Journal of General Virology(1991), 72, 259-266. Printedin Great Britain
259
Shortened forms of beet necrotic yellow vein virus RNA-3 and -4:
internal deletions and a subgenomic RNA
S. Bouzoubaa, U. Niesbach-KlOsgen,t I. Jupin,~ H. Guilley, K. Richards* and
G. Jonard
Institut de Biologic Moldculaire des Plantes du CNRS, Universitd Louis Pasteur, 12 rue du Gbn~ral Zimmer,
67000 Strasbourg, France
Beet necrotic yellow vein virus RNA-3 and RNA-4,
produced as full-length biologically active transcripts
in vitro, can undergo spontaneous internal deletions
when inoculated onto Chenopodium quinoa leaves
along with R N A - 1 and -2. The deletion process is
specific, giving rise to only a few major species, and can
be rapid; deleted forms appear after only one or two
passages in leaves. In one of the shortened forms of
RNA-4, the deletion precisely eliminated one copy of a
15 nucleotide (nt) direct sequence repeat from the fulllength prototype sequence, suggesting that 'copychoice' switching of the replicase-template complex
from one repeat to the other during R N A replication
was responsible for the generation of this deletion. The
deletion found in a major shortened form of RNA-3, on
the other hand, did not occur near sequence repeats but
began with G U and ended with AG like a nuclear intron
sequence. Thus it is possible that the deleted sequence
has been removed by splicing. However, two other
deletions that were characterized were not associated
with either of these types of sequence feature. An
approximately 600 nt 5'-terminally truncated nonencapsidated form of RNA-3 was also detected in
infected plant tissue. The evidence suggests that it is a
subgenomic R N A derived from RNA-3.
Introduction
Kuszala et al., 1986; Burgermeister et al., 1986; Quillet et
al., 1989). These observations are consistent with the
idea that the two longest RNAs, RNA-1 and -2, code for
basic functions required in all hosts while RNA-3 and -4
are functional in the natural infection process, that is,
fungus-mediated infection of roots and proliferation
within root tissue (Koenig et al., 1986; Lemaire et al.,
1988; Tamada & Abe, 1989). Recent findings indicate
that RNA-4 is essential for efficient fungal transmission
(Tamada & Abe, 1989) while RNA-3 may facilitate viral
multiplication and/or movement in roots (Koenig &
Burgermeister, 1989; Tamada & Abe, 1989).
During serial propagation in leaves, BNYVV isolates
sometimes lose one or both small RNAs. In other
isolates, the small RNAs are present but appear to have
undergone deletion. This deletion process is of interest
because it results in the appearance of a limited number
of discrete species rather than a heterogeneous population of shortened forms (Bouzoubaa eta/., 1985; Kuszala
et aL, 1986; Burgermeister et al., 1986; Tamada et aL,
1989). This means that there are 'hot spots' or preferred
sites for deletion within RNA-3 and RNA-4 which can
manifest themselves in the absence of selective pressure
for maintenance of the full-length molecules. In previous
studies we have characterized the deleted forms of
Beet necrotic yellow vein virus (BNYVV), a possible
member of the furovirus group (Shirako & Brakke,
1984), has a quadripartite genome consisting of 5'capped, 3'-polyadenylated plus-sense R N A (Bouzoubaa
et al., 1987). BNYVV is transmitted in the field by the
soil-borne fungus Polymyxa betae and is found associated
with roots of sugar beet displaying symptoms of
rhizomania disease (Tamada & Baba, 1973). Under these
conditions all four genome components are invariably
present (Koenig et al., 1986; Bouzoubaa et al., 1988;
Tamada et al., 1989). Isolates of BNYVV maintained in
the laboratory, on the other hand, are customarily
propagated by mechanical inoculation onto leaves of
Chenopodium quinoa or Tetragonia expansa and in these
circumstances neither of the two smallest genomic
RNAs, RNA-3 [approximately 1850 nucleotides (nO
including the 3' poly(A) tail] and RNA-4 (approximately
1550 nt), is required for infection (Koenig et aL, 1986;
t Present address: Max-PlanckInstitut ffirBiochemie,Martinsried,
8033 Miinchen, Germany.
Present address: Institut des SciencesV6g6talesdu CNRS, 91198
Gif-sur-Yvette, France.
0000-9923 © 1991 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52
260
S. Bouzoubaa and others
R N A - 3 and R N A - 4 found in several B N Y V V isolates
(Bouzoubaa et al., 1985, 1988). This work has shown that
the deletions arise from a loss o f a continuous stretch of
sequence rather than several non-contiguous segments,
and that the deletions lie within the R N A - 3 and -4
molecules, leaving the extremities intact. T h e latter
observation is consistent with the finding, based on
deletion mutagenesis of a biologically active R N A - 3
transcript, that the approximately 300 5'-terminal residues and the 70 3'-terminal residues bear sequences
indispensable in cis for successful amplification o f R N A 3 in infected leaves (Jupin et al., 1990).
Study o f the deleted R N A - 3 and R N A - 4 forms in
natural B N Y V V isolates has two drawbacks. First, the
R N A - 3 and R N A - 4 in such isolates almost certainly
consist of a population o f ctosely related sequences
(Holland et al., 1982; D o m i n g o & Holland, 1988) causing
consequent uncertainty as t o the exact nature of the form
that has undergone deletion (sequence analysis of deleted
molecules, of course, provides no information about the
sequence o f the segment deleted from the full-length
form). Second, it is difficult to assess when, and how
frequently, deletions occur in natural isolates, m a n y o f
which have been propagated in leaves for long periods of
time. It is evident that use o f a B N Y V V isolate produced
by transcription of cloned viral c D N A circumvents such
problems by providing a homogeneous starting material
and a means of 'setting the clock to zero' with respect to
the analysis of deletion. In this paper we describe
experiments in which such a transcript-based isolate,
derived from cloned c D N A of R N A o f B N Y V V isolate
F3 (Ziegler-Graff et al., 1988; Quillet et al., 1989), was
used to examine the deletion process. W e have also
characterized a Y-terminally truncatecb I~NA-3 species
found in plants infected with either natural or transcriptderived B N Y V V isolates. This species, termed R N A 3sub, is not encapsidated and corresponds to the
approximately 600 3'-terminal residues o f R N A - 3 . It is
probably a subgenomic R N A .
Methods
Virus. BNYVV isolate F3 and infectious run-off transcripts of
BNYVV RNA-I to -4 have been described elsewhere (Ziegler et al.,
1985; Ziegler-Graff et al., 1988; Quillet et al., 1989). Virus isolates were
propagated by mechanical inoculation of local lesions homogenized in
50 mM-potassiumphosphate pH 7-2, or mixtures of transcripts (Quillet
et al., 1989) onto leaves of C. quinoa. Virus was purified from infected
leaves as described by Putz & Kuszala (1978).
Extraction and analysis of virat RNA. RNA was purified from virions
by phenol extraction. Total RNA was isolated from virus-infected C.
quinoa leaves (Bouzoubaa et al., 1986) or sugar beet roots (Lemaire et
al., 1988) and its viral RNA content was analysed by Northern
hybridization using 32p-labelled antisense RNA transcripts as probes
(Lemaire et al., 1988). The probes were obtained by bacteriophage T3
or T7 RNA polymerase (Stratagene) run-off transcription of appropriate restriction fragments cloned into Bluescribe (Stratagene). The
probes were complementary, in the case of RNA-3, to nt 1 to 740 (probe
3A; see Fig. 3a), nt 754 to 1035 (probe 3B), nt 1036to 1470 (probe 3C)
or nt 1471 to 1773 plus the 3' poly(A) tail (probe 3D) and, in the case of
RNA-4, to nt 35 to 680 (probe 4A; see Fig. 3b), nt 857 to 1186 (probe
4B) or nt 1187 to 1256 (probe 4C). The probes used to detect BNYVV
RNA-I and -2 were as described (Lemaire et al., 1988).
Recombinant pUC plasmids containing inserts of viral cDNA were
prepared by the method of Heidecker & Messing (1983). Sequence
analysis of cDNA inserts with Sequenase (USB) or T7 DNA
polymerase (Pharmacia) was carried out on denatured plasmid DNA
(Zhang et al., 1988) with universal primers (Pharmacia) or custommade primers designed to hybridize to different regions of the cDNA
insert. Other recombinant DNA procedures were essentially as
described (Sambrook et al., 1989).
RNA-3sub was purified by centrifugation of 200 lag of total RNA
from infected leaves through a 5 to 20~ sucrose gradient made up in 10
mM-Tris-HCl pH 7-5, 1 mM-EDTA, 0.1 ~ SDS. Centrifugation was for
14 h at 32000 r.p.m, at 5 °C in a Beckman SW41 rotor. Gradient
fractions enriched in RNA-3 and RNA-3sub were identified by
Northern hybridization analysis of samples. Run-off primer extension
to localize the 5' terminus of RNA-3sub was essentially as described
(Bouzoubaa et al., 1986)using a 5'-3zp-labelled synthetic oligodeoxynucleotide complementary to nt 1356 to 1376 of RNA-3 to prime cDNA
synthesis. Sequence analysis of the 5'-32p-labelled cDNA obtained in
this way and by extension from a second primer (complementary to nt
1646 to I676) was done by using the partial chemical degradation
method (Maxam & Gilbert, 1977).
Results
lnternal deletions in transcript-derived R N A - 3 and
RNA-4
B N Y V V isolates maintained by mechanical inoculation
onto leaves often accumulate specific deleted forms o f
R N A - 3 and R N A - 4 . W e have shown previously that
bacteriophage T7 R N A polymerase run-off transcripts
produced from cloned c D N A and corresponding to
R N A - I to -4 o f B N Y V V isolate F3 are infectious and
induce s y m p t o m s in leaves indistinguishable from those
induced by natural B N Y V V isolates (Quitlet et al., 1989).
In a first experiment, a mixture of R N A - 1 , -2, -3 and -4
transcripts was used to infect C. quinoa leaves. These
leaves in turn served as inocula to initiate four additional
serial passages of the virus; the inoculum for each
passage was an extract of c o m b i n e d local lesions from the
preceding passage. Fig. 1 shows the a m o u n t s o f R N A - 3
(Fig. 1 a) and R N A - 4 (Fig. 1 b) in infected leaf tissue after
each passage. In addition to full-length R N A - 3 and -4,
which were present in all samples, discrete shortened
R N A - 3 - and RNA-4-related forms could be detected
after the first or second passage. T h e deleted R N A - 3
species, termed R N A - 3 a , was about 1500 nt in length,
and the deleted R N A - 4 species were about 1250 nt
( R N A - 4 a ) and 1100 nt ( R N A - 4 b ) in length. Fig. 1 also
shows the R N A - 3 and R N A - 4 content o f C. quinoa leaves
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52
Shortened forms of BN Y V V RNA-3 and -4
(a)
F3
1
2
3
4
5
261
(a)
1
2
3
4
5
6
7
8
9
10
F3
(b)
1
2
3
4
5
6
7
8
9
10
F3
~3a
3at~
3bil~
(b)
F3
1
2
3
4
5
Fig. 2. Viral RNA contents of single local lesions (lanes 1 to 10) on C.
quinoa infected with BNYVV RNA-I to -4 transcripts. Northern blot
4b~
hybridization was as in Fig. 1 using probe 3A (a) or probe 4A (b). Other
details are as in Fig. 1.
Fig. 1. Appearance of deleted forms of RNA-3 (a) and RNA-4 (b)
during serial passage of a BNYVV isolate produced by inoculating
BNYVV RNA-1 to -4 transcripts to C. quinoa leaves. Formaldehydedenatured RNA from leaves infected with natural isolate F3 (lane F3)
and from passages I to 5 (lanes 1 to 5, respectively) of the transcriptderived isolate were subjected to agarose gel electrophoresis. Viral
RNA was detected with the RNA-3-specific antisense RNA-3 probe
3A (a) or the RNA-4-specific probe 4A (b). The positions of full-length
RNA-3 and -4 are indicated by asterisks. The deleted forms of RNA-3
and -4 referred to in the text are indicated by 3a, 3b, 4a and 4b.
tion (Fig. 2). In addition to full-length RNA-3, which
was abundant in all local lesions, a species with a
mobility like that of RNA-3a was present in small
amounts in lesions 3 to 5 and a 1200 nt species (RNA-3b)
was detected in local lesion 2 (Fig. 2a). When the same
RNA samples were probed with an RNA-4-specific
probe, full-length RNA-4 was detected in all cases except
for local lesions 8 and 10 (Fig. 2b). The former contained
no RNA-4-related species but the latter contained the
approximately 1100 nt RNA-4b.
Mapping the internal deletions
infected with BNYVV isolate F3, the isolate initially
used to prepare the cDNA clones employed to synthesize
the RNA-3 and RNA-4 transcripts (Ziegler-Graff et al.,
1988). Since then, isolate F3 has been carried through 30
successive bulk passages in leaves over a 3 year period. It
can be seen that, in addition to full-length RNA-3 and
RNA-4, the isolate has accumulated two deleted forms of
RNA-3 and a deleted RNA-4. One of the RNA-3-related
forms comigrated with RNA-3a from the transcriptderived isolate; the deleted version of RNA-4 had a
mobility similar to that of the shortest transcript-derived
RNA-4 form.
In order to gain further insight into the frequency with
which deletions occur, individual local lesions from
leaves inoculated with the RNA-I to -4 transcript
mixture were extracted separately and their RNA-3 and
RNA-4 content was evaluated by Northern hybridiza-
The above results illustrate that discrete shortened forms
of BNYVV RNA-3 and RNA-4 can appear spontaneously in a transcript-derived BNYVV isolate even after
only one or two passages in leaves. The deleted forms of
RNA-3 and RNA-4 observed previously in several
natural BNYVV isolates have been shown to arise by
elimination of internal sequences. Northern blot hybridization with probes corresponding to different portions of
the RNA-3 and RNA-4 sequence has revealed this also
to be the case for the various shortened transcriptderived species observed in Fig. 1 and 2. Examples of
such analysis are shown in Fig. 3. All of the RNA-3specific probes except probe 3B hybridized with RNA3a (Fig. 3a), indicating that the RNA-3 sequence
eliminated by the deletion process encompasses nt 754 to
1035. Similarly, only probes 4A and 4C hybridize with
the approximately 1100 nt RNA-4b species in BNYVV
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52
262
S. B o u z o u b a a and others
(a)
1
2
3
4
D
RNA 3
3A'
(b) 1
I
2
313 I
3
13D
3c
4
5
6
RNA 4
31K
4A
II
4B
I 4C
isolate F3 and in local lesion 10 of Fig. 2(b), indicating
that the sequence between nt 858 and 1186 was lost in the
deletion process. The R N A preparation shown in Fig.
3(a) also contained an approximately 600 nt RNA-3related species (indicated by S in Fig. 3a) which
hybridized only with the two probes representing the 3"terminal portion of the R N A - 3 sequence. This species is
discussed more fully below.
The boundaries of the deletions in several of the
shortened R N A - 3 and -4 variants were located more
precisely by sequence analysis of cloned c D N A . In one
experiment, virions were purified from passage 4 of the
experiment shown in Fig. 1 and R N A extracted from the
virions was used to generate c D N A clones by the method
of Heidecker & Messing (1983). Clones containing long
RNA-3 c D N A inserts were selected by colony hybridization with probe 3A. The 73 RNA-3-positive clones
detected in this way were then screened with probe 3B to
counterselect for RNA-3 inserts that fail to hybridize
with this segment. In spite of the fact that R N A - 3 a was
almost as abundant as full-length RNA-3 in the R N A
preparation only two of the recombinant c D N A plasmids did not hybridize with probe 3B and were therefore
apparently derived from the deleted form. Sequence
analysis of the c D N A inserts of these two plasmids
revealed that both were identical to full-length RNA-3
except for the loss of a segment between nt 713 and 1051
(Fig. 4a). In the same fashion, virus was prepared from
leaves inoculated with R N A from either local lesion 2 or
local lesion 10 in Fig. 2 and c D N A plasmids generated
from the viral R N A were probed for R N A - 3 or RNA-4specific sequences, respectively. A single clone containing an internal deletion (nt 438 to 1090) was obtained in
the former case and in the latter case six clones
containing internally deleted R N A - 4 sequences were
observed. The deleted R N A - 4 sequences fell into two
classes. Four plasmids lacked the sequence between
approximately nt 724 and 1164 [the exact limits of the
deletion cannot be specified because of sequence
redundancy at the deletion termini (Fig. 4a)] and two
lacked the sequence between nt 697 and 1194.
Fig. 3. Mapping the deletions in RNA-3a (a) and RNA-4b (b) by
Northern hybridization using different antisense RNA probes. The
position of each probe on the RNA-3 and -4 ~quence is indicated in the
lower portion of the figure. (a) Samplesof RNA from passage 5 in Fig.
1 were hybridized with probe 3A (lane 1), probe 3b (lane 2), probe 3C
(lane 3) or probe 3D (lane 4). S indicates the position of RNA-3sub and
the arrow RNA-3a. The band in lane 4 denoted by an open circle arises
from cross-hybridizationof RNA-2 with the Y-terminal homologous
domain (Bouzoubaa et al., 1987) present in probe 3D. (b) RNA from
isolate F3 (lanes l, 3 and 5) and from local lesion 10 from Fig. 2 (lanes 2,
4 and 6) were hybridized with probe 4A (lanes 1 and 2), probe 4B (lanes
3 and 4) or probe 4C (lanes 5 and 6). The arrowhead indicates RNA-4b.
The positions of full-length RNA-3 and -4 are indicated by asterisks.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52
Shortened forms of B N Y V V RNA-3 and -4
(a)
L
2
3
263
4
339 nt
RNA-3a
ACCUURL1U(~(~GACALUC...
702
• . . LV~U(~IJUUAG/g,JG(~JC~UA
1060
652 n t
RNA-3b
CAUCAUr,J A / g 3 U C , ~ ( ~ F , ~ U . . .
427
RNA-4b
•.. ~ G U t G A U G A U G A U G
1099
~98 nt
• . . C~JACI~ I ~ A C
689
721
•
(~"UL~C
1202
•
3a~
1181
(b)
RNA-3sub
25K
I
I
I1239
•. ~GGU(gJAAUCOJCC..
--
I
II
Fig. 4. (a) Sequence near the boundaries of the internal deletions
arising in transcript-derived RNA-3 and -4. The 15 nt sequence repeats
at the deletion boundaries of one of the RNA-4b variants are
underlined with mismatched positions in the repeat indicated by closed
circles. (b) Map of RNA-3sub. The open reading frame for the Mr
25000 protein (Bouzoubaa et al., 1985) and Mr 4600 protein are shown
as open rectangles. The approximate intitiation site for RNA-3sub
synthesis is indicated by an arrow.
Characterization of RNA-3sub
As shown in Fig. 3 (a), total R N A extracted from leaves
infected with a transcript-derived B N Y V V isolate
contains, in addition to full-length and internally deleted
forms of RNA-3, a second RNA-3-related species of
about 600 nt which does not hybridize with probes
specific for the 5'-terminal portion of the R N A - 3
sequence. This species, which we refer to as RNA-3sub,
is similar in size and is presumably identical to an R N A 3-related species previously observed but not characterized by Burgermeister et al. (1986). RNA-3sub appeared
(Fig. 5c) in the first passage after inoculation of fulllength RNA-3 transcript, along with RNA-1 and -2, and
was present in leaves infected with all natural RNA-3containing B N Y V V isolates that we have tested (data
not shown). It was also present in the roots of sugar beet
infected with B N Y V V using viruliferous P. betae (Fig.
5d). RNA-3sub could not be detected in R N A extracted
from purified virions (Fig. 5a) and was sensitive to
degradation by endogenous nucleases in crude homogenates of infected leaves (data not shown) indicating that
RNA-3sub, unlike the various internally deleted RNA-3
and R N A - 4 species, is not encapsidated.
RNA-3sub binds to an oligo(dT)-cellulose column
(data not shown) indicating that, like RNA-3, it
possesses a 3' poly(A) tail. In order to m a p the 5' terminus
of the small R N A with precision, R N A - 3 s u b and fulllength R N A extracted from infected C. quinoa leaves
Fig. 5. Northern hybridization analysis for the presence of RNA-3sub
in purified BNYVV virions (lane 1) and in total RNA extracted from
BNYVV-infected C. quinoa leaves (lanes 2 and 3) or sugar beet roots
(lane 4). The inoculum used to infect the leaves analysed in lane 3 was
an RNA-1 to -4 transcript mixture• The inoculum for lanes 1 and 2 was
a subsequent passage of the transcript-derived isolate and contains the
internally deleted RNA-3a (indicated by 3a) as well as the full-length
RNA-3 (asterisk). S denotes RNA-3sub. Probe 3C was used to visualize
the RNA-3-related species.
were separated from one another by sedimentation
through a sucrose gradient. Fractions highly enriched in
either RNA-3sub or RNA-3 were pooled and used as
templates for primer extension by reverse transcriptase.
The synthetic 5'-32P-labelled oligodeoxyribonucleotide
which served to prime c D N A synthesis was complementary to nt 1356 to 1376 of the RNA-3 sequence. U p o n
extension of the primer with reverse transcriptase two
abundant c D N A run-off species (differing by about one
nucleotide residue in mobility) were produced (Fig. 6,
lane 1) which were absent when full-length RNA-3 was
the template (Fig. 6, lane 2). The lengths of the run-off
transcripts were estimated at 145 or 146 nt by reference
to a D N A sequence ladder. This fixes the 5' terminus of
RNA-3sub at about nt 1230 of RNA-3 provided that
RNA-3sub and RNA-3 are coUinear. Direct evidence for
collinearity between RNA-3sub and R N A - 3 was obtained by sequence analysis of the c D N A synthesized
from the RNA-3sub template. When 5'-32p-labelled runo f f c D N A such as that shown in Fig. 6 was sequenced by
partial chemical degradation (Maxam & Gilbert, 1977)
the resulting sequence was identical to the RNA-3
negative strand sequence up to nt 1234, the point at
which the sequence ladder becomes unreadable because
it merges with the band of non-degraded D N A at the top
of the ladder (data not shown). Similar sequence analysis
of c D N A primed further downstream on RNA-3sub
extended the portion of the sequence which was shown
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52
264
S. Bouzoubaa and others
1
2
Fig. 6. Mapping the 5' extremity of RNA-3sub by primer extension.
The 5'-32p-labelled primer was complementary to nt 1356 to 1376 of
RNA-3. The products of reverse transcription were separated by
electrophoresis through a 6 ~ polyacrylamidesequencing gel alongside
a DNA sequence ladder (not shown) to provide size markers. The
templates for reverse transcription were sucrose gradient-enriched
RNA-3sub (lane l) or full-length RNA-3 (lane 2). Distances in nt from
the primer binding site are indicated on the left. The arrowhead shows
the position of the pair of closely spaced abundant run-off reverse
transcripts of RNA-3sub.
by direct sequence analysis to be collinear with RNA-3 to
within about 110 nt of the 3' poly(A) tail. Determination
of the sequence closer to the 3' extremity in this way was
not possible because the RNA-3sub preparation contained some R N A - 4 and the extensive sequence homology between RNA-3 and -4 near the 3' terminus
(Bouzoubaa et al., 1985) precludes design of a synthetic
oligodeoxynucleotide primer which will bind specifically
to RNA-3 in this region.
Discussion
In this paper we have shown that full-length B N Y V V
R N A - 3 and R N A - 4 produced by in vitro transcription of
cloned c D N A can, when co-inoculated with RNA-1 and
-2, undergo internal deletions after only one or two
passages in C. quinoa leaves. Analysis of the viral R N A
content of infected leaves by gel electrophoresis reveals
discrete bands of deleted R N A - 3 and -4 rather than a
continuum of shortened forms. In the case of R N A - 3 the
deleted species R N A-3a and -3b correspond in size to the
abundant deleted forms present after extensive passage
in leaves of B N Y V V isolate F3, the isolate from which
the c D N A clones used to generate the prototype fulllength RNA-3 and -4 transcripts were derived. Thus our
observations suggest that the rapid deletion events
observed with transcript-derived R N A - 3 can account for
the appearance of the similar forms in the natural isolate.
Comparable findings have been reported for D-satellite
(D-sat) R N A of cucumber mosaic virus where a point
sequence heterogeneity, appearing at the same site at
which heterogeneity occurs in natural D-sat populations,
can develop within two to four passages after inoculation
with transcript R N A (Kurath & Palukaitis, 1990).
In the case of B N Y V V R N A - 4 the 1100 nt R N A - 4 b
was the only RNA-4-related species detected in one local
lesion from a leaf infected with the transcript-derived
inoculum. Bulk serially passaged material on the other
hand contained in addition to RNA-4b, a deleted
R N A - 4 species of about 1250 nt which clearly had no
counterpart in the natural isolate. Perhaps this 1250 nt
species is a quasi-stable intermediate which is subject to
further deletion to produce the 1100 nt form.
In order to fix the boundaries of some of the deletions
more precisely, two clones with deletion-bearing c D N A
inserts were obtained from viral R N A containing
RNA-3a, a single clone for R N A - 3 b and six clones for
RNA-4b. The limits of the deletions, defined by
sequence analysis of such clones (Fig. 4a), were
consistent with the observed lengths of the deleted forms.
It is evident, however, that not enough clones were
analysed in the case of RNA-3 to allow conclusions to be
drawn about possible fine-scale heterogeneity at the
deletion boundaries. Indeed, such heterogeneity was
detected for the R N A - 4 b where four of six clones
examined were of one type (nt 724 to 728 to nt 1164 to
1168 deleted; the uncertainty concerning the exact site of
the deletion in the prototype sequence is due to the
presence of a 4 nt sequence redundancy at the deletion
boundaries) whereas the sequence deleted in the other
two spanned the same region but was slightly longer (nt
697 to 1194 deleted).
We regard copy choice during replication, i.e. switching of the replicase-nascent strand complex from one
point on the template strand to another point further
downstream on the same or another R N A molecule, as
the most plausible mechanism for generation of the
internal deletions in RNA-3 and -4. Such a mechanism is
the currently favoured mechanism for viral R N A
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52
Shortened forms of B N Y V V RNA-3 and -4
recombination (Kirkegaard & Baltimore, 1986) and the
formation of defective interfering RNA (Holland, 1986).
In view of the apparent specificity of the deletion process
in BNYVV we have examined the prototype RNA-3 and
RNA-4 sequences for unusual features at the deletion
boundaries. In the case of the more abundant of the two
characterized RNA-4 deletion variants, a 15 nt direct
sequence repeat (with two mismatches) is found at the
boundaries, with the deletion having occurred in such a
way that one of the two repeats is eliminated (Fig. 4a).
Evidently, a sequence repetition of this sort would aid
reassociation of a detached replicase-nascent strand
complex to the ne.w site on the template, although it does
not explain why disassociation occurred at this point in
the first place. It is also noteworthy that the copy of the
sequence repeat conserved during the deletion process is
the one nearer the 3' terminus, indicating that the
putative template switch occurred during negative
strand RNA synthesis.
No extensive sequence duplications were present at
the limits of the other RNA-3 and -4 deletions
characterized. We have noted, however, that the
sequence deleted in RNA-3a resembles a nuclear premRNA intron in that it begins with GU and ends with
AG (Fig. 4a). Certain other features at the deletion
boundaries of RNA-3a also resemble splice junction
consensus sequences (Brown, 1986), most notably the Urich tract preceding the 3' boundary. Thus a postreplicational mechanism involving splicing cannot be
ruled out for the generation of RNA-3a. There is no
evidence, however, for a nuclear phase in the BNYVV
replication cycle or for any particular association with
subcellular organelles such as chloroplasts or mitochondria. Thus encounters between RNA-3 and the nuclear
or organelle splicing machinery, if they occur, would
presumably be infrequent. In any event, it will now be
possible to test the role of the boundary sequences and
other features of RNA-3 and -4 in the deletion process by
altering the sequence by site-directed mutagenesis at the
cDNA level and exploring the effects of such changes
when RNA-3 and -4 molecules carrying such alterations
are propagated on plants.
In this paper we have shown that, in addition to fulllength and internally deleted forms of RNA-3 and
RNA-4, BNYVV-infected tissue also contains another
type of RNA-3-related species, termed RNA-3sub.
RNA-3sub is collinear with most of (presumably all of)
the T-terminal portion of RNA-3 with its 5' terminus
mapping at about nt 1230 (Fig. 4b). RNA-3sub is distinct
from the internally deleted forms of RNA-3 in that it
always appeared in infected tissue whenever natural
RNA-3 or RNA-3 transcript was present in the
inoculum, it lacks the 5'-terminal portion of the RNA-3
sequence and it was not encapsidated. We have shown
265
elsewhere (Jupin et al., 1990) that a 5'-terminal domain of
about 300 nt is necessary for productive replication of
RNA-3 in plants. Thus it appears likely that RNA-3sub
cannot exist independently but is synthesized de novo
from RNA-3 in the course of infection, i.e. that RNA3sub is a subgenomic RNA. Site-directed mutagenesis of
the RNA-3 prototype sequence should allow us to map
sequences important in cis for its synthesis.
The failure of RNA-3sub to undergo encapsidation
suggests that the 5'-terminal portion of RNA-3 which is
absent from the subgenomic RNA must contain a
sequence indispensable in cis for packaging. Recent
analysis of RNA-3 variants produced by site-directed
mutagenesis of the prototype cDNA clone has mapped
such a signal to near nt 200 (unpublished observations).
An open reading frame for an Mr 4600 protein begins
near the RNA-3sub 5' terminus (Fig. 4b) but there is no
evidence as yet either for or against its expression in vivo.
Thus the role, if any, of RNA-3sub in the viral infection
cycle remains to be established.
The authors would like to acknowledge the technical assistance of
Dani61e Scheidecker and financial support from the ITB.
References
BOUZOUBAA,S., GUILLEY, H., JONARD, G., RICHARDS,K. & PUTZ, C.
(1985). Nucleotide sequence analysis of RNA-3 and RNA-4 of beet
necrotic yellow vein virus, isolates F2 and G1. Journal of General
Virology 66, 1553-1564.
BOUZOUBAA,S., ZIEGLER, V., BECK, D., GUILLEY, H., RICHARDS, K. &
JONARD,G. (1986). Nucleotide sequence of beet necrotic yellow vein
virus RNA-2. Journal of General Virology 67, 1689-1700.
BOUZOUBAA,S., QUILLET, L., GUILLEY, H., JONARD,G. t~ RICHARDS,
K. (1987). Nucleotide sequence of beet necrotic yellow vein virus
RNA-1. Journal of General Virology 68, 6154526.
BOUZOUBAA,S., GUILLEY, H., JONARD, G., JUPIN, I., QUILLET, L.,
RICHARDS, K., SCHEIDECKER, D. & ZIEGLER-GRAFF, V. (1988).
Genome organization and function of beet necrotic yellow vein
virus. In ViruseswithFungal Vectors,Developmentsin Applied Biology,
vol. 2, pp. 99-110. Edited by J. I. Cooper & M. J. C. Asher.
Wellesbourne : Association of Applied Biologists.
BROWN, J. W. S. (1986). A catalogue of splice junctions and putative
branch point sequences from plant introns. Nucleic Acids Research
14, 9549-9559.
BURGERMEISTER, W., KOENIG, R., WEICH, H., SEBALD, W. 8/;
LESEMANN,D. E. (1986). Diversity of the RNAs in thirteen isolates of
beet necrotic yellow vein virus in Chenopodium quinoa detected by
means of cloned cDNAs. Journal of Phytopathology 115, 229-242.
DOMINGO, E. & HOLLAND,J. J. (1988). High error rates, population
equilibrium, and evolution of RNA replication systems. In RNA
Genetics, vol. 3, pp. 3-36. Edited by E. Domingo, J. J. Holland & P.
Ahlquist. Boca Raton: CRC Press.
HEIDECKER,G. & MESSING,J. (1983). Sequence analysis ofzein cDNAs
obtained by an efficient mRNA cloning method. Nucleic Acids
Research 11, 4891-4906.
HOLLAND, J. J. (1986). Generation and replication of defective viral
genomes. In Virology, pp.77-99. Edited by B. N. Fields & D. M.
Knipe. New York: Raven Press.
HOLLAND,J., SPINDLER,K., HORODYSKI,F., GRABAU,E., NICHOL,S. 8£
VAN DE POL, S. (1982). Rapid evolution of RNA genomes. Science
215, 1577-1585.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52
266
S. Bouzoubaa and others
JuPIN, I., RICHARDS,K., JONARD,G. GUILLEY,H. & PLEIJ, C. W. A.
(1990). Sequences required for BNYVV RNA 3 amplification.
Virology 178, 273-280.
KIRKEGAARD,K. & BALTIMORE,D. (1986). The mechanism of RNA
recombination in poliovirus. Cell 47, 433-443.
KOENIG, R. & BURGERMEISTER,W. (1989). Mechanical inoculation of
sugar beet roots with isolates of beet necrotic yellow vein virus
having different RNA compositions. Journal of Phytopatbology 124,
249-255.
KOENIG, R., BURGERMEISTER,W., WEICH, H., SEBALD,W. & KOTHE,
C. (1986). Uniform RNA patterns of beet necrotic yellow vein virus
in sugar beet roots, but not in leaves from several plant species.
Journal of General Virology 67, 2043-2046.
KURATH, G. & PALUKAITIS,P. (1990). Serial passage of infectious
transcripts of a cucumber mosaic virus satellite RNA clone results in
sequence heterogeneity. Virology 176, 8-15.
KUSZALA,M., ZIEGLER,V., BOUZOUBAA,S., RICHARDS,K., PUTZ, C.,
GUILLEY, H. & JONARD,G. (1986). Beet necrotic yellow vein virus:
different isolates are serologically similar but differ in RNA
composition. Annals of Applied Biology 109, 155-162.
LEMAIRE, O., MERDINOGLU,D., VALENTIN, P., Ptrrz, C., ZIEGLERGRAFF, V., GUILLEY,H., JONARD,G. & RICHARDC$,K. (1988). Effect
of beet necrotic yellow vein virus RNA composition on transmission
by Polymyxa betae. Virology 162, 232-235.
M A x , , A. M. & GILBERT,W. (1977). A new method for sequencing
DNA. Proceedingsof the National Academy of Sciences, U.S.A. 74,
560-564.
PUTZ, C. & KtISZ_ALA,M. (1978). La rhizomanie de la betterave sucri&e
en Alsace. Recherche d'une novelle re&bode de purification du beet
necrotic yellow vein virus. Annals of Phytopathology 10, 247-262.
QVlLLET,L., GUmLEY,H., JON~XD, G. & RICHARDS,K. (1989). In vitro
synthesis of biologically active BNYVV RNA. Virology 172, 293301.
SAMBROOK, J., FRrrSCH, E. F. & MANIATIS, T. (1989). Molecular
Cloning: A Laboratory Manual, 2nd edn. New York: Cold Spring
Harbor Laboratory.
SHIRAKO, Y. & BRAKKE, M. K. (1984). Two purified RNAs of soilborne wheat mosaic virus are needed for infection. Journalof General
Virology 65, 119-127.
TAMADA,T. & ABE, H. (1989). Evidence that beet necrotic yellow vein
virus RNA-4 is essential for efficient transmission by the fungus
Polymyxa betae. Journal of General Virology 70, 3391-3398.
TAMADA,T. & BAnA, T. (1973). Beet necrotic yellow vein virus from
rhizomania-affected sugar beet in Japan. Annals of the Phytopathological Society of Japan 39, 325-332.
TAMADA,T., SmRAKO,Y., AaE, H., SAITO,M., Kmucm, T. & HARADA,
T. (1989). Production and pathogenicity of isolates of beet necrotic
yellow vein virus with different numbers of RNA components.
Journal of General Virology 70, 3399-3409.
ZHANG,H., SCHOLL,R., BROWSE,J. & SOMERVILLE,C. (1988). Doublestranded DNA sequencing as a choice for DNA sequencing. Nucleic
Acids Research 16, 1220.
ZIEGLER, V., RICHARDS,K., GUILLEY, H., JONARD, G. & Ptrrz, C.
(1985). Cell-free translation of beet necrotic yellow vein virus:
readthrough of the coat protein cistron. Journal of General Virology
66, 2079-2087.
ZIEGLER-GRAFF,V., BOUZOUBAA,S., JUPIN, I., GUILLEY,O., JONARD,
G. & RlCHARDS,K. (1988). Biologically active transcripts of beet
necrotic yellow vein virus RNA-3 and RNA-4. Journal of General
Virology 69, 2347-2357,
(Received 17 September 1990; Accepted 6 November 1990)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 18:20:52