost a destabilized promoter region and that destabilization at the downstream end is
largely reduced. Between IFNA10 and IFNA7 a potential regulatory region is found around map position 127 000.
(D) Repetitive signals (UEs) in an extended non-coding region. Note the apparent 2800 and 5000 periodicities of
sharp destabilized sites in this range. Although all fragments (F20 ^F275) have a considerable affinity to the nuclear
matrix (57^96% association in an equilibrium situation), competition experiments demonstrate that these UEs do not
represent classical S/MARs [16].
types [14, 15] for which it provides the following
properties:
(i) it insulates integrating vectors from the mostly
negative influences of the surrounding (hetero-)
chromatin [16];
(ii) it provides a domain-opening function as it
mediates the acetylation of histones [11];
(iii) it provides long-term stabilization of gene
expression because it prevents methylation
dependent inactivation processes as they occur
on Lys-9 of histone H3 and at CpG tracts in
the DNA;
(iv) as a consequence of their strand separation
potential S/MARs are recombinogenic
elements promoting recombination processes
Positive aspects of stress: SIDD
that may be desired for certain systematic
modifications of target cells (homologous and
site-specific recombination protocols [17]).
These activities motivated a detailed investigation
of this and related elements by the combination of
biomathematical and in vitro approaches. Figure 1B00
illustrates a distinct architecture across the BUR,
which is composed of a regular succession of minima
(unpairing elements, UEs) within the SIDD profile.
This architecture is responsible for the association
of prominent proteins of the nuclear matrix most
notably SAF-A (scaffold-attachment factor A, otherwise known as hnRNP-U). Its parameters (over-all
destabilization, extension and spacing of the UEs)
have recently been amalgamated into a model that,
for the first time, permits the precise prediction of
interaction strength with the nuclear matrix and
thereby its biological activity [3]. It should be noted
that the over-all destabilization of an insert can be
quantified in the context of a plasmid using the
mentioned ampicillin-gene associated UEs as an
internal standard (competition principle [3],
Figure 1B00 ).
In contrast to the domain borders, the facultative
S/MAR-region of IFNB1 consists of three strongly
destabilized, widely separated elements at positions
500, 2000 and 3000 bp relative to the
transcriptional start site, which could be correlated
with regions of DNAse I hypersensitivity (HS3, HS2,
HS1 in the above order, Figure 1B0 and [12]).
Electrophoretic mobility shift analyses demonstrated
that either the known transcription factor YY1 or its
recently discovered analogue YY2 [18] can be accommodated by an aaATGGt motif, which, within 5kb
of upstream sequence, only exists twice at the flanks
of the HS1 and HS2-associated peaks [12]. Both of
these sites have to be occupied simultaneously to
enable activation by two molecules of YY2, as these
recruit a histone-acetyltransferse (GCN5) to the
enhanceosome. It should be noted that, until
recently, the enhanceosome-binding sequences
upstream from position 110 constituted the only
known control element that was implicated in the
control of the IFN-b genes in humans and in mice.
Based on the present data, a refined mechanism
for gene induction was proposed in which the
ubiquitous factor YY1 provides inactivity by its
competition with YY2 [13].
In spite of the apparent space limitation within
the gene subclusters II and III, all functional gene
27
members appear to be separated by efficient, though
restricted S/MARs as well (Figure 1A). Figure 1C
covers two functional IFN-alpha genes A10
and A7 in addition to a pseudogene (WP18).
Interestingly, in this and several other cases the
pseudogenes have lost flanking destabilized regions
indicating that there is a selective pressure acting
on the physicochemical properties investigated
here [19].
In an extended non-coding region between the
IFNB1- and IFNW1-genes, we detected a striking
periodicity of rather restricted SIDD minima, which
seem to obey a periodicity of roughly 2500 bp
(Figure 1D). These elements had escaped prior
in vitro S/MAR mapping efforts. Only their precise
localization in the SIDD profile enabled a subsequent
investigation, which led to the conclusion that the
signals correspond to a new class of S/MARs with
transcriptional augmentation- but no insulationpotential [16]. Initially, it has been suggested that
this register of elements might be involved in levels
of chromatin organization above the periodicity of
DNA bend sites, which occur once per four
nucleosomes [8]. Regarding the increasing number
of reports dealing with TUFs (transcripts of
unknown function) that are presently discovered
across large sections of the human genome, this
region may alternatively serve yet unknown regulatory functions (report by T. Gingeras at the 2005
BITS meeting and [20]). Also, in this case the SIDD
concept would have opened the door for novel
insights into DNA structure–function relationships.
Intronic S/MARs
The example of a facultative element has already
shown that S/MARs cannot simply be considered to
be static delimiters of functional domains. This view
is supported by the detection of S/MARs within
the first intron of a number of genes, where they
have regulatory potential [21]. Such a relationship
has recently also been confirmed in a wholegenome screen on Arabidopsis [22] (communicated
by T. Werner at the 2005 BITS meeting; Tetko et al.
2005, submitted).
Classical examples of this category are S/MARs
that are associated with the immunoglobulin
k- and m-chain genes [23, 24]. Since, by definition,
intronic S/MARs are transcribed, and since they do
not impede passage of RNA polymerase II, their
occupation must be regulated. Ig-m genes are formed
by the joining of three gene segments that are
28
Winkelmann et al.
L1
Vλ1
Jλ3 Cλ3
Jλ1 Cλ1
G(x) [kcal/mol]
10
8
6
Jλ3
4
Cλ3
Jλ1
Cλ1
2
0
0
1000
2000
3000
4000
5000
6000
7000
8000
[bp]
Rearr.
Vλ1 Jλ1
L1
Cλ1
G(x) [kcal/mol]
10
8
6
Vλ1
Jλ1
Cλ1
4
2
0
1000
[bp]
2000
3000
Figure 2: An S/MAR in the un-rearranged Igl light-chain locus. The S/MAR character of a destabilized site in the
Cl3-Jl1 (map position 4700) and the analogous Cl2-cJl1 region (not shown) was verified by wet-lab techniques [29].
In contrast to the situation for the m- and k-chain genes the element is situated at a position that is lost during the
rearrangement process. Its function must thereby be restricted to germline transcription, where it may set the stage
for rearrangement. New dominant sites appear after this process (lower part).
separated in the germline, i.e. the variable (V-),
diversity (D-) and joining (J-) regions. Two segments
(V and J) are linked to generate k and l light chain
genes [25]. Key to this control is the regulated access
of the lymphoid-specific RAG-recombinase proteins
to the recombining loci. The process commences at
the m-chain locus at the pre-B I stage, followed
by rearrangement at the k-locus. Finally, the l-locus
is rearranged, unless the expression of a functional
k-chain has ablated this process. Prior to rearrangements, germline transcription is observed, which
serves to create accessibility as it is associated with
changes in the methylation status and over-all
chromatin structure.
For the immunoglobulin m- and k-genes,
transcription before and after rearrangement depends
on an enhancer within one of the introns.
Experiments with transgenic mice have shown that
promoter activity also requires the S/MARs at one
(k) or both (m) flanks of the enhancer. While the
m-enhancer alone permits access to its vicinity,
accessibility is significantly extended by the
S/MAR(s) in a process that correlates with extended
demethylation [26].
Positive aspects of stress: SIDD
The juxtaposition of S/MARs with transcriptional enhancer elements has been evolutionarily
conserved within the Igk genes of the mouse, rabbit
and human. Mouse k-constructs lacking the S/MAR
have a lower and erratic expression in transgenic
animals and, particularly, in cultured cells. Just as its
m-counterpart, the k-gene becomes demethylated
during B-cell maturation. While any S/MAR
sequence is sufficient for this reaction, tissue
specificity is mediated by NF-kB binding sequences
within the k-enhancer [27]. Deletion of the Igkintronic S/MAR results in hyperrecombination of
closeby k-genes as well as to a decreased level of
somatic hypermutation [28]. These observations
suggest that its sequences contribute to changes in
chromatin structure.
In mice, l-chains form only about 5% of the total
serum immunoglobulin light chains and they are
much less heterogenous than k-chains. Although no
S/MAR has been found in the J–C introns of the
l clusters, the SIDD approach led to the detection of
an exceptionally strong element in the Cl2-Jl4
and Cl1-Jl1 introns [29]. In contrast to the m- and
k-genes, where the loops are anchored at the same
site before and after rearrangement due to the
persistence of the S/MAR between the J- and
C-fragments, the rearrangement at the l loci leads
to the loss of the respective introns (Figure 2). The
elements upstream from l1 and l4 might therefore
be involved in the long-distance interaction during
activation or in the rearrangement process itself and
their function may later be taken over by destabilized
sites that arise downstream from Cl1 (Figure 2,
lower part). To what extent these differences between
the light-chain loci contribute to the apparent
dominance of the k-chains remains to be determined.
S/MARs by design
Small circular episomes of DNA viruses such as
SV40, BPV and EBV are well-studied minimal
models of a chromatin domain the more, as their
function and maintenance has been associated
with the nuclear matrix [9]. The practical application
of these vehicles is limited, however, by the fact
that their function depends on virally encoded,
oncogenic factors. The design and verification of a
non-viral circular episome that is in the position to
recruit components from the replication apparatus
of the host cell via a S/MAR came therefore as a
breakthrough [9]. Work in the subsequent years
clearly showed that there is an inverse correlation
29
Figure 3: Replication of non-viral episomes depends
on a S/MAR element. While the original vector contains
the S/MAR element (‘SAR-E’) shown in Figure1B00 , a minimal version, which provides full replication capacity was
obtained when this element was replaced by an artificial
tetramer containing four identical copies of the CUE
[30]. Current evidence shows that only the eukaryotic
elements of the vector (bottom part of the box) are
required for replication as an episome.
between an episome’s molecular size (4–12 kb) and
its extrachromosomal maintenance. These observations triggered various efforts to minimize these
vehicles by a rigorous definition of their molecular
components.
An early approach in this direction concerned the
replacement of the original 2.0 kb S/MAR segment
from the IFNB1-upstream domain border by a
minimal element that had been obtained by
oligomerizing an UE (bottom parts of Figure 3).
Binding parameters and transcriptional activities of
30
Winkelmann et al.
various oligomers have been determined and the
620 bp tetrameric insert that had been derived from
the 155 bp CUE was found to maintain episomal
replication [30] (details under http://www.pnas.org/
cgi/content/full/0401355101/DC1/8).
It is particularly this series of experiments, which
demonstrates the utility of the SIDD algorithm for
a deeper molecular understanding of the molecular
functions of S/MAR modules and their combinations. The door is therefore open for the rational
design of autonomous chromatin domains, which,
in the extreme case, are circular, non-viral episomes
with a single domain boundary (S/MAR). Current
evidence suggests that such systems will efficiently
circumvent the drawbacks that are otherwise due to
suppression mechanisms acting on transgenes.
2.
3.
4.
5.
6.
7.
PERSPECTIVE
It appears that distinct patterns of UEs are associated
with major aspects of both replication and transcription. Our work indicates that the SIDD method
may supplement or even partially replace current
efforts such as that reported by Crawford et al. [31]
to identify the location of all cis-acting regulatory
elements by a genome-wide recovery of DNAse I
hypersensitive sites. Monitoring the ‘stress-induced
duplex destabilization’ (SIDD) therefore qualifies as a
method to indicate the sites where ‘strain initiates
domain decondensation’ (SIDD).
8.
9.
10.
11.
12.
Key Points
Gene regulation is not only explained by the sequence but
also by the higher structure of DNA elements.
An important prerequisite is the strand separation
potential under superhelical strain (SIDD).
Independently regulated chromatin domains can be
localized in an SIDD profile together with the regulatory
elements (DNAse I hypersensitive sites) they contain.
We apply these principles to demonstrate a remotecontrol mechanism for the human interferon-beta gene,
to localize fragile genomic sites and to unravel the prerequisites for the rearrangement of immunoglobulin genes.
Modular S/MAR-elements with predictable properties
can be designed based on their SIDD properties.
13.
14.
15.
16.
References
1.
Kowalski D, Natale DA, Eddy MJ. Stable DNA unwinding,
not ‘breathing,’ accounts for single-strand-specific nuclease
17.
hypersensitivity of specific AþT-rich sequences. Proc Natl
Acad Sci USA 1988;85:9464–8.
Benham C, Kohwi-Shigematsu T, Bode J. Stress-induced
duplex DNA destabilization in scaffold/matrix attachment
regions. J Mol Biol 1997;272:181–96.
Bode J, Winkelmann S, Goetze S, et al. Correlations
between scaffold/matrix attachment region (S/MAR)
binding activity and DNA duplex destabilization energy.
J Mol Biol 2005, (in press).
Norton VG, Imai BS, Yau P, et al. Histone acetylation
reduces nucleosome core particle linking number change.
Cell 1989;57:449–57.
Bode J, Kohwi Y, Dickinson L, et al. Biological significance
of unwinding capability of nuclear matrix-associating
DNAs. Science 1992;255:195–7.
Bode J, Schlake T, Rı̀os-Ramı̀rez M, et al. Scaffold/matrixattached regions: structural properties creating transcriptionally active loci in International Review of Cytology. In:
Berezney R, Jeon KW (eds), On structural and functional
organization of the nuclear matrix, 1995; 162A, Academic
Press, 389–453.
Bode J, Benham C, Ernst E, et al. Fatal Connections: when
DNA ends meet on the nuclear matrix. J Cell Biochem 2000;
35(Suppl):3–22.
Bode J, Goetze S, Heng H, et al. From DNA structure to
gene expression: mediators of nuclear compartmentalization
and dynamics. Chromosome Res 2003;11:435–45.
Bode J, Fetzer CP, Nehlsen K, et al. The Hitchhiking
principle: optimizing episomal vectors for the use in gene
therapy and biotechnology. Int J GeneTher Mol Biol 2001;6:
33–46.
Schübeler D, Mielke C, Maaß K, et al. Scaffold/matrixattached regions act upon transcription in a contextdependent manner. Biochemistry 1996;35:11160–9.
Bode J, Benham C, Knopp A, et al. Transcriptional
augmentation: modulation of gene expression by scaffold/
matrix attached regions (S/MAR Elements). Crit Rev
Eukaryot Gene Expr 2000;10:73–90.
Klar M, Stellamanns E, Ak P, et al. Dominant
genomic structures: detection and potential signal
functions in the interferon-beta chromatin domain. Gene
2005;364:79–89.
Klar M, Bode J. Enhanceosome formation over the
interferon-beta promoter underlies a remote-control
mechanism mediated by YY1 and YY2. Mol Cell Biol
2005;25:10159–70.
Ma Y, Ramezani A, Lewis R, et al. High-level sustained
transgene expression in human embryonic stem cells. Stem
Cells 2003;21:111–7.
Bode J, Goetze S, Ernst E, et al. Architecture and utilization
of highly-expressed genomic sities. In: Bernardi G (ed.),
New Comprehensive Biochemistry, 38, Makrides S (Volume
ed.). Gene Transfer and Expression in Mammalian Cells, 2003.
Elsevier, Amsterdam, Chap. 20; 551–72.
Goetze S, Baer A, Winkelmann S, et al. Genomic bordering
elements: their performance at pre-defined genomic loci.
Mol Cell Biol 2005;25:2260–72.
Oumard A, Qiao J, Jostock T, et al. Recommended method
for chromosome exploitation: RMCE-based cassetteexchange systems in animal cell biotechnology.
Cytotechnology, 2005, in press.
Positive aspects of stress: SIDD
18. Nguyen N, Zhang X, Olashaw N, et al. Molecular cloning
and functional characterization of the transcription factor
YY2. J Biol Chem 2004;279:25927–34.
19. Goetze S, Gluch A, Benham C, et al. Computational
and in vitro analysis of destabilized DNA regions in the
interferon gene cluster: the potential of predicting functional gene domains. Biochemistry 2003;42:154–66.
20. Mattick JS. The functional genomics of noncoding RNA.
Science 2005;309:1527–8.
21. Bode, J, Bartsch J, Boulikas T, et al. Transcriptionpromoting genomic sites in mammalia: their elucidation and
architectural principles. GeneTher Mol Biol 1998;1:551–80.
http://www.gtmb.org/volume1/29_Bode.htm.
22. Rudd S, Frisch M, Grote K, et al. Genome-wide in silico
mapping of scaffold/matrix attachment regions in
Arabidopsis suggests correlation of intragenic scaffold/
matrix attachment regions with gene expression. Plant
Physiology 2004;135:715–22.
23. Cockerill PN, Garrard WT. Chromosomal loop anchorage
of the kappa immunoglobulin gene occurs next to the
enhancer in a region containing topoisomerase II sites. Cell
Vol 1986;44:273–82.
24. Cockerill PN, Yuen MH, Garrard WT. The enhancer of
the immunoglobulin heavy chain locus is flanked by
presumptive chromosomal loop anchorage elements. J Biol
Chem 1987;262:5394–7.
31
25. Tonegawa S. Somatic generation of antibody diversity.
Nature 1983;302:575–81.
26. Jenuwein T, Forrester WC, Fernandez-Herrero LA, et al.
Extension of chromatin accessibility by nuclear matrix
attachment regions. Nature 1997;385:269–72.
27. Kirillov A, Kistler B, Mostoslavsky R, et al. A role for
nuclear NF-KB in B-cell specific demethylation of the
Ig-kappa locus. Nature Genet 1996;13:435–41.
28. Yi M, Wu P, Trevorrow KW, et al. Evidence that the Ig
kappa gene MAR regulates the probability of premature V-J
joining and somatic hypermutation. J Immunol 1999;162:
6029–39.
29. Engel H, Rühl H, Benham CJ, et al. Germ-line transcripts
of the immunoglobulin J-C clusters in the mouse:
characterization of the initiation sites and regulatory
elements. Mol Immunol 2001;38:289–302.
30. Jenke AW, Stehle IM, Eisenberger T, et al. Nuclear
scaffold/matrix attached region modules linked to a
transcription unit are sufficient for replication and maintenance of a mammalian episome. Proc Natl Acad Sci USA
2004;101:11322–27. Supporting text: http://www.pnas.org/
cgi/content/full/0401355101/DC1/8.
31. Crawford GE, Holt IE, Mullikin JC, et al. Identifying
gene regulatory elements by genome-wide recovery
of DNase hypersensitive sites. Proc Natl Acad Sci 2004;101:
992–7.