| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, September 27,
1994; and in revised form, January 4, 1995) From the
More than 70% of the RNA synthesized by T7 RNA polymerase during
run-off transcription in vitro can be incorrect products, up
to twice as long as the expected transcripts. Transcriptions with model
templates indicate that false transcription is mainly observed when the
correct product cannot form stable secondary structures at the 3`-end.
Therefore, the following hypothesis is tested: after leaving the DNA
template, the polymerase can bind a transcript to the template site and
the 3`-end of the transcript to the product site and extend it, if the
3`-end is not part of a stable secondary structure. Indeed, incubation
of purified transcripts with the polymerase in transcription conditions
triggers a 3`-end prolongation of the RNA. When two RNAs of different
lengths are added to the transcription mix, both generate distinct and
specific patterns of prolonged RNA products without any interference,
demonstrating the self-coding nature of the prolongation process.
Furthermore, sequencing of the high molecular weight transcripts
demonstrates that their 5`-ends are precisely defined in sequence,
whereas the 3`-ends contain size-variable extensions which show
complementarity to the correct transcript. Surprisingly, a reduction of
the UTP concentration to 0.2-1.0 mM in the presence of
3.5-4.0 mM of the other NTPs leads to faithful
transcription and good yields, irrespective of the nucleotide
composition of the template.
Run-off transcription in vitro using bacteriophage RNA
polymerases (Chamberlin and Ryan, 1982) was developed during the last
decade as an efficient method for the synthesis of biologically active
RNA (Melton et al., 1984; Krieg and Melton, 1984). Among the
known variants of this technique the systems based on T7 RNA polymerase
are the most widely used. This enzyme can be isolated in large amounts
from an overproducing Escherichia coli strain (Davanloo et
al., 1984), and a number of suitable plasmid vectors containing
the strong class III T7 promoters (Dunn and Studier, 1983) or synthetic
DNA (Milligan et al., 1987) can be used as templates. The
transition from small scale transcription assays as described in the
early reports (Chamberlin et al., 1970; Chamberlin and Ring,
1973) to large scale synthesis of RNA requires an increase in the input
of nucleoside triphosphates to the millimolar level (Milligan and
Uhlenbeck, 1989; Weitzmann et al., 1990). The optimal
concentration of nucleotides can depend on the particular template in
use (Milligan and Uhlenbeck, 1989), and it has been shown that
an excess of nucleotides (more than 4 mM each) can inhibit the
polymerization (Draper et al., 1988). In general, an upper
value of 2-4 mM for all four NTPs combined with a
magnesium concentration of 14-22 mM gives satisfactory
yields. The fidelity of the in vitro transcription process
is usually sufficient for low scale synthesis. However, non-coded 3`
prolongation of run-off transcripts has been reported as one or two
extra nucleotides in the case of synthetic DNA templates (Milligan et al., 1987), or as very long RNA chains in the case of
linearized plasmid templates with 3` overhanging ends (``snap
back'' effect; Schenborn and Mierendorf, 1985, Weitzmann et
al., 1993). Using in vitro run-off transcription with
T7 RNA polymerase, we found that more than 70% of the products could be
aberrant transcripts which were longer than the coded RNA (>15
additional nucleotides). We analyzed the reasons for this dramatic
mis-synthesis and observed that the processivity of the T7 RNA
polymerase is specifically very sensitive to the UTP concentration.
Under conditions where the initial concentrations of all four NTPs were
the same, a new phenomenon was revealed: after a faithful run-off
transcription, the T7 RNA polymerase can accept a former transcript as
second template and extend this RNA at the 3`-end in a self-coded
fashion, if the 3`-end is not folded into a stable secondary structure.
Reduction exclusively of the UTP concentration enables an efficient
synthesis of correct products but effectively prevents that of false
products.
The sequence of a
tRNA
The sequencing of the
27-mer product obtained from transcription of MF-mRNA was performed
using the phosphorothioate method of Gish and Eckstein(1988) with
slight modifications as described in Schatz et al.(1991). Five
transcription reactions were performed in parallel, in four of which
20% of one nucleotide was replaced by the corresponding
[
The gel analysis of these transcripts
showed a second anomaly which might be the reason for the mis-estimate.
The RNA synthesized ran in the sequencing gel as a smear starting at
the position of the full-length transcript and going to higher
molecular weights (Fig. 1A, lanes 4-6).
The track corresponding to the MF-mRNA transcript (Fig. 1A, lane 7) also showed a smear ranging
from the expected 46-mer to the region corresponding to about 75
nucleotides. The RNA obtained from the synthetic DNA templates
similarly showed a non-homogeneous electrophoretic pattern (Fig. 1C, lane 2). It is clear that the
assumption of exact transcription was not justified. Since the
nucleotide composition of the longer transcripts is not known, the
transcription efficiency calculation based on one radioactive
nucleotide is not reliable. In contrast, the unmodified tRNA
Figure 1:
Gel electrophoresis analysis of
different in vitro transcription products. 4-µl samples
from the indicated standard transcription assays were applied to
appropriate sequencing gels. After separation the gels were stained
with toluidine blue or ethidium bromide and, when indicated, exposed
for 12-48 h at -80 °C to an x-ray film. A, 10%
sequencing gel, ethidium bromide staining: lane 1,
0.16-1.77-kilobase RNA molecular weight standards; lane
2, tRNA
Kinetics of the MF-mRNA synthesis were
performed by subjecting the samples withdrawn from the transcription
assay at various times to gel electrophoresis and analyzing the
resulting autoradiogram by densitometry (Fig. 1B). At
the onset of incubation predominantly correct products were
synthesized, as evidenced by the main band product with the expected 46
nucleotides. However, the longer the incubation for transcription lasts
the more false, overlong products are produced, reaching a comparable
level with the expected product after 180 min of incubation (Fig. 1B). The apparent lag period between the
incubation start and the efficient synthesis of high molecular weight
products indicates that the synthesis of longer RNA is a secondary
process which needs an induction time. This induction time appears to
be inversely related to the speed of RNA synthesis (i.e. with
more enzyme the lag period is smaller, data not shown) and seems to be
different for every template. These observations suggest that a minimal
concentration of transcripts is required for the onset of aberrant
transcription.
The template for the
decacodon LS-mRNA offered a convenient system to test the importance of
secondary structure at the 3`-end of the correct transcript with regard
to the quality of transcription. Upon restriction with BamHI,
this plasmid serves as a template for the synthesis of the long form of
the decacodon LS-mRNA, which contains a poly(A) tail of 59 adenyl
residues and is thus free of secondary structure at the 3`-end (Fig. 2A). When restricted with BglII the
plasmid codes a short form of the mRNA which lacks the poly(A) tail and
carries a secondary structure motif at the 3`-end (Fig. 2B). The results of parallel transcription assays
with both forms of the template were striking: products with 3`-end
secondary structure show one main product corresponding to the expected
length practically free of longer transcripts (Fig. 2B), whereas large amounts (
Figure 2:
Predicted secondary structures and in
vitro transcription with two forms of the pTDLS template. The
optimal predicted secondary structures and their
One possible explanation for the
appearance of high molecular weight products is that after running off
the T7 RNA polymerase can rebind the previously synthesized RNA and
continue the RNA synthesis thus making longer transcripts. A
prerequisite for an RNA to function in that way is a loose 3`-end free
of secondary structure. It is not yet clear at this point whether one
RNA molecule can serve as template for the prolongation of another RNA
molecule or the template RNA is prolonged at its own 3`-end
(self-coding; Fig. 3C).
Figure 3:
RNA dependent in vitro transcription with T7 RNA polymerase, MF-mRNA as template. A, autoradiograms from 15% sequencing gels. Left
panel, lanes 1-3: 125, 250, and 500 pmol of
non-labeled MF-mRNA, respectively, were incubated in 25 µl of
standard transcription mix containing 3.75 mM of each ATP,
GTP, CTP, and [
When a small RNA practically
lacking predictable secondary structure was used as template (MF-mRNA,
for sequence see Table 1) the polymerization was also efficient
(up to 1600 nt ( In order to determine the sequence of the strong band with about 27
nt, the transcription was repeated, but now allowing for random
incorporation of a few phosphorothioated nucleotides at either A, G, C,
or U positions. The oligonucleotides corresponding to the 27-mer were
isolated, labeled at their 3`-ends with
Figure 4:
The sequence of the short 27-nt transcript
of the MF-mRNA (46 nt). The sequence was assessed with the
phosphorothioate method (Gish and Eckstein, 1988). A, G, U, and C, transcripts containing the
respective phosphorothioated nucleotides. O, transcript
without phosphorothioated nucleotides but otherwise treated as the
thioated transcripts. Bottom, the sequence of the transcript
is compared to that of the MF-mRNA used as a template. For details see
``Experimental Procedures.''
The fact that the addition of the non-labeled
MF-mRNA yields two more or less defined product families (Fig. 3A), one shorter (about 27 nt) and one longer
(50-80 nt) than the added MF-mRNA with 46 nt, could be explained
as follows. The MF-mRNA binds around its 27th nucleotide to the
template site of the T7 RNA polymerase. Then either de novo synthesis (Fig. 3C, left) occurs yielding
the 27-nt long products; in other cases the polymerase prolongs added
MF-mRNAs for about 27 nucleotides yielding the 75-nt products. It is
possible that a template mRNA serves to prolong its own 3`-end at the
later process. In the next experiment we wanted to analyze the fate
of the input MF-mRNA, which therefore carries a A rigorous test of this
view is made in the next experiment. [
After transcription under standard
conditions, using [
Figure 5:
The sequence of the T7 transcripts is
prolonged at the 3`-end. 5 pmol of the MF-mRNA DNA template (pTMF) were
incubated in standard conditions (all NTPs at 3.75 mM) in the
presence of [
A second
parameter is the concentration of T7 RNA polymerase. With the template
for the decacodon LS-mRNA used in the experiment shown in Fig. 6, one should perform the reaction with an enzyme
concentration of less than 10 µg/ml to avoid most of the smear of
higher molecular weight products. However, adjusting the enzyme
concentration for every individual template in order to achieve precise
transcription is not always advisable, since for a number of templates
the enzyme concentration had to be reduced below 4 µg/ml, which is
too low for preparative purposes.
Figure 6:
The concentration of T7 RNA polymerase
affects the quality of the transcripts. 1 µg (0.5 pmol) samples of
the decacodon LS-mRNA DNA template (pTDLS) were incubated in 25 µl
of standard transcription mix containing [
The third parameter was a
surprise. A strong and specific effect of the UTP concentration on the
accuracy of transcription was found. Fig. 7A shows the
transcription of the MF-mRNA (46 nucleotides) under conditions where
the concentration of the nucleotides was systematically changed (one by
one) in a range covering preparative and analytical conditions (3.75
mM to 188 µM). The autoradiogram reveals that a
reduction of ATP or GTP lowers the efficiency of transcription with
apparently no effect on the proportion of correct/incorrect products.
The reduction of CTP seems to have no effect in the range of
concentrations used, probably due to the fact that the transcript has
only 1 cytosine. In sharp contrast, reduction of UTP predominantly
reduces the synthesis of incorrect products. When we evaluated the gels
quantitatively (Fig. 7B), the following results were
obtained. At the lowest UTP concentration applied (188 µM)
the most accurate transcription was observed and the correct
transcripts amounted to about 80% of the total transcribed products,
whereas the maximal amount of correct products (100%) was found at 0.7
mM UTP. Variation of the CTP concentration did not affect the
accuracy of transcription; only 25% of the total transcripts were
correct, whereas the yield of correct transcripts showed a shallow
maximum at 0.7 mM CTP reaching 86% of the best yields found in
the UTP series. Reduction of the ATP or GTP had a disastrous effect.
The yield of correct transcripts went down to 3%, whereas the synthesis
of incorrect transcripts increased to over 90%.
Figure 7:
Influence of the concentrations of
nucleotides on the accuracy of the transcription. 0.5 pmol of the
MF-mRNA template (pTMF) were incubated in 25-µl aliquots under
standard conditions adjusting the concentration of one nucleotide at a
time to the following values: 3.75 mM (standard conditions);
1.4 mM; 0.7 mM; 0.35 mM and 0.188
mM. 17 parallel incubations changing the concentration of ATP,
GTP, CTP, and [
Qualitatively, the
same results were observed with various plasmid templates and synthetic
DNA templates coding for transcripts with weak secondary structure at
the 3`-end and with different nucleotide compositions For example, the
oligo(DNA) containing the fragment of a tRNA gene corresponding to
positions 1-40 of tRNA The run-off in vitro transcription system with T7
RNA polymerase has been used for the synthesis of short (less than 100
nucleotides; Groebe and Uhlenbeck, 1988; Sampson and Uhlenbeck, 1988;
Milligan and Uhlenbeck, 1989) and long RNAs (e.g. 16 S and 23
S rRNA; Nègre et al., 1989; Weitzmann et al., 1990). The transition from small scale to large scale
assays required some changes of the transcription systems to improve
the yield of synthetic RNA. Major changes are the inclusions of
ribonuclease inhibitor and inorganic pyrophosphatase, which
significantly improve the quality and the yield of products (Cunningham
and Ofengand, 1990); minor improvements are the additions of
polyethyleneglycol and Triton X-100 (Milligan and Uhlenbeck, 1987).
Furthermore, the concentrations of nucleotides have to be increased,
and it is normally assumed that an increase in the nucleotide
concentration has no major effect on the quality of the RNA
synthesized. However, it is frequently seen that the synthesis of
full-length transcript is accompanied by production of higher and lower
molecular weight products during large scale transcription assays using
T7 and also other bacteriophage RNA polymerases (Milligan et
al., 1987, Ling et al., 1989, Paddock, 1989). The
transcription system previously optimized by us for the synthesis of a
model mRNA coding for a decapeptide (see ``Experimental
Procedures'') was used to test the templates for different
heteropolymeric mRNAs. The system works efficiently with respect to the
incorporation of ribonucleotides in RNA, but the fidelity of the
synthesis is far from optimal. In all cases presented in Fig. 1,
the synthesis of the expected transcripts was accompanied by a
significant production of RNA with higher molecular weights than
expected, the only exception being the transcription with the
tRNA Several observations reported in the literature as well as the
results presented here led to a possible explanation for the aberrant
transcription. First, it is known that the T7 RNA polymerase (as well
as many other RNA polymerases) can accept RNA as template (Chamberlin
and Ring, 1973; Kornaska and Sharp, 1989, 1990). Second, as shown in Fig. 1B, the production of high molecular weight RNA
appears to be a late event during the transcription incubation. Third,
in all cases where the phenomenon is visible the expected transcript
has a 3`-end region which appears to be free from stable secondary
structure (Fig. 2). Fourth, the presence of two different RNAs
both with 3`-ends lacking stable secondary structures triggers a
pattern of overlong products independently from each other. And
finally, sequencing of the high molecular weight RNA revealed that it
consists of a 5` portion identical to the expected transcript plus a
variable 3` prolongation, which shows apparent complementarity with the
transcript itself (Fig. 3D). It follows that the
synthesis of aberrant transcripts is not related to abnormal
transcription initiation. The problem starts when the
``run-off'' process is completed. The data suggest the
following. During the run-off transcription and after the production of
a certain amount of transcript, the RNA can serve as substrate for a
self-coded 3`-end prolongation. A certain amount of transcript is
required because the RNA has to compete with the DNA for the template
site, which has a higher affinity for DNA. This explains the lag of
appearance of false transcripts in the kinetics shown in Fig. 1B. In order to be fully efficient in this
process, the RNA must have a 3`-end free of secondary structures. This
means that at sufficient concentrations of RNA the T7 RNA polymerase
swap from a DNA to an RNA template. A stable secondary structure at the
3`-end of a transcript obviously prevents the binding to the product site and possibly also to the template site
of the polymerase ( Fig. 2and Fig. 3C). Since a
secondary structure involving the 3`-end will form only after the
release of the respective transcript, the RNA has probably to be
released before it can occupy the template site and with its
3`-end the product site of the T7 RNA polymerase. After the
self-coding process, the overlong products have now 3`-ends which are
self complementary, are therefore involved in stable secondary
structures, and thus cannot serve as template again preventing a
further prolongation. This explains why overlong products have a length
of less than twice that of the correct transcript. Our observations
concerning the importance of a secondary structure at the 3`-end of a
transcript agree well to what is known about the requirements for
termination. T7 RNA polymerase terminates efficiently in vitro at a G:C rich hairpin/U stretch structure (Dunn and Studier,
1983). The hairpin formation is important for disrupting DNA
(template)-RNA (transcript) interactions as well as interactions
between RNA and polymerase (Sousa et al., 1992, and references
therein). Three parameters of the transcription assay were shown to
affect the relative amount of aberrant transcription: the time of
transcription incubation (Fig. 1B), the stoichiometry
of T7 RNA polymerase used relative to the DNA template (Fig. 6),
and the concentrations of nucleotides. Fig. 6demonstrates that,
when the amount of enzyme surpasses a critical level, the synthesis of
``smeared'' products with a high molecular weight is
exacerbated. It follows that an adjustment of the enzyme amount could
be sufficient for a synthesis with tolerable amounts of secondary
products. However, the enzyme reduction is not a practical solution in
general, since the enzyme concentration must be reduced to inefficient
low levels for many templates. Since higher concentration of polymerase
leads to enhanced synthesis resulting in high concentration of RNA in
the transcription assay, a transcript-coded transcription is favored
under these conditions. Thus, lowering the polymerase concentration
improves the transcription quality not by influencing the mode of
action of the polymerase, but rather by lowering the synthesis of a
possible RNA template and therefore does not provide a real
improvement. According to the results in Fig. 7, manipulation of
the nucleotide concentrations dramatically affects the production of
high molecular weight RNA. The reduction of purine nucleotides and CTP
caused a decrease in the yield of transcripts without improving the
fidelity of transcription. In contrast, when the UTP concentration was
decreased, a strong reduction in the synthesis of high molecular weight
RNA was observed without a severe effect on the yield of the expected
transcript. The same phenomenon was observed with different templates
coding for RNAs with unrelated sequences and nucleotide composition
(with or without homopolymeric tracks) but all preserving the lack of
secondary structure at the 3`-end. Since there is no clear
correlation between the composition of the transcripts and the effects
of UTP, the data suggest that at high concentrations the UTP is an
allosteric effector of the T7 RNA polymerase inducing a conformation
which facilitates the acceptance of RNA as a template. This hypothesis
is still speculative but has some support from the following
observations. In analytical assays using micromolar concentrations of
all four nucleotides, it has been shown that a reduction of the
pyrimidines, especially the UTP, triggers an increased production of
abortive transcripts (Ling et al., 1989) indicating loose
enzyme-RNA complexes, i.e. the processivity of the enzyme is
reduced when the UTP concentration is lower than that of the other
nucleotides. A reduced processivity at decreased UTP conditions would
certainly counteract the tendency to self-coded transcript
prolongation. It is not clear whether RNA-coded de novo synthesis or self-coded 3`-end extension events occur in
vivo. However, the intracellular molar ratio of ATP/UTP in E.
coli is similar to that which we found optimal for an accurate
run-off transcription (the ATP concentration in vivo is about
3 mM, while that of UTP is about 0.8 mM; Neuhard and
Nygaard, 1987), suggesting that these types of activities of the T7 RNA
polymerase are not common events (or constitute part of well controlled
processes) inside the living cell. In vitro, a reduction of
the UTP concentration to 0.2-1.0 mM while keeping the
other nucleotides in the range of 3.5-4 mM (here the
base composition of the expected transcript should be taken into
account) is generally applicable and improves the fidelity of
transcription, as well as giving a high yield of correct RNA
synthesized with T7 RNA polymerase.
Volume 270,
Number 11,
Issue of March 17, 1995 pp. 6298-6307
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
Materials
Deoxynucleoside triphosphates as well
as all the restriction enzymes (except SspI), calf intestine
alkaline phosphatase, T4 polynucleotide kinase, T4 DNA ligase, poly(U)-
and oligo(dT)-cellulose were purchased from Boehringer Mannheim. SspI and the Klenow fragment of E. coli DNA
polymerase were from New England Biolabs and mung bean nuclease from U.
S. Biochemicals. The ribonucleoside triphosphates and the inorganic
pyrophosphatase were from Sigma. Plasmid pSP65 and RNasin (40
units/µl) were obtained from Promega Biotec,
[
P]UTP, [P]CTP,
[P]ATP, and [P]GTP from
Amersham. The oligodeoxyribonucleotides were synthesized on an Applied
Biosystems 380B DNA synthesizer and purified by high performance liquid
chromatography on a Hypersil ODS 5-C18 column. Chemical synthesis of
RNA was performed on an Applied Biosystems 392 DNA/RNA synthesizer
using phosphoramidite chemistry (Scaringe et al., 1990). The
plasmid pTZDec-m, a derivative form pTZ18R (Pharmacia) coding for the
decacodon O-mRNA (see Table 1), was a kind gift from Drs. H.-J.
Rheinberger and B. Lewicki, MPI Berlin. The T7 RNA polymerase was
isolated from the E. coli strain BL21 containing the plasmid
pAR1219, according to published procedures (Davanloo et al.,
1984).
Construction of Plasmid DNA
Templates
Oligodeoxynucleotides were synthesized containing the
sequence to be cloned (50-60 nucleotides, see corresponding RNA
sequences in Table 1). A short complementary oligo(DNA) served as
a primer for double-stranded DNA synthesis (Cobianchi and Wilson,
1987). The sequence corresponding to the O-mRNA (Table 1) was
deleted by successive digestions of pTZDec-m plasmid with EcoRI, mung bean nuclease, and HpaI, respectively;
and the 5`-phosphorylated double-stranded sDNAs were inserted into the
pTZDec-m plasmid by blunt-end ligation.
gene from E. coli together with a T7
promoter sequence was cloned at the SmaI site of the plasmid
pSP65. The double-stranded sDNA insert was obtained in the following
way. Two oligo(DNA)s were synthesized, the first contained the sequence
of the T7 promoter and the sequence corresponding to that of
tRNA from nucleotide 1 to 40, the second contained the
sequence complementary to that of tRNA from position 22
to 76 (18 nucleotides are complementary to the first oligo) plus a BstNI site for linearization. The recombinant plasmids were
used for transformations of E. coli DH5 according to standard
procedures. The positive clones were sequenced in both directions in
the region of the insert.Large Scale Plasmid Preparations
Large scale
preparation of plasmid was done by means of a protocol combining
alkaline lysis (Birnboim, 1983) and polyethylene glycol precipitation
(Lis, 1980), followed by anion-exchange chromatography purification on
QIAGEN columns. This combination of methods yields about the same
amount and quality of plasmid DNA as the standard protocols (Sambrook et al., 1989) and has the advantage of not requiring the use
of RNase A during the preparation, thus avoiding the risk of
contamination of the transcription templates. The run-off transcription
templates were then prepared by incubation of the purified plasmids
with the appropriate restriction enzymes.In Vitro Transcription Assays
The assays were
performed using a set of conditions optimized according to previous
reports (Milligan and Uhlenbeck, 1989; Weitzmann et al.,
1990). The standard analytical assays were performed in 25 µl of a
transcription mix containing 40 mM Tris-HCl, pH 8.0 (37
°C), 22 mM MgCl
, 1 mM spermidine, 5
mM dithioerythitol, 100 µg/ml of bovine serum albumin
(RNase and DNase free), 1 unit/µl ribonuclease inhibitor (RNasin),
5 units/ml of inorganic pyrophosphatase, 0.5 pmol (
1 µg) of
linearized plasmid DNA template (or 1 pmol of synthetic DNA templates),
3.75 mM (or the indicated values) each of ATP, GTP, CTP, and
UTP, and 40 µg/ml (or the indicated amount) of purified T7 RNA
polymerase. The components were mixed at room temperature and incubated
at 37 °C for the times indicated (3 h standard). The transcription
products were analyzed by cold trichloroacetic acid precipitation in
order to estimate the extent of reaction (when a P label
was included), and by electrophoresis in denaturing polyacrylamide gels
to determine the quality of the RNA synthesized. The electrophoretical
patterns were visualized by autoradiography and/or via staining with
ethidium bromide or toluidine blue. The synthetic DNA templates were
assembled by annealing a 17-mer oligodeoxynucleotide with the sequence
of the region -17 to -1 of the consensus class III T7
promoter to a template DNA strand containing complementary sequence of
the promoter followed by that of the RNA to be synthesized. In assays
where purified RNA was used as sole template, it was added in amounts
ranging from 0.25 to 20-fold molar excess over polymerase.Transcript Purification
The transcription products
of the plasmids pTZDec-m, pTDMS, and pTDLS (linearized with BamHI) were purified after phenol-chloroform extraction of the
transcription mix by chromatography on oligo(dT)-cellulose (Sambrook et al., 1989) followed by ion-exchange chromatography on
QIAGEN columns. Transcripts obtained from the pTMF template (linearized
with SspI) or synthetic DNA templates were purified by
polyacrylamide gel electrophoresis under denaturing conditions.RNA Sequencing
The run-off transcript of the
plasmid pTZDec-m was sequenced with the chain termination method (Hahn et al., 1989) using avian myeloblastosis virus reverse
transcriptase (from Pharmacia-PL Biochemicals) and a synthetic DNA
primer. The transcripts derived from the pTMF template were directly
labeled at the 5`-end by including [
P]GTP
in the in vitro transcription reaction. The sequences were
determined with the enzymatic sequencing method using the RNA
Sequencing Kit from Pharmacia-PL Biochemicals.
S]NTP. The fifth reaction was lacking
[S]NTPs, and the corresponding product was used as
control for detecting spontaneous chain disruptions. The transcription
reactions were performed under standard conditions (see above) in a
total volume of 50 µl containing 200 pmol of MF-mRNA as template
and 20 pmol of T7 polymerase. The transcription products were separated
on a polyacrylamide gel, the bands visualized by UV shadowing, and the
transcripts corresponding to the ``27-mer'' band were
extracted. After 5`-labeling, 10 µl (200,000
disintegrations/min) were treated with 1 µl of a 10 mM J solution in ethanol, incubated for 1 min at room
temperature, precipitated with ethanol, and analyzed by polyacrylamide
gel electrophoresis.Secondary Structure Predictions
The secondary
structure of minimum free energy for the synthetic RNA molecules was
determined with the program FOLD (Zuker and Stiegler, 1981) using the
energy parameters defined by Turner and co-workers (Freier et
al., 1986), and/or the program MFOLD (Zuker, M., 1989).
Transcription under Standard Conditions
The
startling observation was derived from analytical in vitro transcription assays performed with templates designed for the
synthesis of small model mRNAs devoid of strong secondary structure
(see Table 1). Under standard conditions (all NTPs at 3.75
mM), the products obtained with the plasmid templates revealed
a surprising phenomenon. In the cases of pTZDec-m, pTDMS and pTDLS
(decacodon template series) the calculated transcription efficiency
(RNA copies synthesized per template) suggested more than quantitative
utilization of the ATP added to the assay, if one assumes exclusive
synthesis of the expected transcript; the calculated ATP consumption
was 12-15% higher than the total input of non-labeled ATP (the
label was incorporated as an [P]NTP different
from ATP in these cases). transcript consisted of a single product of the right size (Fig. 1C, lane 4). At first glance it appeared
that the partial homopolymeric nature of the model mRNAs could be the
cause of the aberrant transcription as indicated in previous reports of
``slippage'' of T7 RNA polymerase on poly(A) tracks (Milligan
and Uhlenbeck, 1989). However, a second series of templates with
balanced base composition and lacking homopolymeric tracks gave similar
results (data not shown).
(0.02 A
units); lane 3, 5 S rRNA (0.02 A units); lane
4, pTZDec-m transcript; lane 5, pTDMS transcript; lane 6, pTDLS transcript; lane 7, pTMF transcript.
Autoradiogram (lanes 4-7). B, scans of an
autoradiogram. 1 pmol (2 µg) of SspI-restricted pTMF was
incubated at 37 °C in 50 µl of standard transcription mix
containing [P]ATP (30 cpm/pmol) with T7 RNA
polymerase at a final concentration of 4 µg/ml. At the indicated
times (0, 15, 30, 60, and 180 min), 5-µl samples were taken, frozen
in liquid nitrogen, and dried under vacuum. The material was then
dissolved in 5 µl of RNA sample buffer and run in a 15% sequencing
gel. The quality of the RNA synthesized was analyzed by densitometry of
the corresponding autoradiograms (B, left panel). The
area in arbitrary units, (A.U.) from the peaks corresponding
to the main band (*) and the longer RNAs, 
, respectively, are
plotted against the time of incubation (B, right). C, 15% polyacrylamide-urea gel. Toluidine blue staining: lane 1, 5 S rRNA (0.02 A units); lane 2, 14-mRNA transcript (65 nt), lane 3, pTDLS
transcript; lane 4, pSTtPhe transcript (76 nt); lane
5, tRNA from E. coli (0.03 A units).
Secondary Structure at the 3`-Ends of Products and
Aberrant Transcription
The observation that some templates
reproducibly yielded exact products whereas others provided large
amounts of overlong products was a surprise. A possible reason could be
seen in the fact that the transcripts systematically differed in their
capability to form secondary structure at the 3`-end region. For
example, unmodified tRNA is obviously able to form the
known secondary structures of the wild type tRNA since the transcript
was functional in aminoacylation assays and in poly(Phe) synthesis with
poly(U) programmed 70 S ribosomes from E. coli (data not
shown). It follows that the 3`-end of the tRNA transcript
should have only four nucleotides which are not involved in secondary
structure. In contrast, the other transcripts listed in Table 1(except the short form of the decacodon LS-mRNA) show a
low overall tendency to form double helices and their 3`-end regions
are not involved in stable secondary structures (as predicted with the
programs FOLD and MFOLD; data not shown).60% of the
toluidine blue-stainable material) of prolonged false transcripts were
found with the RNA containing the poly(A) tail (Fig. 2A).
G
were determined for the
long (A) and the short (B) forms of the decacodon
LS-mRNA (see Table 1) using the program MFOLD (S.D. = Shine-Dalgarno sequence). The scans of autoradiograms
correspond to analytical assays where 0.5 pmol of pTDLS digested with BamHI (A) or BglII (B) were
incubated in standard transcription conditions during 3 h at 37 °C.
5-µl samples of the transcription mix were applied to a 10%
sequencing gel, and after separation the nucleic acids were stained
with toluidine blue.
P]UTP (20 cpm/pmol) and 25 pmol
of T7 RNA polymerase during 4 h at 37 °C. Lane 4 corresponds to a control incubation omitting the MF-mRNA. Right panel, lanes 6-9: 250, 62, 25, and 12.5
pmol of 5`-P-labeled MF-mRNA (390 cpm/pmol), respectively,
were incubated as indicated above in the presence of non-labeled NTPs. Lane 5 corresponds to a control incubation containing 25 pmol
of [P]MF-mRNA and omitting the T7 RNA
polymerase. B, left panel, autoradiogram from 10%
sequencing gels. Lanes 1-4, 100 pmol of
5`-P-labeled MF-mRNA (260 cpm/pmol) were incubated in
standard transcription conditions with non-labeled NTPs, 100 pmol of T7
RNA polymerase, and 0 (no addition), 50, 100, and 200 pmol of
gel-purified non-labeled decacodon O-mRNA, respectively. Lane 5 corresponds to a control incubation omitting NTPs and decacodon
O-mRNA. Right panel, ethidium bromide staining. Lane 6 corresponds to lane 5, and lanes 7-10 correspond to lanes 1-4 in the autoradiogram (left panel); lane 11, untreated decacodon O-mRNA. C, three models for RNA-coded RNA
synthesis.
Self-coded Transcript Extension
The possibility of
using RNA as template for the DNA dependent T7 RNA polymerase has in
fact already been described by Chamberlin and Ring(1973) and Kornaska
and Sharp (1989), among others. We confirmed that poly(U) could indeed
be used as template, yielding a rate of polymerization of 0.3-0.5
nucleotides/s and enzyme (at saturation; data not shown), which is
about 100-fold slower than in the presence of a DNA template under the
conditions used. Furthermore, the fact that for detectable
transcription the concentration of RNA template must be approximately
20 times higher than that of DNA templates indicates a higher affinity
of the enzyme for DNA templates.
)polymerized/µg of T7 RNA polymerase in
standard conditions, data not shown). The products of a transcription
incubation including [P]UTP as label were
analyzed by gel electrophoresis. The corresponding autoradiogram (Fig. 3A, lanes 1-3) showed no products
with the length of the starting template (46 nt) but rather a group of
short aberrant products 5-26 nucleotides long (about 20% of the
synthesized RNA); a strong band corresponding to a 27-28-mer
(40% of the total RNA) is also present, suggesting that the RNA is
mainly used as template at about the middle of the sequence, and
relatively large amounts of long transcripts (40%) also appear
which are longer than the starting template. Similar results with
respect to nucleotide incorporation and size distribution of the
transcripts were obtained using chemically synthesized RNAs having
totally unrelated sequences but preserving the size and the lack of
predictable secondary structure as common features (data not shown). P, and then
cleaved with iodine at the phosphorothioated positions and separated on
a sequence gel (Gish and Eckstein, 1988). Fig. 4demonstrates
that the oligonucleotides were complementary to the 5`-end of the
MF-mRNA. The prolongation of the transcripts exceeded the 23 nt
identified on the sequence gel, since the last one (3`-C, see Fig. 4) and the first 2-3 nucleotides (AAG-5`)
could not be detected for technical reasons. For convenience we call
the respective band the ``27-nt'' product. The fact that a
clean sequence pattern is obtained indicates that the 27 nt band
contains one defined transcript. It is therefore clear that the 27-nt
oligomer is the product of a defined de novo synthesis
starting at a position which is about 19 nucleotides from the 3`-end of
the template MF-mRNA.
P label at
the 5`-end. The transcription was performed with non-labeled NTPs (Fig. 3A, lanes 5-9). The radioactivity
was distributed between the original mRNA band (see lane 5)
and a smear of overlong products of up to 80 nt comprising about 60% of
the original input (determined by densitometry of the autoradiogram and
by direct counting of gel slices) and showing a pattern of the main
products (Fig. 3A, lane 6) very similar to
that observed when a pTMF DNA template is used. These results again
indicate that the T7 RNA polymerase uses the RNA as transcription
template and is also able to prolong this RNA at the 3`-end.
Interestingly, the input of RNA template, which varied from 10- to
0.5-fold molar excess over polymerase, had practically no effect on the
pattern of 3`-end prolonged products (Fig. 3A, lanes 6-9). The fact that the same pattern is seen even
at the lowest input of MF-mRNA is in favor of the view that the
prolongation process is a self-coding mechanism (Fig. 3C, right).P]MF-mRNA
(46 nt) is mixed with various amounts of non-labeled decacodon O-mRNA
(143 nt). Both mRNAs can trigger different patterns of overlong
products in a transcription assay. A self-coding mechanism (Fig. 3C, right) will show in the mixture
experiment the two distinct patterns without any interference, whereas
the possibility that ``one RNA molecule serves as template for the
prolongation of a second RNA molecule'' (intermolecular mechanism; Fig. 3C, center) implies a significant change
in the pattern of radioactive prolonged RNAs. Fig. 3B, lanes 1-4, shows that increasing amounts of decacodon
O-mRNA do not at all interfere or disturb the distribution of the
overlong products derived from MF-mRNA. Ethidium bromide staining of a
second sequencing gel reveals that the decacodon O-mRNA also triggered
its distinct pattern of prolonged RNA (Fig. 3B, lanes 7-10) independently of the presence of MF-mRNA. It
follows that the 3`-end prolongation is most probably an
intramolecularly coded, i.e. a self-coded process.Sequencing the Products
According to the
conclusions of the preceding section the extended transcripts must all
have a defined 5`-end but various 3` extensions, which should be
complementary to the RNA template. This expectation was checked by
sequencing the transcripts produced in a preparative assay (3.75 mM NTPs were used in these cases) and purified as described under
``Experimental Procedures.'' The DNA template for the MF-mRNA
was selected for this purpose.P]GTP as label and pTMF (SspI restricted) as template, the full-length transcript and
one sample from the area containing products with a higher molecular
weight were isolated after polyacrylamide gel electrophoresis (Fig. 5A) and sequenced using enzymatic techniques. The
sequencing patterns (Fig. 5, B and C)
confirmed the expected sequence of the full-length transcript and
revealed that the secondary bands (in a range of 50-70
nucleotides) contain a 5`-end with a sequence identical to the correct
transcript (Fig. 5B) and 3` prolongations with apparent
complementarity to the correct transcript (Fig. 5, C and D). It follows that the aberrant transcripts are not
the result of false transcription initiation or errors in the
elongation process but rather of an extended synthesis (Fig. 5D). Sequencing of the heterogeneous transcripts
from the template for the O-mRNA ( Table 1and Fig. 1A, lane 4) also revealed defined 5`-ends
expected for the correct transcript (data not shown).
P]GTP (40 cpm/pmol). The
transcription products were then separated in a 17% sequencing gel
(autoradiogram in A) from which two sections were cut and the
RNA extracted. One section corresponded to the position of the expected
transcript (46 nt; B) and the other corresponded to an RNA
about 20 nt longer (C). The sequences of these two RNAs were
determined enzymatically showing that the 46-nt long RNA corresponds to
the expected transcript (B), and the longer fragment contains
an identical 5` region followed by a 3` prolongation (C) which
shows complementarity to the expected transcript. D,
self-coding model showing the complementarity between the MF-mRNA
sequence and the 3` extension.
What Avoids Self-coded 3`-End Prolongation of
Transcripts?
We have identified three parameters which influence
the extent of false transcription. One parameter already mentioned is
the time of transcription incubation. At the beginning predominantly
correct products are produced, whereas false transcripts accumulate at
the end of long incubations (Fig. 1B).
P]ATP
(2 cpm/pmol) with the indicated amounts of T7 RNA polymerase
(2.5-40 µg/ml) during 3.5 h at 37 °C. Samples of 6 µl
from every incubation were applied to a 10% sequencing gel. Lane
1, 0.16-1.77-kilobase RNA molecular weight standards; lane 2, tRNA (0.02 A units); lane 3, 5 S rRNA (0.02 A units); lanes 4-8, incubations with 40, 20, 10, 5,
and 2.4 µg of T7 RNA polymerase/ml of incubation mix, respectively.
The autoradiogram shows lanes
4-8.
P]UTP (2 cpm/pmol) as well as a
control omitting the plasmid were performed. After 3.5-h incubation at
37 °C the reaction was stopped, and samples from every incubation
were applied to a 10% sequencing gel. The corresponding autoradiograms
were analyzed by densitometry. A, autoradiograms: lanes
1-4, UTP increasing from 0.188 mM to 1.4
mM; lanes 5-8, corresponding variation of CTP; lanes 9-12, variation of ATP; lanes
13-16, variation of GTP; lane 17, starting
conditions (all nucleotides at 3.75 mM). B, results
of densitometry: 
, fraction of correct transcripts in % of the
total RNA synthesized; 
, amounts of correct transcript in
arbitrary units.
plus the T7 promoter was
transcribed. The transcription yielded mainly overlong products of
about 60 nt. The prolongation complementary to the first 20 nucleotides
of tRNA contained only 3 U residues. Nevertheless, a
reduction of the UTP concentration during transcription grossly
prevented the synthesis of overlong transcripts (data not shown). It
follows that the strong and specific effect observed with decreasing
the UTP concentration is not related to the uridine content or
nucleotide composition of the transcripts.
template. Since all the plasmid templates used in
this experiment had 5`-protruding or blunt ends, the observed
phenomenon is different from the snap back effect observed with DNA
templates having 3`-protruding ends. This term describes the fact that
an RNA polymerase engaged in transcription can, upon reaching the
5`-end of the template strand, turn to use the 3`-protruding end as a
second template (a DNA to DNA jump) continuing the synthesis toward
plasmid-sized transcripts (Schenborn and Mierendorf, 1985).
Furthermore, the RNA extensions described in this paper were also
produced with single-stranded DNA templates, which cannot trigger a
snap back event. The template-independent addition of a few nucleotides
at the 3`-end of small transcripts observed by Uhlenbeck and colleagues
(Milligan et al., 1987) can also not explain the results,
since the additions found in the cases reported here are much longer.
)
We thank Dr. Richard Brimacombe and Ralf
Jünemann for discussions and Cordula Stiege and Uwe
Fehner for expert technical assistance.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  V. Di Giacco, V. Marquez, Y. Qin, M. Pech, F. J. Triana-Alonso, D. N. Wilson, and K. H. Nierhaus Shine-Dalgarno interaction prevents incorporation of noncognate amino acids at the codon following the AUG PNAS, August 5, 2008; 105(31): 10715 - 10720. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. V. Budkevich, A. V. El'skaya, and K. H. Nierhaus Features of 80S mammalian ribosome and its subunits Nucleic Acids Res., August 1, 2008; 36(14): 4736 - 4744. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. P. Dinos, S. R. Connell, K. H. Nierhaus, and D. L. Kalpaxis Erythromycin, Roxithromycin, and Clarithromycin: Use of Slow-Binding Kinetics to Compare Their in Vitro Interaction with a Bacterial Ribosomal Complex Active in Peptide Bond Formation Mol. Pharmacol., March 1, 2003; 63(3): 617 - 623. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Schafer, A. O. Tastan, S. Patzke, G. Blaha, C. M. T. Spahn, D. N. Wilson, and K. H. Nierhaus Codon-Anticodon Interaction at the P Site Is a Prerequisite for tRNA Interaction with the Small Ribosomal Subunit J. Biol. Chem., May 17, 2002; 277(21): 19095 - 19105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. I. Zvereva, P. V. Ivanov, Y. Teraoka, N. I. Topilina, O. A. Dontsova, A. A. Bogdanov, M. Kalkum, K. H. Nierhaus, and O. V. Shpanchenko Complex of Transfer-messenger RNA and Elongation Factor Tu. UNEXPECTED MODES OF INTERACTION J. Biol. Chem., December 7, 2001; 276(50): 47702 - 47708. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. K. Agrawal, C. M.T. Spahn, P. Penczek, R. A. Grassucci, K. H. Nierhaus, and J. Frank Visualization of tRNA Movements on the Escherichia coli 70S Ribosome During the Elongation Cycle J. Cell Biol., August 7, 2000; 150(3): 447 - 460. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. P. Gmyl, E. V. Belousov, S. V. Maslova, E. V. Khitrina, A. B. Chetverin, and V. I. Agol Nonreplicative RNA Recombination in Poliovirus J. Virol., November 1, 1999; 73(11): 8958 - 8965. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-W. Oh, T. Ito, and M. M. C. Lai A Recombinant Hepatitis C Virus RNA-Dependent RNA Polymerase Capable of Copying the Full-Length Viral RNA J. Virol., September 1, 1999; 73(9): 7694 - 7702. [Abstract] [Full Text]
|
|
|
|
|
|
R. K. Agrawal, P. Penczek, R. A. Grassucci, N. Burkhardt, K. H. Nierhaus, and J. Frank Effect of Buffer Conditions on the Position of tRNA on the 70 S Ribosome As Visualized by Cryoelectron Microscopy J. Biol. Chem., March 26, 1999; 274(13): 8723 - 8729. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Dabrowski, C. M. T. Spahn, M. A. Schafer, S. Patzke, and K. H. Nierhaus Protection Patterns of tRNAs Do Not Change during Ribosomal Translocation J. Biol. Chem., December 4, 1998; 273(49): 32793 - 32800. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Smith, J. Baeten, and J.-S. Taylor The Ability of a Variety of Polymerases to Synthesize Past Site-specific cis-syn, trans-syn-II, (6-4), and Dewar Photoproducts of Thymidylyl-(3'right-arrow5')-thymidine J. Biol. Chem., August 21, 1998; 273(34): 21933 - 21940. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Cullen, C. K. Kaufman, W. C. Bowman, and R. G. Kranz Characterization of the Rhodobacter capsulatus Housekeeping RNA Polymerase. IN VITRO TRANSCRIPTION OF PHOTOSYNTHESIS AND OTHER GENES J. Biol. Chem., October 24, 1997; 272(43): 27266 - 27273. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. S. Sastry and B. M. Ross Nuclease Activity of T7 RNA Polymerase and the Heterogeneity of Transcription Elongation Complexes J. Biol. Chem., March 28, 1997; 272(13): 8644 - 8652. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M.T. Spahn, J. Remme, M. A. Schafer, and KnudH. Nierhaus Mutational Analysis of Two Highly Conserved UGG Sequences of 23S rRNA from Escherichia coli J. Biol. Chem., December 20, 1996; 271(51): 32849 - 32856. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M.T. Spahn, M. A. Schafer, A. A. Krayevsky, and KnudH. Nierhaus Conserved Nucleotides of 23S rRNA Located at the Ribosomal Peptidyltransferase Center J. Biol. Chem., December 20, 1996; 271(51): 32857 - 32862. [Abstract] [Full Text] [PDF]
|
|
