J. gen. ViroL (1982), 58, 429-441.
429
Printed in Great Britain
Key words: VSV/mRNA synthesis~transcription/5' termini
The Metabolic Fate of Independently Initiated VSV mRNA Transcripts
By R O B E R T A. L A Z Z A R I N I , 1. I. C H I E N , 1 F U N M E I
J A C K D. K E E N E 2
Y A N G 1 AND
iLaboratory of Molecular Genetics, National Institute of Neurological and Communicative
Disorders and Stroke, NIH, Bethesda, Maryland 20205 and 2Department of Microbiology
and Immunology, Duke University Medical Center, Durham, North Carolina 27710, U.Sdt.
(Accepted 28 September 1981)
SUMMARY
The kinetics of synthesis and the metabolic stability of uncapped vesicular
stomatitis virus (VSV) mRNA transcripts have been studied using techniques which
clearly differentiate them from other RNA species. The triphosphate-initiating
mRNA transcripts accumulate for at least 5 h during a typical in vitro transcription
reaction. The great majority of these transcripts are smaller than a functional
message and have been released from the template-transcriptase complex. Label that
accumulates in them is stable and is not detectably diminished after a 1 h chase with
unlabelled precursor. These kinetic properties are not those expected for active
intermediates in mRNA synthesis and suggest that the uncapped transcripts are
products of aborted transcription that accumulate during the transcriptive process.
However, we cannot rule out that a small subset of these transcripts may be
elongated into mature mRNA. Initiation of transcription at internal positions along
the VSV genome is both frequent (one-half to one-sixth as frequent as initiations at
the leader RNA gene) and site-specific (occurring only at the beginning of cistrons).
The relevance of these results to the models for VSV transcription is discussed.
INTRODUCTION
The genome of vesicular stomatitis virus (VSV) is a single strand of RNA which contains
five cistrons. Transcription of this negative (anti-sense) -strand RNA yields, as a final
product, five monocistronic mRNAs that are capped and methylated at their 5' termini and
polyadenylated at their 3' termini (for reviews, see Banerjee et aL, 1977; Pringle, 1977;
Wagner, 1975). Although details of this transcription process are largely unknown, studies
from a number of laboratories have illuminated several aspects. U.v. light-inactivation studies
(Abraham & Banerjee, 1976; Ball & White, 1976) indicate that transcription is processive,
since the target-size for distal genes is proportional to the distance between the 3' terminus of
the genome and the 5' terminus of the gene in question. These results have been interpreted as
meaning that the 3' terminus of the genome plays an indispensable role in either the binding of
the transcriptase to the template or the initiation of RNA synthesis. The major 5'
triphosphate-bearing transcript is the 47 nucleotide-long leader RNA which is encoded by the
extreme 3' terminus of the genome (Colonno & Banerjee, 1976, 1978). This RNA is
synthesized in the greatest molar abundance of all the transcripts. There is a decreasing
gradient in the molar amounts of transcripts encoded by the ¢istrons as a function of their
distance from the 3' terminus (Villarreal et aL, 1976). Two fundamentally different models for
transcription have been proposed to explain and integrate these results. In the first, the
processing model (Banerjee et al., 1977, 1978), the transcriptase initiates synthesis at the 3'
terminus of the template and has the potential to transcribe the genome in a single
Downloaded from www.microbiologyresearch.org by
0022-1317/82/0000-4858 $02.00 © 1982 SGMIP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
430
R.A.
LAZZARINI AND OTHERS
uninterrupted synthetic event. The polycistronic transcript, perhaps while it is still incomplete,
is then processed into monocistronic capped and polyadenylated mRNAs. The gradient of
transcripts, presumably, results from the obligate initiation at the precise 3' terminus of the
genome and incomplete transcription by some of the polymerase molecules. The recent
identification by Herman et al. (1978, 1980) of VSV mRNAs linked together by tracks of
poly(A) strongly indicates that the VSV polymerase can both polyadenylate nascent
transcripts and read through the intercistronic boundary into the adjacent cistron. However,
there has been no demonstration that these linked transcripts can be processed into
monocistronic messages. In the second model, the stop-start model, the VSV polymerase
enters the template at the precise 3' terminus but initiates synthesis at the 3' terminus and at
the beginning of each cistron and terminates synthesis at the end of each cistron. Evidence
supporting this model comes principally from studies of the 5' termini of transcripts of VSV.
Bishop and his collaborators (Hefti & Bishop, 1976; Roy & Bishop, 1973) have identified
several in vitro VSV transcripts other than leader RNA that bear a 5' terminal triphosphate.
Similarly, Rose (1975) observed that 10 to 15% of in vivo synthesized mRNAs bore a
triphosphate rather than a cap at their 5' terminus. Recently, Testa et al. (1980a) have
identified several short (28 to 70 nucleotides) transcripts that bear di- or triphosphate termini
and correspond to the sequence at the 5' end of VSV mRNAs. The possibility that these short
transcripts are incomplete, nascent mRNAs prompted us to investigate the kinetics of
synthesis and metabolic stability of these RNAs. In this communication we show that during
in vitro transcription of the VSV genome these RNAs accumulate linearly for at least 5 h,
that greater than 90% of them are not associated with the transcription complex and that
label in them cannot be chased, but rather, is stable. These characteristics are not those
expected of intermediates in the synthesis of capped and methylated mRNAs and suggest
that the majority of the short triphosphate-bearing mRNA sequences are products of aborted
transcriptions. However, we cannot rule out that a very small subset of the triphosphateinitiating RNAs can mature into functional mRNA.
METHODS
Virus. The Mudd-Summers strain of VSV (Indiana serotype) was used throughout
(Lazzarini et al., 1975). The virus was cultured in monolayers of BHK-21 cells and purified
as previously described (Lazzarini et al., 1975).
Purification o f thio-A TP. Solutions containing 4 pmol thio-ATP (Boehringer-Mannheim)
were adsorbed on to a 30 ml column of mercury-agarose (Affl-gel 501, Bio-Rad). The column
was washed with 0.01 M-tris pH 7.9, 1 mM-EDTA, 0.1 M-NaC1 until the absorbance at 260
nm of the effluent fell below 0.01, then washed with deionized water. The bound thio-ATP
was eluted with 0.01 M-fl-mercaptoethanol and lyophilized to dryness. The recovery from
most lots ofthio-ATP was approx. 50%.
Viral transcription. VSV transcripts were prepared as previously described (Colonno &
Banerjee, 1976) but modified so that the nucleoside triphosphate concentrations were as
follows: 1 mM-thio-ATP or ATP; 0.05 mM-[~-32p]CTP (10 to 60 mCi/pmol); 0.1 mM-UTP
and 0.1 mM-GTP. The transcription was allowed to proceed for 1 to 5 h depending upon the
experiment. When the transcription template was removed from the product by filtration, the
transcription mix was diluted to 1 ml with transcription mix lacking virus and triphosphates
but containing 50/~g tRNA, and slowly passed through a 13 mm Swinnex-Millipore filter unit
carrying a HA 0.45 pm nitrocellulose filter (1 ml/min). The filter was washed by filtering an
additional 1 ml portion of the transcription buffer. Under these conditions the proteins from
300 /ag of detergent-disrupted VSV can be retained by the filter. The template was also
removed by centrifugation at 45 000 rev/min for 1 h at 4 °C in a Beckman SW50.1 rotor
adapted for use with 3/16 x 1 21/32 inch cellulose nitrate tubes. Supernatant fluids or filtrates
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
Initiation o f VS V m R N A s
431
were adjusted to 1% SDS, and deproteinized with phenol and chloroform. Filters and
centrifugation pellets containing the template-bound transcripts were suspended in 1% SDS
in SET (0.1 M-NaC1, 0.05 M-tris pH 7.5, 1 mM-EDTA) and extracted with phenol and
chloroform. The transcripts in each aqueous phase were separated from the low mol. wt.
material with Sephadex G-50 equilibrated with 0.1% SDS in SET. In some experiments the
phenol extraction of the filtrate was omitted. Transcripts excluded from the Sephadex
G-50 were precipitated with NaCI and ethanol, rinsed once with 80% ethanol and dried in
vacuo.
Analysis of transcript. Transcripts were dissolved in 1 ml 0.01 M-tris-HCl pH 7.5 and
digested with 200 units RNase T 1 for 1 h at 37 °C. The digested transcripts were diluted with
column buffer and the thio-triphosphate-bearing oligonucleotides were separated by affinity
chromatography on a 5 ml column of mercury-agarose according to the method of Smith et
al. (1978). Fractions of the eluted oligonucleotides containing most of the radioactivity were
combined, adjusted to 1.1 M-LiC1, 0.25 mM-Na2HPO4, 10/tg/ml tRNA, 0.017 M-tris-HC1
pH 8.8. The oligonucleotides were co-precipitated with the lithium phosphate by adding 2.5
vol. ethanol and chilling the mixture to - 7 0 °C for 2 h. The precipitate was collected by
centrifugation, washed once with 80% ethanol and dried in vacuo. The precipitate was
dissolved in a small volume of 5.5 M-urea in tris-EDTA-borate electrophoresis buffer, heated
to 100 °C for 30 s, quick-cooled and fractionated on 20% polyacrylamide gels containing 8
M-urea according to Donis-Keller et al. (1977).
When VSV transcripts were not digested with RNase T 1 prior to affinity chromatography
(as in Fig. 1 a), mercury-Sepharose rather than mercury-agarose was employed. The former
was obtained from R. C. C. Huang, and is capable of retaining RNAs larger than 28S (Smith
et al., 1978).
Analysis of cap structures. Transcripts to be analysed for the presence of caps were
digested in 50 ~tl 0.01 M-NH4OAc pH 6, containing 30/tg nuclease P1 for 2 h at 37 °C. A 4
/tl amount of 0.2 M-tris-HC1 pH 8 and 4/tl 100 units/ml calf intestinal alkaline phosphatase
were then added and the incubation continued for 1 h at 37 °C to remove all terminal
phosphates. The digest was applied to 20 × 40 cm sheets of either DE-81 paper or PEI thin
layers. Ionophoresis was carried out either in 7 % formic acid (DE-81 paper) as described by
Brownlee (1972) or in 7 M-urea, 0.1 M-phosphate buffer pH 7.5 (PEI sheets) according to
Keller & Crouch (1972).
RESULTS
Identification of triphosphate-terminating transcripts with thio-A TP
The chemical ~-thio-ATP (Goody & Eckstein, 1971) is an ATP analogue in which one of
the oxygen atoms on the y-phosphate has been replaced with a sulphydryl group. This
analogue is distinguished in that the bridge atoms between the ribose and the a-phosphate and
between the a-, fl- and 7-phosphates are oxygens, as in the natural compound, and these
phosphates are potentially hydrolysable with free energies not likely to be dramatically
different from those of the natural compound. This analogue has proven very useful in the
study of initiation of RNA on synthetic templates, bacteriophage DNA and isolated mouse
myeloma nuclei (Reeve et al., 1977; Smith et al., 1978). Recently, Carroll & Wagner (1979)
have demonstrated that the transcriptase of VSV can also use this analogue in the synthesis of
RNA. The thio-triphosphate-bearing transcripts are easily separated from other RNAs by
affinity chromatography using mercury-agarose. We have confirmed Carroll & Wagner's
observations that the VSV transcriptase is not inhibited by the complete replacement of ATP
with thio-ATP in the transcription reaction and that the major, easily recognized
thio-triphosphate-bearing transcript is leader RNA. Fig. 1 (lane 1) shows a polyacrylamide
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
432
R. A . L A Z Z A R I N I
1
o-
m
AND
OTHERS
2
........
o
~i••• •i• i'~!~ii
m
mBp
XC--
F i g . 1. Gel electrolJhoresis of thio~triphosphate-terminating transcripts and thio-triphosphateterminating T t oligonucleotides derived from them. Lane 1, undigested thio-triphosphate-terminating
transcripts were prepared with [32p]CTP, purified by adsorption to mercury-Sepharose, heat-denatured
and electrophoresed into a 12 % polyacrylamide gel as described in Methods; lane 2, transcripts labelled
with [32p]CTP were digested with RNase T~ and the thio-triphosphate-bearing oligonucleotides were
purified with mercury-agarose. Oligonucleotides were separated by electrophoresis in a 20%
polyacrylamide gel. O, Origin; XC, xylene cyanol; BP, bromophenol blue (see also Fig. 5 and 6).
gel profile of 32p-labelled VSV transcripts that were synthesized with thio-ATP, bound to
mercury-Sepharose and eluted with dithiothreitol. This affinity column matrix is capable of
retaining RNAs as large as 28S with 75 to 80% recovery, so that little size selection should
be exerted by this procedure. The major component is a family of RNAs that migrate slightly
slower than the xylene cyanol dye. These are shown by hybridization to contain leader RNA
(R. C. Herman & R. A. Lazzarini, unpublished results). The precise size of the RNA that
remained at the top of the gel in Fig. 1 (lane 1) is not known, but it is single-stranded and
larger than 600 bases.
It is clear from the work of others that some of the transcripts shown in Fig. 1 should
contain mRNA sequences (Hefti & Bishop, 1976; Roy & Bishop, 1973; Testa et aL, 1980a).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
Initiation of VS V mRNAs
433
The relative number of these transcripts can be estimated by digesting the mixed transcripts
with RNase T~ prior to affinity selection on mercury-agarose columns. Since all VSV mRNA
begin with the pentamer 5'-AACAG ... 3' (Rhodes & Banerjee, 1976), transcripts initiated at
the beginning of any cistron should have the sequence HSpppApApCpApGp... whereas
transcripts initiated at the leader gene begin HSpppApCpGp... (Colonno & Banerjee, 1978).
Since each initiating sequence contains a single C and G, the radiolabel incorporated into
these fragments can be directly compared if the source of the label was CTP or GTP.
Fig. 1 (lane 2) shows the electrophoretic pattern obtained on a 20 % polyacrylamide gel of
thio-triphosphate-bearing RNase T 1 oligonucleotides. These oligonucleotides were prepared
from [32P]CTP-labelled transcripts like those shown in lane 1, but digested with RNase T 1
prior to affinity chromatography on mercury-agarose. The two bands obtained by
electrophoresis correspond to HSpppApApCpApGp (upper band) and HSpppApCpGp
(lower band) as discussed below. These oligonucleotides contain approx. 0.1% of the
radiolabel incorporated into the total transcripts.
Identical electrophoresis results were obtained when the transcripts were labelled with
[32p]GTP. Since, under the latter conditions, all T 1 oligonucleotides except those from the
extreme 3' terminus of the transcript, must be labelled, these results show that there are only
two initiating sequences that begin, pppA . . . .
The sequence of both triphosphate-bearing T~ oligonucleotides was confirmed by eluting
the fragments from the polyacrylamide gel, digesting them with RNases U2, T2 and A and
analysing the products by thin layer chromatography on PEI sheets. When labelled with
[a-32p]GTP, the pentamer yielded ApGp when digested with RNase A. When labelled with
[a-32p]CTP, it yielded Ap when digested with RNases T 2 or U2. Consequently, the fragment
must have the sequence ... ApApCpApGp. Similarly, the trimer when labelled with
[a-a2p]GTP yielded Cp and CpGp when digested with RNases T 2 and U 2 respectively. Its
sequence must be... ApCpGp.
Transcripts prepared with thio-A TP are capped and methylated
Since the purpose of this study was to explore the possibility that triphosphate-bearing
mRNA sequences are the direct precursors of the capped and methylated mRNAs, it was
important to demonstrate that RNA synthesis, including capping and methylation, proceeds
normally when thio-ATP replaces ATP in the transcription mixture. The amount of synthesis
of capped structures was determined by nuclease P1 and calf intestine alkaline phosphatase
digestion of [ct-32p]GTP-labelled transcripts prepared with either ATP or thio-ATP. Digestion
products were separated by electrophoresis on either DE-81 paper or PEI thin layer sheets as
described in Methods. Table 1 shows the incorporation of GTP into the GpppA caps of
mRNA synthesized with ATP or thio-ATP. In these experiments, the caps are unmethylated
since SAM was omitted from the transcription reaction. The incorporation into the cap,
whether measured directly or after normalization to the total GTP incorporation, was the
same or slightly higher when thio-ATP completely replaced ATP.
The synthesis of the cap structure, 7mGpppAm, is also unaffected by the substitution of
thio-ATP for ATP. Table 2 shows the incorporation of 3H-labelled SAM and [a2p]GTP into
RNA was slightly higher when thio-ATP was substituted for ATP. Both labelled compounds
were incorporated linearly over the 4 h period. That the 3H-labelled methyl groups of SAM
were, in fact, incorporated into cap structures is shown in Fig. 2. Portions of the transcripts
from the 4 h incubation described in Table 2 were digested with nuclease P~ and calf intestine
phosphatase, and the products separated by ionophoresis at pH 7.5 on PEI sheets or at pH
1.9 on DE-81 paper. Since the specific activity of the [32p]GTP was small compared to that
of 3H-labelled SAM, the isolate cap is labelled only with 3H and virtually all of the detectable
32p was recovered as inorganic phosphate. In both of these ionophoresis systems cap I and
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
434
R. A. L A Z Z A R I N I AND O T H E R S
(a)
I
I
I
'
'
I
200
100
t
(b)
200
100
(c)
.~ 1200
8O0
1000 .~
400
1000~
I
I
I
(d)
600
1000
i
2001
,
10
20
30
40
Fraction no.
50
Fig. 2. Ionophoresis of VSV cap structures. Doubly labelled (3H-labelledSAM, [32p]GTP)transcripts
were prepared with either thio-ATP or ATP, and digested with the nuclease P1 and alkalinephosphatase
as described in Methods. Samples of these digests were ionophoresed on DE-81 paper (a, b) or PEI
sheets (c, d). The ionopherograms were cut into 0-5 cm strips and counted. Horizontal bars indicate
the position of standard compound (7mGpppAm). Hatched area represents 3Zp-radioactivity;open area
represents ~H-radioactivity.(a, e) Thio-ATP; (b, d) ATP.
Table 1. Synthesis of capped VSV mRNAs in the presence of the thio-ATP
Chemical
ATP
Thio-ATP
CAP (ct/min)
626
664
GMP (ct/min)
109092
87416
CAP/GMP
5-7 x 10 3
7.5 × 10-9
Table 2. Incorporation of3H-labelled SAM and [32p]CTP into VSV mRNAs in the presence
of thio-A TP
Chemical
Thio-ATP
ATP
Time (]1)
2
4
2
4
3H (ct/min)
24 660
40995
11452
29 302
32p (ct/min)
51 165
95657
35 752
105 750
cap II structures are poorly resolved from one another and the positions of these materials, as
determined with u.v. standards, are shown in the figure. In both ionophoresis systems, > 85 %
of the 3H radioactivity co-migrated with the cap structures. Thus, both the incorporation of
the nucleoside triphosphates into R N A and the synthesis of methylated and capped R N A
proceeds normally when A T P is replaced with thio-ATP.
The conversion of thio-ATP to A T P during the transcription reaction by an enzymic
exchange with the y-phosphates of other nucleoside triphosphates or by desulphurization was
also examined and shown to be negligible. The exchange of y-phosphates between positions of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
Initiation of VS V mRNAs
43 5
thio-ATP and [~-32p]CTP were measured in the presence of the detergent-disrupted VSV
under conditions identical to those used for transcription except that UTP and GTP were
omitted to prevent RNA synthesis. Less than 0.004% of [a-3zP]CTP acquired a
7-thio-phosphate during a 1 h incubation, as estimated by the ability of labelled nucleotides to
bind to mercury-agarose. Similarly, the exchange between [y-32p]GTP and thio-ATP was
measured in reaction mixtures identical to those used for transcription but lacking UTP and
CTP. Under these conditions only 1.1% of the thio-ATP was converted to [~-32p]ATP during
a 2 h incubation. Desulphurization of thio-ATP to yield ATP was also estimated in similar
incubation mixtures which lacked all nucleoside triphosphates except thio-ATP. The possible
production of ATP from thio-ATP was examined by thin layer chromatography on PEI
sheets using a 0.75 M-KH2PO 4, a chromatography system that easily separates ATP from
thio-ATP. No detectable ATP was observed during a 2 h incubation of this ATP with
detergent-disrupted virus. These results indicate that less than 5 to 10% of the thio-ATP
could have been converted to ATP. In a fourth set of experiments, possible desulphurization
of thio-ATP was examined by comparing the species of RNA that bind to mercury-agarose
to those that pass through the column. In these experiments, transcripts prepared with a
defective-interfering particle (DI-T) and thio-ATP were used since these are less complex,
consisting only of DI particle product RNA. Polyacrylamide gel electrophoresis of the bound
and flow-through fractions of the mercury-agarose column showed that only 25 % of the DI
particle product RNA failed to bind to mercury-agarose. Had significant amounts of ATP
been generated during the transcription reaction, DI particle products with triphosphate and
thio-phosphate termini would have been synthesized and resolved by the mercury-agarose
column. This type of experiment yields an estimate of the maximum incorporation of ATP in
RNA termini since there are numerous reasons why some transcripts might not bind to the
column (oxidation of the thio-group, dephosphorylation of the terminus or mechanical failure
of the column, etc.). These results are in complete agreement with those of Carroll & Wagner
(1979) who found that at least 80% of the leader RNA synthesized in the presence of
thio-ATP possessed a thio-triphosphate terminus.
We conclude from these experiments that VSV transcripts are synthesized, capped and
methylated as efficiently in the presence of thio-ATP as with ATP. If triphosphate-bearing
transcripts are the precursors to capped and methylated mRNAs, then thio-triphosphatebearing transcripts are capped and methylated efficiently by detergent-disrupted VSV.
Triphosphate-terminating transcripts accumulate during in vitro transcription of VSV
The appearance of triphosphate-terminating transcripts, in a VSV transcription reaction
was followed as a function of time. For these experiments, use was made of the simple
procedure for separately quantifying sequences that begin HSpppApApCpApG... and
HSpppApCpG .... The transcription of VSV was carried out with thio-ATP and [a-32p]CTP,
and portions were removed during the course of the reaction. These were treated with SDS
and the transcripts were separated from unincorporated triphosphates by chromatography on
G-50, digested with RNase T 1 and the thio-triphosphate-bearing oligonucleotides were
recovered by affinity chromatography with mercury-agarose. The T 1 oligonucleotides were
concentrated by co-precipitation with lithium phosphate and tRNA and finally separated by
electrophoresis on 20% polyacrylamide gels. The radiolabel in the pentamer and trimer were
estimated by Cherenkov counting of the excised pieces of gel. The results of a typical
experiment are shown in Fig. 3. Both leader and mRNA termini accumulate linearly for at
least 5 h. In several such experiments the pentanucleotide was one-fifth to one-half the concentration of the trinucleotide. The cause of this variability in relative concentrations is not
known but within any single experiment the relative concentration did not change with time.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
436
R.
A°
, '/30
AND OTHERS
LAZZARINI
7.5
m
x
.9
5o
5
~
X
)<
= 4
4 .9
~3
3
5
.
~20
•
x
e-
_
;z~
"~
10 g
2 "~-
1
2
3
4
5
1
2
Time (h)
Time (h)
Fig. 3
Fig. 4
Fig. 3. Time course of 32p incorporation into total transcripts and triphosphate-terminating RNase T 1
oligonucleotides derived from them (see text and Methods for details). O, Total incorporated a2p
(ct/min x 10-~); O, HSpppACG and A, HSpppAACAG 32p incorporated (ct/min x 10-3).
Fig. 4. Kinetics of labelling RNA in transcriptive complexes. The transcriptive complexes from equal
portions of a transcription reaction incorporating 3H-labelled UTP (1 mCi//~mol) were collected by
filtration at the indicated times. At 1 h (arrow) a 20-fold excess of unlabelled UTP and tracer amounts
of [32P]CTP (final sp. act. 0.33 mCi//~mol) were added to the reaction mixture. The results shown
have been corrected for the small dilution by the addition of the nucleotides. RNA was extracted
from filters with 1% SDS and precipitated with trichloroacetic acid. The 3H and 32p content of RNA
was determined by liquid scintillation counting. O, 3H incorporation in filter-bindable RNA; A, 32p
incorporation in filter-bindable RNA;/x, a2p incorporation into total RNA.
The linear incorporation into leader R N A termini shown in Fig. 3 had been expected since
leader R N A is believed to be a final product and not an intermediate in the transcription
process. The linear accumulation of the pentamer suggests that the transcripts that contain
them were also metabolically stable and not further metabolized.
Most triphosphate-bearing transcripts are not associated with the transcriptive complex
VSV transcriptive complexes containing newly initiated R N A have been isolated by
centrifugation of transcription reaction mixtures (Testa et al., 1980 b). Since these complexes
contain protein firmly associated with the nucleic acid templates, they can also be collected by
filtration through nitrocellulose filters under conditions in which proteins and nucleic acids
bound to them adhere to the filter, while most free nucleic acids pass through the filter (Yams
& Berg, 1967). That the filter-bound transcriptive complexes contain newly synthesized,
metabolically unstable R N A is shown in Fig. 4. In that experiment, equal amounts of
complexes were collected from a 3H-labelled transcription reaction at the indicated times.
The incorporation of 3H into the complexes is rapid until a 20-fold excess of unlabelled U T P
and trace amounts of [32P]CTP were added to the reaction (indicated by the arrows). The
addition of the unlabelled U T P chased one-third of the all-labelled R N A from the complex.
Reconstruction experiments suggest that the remaining two-thirds of the all-labelled R N A
represents stably and non-specifically associated R N A . The continued synthesis of R N A
during the chase is verified by the rapid incorporation of 32p into filter-bound and total R N A .
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
Initiation o f V S V m R N A s
(a)
(b)
1
(a)
2
2
m
437
(b)
1
2
:
--o
O
~XC
~XC
BP
P
Fig. 5
Fig. 6
Fig. 5. Analysis of template-bound and free triphosphate-bearing transcripts. 32p-labelled transcripts
were prepared with thio-ATP and fractionated into template-bound (lane 1) and free (lane 2) fractions
by filtration (a) or centrifugation (b). Triphosphate-bearing RNase T x oligonucleotides were prepared
and analysed as described in Methods. O, Origin.
Fig. 6. Metabolic stability of radiolabel in triphosphate-bearing oligonucleotides. The VSV transcription reaction was set up using [32p]CTP and thio-ATP. A sample was withdrawn after 2 h of
syn'thesis and enough unlabelled CTP added to the balance to lower the CTP specific activity 20-fold.
After a 1 h chase period the balance of the reaction mixture was withdrawn. Both samples were
separated into bound and unbound transcripts by filtration and RNA prepared from all four samples.
Triphosphate-bearing oligonucleotides derived from these samples were separated by electrophoresis in
a 20% acrylamide gel. (a) Filtrate; (b) bound; lane 1, samples drawn before the chase; lane 2, samples
drawn after the chase. O, Origin.
The i n c o r p o r a t i o n o f 3H before the chase a n d i n c o r p o r a t i o n of 32p d u r i n g the chase, m o s t
likely do n o t reach s a t u r a t i o n b e c a u s e o f the long transit time for the VSV t r a n s c r i p t a s e
( I v e r s o n & Rose, 1981) a n d the a c c u m u l a t i o n o f non-specifically a n d stably associated
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
438
R.A.
LAZZARINI AND OTHERS
labelled RNAs. Nonetheless, the experiment shows that the filter-bound RNAs are greatly
enriched in the nascent species.
The possibility that the thio-triphosphate-terminating RNAs might be released from the
template and not associated with the transcriptive complex was examined in two ways. In
the first, a transcription reaction was allowed to proceed for 3 h and then centrifuged to
sediment the transcriptive complex. RNA was prepared from the supernatant and pellet
fraction and each was analysed for the presence of the triphosphate pentamer. In a second
series of experiments the transcriptive reaction was fractionated by filtration through
nitrocellulose filters. RNA was prepared from the filter-bound material a n d filtrate, and
similarly analysed for the content of the pentamer. The results of these two experiments are
shown in Fig. 5. In each case the RNA from the soluble fraction yielded the trimer and
pentamer. RNA from the particulate or protein-bound fraction yielded only small amounts of
the trimer and traces of the pentamer. The material that migrated more slowly than
xylene cyanol dye in Fig. 5 is leader RNA which has escaped digestion by RNase T 1 because
of it forming a duplex with the template RNA during phenol extraction. This RNA has the
precise size of leader RNA and when extracted from the gel, denatured and redigested with
RNase T 1yields the trimer, pppACG. The molar amount of the undigested leader RNA in the
filter-bound material (Fig. 5, lane 1) is small when compared to the trimer obtained from the
filtrate fraction, since the film darkening by it is less than that given by the trimer, although it
contains nine times as many cytidylate residues. Both the centrifugation and filtration
experiments lead to the same conclusion: that approx. 95 % of each class of triphosphatebearing RNAs are found in the supernatant or filtrate and, therefore, are not associated with
the transcriptive complex.
The triphosphate-bearing transcripts are metabolically inert
Two mechanisms have been described for the in vitro formation of the cap on rhabdovirus
mRNAs. When the natural nucleoside triphosphates are used, a guanosine diphosphate is
transferred to form the guanylate cap; the remaining phosphate of the cap structure is derived
from the a-position of ATP (Abraham et al., 1975; Gupta & Roy, 1980). Cap formation
proceeds by a different route when an analogue of GTP with a non-hydrolysable fl-? bond
such as fl-?-imido-GTP is used (Banerjee et al., 1978; Gupta & Roy, 1980). Under these
circumstances only the a-phosphate of the capping GTP enters the final product and the two
remaining phosphates of the cap are derived from the a and fl position of ATP. However, in
both mechanisms the y-phosphate of ATP is lost and not incorporated into the cap.
Consequently, if the ~,-thio-triphosphate-bearing mRNAs are precursors to the cap mRNAs,
the thio-phosphate group would be lost in the reaction and the capped RNAs not recoverable
by mercury-agarose chromatography. Thus, there would be a continuous flux of material
through the mercury-agarose bindable fraction that would be easily visualized with the
radiochemical pulse-chase experiment. Fig. 6 shows the results of such an experiment. A VSV
transcription reaction using thio-ATP and [a-32p]CTP was allowed to proceed for 2 h at
30 °C. At that time a sample was withdrawn and 20-fold more unlabelled CTP was added to
the remainder of the incubation mixture. Control experiments have demonstrated that this
sudden increase in CTP concentration does not perturb the Synthesis of the transcripts
although it leads to an abrupt halt of labelling. After a 1 h chase period, the remainder of the
reaction mixture was withdrawn. All samples were fractionated into a template-bound and
unbound fraction by filtration. RNA was prepared from each and digested with RNase T 1.
Oligonucleotides bearing thio-triphosphate termini were purified with mercury-agarose and
analysed on 20% polyacrylamide gels. Fig. 6 shows an autoradiogram of a typical
polyacrylamide gel. As seen earlier (Fig. 5) most of thio-triphosphate-terminating RNAs are
not associated with the template and are found in the filtrate. The radiolabel in the trimer and
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
Initiation o f V S V m R N A s
439
pentamer from both the template-bound and filtrate fractions were not diminished by the 1 h
chase with unlabelled CTP" the radioactivity remaining in each oligonucleotide after the chase
was within experimental limits of the amounts of radioactivity present in the same
ofigonucleotide before the chase. We conclude from these experiments that most, if not all,
of the triphosphate-bearing RNAs are metabolically inert and that no more than 15 % of the
label in these materials (the limits of resolution of our experiments) are 'chaseable'.
DISCUSSION
Among the variety of products of in vitro VSV transcription are transcripts that bear
triphosphate termini (Hefti & Bishop, 1976; Rose, 1975; Roy & Bishop, 1973; Testa et al.,
1980a). In the present communication we have examined the metabolic stability of those
transcripts that begin with pppA .... These have been shown to contain leader RNA and
RNA species templated by the cistrons that are efficiently transcribed in vitro, the N and NS,
cistrons (Testa et al., 1980a). In this analysis, we have made use of the difference in sequence
of the 5' termini of leader RNA (Colonno & Banerjee, 1978) on the one hand and all of the
VSV mRNAs (Rhodes & Banerjee, 1976) on the other. Thus, the radioactivity contained in
the RNase T 1 oligonucleotides pppACG and pppAACAG were used to estimate the relative
concentration of leader RNA and all mRNA bearing a pppA terminus respectively.
By substituting ?-thio-ATP for ATP in the transcription reaction we are able to purify the
RNase T 1 oligonucleotides of interest in a single affinity chromatography step using
mercury-agarose. The resolution of the purified triphosphate-bearing oligonucleotides into a
trimer (ppACG) and a pentamer (pppAACAG) by polyacrylamide gel electrophoresis
completed the simple assay procedure.
The results of this study indicate that initiation of transcription at the beginning of mRNA
cistrons is quite frequent: one-half to one-sixth as frequent as initiations at the beginning of
the leader region. The triphosphate-bearing mRNA transcripts accumulate for at least 5 h
during a standard transcription reaction. The majority of these are released from the template
and are recoverable from the soluble fraction of the transcription reaction mixture. If these
transcripts are precursors to full-length mRNA, one would expect them to remain
template-bound. Finally, label that accumulates both in the triphosphate-bearing trimer and
pentamer is stable and is not chaseable, indicating that less than 15 % of these transcripts
could have been capped during the chase. In the case of transcripts beginning HSpppAACAG,
it is unlikely that the failure to cap detectable amounts of these RNAs results from
differences between their sequences and those of the putative natural intermediate. The
transcripts begin with the pentamer found at the 5' terminus of all VSV mRNAs and it is very
likely that the sequence homology extends for their full length (Testa et al., 1980a). These
results suggest that triphosphate-bearing RNA, large enough to be excluded by Sephadex
G-50, is not capped by the VSV enzymes.
Polyacrylamide gel electrophoresis indicates that the vast majority of the triphosphateterminating transcripts are smaller than the smallest mRNAs (Fig. 1). The molar amount of
the high mol. wt. material at the origin of the gel is very small when compared to the RNAs
less than 200 nucleotides in length. Consequently, the majority of the triphosphateterminating mRNAs are most likely incomplete transcripts. This conclusion is in agreement
with the recent finding (Testa et al., 1980a) that some triphosphate-bearing mRNA
transcripts are 28 to 70 nucleotides long. Taken together, these results suggest that the
triphosphate-bearing mRNAs may be viral debris that accumulate during in vitro
transcription and are not precursors to functional mRNA. Alternatively, these triphosphatebearing transcripts may be specific products of an attenuation mechanism controlling
mRNA synthesis.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
440
R.A.
L A Z Z A R I N I AND O T H E R S
Our results, however, do not lead to the conclusion that independent initiation of functional
(capped and methylated) m R N A s does not occur. Firstly, we observed only two types of
triphosphate-terminated transcripts - those beginning p p p A C G and those beginning
p p p A A C A G - although our method is capable of detecting all types of transcripts that begin
with A T P and contain a guanosine or cytosine residue. This restriction of initiation to the
precise 3' terminus of the leader and the beginning of virus m R N A cistrons suggests a
sequence specificity that is most easily reconciled with the independent initiation model.
Secondly, our procedures would not detect the capping of very short transcripts since only
material that is excluded by Sephadex G-50 was analysed. These considerations suggest to us
that if capped and methylated m R N A s are independently initiated, the capping and methylation m a y take place when the transcripts are very short. It is conceivable that capping and
methylation form an integral part of the initiating event.
The results presented here are equally compatible with the processing model for m R N A
biogenesis. In that model the sequences to be capped are internal in the polycistronic R N A
transcript, and the capping and methylation form part of a post-transcriptional cleavage and
processing of the transcript. Consequently, the triphosphate-terminating transcripts might not
be suitable substrates for the capping enzyme, thus accounting for their metabolic stability.
We thank Dr B. Moss for his assistance and advice on the electrophoretic separation of cap
structures, and R. C. C. Huang for samples of mercury-Sepharose. J.D.K. was supported by a grant
from the National Institute of Allergy and Infectious Diseases (AI 16099). A preliminary account of
some of this research was presented at the Conference of Negative Strand RNA Viruses, St Thomas,
Virgin Islands, November 1980.
REFERENCES
ABRAHAM, G. & BANERJEE, A. K. (1976). Sequential transcription of the genes of vesicular stomatitis virus.
Proceedings of the National Academy of Sciences of the United States of America 73, 1504-1508.
ABRAHAM,G., RHODES,D. F. & BANERJEE,A. K. (1975). Novel initiation of RNA synthesis in vitro by vesicular
stomatitis virus. Nature, London 255, 37-40.
BALL,L. A. & WHITE,(2. N. (1976). Order of transcription of genes of vesicular stomatitis virus. Proceedings of the
National Academy of Sciences of the United States of America 73, 442-446.
BANEFJEE,A. K., ABRAHAM,G. & (2OLONNO,R. J. (1977). Vesicular stomatitis virus: mode of transcription. Journal
of General Virology 34, 1-8.
BAt~EPJEE,A. K., (2OLOrCNO,R. J., TESTA,D. & FRANZE-FERNANDZ,M. T. (1978). Mechanism of RNA synthesis in
vitro. In Negative Strand Viruses and the Host Cell, pp. 249-259. Edited by B. W. J. Mahy & R. D. Barry.
London: Academic Press.
BROWNLEE, G. (3. (1972). Determination of Sequences in RNA. Amsterdam, London & New York:
Elsevier/North-HoUand.
CARROLL,A. R. & WAGNER,R. R. (1979). Adenosine-5'-O-(3-thiotriphosphate)as an affinity probe for studying
leader RNAs transcribed by vesicular stomatitis virus. Journal of Biological Chemistry 254, 9339-9341.
COLONNO,R. J. & ~ANERJEE,A. K. (1976). A unique RNA species involvedin initiation of vesicular stomatitis virus
RNA transcription in vitro. Cell 8, 197-204.
(2OLONNO,R. J. & BANERJEE,A. K. (1978). Complete nucleotide sequenceof the leader RNA synthesizedin vitro by
vesicular stomatitis virus. Cell 15, 93-101.
DONIS-KELLER, H., MAXAM, A. M. & G I L B E R T , W. (1977). Mapping adenines, guanines and pyrimidines in RNA.
Nucleic Acids Research 4, 2527-2538.
GOODY, R. S. & ECKST~, F. (1971). Thiophosphate analogs of nucleoside di and triphosphates. Journal of the
American Chemical Society 93, 6252-6257.
GUPTA,K. C. & ROY,P. (1980). Alternate capping mechanisms for transcription of spring viremia of carp virus:
evidencefor independentmRNA initiation.Journal of Virology 33, 292-303.
I-IEFTI, E. & msnop, o. H. L. (1976). The sequences of VSV in vitro transcription product RNA (_+SAM).
Biochemical and Biophysical Research Communications 68, 393-400.
HERMAN, R. (2., ADLER, S., LAZZARINI, R. A., (2OLONNO, R. J., BANER.IEE, A. K. & WESTPHAL, H. (1978). Intervening
polyadenylate sequencesin RNA transcripts of vesicular stomatitis virus. Celt 15, 587-596.
HERMAN, R. (2., SCHUBERT, M., KEENE, J. D. & LAZZAKINI, R. A. (1980). Polycistronic vesicular stomatitis virus
RNA transcripts. Proceedings of the National Academy of Sciences of the United States of America (in
press).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32
Initiation of VS V mRNAs
441
IVERSON, L. E. & ROSE, J. K. (1981). Localized attenuation and discontinuous synthesis during vesicular stomatitis
virus transcription. Cell 23, 477-484.
KELLER, W. • CROUCh, R. (1972). Degradation of D N A : R N A hybrids by ribonuclease H and D N A polymerases
of cellular and viral origins. Proceedings of the National Academy of Sciences of the United States of
America 69, 3360-3364.
LAZZARINI, n. A., WEBER, 6. H., JOHNSON, L. D. & STAM~INtER, C. M. (1975). Covalently linked message and
anti-message (genomic) R N A from a defective vesicular stomatitis virus particle. Journal of Molecular
Biology 97, 289-307.
PRXNGLE, C. R. (1977). Genetics of rhabdoviruses. In Comprehensive Virology, vol. 9, pp. 239-290. Edited by H.
Fraenkel-Conrat & R. R. Wagner. New York: Plenum Press.
REEVE, A. E., SMITh, M. M., PIGmT, V. & HUANG, a. c. c. (1977). Incorporation of purine nucleoside 5'-[;e-s]
triphosphates as affinity probe for initiation of R N A synthesis in vitro. Biochemistry 16, 4464-4469.
Rr~OD~S, D. P. & BANEgJEE, A. K. (1976). 5'-Terminal sequence of vesicular stomatitis virus m R N A s synthesized in
vitro. Journal of Virology 17, 33-42.
ROSE, S. K. (1975). Heterogeneous Y-terminal structures occur on vesicular stomatitis virus m R N A s . Journal of
Biological Chemistry 250, 8098-8104.
ROY, o. & BISHOP, D. rx. I.. (1973). Initiation and direction of R N A transcription by vesicular stomatitis virus virion
transcriptase. Journal of Virology 11, 487-501.
S~ITH, M. M., ~EVE, A. E. & HUANG, R. C. C. (1978). Analysis of R N A initiated in isolated mouse myeloma nuclei
using purine nucleotide 5' [?-s] triphosphates as affinity probes. Cell 15, 615-626.
TESTA, D., CHANDA, P. K. & BANE~EE, A. K. (1980a). Unique mode of transcription in vitro by vesicular
stomatitis virus. Cell 21, 267-275.
TESTA, D., CHANDA, P. K. & BANE~EE, A. K. (1980b). In vitro synthesis of the full-length complement of the
negative strand genome R N A of vesicular stomatitis virus. Proceedings of the NationalAcademy of Sciences
of the United States of America 77, 294-298.
VILLARREAI., L. P., BREINDL, M. & HOLLAND,J. J. (1976). Determination of molar ratios of vesicular stomatitis virus
induced R N A species in BHK2~ cells. Biochemistry 15, 1663-1667.
WAGNER, R. R. (1975). Reproduction of rhabdoviruses. In Comprehensive Virology, vol. 4, pp. 1-94. Edited by H.
Fraenkel-Conrat & R. R. Wagner. New York: Plenum Press.
~'ARUS, M. & BERG, P. (1967). Recognition of t R N A by amino acyl t R N A synthetase. Journal of Molecular
Biology 28,479-489.
(Received 5 August 1981)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 15:04:32