 1996 Oxford University Press
4890–4894 Nucleic Acids Research, 1996, Vol. 24, No. 24
T7 RNA polymerase cannot transcribe through a
highly knotted DNA template
José Portugal and Antonio Rodríguez-Campos1,*
Departamento de Biología Molecular y Celular, Centro de Investigación y Desarrollo, CSIC and 1Departamento de
Patología Molecular y Terapéutica, Instituto de Investigaciones Biomédicas de Barcelona, CSIC, J. Girona 18–26,
08034 Barcelona, Spain
Received September 9, 1996; Revised and Accepted November 7, 1996
ABSTRACT
The ability of T7 RNA polymerase to transcribe a
plasmid DNA in vitro in its linear, supercoiled, relaxed
and knotted forms was analysed. Similar levels of
transcription were found on each template with the
exception of plasmids showing varying degrees of
knotting (obtained using stoichiometric amounts of
yeast topoisomerase II). A purified fraction of knotted
DNA with a high number of nodes (crosses) was found
to be refractory to transcription. The unknotting of the
knotted plasmids, using catalytic amounts of topoisomerase II, restored their capacity as templates for
transcription to levels similar to those obtained for the
other topological forms. These results demonstrate
that highly knotted DNA is the only topological form of
DNA that is not a template for transcription. We suggest
that the regulation of transcription, which depends on
the topological state of the template, might be related
to the presence of knotted DNA with different number
of nodes.
INTRODUCTION
It is widely accepted that the topological state of DNA has an
outstanding role in processes such as recombination, replication,
gene expression and DNA packaging (1–4). Since DNA is
supercoiled inside living cells, any passage of segments of DNA
within a topological domain by topoisomerases (5,6), site-specific
recombinases such as Tn3 resolvase (7), or during the integrative
recombination in phage lambda (1) leads to knotting or catenation
of the polynucleotide. Indeed, both the natural requirement for
powerful topoisomerases in living cells and for the packing of
DNA in viruses results in the knotting of DNA (8). The
arrangement of the tailless capsid of the bacteriophage P2 is such
that it leads to the formation of a complex knot (8). DNA knotting
obliterates chromatin assembly in vitro, though the presence of
knotted template does not inhibit the assembly over the relaxed
plasmid (9).
While several studies have examined the effects of knotted
DNA on site specific recombination and DNA replication, there
have been no studies which have analysed the effect of a knotted
DNA substrate on transcription. Notwithstanding, large differences
in transcription rates have been observed depending on the substrate
topology, other than those of the knotted form, which suggests
that the topology of DNA molecules has a fundamental role in the
transcription process in vivo and in vitro (10–12). Circular
plasmids appear to be more efficiently transcribed than linear
ones, and transcription is substantially favoured by the torsional
stress. Eukaryotic RNA polymerases prefer to elongate RNA on
templates which are found under torsional stress, using a minimal
set of transcription factors (13). There are reasons to believe that
the sequence dependent flexure of DNA has little influence over
the activity of the T7 promoter in Escherichia coli, but it has
incidence on transcription with the E.coli polymerase. Moreover,
the position of promoter sequences with respect to the confines of
a nucleosome displays large effects in the initiation or elongation
steps of transcription by the T7 RNA polymerase (10,14).
In this study, we have used a plasmid with different degrees of
topological knotting (obtained by treatment with stoichiometric
amounts of yeast topoisomerase II) and containing a T7 promoter
and transcription stop signal, to gain new insights into the effect
of this topological form on DNA transcription. The T7 RNA
polymerase was chosen since, unlike the eukaryotic RNA polymerase II or the E.coli RNA polymerase, it is a single-subunit enzyme
(15) that performs all the functions required for transcription,
including promoter recognition, initiation and termination, without
the need for auxiliary protein factors (15–17).
When highly-knotted plasmid, obtained by treatment of either
supercoiled or relaxed plasmid DNA with stoichiometric amounts
of topoisomerase II, was employed as a template for transcription
in vitro by the T7 RNA polymerase, there was a total absence of
transcript formation. However, when it was unknotted, by
treatment with topoisomerase II, it recovered its capability of
acting as a template for transcription.
MATERIALS AND METHODS
Enzymes and DNA
Topoisomerase II was purified from Saccharomyces cerevisiae as
described elsewhere (18). The plasmid pET14b, which contains the
T7 bacteriophage promoter and terminator signals, was purchased
from Novagen. T7 RNA polymerase and ribonuclease inhibitor
(RNasin) were purchased from Promega.
*To whom correspondence should be addressed. Tel: +34 3 400 6271; Fax: +34 3 204 5904
4891
Nucleic Acids
Acids Research,
Research,1994,
1996,Vol.
Vol.22,
24,No.
No.124
Nucleic
4891
Knotting and unknotting reactions
Knotting reactions were carried out as described previously (9)
using a topoisomerase II/DNA mass ratio of 4. The final
concentration of closed circular DNA (either relaxed or supercoiled)
was around 10 ng/µl. The time reaction varied from 1 h to overnight
and the reactions were stopped and deproteinized using
proteinase K and phenol extraction. Unknotting/relaxing reactions
with topoisomerase II were performed using smaller topoisomerase
II to DNA ratios (9).
Isolation of highly knotted DNA
In order to obtain a homogeneous population of highly knotted
DNA, ∼25 µg of pET14b were knotted as described above, run
for 2 h at 60 V on a 0.8% agarose gel, containing TBE Buffer
[90 mM Tris-borate (pH 8.3), 2 mM EDTA]. The highly knotted
DNA fraction was recovered (cut) from the gel and purified using
the Geneclean kit (Bio101, Inc.).
In vitro transcription assays
Transcription was performed in a buffer containing 40 mM
Tris–HCl (pH 7.5), 10 mM NaCl, 6 mM MgCl2, 2 mM
spermidine, 2 U RNAsin and 10 mM dithiothreitol; and 2.5 mM
NTPs (500 µM each of CTP, UTP, GTP and 1 mM ATP), together
with 50 ng of the different topological forms of plasmid pET14b,
and 5 U T7 RNA polymerase (Promega). The samples were
incubated at 37C for 30 min and stopped by addition of phenol.
The aqueous phase was ethanol precipitated and the template was
sometimes subsequently linearized by treatment with HindIII.
To analyze the presence of short (aborted) transcripts, we
performed additional experiments using radiolabelled RNA (20 µCi
[α-32P]UTP in the transcription mix), and denaturing polyacrylamide gel electrophoresis (see below).
Gel electrophoresis and quantitative analysis
The different topological forms obtained using topoisomerase II
(see above) were resolved on 0.8% agarose gels in the absence of
ethidium bromide. The products of the transcription in vitro were
analyzed on 2% agarose gels in TBE buffer. The gel and the
running buffer contained 0.5 µg/ml ethidium bromide. All the
gels were photographed under a UV light using Polaroid 665 film.
Radioactive transcripts were heated at 90C for 2 min prior to
electrophoresis on 6% polyacrylamide gels containing 7 M Urea.
Gels were subjected to autoradiography at 70C with an intensifying
screen.
Densitometric scans of the negatives were carried out in a
Molecular Dynamics densitometer to produce profiles from which
the relative intensity of the full length transcript was quantified.
RESULTS AND DISCUSSION
In this article we assess the relationship between the topoisomerase
II-mediated knotting of plasmid pET14b (which contains initiation
and termination signals for the phage T7 RNA polymerase) and
its usefulness as a template for transcription. Figure 1 shows that
the supercoiled form of the plasmid incubated with stoichiometric
amounts of yeast topoisomerase II became knotted. Each band of
the extended ladder (lane 2) corresponded to a population of knots
with a different number of crosses (nodes), in which the upper
band of the knotted series was the trefoil, the simplest knot.
Figure 1. Topoisomerase II produces knots in plasmid DNA with different
degrees of compactness. Supercoiled pET14b was knotted using topoisomerase
II at an enzyme to DNA mass ratio of four. In unknotting experiments, the
knotted DNA was treated with a smaller amount of enzyme (catalytic
conditions). The figure displays a 0.8% agarose gel stained with ethidium
bromide. Lane 1, 0.1 µg supercoiled pET14b plasmid; lane 2, 0.3 µg knotted
DNA; lane 3, 0.3 µg unknotted DNA; lane 4, 0.3 µg relaxed pET14b DNA; lane
5, 50 ng knotted DNA digested with HindIII. Forms I, II and III represent
supercoiled, relaxed and linear DNA respectively. Lane 2 shows that the
knotted species run electrophoretically displaying a broad range of mobility
according to its compactness.
On the other hand, the knotted DNA can be unknotted (relaxed)
by using smaller amounts of topoisomerase than those used in the
initial knotting experiments (9). Here, any relaxed DNA originating
from a previously knotted substrate is described as ‘unknotted’,
since the unknotted species run as a typically relaxed DNA during
electrophoresis (cf. lanes 3 and 4 in Fig. 1).
As with other topological forms, knots can show varying levels
of compactness. Thus, a minimum of three crosses were required
to produce the simplest knot (trefoil), and as the number of nodes
increases the compactness is progressively enhanced (6,9,19).
For the sake of comparison, Figure 1 displays the linear (form III),
supercoiled (form I) and relaxed (form II) forms of pET14b. The
linear form was obtained by treatment of a knotted sample with
HindIII.
Figure 2 shows the full transcripts (191 nucleotides long)
produced by the T7 RNA polymerase using pET14b as a template,
under the different topological states displayed in Figure 1. The
knotted template, Figure 2 lane 3, generated a marked decrease in
the RNA synthesis, while qualitatively similar amounts of transcript
were observed for the other ‘topological templates’. These results
suggest that knotted DNA acts as an impediment to transcription,
even for a highly processive enzyme such as the T7 RNA
polymerase (16). The different templates were also visible at the
top of the agarose gels.
Since this partial inhibition of the transcription procedure could
be related to the compactness of the template, we sought to gain
further insights into the inhibition of transcription by the phage
RNA polymerase by analyzing the transcription procedure using
a highly knotted template. Therefore, highly knotted plasmid
(Fig. 3, lane 4) was isolated from a preparative agarose gel as
described above (see also Material and Methods). The most highly
knotted fraction of the heterogeneous population (indicated by a
4892 Nucleic Acids Research, 1996, Vol. 24, No. 24
Figure 2. The knotting of the template interferes with the transcription in vitro.
The pET14b plasmid was prepared in the different topological forms shown in
Figure 1, and transcribed using T7 RNA polymerase (see Materials and
Methods) for 30 min. The figure displays a 2% agarose gel agarose containing
0.5 µg/ml ethidium bromide in both the gel and the running buffer. Both the
transcript and the template (∼50 ng) can be observed. Lane 1, supercoiled
template; lane 2, unknotted template; lane 3, knotted template; lane 4, relaxed
template; lane 5, linear DNA.
Figure 3. The highly knotted DNA fraction was purified from an agarose gel.
About 25 µg of supercoiled pET14b were knotted using yeast topoisomerase
II as described in Figure 1, and the highly knotted fraction (h-kn) was purified
from a 1% agarose gel (see Materials and Methods). Lane 1, supercoiled
pET14b plasmid; lane 2, relaxed DNA; lane 3, linear DNA (knotted sample
cleaved with HindIII); lane 4, highly knotted DNA as used in the purification
protocol; lane 5, unknotted DNA (knots treated with topoisomerase II under
catalytic conditions). No other topological forms than the highly knotted ones
were present in the sample in lane 4 used for the purification procedure. Forms
I, II and III represent supercoiled, relaxed and linear DNA respectively. Other
details as in legend to Figure 1.
bracket in Fig. 3) exhibited the highest electrophoretic mobility
concomitant with its dense compactness (9). Lane 5 in Figure 3
corresponds to knotted DNA treated with topoisomerase II. This
gel lane shows that after treatment with topoisomerase II, the
highly knotted species became untied DNA, as already observed
with the less-dense knots shown in Figure 1.
Figure 4 displays transcription experiments performed in vitro
with the templates displayed in Figure 3. It is evident, even to the
unaided eye, that highly knotted DNA was totally refractory to
transcription, which appeared to be blocked (Fig. 4, lane 4),
whereas the unknotted template (lane 5) recovered its capacity as
a template. The other topological substrates also rendered
qualitatively similar amounts of RNA transcripts. The differences
between the knotted substrate and the others are not due to any
kind of loading artefact, since at the top of the gel there are very
similar amounts of the HindIII-linearized template (after the
transcription assay) which acts as a recovering (loading) control.
Figure 4. T7 RNA polymerase cannot transcribe in vitro through the highly
knotted DNA template. The pET14b plasmid prepared in the different
topological forms shown in Figure 3, including the purified highly knotted
substrate (50 ng) were transcribed using T7 RNA polymerase (see Materials
and Methods) for 30 min. The figure displays a 2% agarose gel containing
0.5 µg/ml ethidium bromide in both the gel and the running buffer. Both the
transcript and the template (linearized with HindIII) can be observed. The
linearized template confirms that about the same amounts of DNA were used
in the transcription of each sample, and that all the gel tracks were loaded with
the same amount of material. Hence, the absence of transcript in lane 4 is
unambiguously the result of the inhibition of transcription by the presence of
highly knotted template. Other details as in legend to Figure 2.
Therefore, the differences in transcription between the two
populations of knotted DNAs (cf. lane 3 in Fig. 2 and lane 4 in
Fig. 4) might be understood if we consider the inability of T7
RNA polymerase to transcribe through the densely knotted
template, while this processive enzyme still worked on the other
knotted forms, in which the low compactness of the knots allowed
their displacement by the enzyme. T7 RNA polymerase has
previously been used to assay the ability of transcription to
overtake the blockage induced by obstacles as heterogeneous as
nucleosomes (14), triplex-forming oligonucleotides (20) or some
small ligands (21). In all these cases the location of the obstacle
in respect to the initiation site was the limiting factor that
influenced the ability of the polymerase to progress through any
of them, yet the amount of enzyme used can influence the
efficiency of transcription over some strongly phased nucleosomes
(14). However, for the knotted DNA (Figs 2 and 4) it seems that
the ability to overtake the topological obstacle (the nodes)
depends only on the compactness of knotting in the plasmid, since
the knots constitute a fairly mobile element rather than a fixed
translationally-phased stop to the activity of the enzyme.
Figure 5 shows the results of an experiment performed using
radioactive nucleotides. The full-length transcript is mostly absent
when a knotted template was employed, even in this over-exposed
autoradiography (cf. lanes 2 or 3 and lane 4 in Fig. 5). Using a
denaturing polyacrylamide gel and radiolabelled sample, we only
observe a few incomplete transcripts (Fig. 5, lanes 2 and 3), which
are not present in the transcription experiment performed using
knotted DNA as a substrate (lane 4 in Fig. 5). The absence of short
transcripts strengthens our interpretation that knots are dynamic
obstacles to transcription, rather than well-phased ones. Therefore,
aborted transcripts of well-determined size did not accumulate.
The results in Figures 2, 4 and 5 clearly suggest that highlycompacted knots in DNA are able to form effective blocks that
impede transcription, regardless of whether they are a dynamic
rather than a well-positioned obstacle. A quantitative comparison
of the relative amounts of the 191 nt transcript is displayed in
Figure 6. It shows that in both experiments using knotted and
highly-knotted DNA, the supercoiled, relaxed or linear DNA
produced almost the same quantities of transcript (cf. Figs 2 and
4893
Nucleic Acids
Acids Research,
Research,1994,
1996,Vol.
Vol.22,
24,No.
No.124
Nucleic
Figure 5. A high-resolution denaturing polyacrylamide gel of the transcription
by T7 RNA polymerase of different topological forms of the PET14b plasmid
DNA. Lane 1, labelled ΦX 174 DNA/HinfI marker (Promega). Lane 2, relaxed
template; lane 3, supercoiled template; lane 4, knotted template. The full-length
transcript is indicated by a black arrow. Only a few incomplete transcripts are
observed in this over-exposed gel (white arrows). They are not evident in the
transcription on the knotted template.
4). The knotted substrates were not a good template, and the
highly-knotted DNA was not transcribed at all. The unknotted
substrate recovered its role as template for the T7 RNA
polymerase, yet partially. The quantification of these transcripts
shows that the transcription falls ∼60% with the knotted DNA
(shown in Fig. 1). Using the highly knotted template, characterized
in Figure 3, the inhibition is near 99%. Since, as described in
detail in the introduction, knotted DNAs can be formed under
several biological contexts (1–3,7–9), we tentatively consider
that the ability of topoisomerase II to knot/unknot the templates,
so they can become substrates for transcription when required,
might have a crucial biological role.
Two key points emerge from our experimental approach. First,
to our knowledge, and in agreement with the results shown in
Figures 2, 4, 5 and 6, the knotted DNA template is the sole
topological form which cannot be transcribed in vitro. This
situation is likely to occur also in vivo, though it would be very
difficult to demonstrate, because of the potential ability of the
constitutive topoisomerase II to unknot the knotted substrate (19),
and the enormous difficulty involved in obtaining any viable
organism without this enzymatic activity. Here, we have shown
the inhibition of transcription using the highly processive T7
RNA polymerase. In fact, this enzyme is quite small (15)
compared to the larger eukaryotic transcription systems, which
we tentatively expect to be virtually incapable of transcribing
knotted templates, because of its size and multiprotein complexity
(22). Secondly, knotted DNA is easily untied to relaxed DNA.
When this occurs, transcription takes place as with any other
template (see Figs 2 and 4), without loss of DNA linkage. While
two enzymes, topoisomerases I and II, are needed for the various
topological manipulations of DNA in vivo, it is worth mentioning
that one of them, the yeast topoisomerase II, is able to produce
knotted plasmids with equivalent efficiency, using either supercoiled or relaxed DNA. The knotted DNA can be unknotted with
4893
Figure 6. Plot of the relative full-length transcription obtained using different
topological DNA templates. The quantification was performed from densitometric scans of Figures 2 (J) and 4 (j), as described in Materials and Methods.
While the transcription of the knotted template is clearly diminished and
depends on the density (amount) of knots, the other topological templates retain
about the same capacity of acting as templates in any experiment.
the same enzyme, though under different experimental conditions,
but while the knotted template does not transcribe efficiently,
depending on the level of knotting (Figs 2 and 4), the unknotted
template is transcribed. An equivalent effect of the knotting on
chromatin assembly has been described (9). In summary, our
results show how the presence of knots in DNA might decrease
and even abolish transcription from strong promoters, and how their
formation can strongly affect transcription from more complex
systems, such as eukaryotic RNA polymerases which are large
protein complexes (22).
In general, gene activation/repression is understood as being
controlled by proteins which work as trans-acting factors. Therefore,
repressors, activators and nucleosome displacement among other
events (12,22–24) have been well characterized. Nevertheless,
knotting/unknotting appears to be able to exercise fine tuning over
biological processes as replication, site-specific recombination or
DNA package (1–4), and the control of transcription depending
on the compactness of the knotted DNA, as described in this
article. The present results should be considered as a complementary
point of view towards the comprehensive understanding of the
effect of the sequence-dependent flexure of DNA and its topology
(2,10–12) on gene regulation in living cells.
ACKNOWLEDGEMENTS
We thank Drs F. Azor’n, J. Bernus and B. Piña for their help and
encouragement. This work was financed by grants from the
‘Dirección General de Enseñanza Superior’ (PB 95-0065) and the
DGICYT (PB 93/0102). The support of the CIRIT of the
‘Generalitat de Catalunya’ through its ‘Centre de Referencia en
Biotecnologia’ is also acknowledged.
REFERENCES
1
2
3
4
Spengler,S.J., Stasiak,A. and Cozzarelli,N.R. (1985) Cell, 42, 325–334.
Wasserman,S.A. and Cozzarelli,N.R. (1986) Science, 232, 951–960.
Fishel,R.A. and Warner,R.C. (1986) Virology, 148, 198–209.
Vologodskii,A.V. and Cozzarelli,N.R. (1994) Annu. Rev. Biophys. Biomol.
Struct., 23, 609–643.
4894 Nucleic Acids Research, 1996, Vol. 24, No. 24
5 Dean,F.B. and Cozzarelli,N.R. (1985) J. Biol. Chem., 260, 4984–4994.
6 Wasserman,S.A. and Cozzarelli,N.R. (1991) J. Biol. Chem., 266,
20567–20573.
7 Droge,P. and Cozzarelli,N.R. (1989) Proc. Natl. Acad. Sci. USA, 86,
6062–6066.
8 Liu,L.F., Perkocha,L., Calendar,R. and Wang,J.C. (1981) Proc. Natl. Acad.
Sci. USA, 78, 5498–5502.
9 Rodríguez-Campos,A. (1996) J. Biol. Chem., 271, 14150–14155.
10 Collis,C.M., Molloy,P.L., Both,G.W. and Drew,H.R. (1989) Nucleic Acids
Res., 17, 9447–9468.
11 Harland,R.M., Weintraub,H. and McKnight,S.L. (1983) Nature, 302,
38–43.
12 Van Holde,K. and Zlatanova,J. (1994) BioEssays, 16, 59–68.
13 Parvin,J.D. and Sharp,P.A. (1993) Cell, 73, 533–540.
14 Wolffe,A.P. and Drew,H.R. (1989) Proc. Natl. Acad. Sci. USA, 86,
9817–9821.
15 Sousa,R., Chung,Y. J., Rose,J.P. and Wang,B.C. (1993) Nature, 364,
593–599.
16 Sousa,R., Patra,D. and Lafer,E.M. (1992) J. Mol. Biol., 224, 319–334.
17 Macdonald,L.E., Durbin,R.K., Dunn,J.J. and McAllister,W.T. (1994) J.
Mol. Biol., 238, 145–158.
18 Worland,S.T. and Wang,J.C. (1989) J. Biol. Chem., 264, 4412–4416.
19 Liu,L.F., Liu,C.-C. and Alberts,B.M. (1980) Cell, 19, 697–707.
20 Skoog,J.U. and Maher,L.J. (1993) Nucleic Acids Res., 21, 4055–4058.
21 Portugal,J. (1995) Anti-Cancer Drug Des., 10, 427–438.
22 Tjian,R. and Maniatis,T. (1994) Cell, 77, 5–8.
23 Kornberg,R.D. and Lorch,Y. (1991) Cell, 67, 833–836.
24 Felsenfeld,G. (1992) Nature, 355, 219–224.