Virology 292, 114–126 (2002)
doi:10.1006/viro.2001.1268, available online at http://www.idealibrary.com on
The 3⬘-Terminal Structure Required for Replication of Barley Yellow Dwarf Virus RNA
Contains an Embedded 3⬘ End
Gennadiy Koev, 1 Sijun Liu, Randy Beckett, and W. Allen Miller 2
Plant Pathology Department, 351 Bessey Hall, Iowa State University, Ames, Iowa 50011-1020
Received September 7, 2001; returned to author for revision October 19, 2001; accepted October 31, 2001
We determined the 3⬘-terminal primary and secondary structures required for replication of Barley yellow dwarf virus
(BYDV) RNA in oat protoplasts. Computer predictions, nuclease probing, phylogenetic comparisons, and replication assays
of specific mutants and chimeras revealed that the 3⬘-terminal 109 nucleotides (nt) form a structure with three to four
stem-loops followed by a coaxially stacked helix incorporating the last four nt [(A/U)CCC]. Sequences upstream of the 109-nt
region also contributed to RNA accumulation. The base-pairing but not the sequences or bulges in the stems were essential
for replication, but any changes to the 3⬘-terminal helix destroyed replication. The two 3⬘-proximal tetraloops tolerated all
changes, but the two 3⬘-distal tetraloops gave most efficient replication if they fit the GNRA consensus. A mutant lacking the
3⬘-proximal stem-loop produced elevated levels of less-than-full-length minus strands, and no (⫹) strand. We propose that
a “pocket” structure is the origin of (⫺)-strand synthesis, which is negatively regulated by the inaccessible conformation of
the 3⬘ terminus, thus favoring a high (⫹)/(⫺) ratio. This 3⬘ structure and the polymerase homologies suggest that genus
Luteovirus is more closely related to the Tombusviridae family than to other Luteoviridae genera. © 2002 Elsevier Science
Key Words: tetraloop; RNA replication; minus-strand promoter; origin of replication; Luteoviridae; Tombusviridae; plant RNA
virus.
naviridae (Hsue and Masters, 1997; Williams et al., 1999),
Pestivirus (Yu et al., 1999), Rubivirus (Chen and Frey,
1999), and others contain stable secondary and tertiary
structural elements that are indispensable for replication. Host factors and viral replication proteins can bind
specifically to such structures, and often their binding
ability correlates with replication competence (reviewed
by Lai, 1998).
Many viral RNAs contain poly(A) tails at their 3⬘ termini. A poly(A) tail is required for stability and translation
initiation of nonviral eukaryotic mRNAs (Sachs, 2000).
Poly(A) tails on viral RNAs can also play a role in replication. The RNA-dependent RNA polymerase (RdRp) of
Zucchini yellow mosaic potyvirus was shown to bind
poly(A) binding protein, indicating a direct role for the
poly(A) tail in recruiting the replicase to the 3⬘ end of
potyviral RNA (Wang et al., 2000). The poly(A) tail of
Bamboo mosaic potexvirus RNA participates in a
pseudoknot that serves as a replicase binding site
(Huang et al., 2001). The polyadenylated genomes of
coronaviruses and potyviruses also contain stem-loops
in their 3⬘-UTRs that are required for replication (Haldeman-Cahill et al., 1998; Hsue and Masters, 1997). Specific pseudoknots or kissing stem-loops are necessary in
the 3⬘-UTRs for efficient replication of picornaviruses
(Melchers et al., 1997; Mirmomeni et al., 1997; Pilipenko
et al., 1996). On the other hand, deletion of entire 3⬘-UTRs
in poliovirus and human rhinovirus, while severely inhib-
INTRODUCTION
Initiation of RNA replication of positive-strand RNA
viruses takes place at the 3⬘ terminus of the genomic
RNA. Thus, a variety of 3⬘-terminal structures have
evolved that serve as promoters for minus-strand RNA
synthesis (Buck, 1996; Dreher, 1999). These include
tRNA-like structures (TLSs) in the Bromoviridae family
and Tobamovirus, Tymovirus, and Hordeivirus genera
(Dreher, 1999) and other less-characterized structures in
other RNA viruses. The TLS is required for replication of
Bromoviridae RNA because it contains promoter elements for replicase binding (Chapman and Kao, 1999)
and initiation (Dreher and Hall, 1988) of (⫺)-strand RNA
synthesis. The specific structure bound by BMV replicase has been determined at high resolution (Kim et al.,
2000).
RNAs of many other viruses have 3⬘-termini that do not
mimic tRNAs, but contain various conserved structural
elements that play a role in RNA replication. The 3⬘untranslated regions (UTRs) of viruses in such unrelated
groups as Tombusviridae (Song and Simon, 1995), Coro1
Current address: Howard Hughes Medical Institute, Dept. of Molecular Microbiology and Immunology, University of Southern California
School of Medicine, 2011 Zonal Ave., HMR-401, Los Angeles, CA 900331054.
2
To whom correspondence and reprint requests should be addressed. Fax: 515-294-9420. E-mail: wamiller@iastate.edu.
0042-6822/02 $35.00
© 2002 Elsevier Science
All rights reserved.
114
BYDV RNA 3⬘ END STRUCTURE
iting viral growth, did not abolish replication, suggesting
that template recognition by the replicase might not be
as stringent as originally believed (Todd et al., 1997).
Subcellular compartmentalization of the replicase with
the template RNA may also play an important role in
specificity of RNA amplification (Todd et al., 1997).
Sequences required for replication initiation are not
confined to the 3⬘ region. The replicase binding site of
bacteriophage Q␤ is 1.5 kb upstream of the 3⬘ terminus.
Long-distance base-pairing then brings the RdRp to the
3⬘ end where initiation takes place (Klovins et al., 1998).
The intergenic region in the middle of BMV RNA 3 harbors an internal replication enhancer (Sullivan and Ahlquist, 1999). In a very different mechanism, a central
stem-loop of poliovirus RNA guides uridylylation of the
genome-linked protein (VPg) which then primes (⫺)strand synthesis at the poly(A) tail (Paul et al., 2000;
Rieder et al., 2000).
In contrast to the concept of specific protein–RNA
recognition by complex interactions of higher order
structures, Dreher and colleagues proposed that, in
some viral RNAs, secondary structure in the promoter
serves mainly to block initiation at the wrong sites. The
promoter within the TLS of Turnip yellow mosaic virus
(TYMV) can be reduced to a single-stranded, 3⬘-terminal
CC(A/G) repeat (Singh and Dreher, 1998). Similarly minimal initiation sites function on TCV and Q␤ RNAs (Yoshinari et al., 2000). The TLS of TYMV, and the very
different secondary structures of TCV and Q␤, were proposed to simply prevent initiation at internal CCA elements and also to ensure that the 3⬘-terminal CCA is
single-stranded (and thus available for recognition by the
replicase) by tying up the adjacent sequence in secondary structure.
Many viruses produce subgenomic RNAs (Miller and
Koev, 2000). This also requires specific interactions between the replicase and the viral RNA to initiate (⫹)strand synthesis within the (⫺) strand (Siegel et al.,
1997), or possibly to terminate at a defined site during
(⫺)-strand synthesis (Sit et al., 1998). The promoters for
subgenomic RNA synthesis also consist of specific
structures, yet they rarely resemble the (⫺)-strand promoter, despite RNA synthesis by the same RdRp.
Barley yellow dwarf virus (BYDV) is a single-stranded,
positive-sense RNA virus, the type member of genus
Luteovirus in the family Luteoviridae (D’Arcy et al., 2000).
It has an uncapped, nonpolyadenylated, 5.7-kb genomic
RNA (gRNA) that contains six open reading frames
(ORFs) (Fig. 1A). During infection, BYDV generates three,
3⬘-coterminal subgenomic RNAs (sgRNAs). Viral gRNA
serves as the mRNA for ORFs 1 and 2; ORF 2 is expressed fused to ORF1 by ribosomal frameshifting (Di et
al., 1993). The role of the 39-kDa protein encoded by
ORF1 is unclear, whereas the 99-kDa frameshift product
of ORFs 1 and 2 is the viral RdRp. Although classified in
115
FIG. 1. (A) Genome organization of BYDV. Horizontal lines
represent genomic and subgenomic RNAs (sgRNA). Open reading
frames (ORFs) are indicated by number (1 through 6), molecular
mass (K), and encoded functions (RdRp, RNA-dependent RNA polymerase; CP, coat protein; MP, putative movement protein; AT, aphid
transmission protein). (B) Sequence alignment of the 3⬘ termini of
BYDV strains: PAV-Australia (PAV-Aus, GenBank Accession No.
X07653), MAV, PAV-Japan (PAV-Jpn, Accession No. D85783), and
PAV-129 (Accession No. NC_002160). The PAV-Aus sequence is
shown with dots indicating gaps. Hyphens indicate bases in the
other isolates that are identical to PAV-Aus. (C) Potential secondary
structures of the 3⬘ termini of PAV-Aus (left) and PAV-129 (right)
isolates as predicted by MFOLD (Mathews et al., 1999; Zuker et al.,
1999). Nucleotides that vary among non-PAV-129 isolates are in
boldface italics on the PAV-Aus structures, with arrow pointing to
variant base. Base changes that preserve proposed base-pairing or
tetraloop motif are boxed. Nucleotide variations that disrupt secondary structure are circled. On the PAV-129 structure, long blocks
of sequences that differ from PAV-Aus are boxed. Other varying
bases are in italics.
116
KOEV ET AL.
Luteoviridae, the RdRp gene of BYDV shares significant
homology to those of Dianthovirus and Necrovirus genera of the Tombusviridae, and no homology to the those
in the Polerovirus genus of the Luteoviridae (Koonin and
Dolja, 1993; Miller et al., 1995). Moreover, similar to other
Tombusviridae, the 3⬘ end of Luteovirus RNA terminates
in CCC, while the Polerovirus genomes terminate in GU.
Previous research on origins of replication of Tombusviridae RNAs revealed specific structures (Song and Simon, 1995; Turner and Buck, 1999) and in some cases
surprisingly high tolerance of variation (Carpenter and
Simon, 1998). Our previous studies have shown that
certain sequences in the 869-nt 3⬘-UTR of the BYDV
genome control translation (Miller et al., 1997; Wang et
al., 1999), but so far they have been located hundreds of
bases upstream of the 3⬘ terminus. We also mapped the
promoters for synthesis of all three sgRNAs (Koev and
Miller, 2000; Koev et al., 1999). They show surprisingly
little sequence or structural similarity. How BYDV or any
virus balances genomic replication with subgenomic
RNA synthesis and translation is unclear. One thing certain is that RNA structures play key roles in these functions. Here, we determine the structures necessary for
genomic RNA accumulation that presumably participates
in (⫺)-strand synthesis.
RESULTS
Predicted folding of the 3⬘ terminus of BYDV RNA
To determine the secondary structure of the BYDV 3⬘
end, we first used computer predictions. Identification of
phylogenetically conserved base-pairing despite sequence variation (covariations) is a powerful tool for
prediction of functional secondary structure. Therefore,
we first aligned the 3⬘ ends of all known BYDV strains.
Most strains show very high conservation throughout
their genomes, but one novel strain, BYDV-PAV-129, varies significantly. All sequenced BYDV isolates other than
PAV-129 share ⬎97% identity in ORFs 1 and 2. PAV-129
(GenBank Accession No. NC_002160), which is very severe and breaks resistance in oat (Chay et al., 1996), has
only 88% amino acid sequence identity in ORF2 (the
RdRp) and 80% in ORF1 to the sequence in our infectious
clone (PAV6), which is a chimera of very similar Illinois
and Australian isolates (Di et al., 1993). 3⬘-end alignment
(Fig. 1B) revealed high conservation among all isolates
except PAV-129. PAV-129 shows similarity in the 3⬘-terminal 30-nt bases, but upstream of that many gaps and
mismatches were present with only small blocks of similarity. This proved to be a powerful guide for secondary
structure prediction.
Using MFOLD (Mathews et al., 1999; Zuker et al.,
1999), we predicted two alternative conformations of the
BYDV (PAV6) 3⬘ terminus (PAV-Aus, Fig. 1C). Both structures harbor four stem-loops (SL1 through SL4 numbered
from the 3⬘ end) that contain terminal tetraloops. The
tetraloops of SL1, SL3, and SL4 fit the GNRA consensus;
the SL2 terminal loop is a UNCG tetraloop. Both GNRA
and UNCG tetraloops confer unusually high stability to
their stem-loop structures due to non-Watson–Crick interactions between the nucleotides in the loops (Cheong
et al., 1990; Heus and Pardi, 1991). The structure with
predicted ⌬G ⫽ ⫺46.9 kcal/mol differs from the “alternative” structure (⌬G ⫽ ⫺45.4) by its shorter SL2 and
base-pairing of the 3⬘-terminal four bases (ACCC 5674–5677)
in a helix that stacks coaxially with SL3 (Fig. 1C).
The slight variation among the BYDV-MAV and BYDVPAV-Jpn strains supported existence of some stems and
tetraloops (Fig. 1C). The base differences in tetraloops of
SL1 and SL3 both maintain the GNRA consensus. The
variable nucleotides in other single-stranded regions do
not change their single-stranded nature. Covariations in
SL3 and SL4 support existence of the base-paired regions, with variations in the internal bulge in SL4. The
closely related sequences did not allow us to distinguish
between the two predicted structures. In contrast, PAV129 3⬘ end can be folded only into the structure with the
embedded 3⬘ terminus (Fig. 1C). The PAV-129 structure
lacks SL2, which perturbed alignment upstream (Fig. 1A),
has a bulge in SL3, and a longer stem and longer loop at
the SL4 position. Except for the proximal end, SL4 bears
no primary sequence similarity to that of PAV6. The PAVAus and PAV-129 structures, from the beginning of SL4 to
the 3⬘ end, end up being exactly the same length.
Nuclease probing of the 3⬘-terminal RNA secondary
structure
To determine the solution structure of the 3⬘ end of
BYDV RNA, we probed a transcript comprising the 3⬘terminal 109 bases with the single-strand-specific, sequence-nonspecific chemical nuclease imidazole and
with ribonuclease T1 (cuts single-stranded guanosine
nucleotides). Most predicted single-stranded bases
were cut strongly with imidazole and RNase T1 (Fig. 2A).
Imidazole failed to cleave the predicted tetraloop nucleotides probably due to the non-Watson–Crick interactions within the loops (Cheong et al., 1990; Heus and
Pardi, 1991). Imidazole is thought to require flexibility at
the phosphodiester bonds to catalyze cleavage (Vlassov
et al., 1995). The GGGU 5619–5622 sequence was insensitive
to imidazole and T1 (Fig. 2), as predicted by both possible structures (Fig. 1C). ACUC 5638–5641 was very sensitive
to imidazole cleavage, which is inconsistent with the
alternative structure in which this region base pairs
with GGGU 5619–5622 (Fig. 1C). Sensitivity of the 3⬘-terminal ACCC 5674–5677 could not be discerned, but it is the
only candidate to base-pair to nuclease-insensitive
GGGU 5619–5622. Therefore, the structural probing data sup-
BYDV RNA 3⬘ END STRUCTURE
117
ported formation of the most stable calculated structure,
which is the conformation shared with PAV-129 RNA.
The 3⬘-terminal stem-loops are required for viral RNA
replication
FIG. 2. Nuclease probing of the 3⬘ terminus of BYDV RNA. (A)
Autoradiograph of an 8% polyacrylamide, 7 M urea gel showing partial
digests of 5⬘ end-labeled transcript comprising the 3⬘-terminal 109 nt of
BYDV (PAV6) RNA. Lane 1 (T1 denatur.), sequencing ladder generated
by partial digestion with the T1 ribonuclease (0.2 u) under denaturing
conditions. Lanes 2–4, partial digests with imidazole (concentration
shown above each lane). Lane 5 (T1 native), partial digest with T1
ribonuclease (0.01 u) under native conditions. Numbers on the left
indicate genomic positions of guanosine residues. Numbers on the
right indicate genomic positions of nucleotides sensitive to T1 or
imidazole. (B) Predicted secondary structure of the 3⬘ end of BYDV
genome with imidazole-sensitive nucleotides indicated by arrows, and
T1-sensitive nucleotides indicated by open triangles.
To test the importance of the 3⬘-terminal stem-loop
structures for viral RNA replication, we constructed
mutants in the full-length infectious clone of BYDV
(PAV6) that contained deletions that spanned each
stem-loop including the larger version of SL2 (Fig. 3A).
Also, to test whether sequence upstream of the four
stem-loops is necessary to support viral RNA replication, we designed an additional mutant, 3⬘SL5D, that
contained a deletion of nts 5503–5567 upstream of SL4
(Fig. 3A). Because portions of the 3⬘-terminal region of
BYDV genomic RNA have been implicated in translational regulation (Miller et al., 1997; Wang et al., 1999),
we first tested these mutants for translation in a wheat
germ extract. All five mutant transcripts translated
similarly to wild-type (PAV6) RNA in vitro, producing
abundant 39-kDa product of ORF1 and low levels of
the 99-kDa frameshift product, the RdRp (Fig. 3B).
None of the full-length viral in vitro transcripts containing any of the stem-loop deletions replicated (Fig. 3C)
in protoplasts. The faint genomic RNA bands in all
lanes are residual inoculum. Absence of subgenomic
RNAs indicates absence of replication. Mutant 3⬘SL5D
replicated at a low level (Fig. 3C), suggesting that all
four 3⬘-terminal stem-loops are necessary and sufficient for basal-level replication, but that upstream sequence (bases 5503–5567) is required for wild-type
activity.
To further assess effects of mutations on replication,
we probed for minus-strand RNA, which is present
only if the origin of replication is functional. In cells
infected with the wild-type (PAV6) RNA, low levels of
(⫺)-gRNA and sgRNA were detected with the minusstrand-specific probe (Fig. 3C). However, we found
that the minus-strand-specific probe can anneal to
BYDV plus-strand RNAs, albeit very inefficiently (data
not shown). Thus, we do not know if the subgenomicsized RNA species are minus-strand sgRNAs or represent inefficient detection of plus-strand sgRNAs.
Interestingly, the mutant 3⬘SL1D, which lacked SL1,
consistently produced very high amounts of negativestrand RNA with the molecular mass lower than that of
the full-length gRNA (Fig. 3C). Owing to the absence of
such a pattern with the plus-strand-specific probe, and
the very strong intensity of the signal, this is not
artifactual detection of plus strand. A smaller amount
of this low molecular weight RNA was detected in cells
inoculated with the deletion mutants 3⬘SL2D through
3⬘SL5D as well (Fig. 3C).
118
KOEV ET AL.
FIG. 3. Translation and replication of stem-loop deletion mutants.
(A) Diagram of mutants with deleted regions in boxes. 3⬘SL2D
includes the entire stem-loop 2 of the alternative conformation (Fig.
1). (B) Translation of wild-type (PAV6) and mutant RNA transcripts in
a wheat germ extract. [ 35S]met-labeled translation products, resolved
by 10% SDS–PAGE, include the 39-kDa product of ORF1 (39K) and
the less abundant 99-kDa frameshift product of ORFs 1⫹2 (99K).
Marker lane, designated BMV, shows translation products of BMV
RNAs. BYDV 39K protein migrates anomalously slow (Di et al., 1993).
(C) Northern blot analysis of total RNA isolated from infected oat
protoplasts ⬃24 hpi. Left, detection of positive-strand viral RNAs
probed with labeled transcript complementary to the 3⬘-terminal 1.5
kb of BYDV genomic RNA. Right, accumulation of the negativestrand viral RNA probed with (⫹)-sense BYDV transcript corresponding to the 3⬘-terminal 1.5 kb. The migration of the viral genomic RNA
(gRNA) and subgenomic RNAs (sgRNA) are indicated. Because a
high amount of the inoculum RNA was used in the experiments
(10–15 ␮g), low-intensity genomic-size RNA is observed in nonreplicating mutants. Therefore, synthesis of sgRNAs is used as the
main criterion of replication.
Helical regions of the 3⬘-terminal stem-loops are
necessary for replication
To test the role of the primary sequence vs the secondary structure of the stems in SL1 through SL4, we
designed a series of mutations that disrupted and restored these helices. Three mutants were generated for
each stem-loop: the 3⬘-proximal strand, the 5⬘-proximal
strand, or both strands were mutated such that each
single-strand mutant disrupted base-pairing and the
double mutant restored base-pairing (Fig. 4A).
Mutants with the stem structures disrupted in SL1,
SL3, and SL4 failed to replicate, whereas restoration of
these stems in the double mutants restored virus replication to the wild-type level in SL3 and to levels lower
than the wild-type in SL1 and SL4 (Fig. 4B). Interestingly, deletion of all four bulged nucleotides in SL1
(Fig. 4A, 3⬘SL14) had no deleterious effect on replication, indicating that unpaired bases in SL1 are not
required for replication (Fig. 4B). These data showed
that the base-pairing of the stems in SL1, SL3, and SL4,
and not their primary RNA sequences, are required for
replication.
We designed a set of mutants in an attempt to determine which conformation involving SL2 or the 3⬘ terminus is necessary. One group (3⬘SL2U1, 3⬘SL2U2,
3⬘SL2U3) disrupted and restored the distal end of SL2
that is present in both predicted structures. The results
clearly supported a requirement for secondary but not
primary structure. The second set disrupted and restored
either the bottom of the longer SL2 that is present only in
the alternative conformation (GGGU 5619–5622:ACUC 5638–5641,
mutants 3⬘SL21, 3⬘SL22, 3⬘SL23, Fig. 4A inset) or the
base-pairing of GGGU 5619–5622 to the 3⬘ terminus
(ACCC 5674–5677, 3⬘SL22, 3⬘SL24, 3⬘SL25, Fig. 4A). To alter
the 3⬘-terminal ACCC, we could not transcribe from plasmid DNA, because the CCC forms part of the SmaI site
required for plasmid linearization prior to run-off transcription. Instead, full-length DNA was prepared by PCR
using a 3⬘ primer containing the appropriate mutation,
and a 5⬘ primer that included the T7 promoter. Full-length
genomic RNA was transcribed directly off the PCR product. To allow fair comparison, all constructs involving
potential base-pairing with the GGGU sequence (3⬘SL21,
3⬘SL22, 3⬘SL24, 3⬘SL25, Fig. 4A), as well as wild-type
RNA, were prepared this way. As seen in Fig. 4C, the
RNA from protoplasts inoculated with transcripts from
PCR products appeared to contain more degraded viral
RNA. However, it is clear that all mutants containing
alterations in the GGGU 5619–5622 (3⬘SL22, 3⬘SL23, 3⬘SL25)
sequence or the 3⬘-terminal ACCC (3⬘SL24, 3⬘SL25) were
dead. The single ACUC 5638–5641 mutant (3⬘SL21) replicated
at a reduced level. Thus, ACUC 5638–5641 does not have to
be base-paired to GGGU for replication, nor is its primary
sequence essential. Either the primary sequence or secondary structures (or both) of GGGU 5619–5622 and
ACCC 5674–5677 are essential. The data cannot distinguish
between these possibilities. However, the data support
the absence of a requirement for base-pairing of
GGGU 5619–5622 to ACUC 5638–5641, consistent with the most
stable computer-predicted structure and that determined
by probing.
BYDV RNA 3⬘ END STRUCTURE
119
FIG. 4. Effect of stem-loop disruption and restoration on virus RNA replication. (A) Secondary structure with mutants indicated (boldface italics).
Disrupted and restored helical regions shown in boxes. The mutants based on the alternative SL2 structure are in the shaded boxes. Mutants 3⬘SL21
and 3⬘SL22 are shown on both structures. (B) Accumulation of positive-sense viral RNAs in protoplasts inoculated with wild-type (PAV6) or mutant
transcripts from SmaI-linearized plasmid was detected by Northern blot hybridization of total RNA isolated from infected oat protoplasts ⬃24 hpi. (C)
Accumulation of viral RNAs in protoplasts inoculated with transcripts derived directly from uncloned, full-length PCR products with the indicated
mutations. Far right lane was from protoplasts inoculated with plasmid-derived RNA. Ethidium bromide staining of the gel (below blot) prior to blotting
shows approximately equal loading as revealed by ribosomal RNA abundance in each lane.
Requirements for the terminal tetraloops
Because of the previously reported importance of
GNRA and UNCG tetraloops in RNA structure and
function (Jaeger et al., 1994; Murphy and Cech, 1994),
we elucidated the necessity of the terminal loop sequences in SL1, SL2, SL3, and SL4 for viral RNA
accumulation. In a series of mutants (Fig. 5A), the
sequence of the loop in SL1 (GAAA) was altered to a
UNCG consensus tetraloop (UUCG), another GNRA
tetraloop (GAGA), and a nontetraloop sequence
(GACA). All these mutants replicated at the wild-type
level, as did the mutant with the SL2 tetraloop (UUCG)
changed to the complementary sequence (AAGC) (Fig.
5B), indicating that the sequences of the terminal
loops in SL1 and SL2 are not important for virus RNA
replication in our assay.
Alteration of the SL3 tetraloop sequence (GCAA) to
its complement (CGUU) abolished virus replication
(Fig. 5B). Changing it to a UNCG tetraloop (UUCG), a
different GNRA tetraloop (GCGA), and a nontetraloop
120
KOEV ET AL.
Replication of chimeric PAV6-PAV129
To test a natural variant that represents extreme
alteration of the 3⬘ end, while conserving key structures, we tested the ability of the 3⬘ end from PAV-129
to support replication of a chimeric RNA. This construct contains a natural deletion of the distal end of
SL2 that is conserved in the two alternative PAV6
structures. The 5⬘ half of the chimeric construct (5⬘
UTR and ORFs 1 and 2) is derived from PAV6, while the
3⬘ half is derived from PAV-129. In the coding regions,
the isolates diverge most in ORF1 (80% amino acid
sequence identity) and ORF 2 (88% identity). It was
possible that the divergent 3⬘ structure of PAV-129
reflects a different template specificity of the replicase
conferred by its different amino acid sequence. In this
case, the PAV6 replicase of the chimera might not
recognize the PAV-129-derived 3⬘-terminus. However,
it is clear that the chimera replicated, albeit at about
10% of the efficiency of full-length PAV6 RNA (Fig. 6).
Thus, while SL2 may contribute to viral RNA accumulation, it is not absolutely necessary. Moreover, the
sequences of stem 3, and stem and loop of SL4 can
vary radically compared to PAV6, while still allowing
significant RNA replication.
FIG. 5. Effect of the terminal tetraloop sequence changes on BYDV
RNA replication. (A) Diagram of the mutants with altered loop sequences shown in boxes. (B) Northern blot showing accumulation of
viral RNAs, ⬃24 hpi, in protoplasts inoculated with indicated transcripts.
sequence with only one base different from the wildtype (GCCA) restored replication to various degrees
but never to the wild-type level (Fig. 5B). The mutant
3⬘SL3GCGA, which had one base different from the
wild-type sequence and fit the GNRA consensus, replicated at the highest level of the mutants. Similar
results were obtained with the SL4 tetraloop mutants
designed in the same fashion as those of SL3 (Fig. 5A).
Although none of the mutations abolished virus replication, they all reduced RNA accumulation by various
degrees (Fig. 5B), indicating that the original wild-type
sequence GAAA is optimal.
FIG. 6. Replication of chimeric PAV6-129 transcript in oat protoplasts.
(A) Map of PAV6-129 showing PAV6 (filled) and PAV-129 (open)-derived
portions, and the common restriction site at which they were joined. (B)
Oat protoplasts were inoculated with transcripts from SmaI-linearized
pPAV6 or pPAV6-129. Viral replication products were detected as in
previous experiments. The two PAV6-129 lanes were from different sets
of protoplasts inoculated with transcripts from independent clones of
the same construct. Ethidium bromide staining of the gel (right) reveals
relative loading of each lane.
BYDV RNA 3⬘ END STRUCTURE
DISCUSSION
Role(s) of the 3⬘-terminal stem-loop structures in RNA
accumulation
Most, if not all, RNA viruses have a specific structure
at the 3⬘ end of the genome that is required for initiation
of (⫺)-strand synthesis. Here, we identified structures in
the 3⬘ end of the BYDV genome that are required for
accumulation of viral RNA. In addition to (⫺)-strand synthesis, the structures may contribute to BYDV RNA stability or translation in vivo. Although translation of the
deletion mutants in wheat germ extract showed no apparent defects (Fig. 3B), some as-yet unmapped BYDV
3⬘-UTR sequences not required for translation in vitro are
required for translation in vivo (Guo et al., 2000; Wang et
al., 1999). These may include sequences that mimic the
poly(A) tail function, functionally resembling the
pseudoknot-rich domains of TMV (Gallie and Walbot,
1990), because wheat germ extracts are insensitive to
the presence of poly(A) tails. Involvement of the 3⬘-terminal structures exclusively in conferring (⫹)-strand stability is unlikely because intact inoculum RNA is still
detectable in cells 24 h after inoculation with replicationdefective mutants (Figs. 3–5).
Features likely to participate directly in (⫺)-strand synthesis are SL1 and the terminal ACCC sequence. Similar
stem-loops that tolerate variation in primary sequence
and are located a few bases upstream of a 3⬘-terminal
CCC sequence have been identified as origins of (⫺)strand synthesis for Turnip crinkle virus (TCV) (Song and
Simon, 1995), Cymbidium ringspot virus (CymRSV) (Havelda and Burgyan, 1995), and Red clover necrotic mosaic
virus (RCNMV) (Turner and Buck, 1999). In the polymerase genes, these viruses, all members of the Tombusviridae family, are closely related to BYDV (Koonin and
Dolja, 1993), so similar replication origins would be predicted in BYDV RNA.
Stem-loops with a relaxed primary sequence requirement in 3⬘ replication elements have also been identified
in Tobacco etch virus (TEV) (Haldeman-Cahill et al.,
1998), Bovine viral diarrhea virus (BVDV) (Yu et al., 1999),
Alfalfa mosaic virus (AMV) (van Rossum et al., 1997), and
others. On the other hand, Beet necrotic yellow vein virus
and Satellite tobacco necrosis virus both require specific
primary and the secondary structures in the 3⬘-helical
regions for replication (Bringloe et al., 1999; Lauber et al.,
1997). The secondary structures of 3⬘-terminal stemloops of RNA viruses have been shown to be required for
protein binding (Lai, 1998). Perhaps failure to bind replication factors rendered the stem disruption mutants nonviable.
An intriguing feature of the comparison of the conserved structures formed by PAV6 and PAV-129 termini (Fig. 1) is that most of the conserved primary sequence is in a “pocket” comprising all of SL1 and imme-
121
diately adjacent bases, the putative GGG(U/A) 5619–5622:(A/
U)CCC 5674–5677 helix, the proximal ends of SL’s 3 and 4, and
the bases connecting them. Although the two-dimensional rendition does not reflect the true 3D structure,
most of these bases must be in fairly close proximity. We
propose that this area of conserved bases comprises a
replicase recognition site. This leads to the question of
why the sequence in SL1 is conserved when it can be
altered drastically without reducing replication, as long
as it can form a stem-loop.
The answer to the above question may lie in the
important caveat for all of these experiments: the functional mutant structures detected here may not be optimal. This is because we are examining replication only
at 24 h postinoculation (hpi) in plant cells that have been
inundated with large quantities of inoculum. In these
conditions, the inoculum RNA has an advantage in sheer
numbers in the competition with host RNAs for host
proteins. Second, the inoculum has no competition with
more fit viral variant RNAs that would normally be
present in a mixed natural population. Third, we are
viewing only a 24 h time point. After longer periods of
infection and many more rounds of replication that take
place in whole plant infections, more fit structures have
time to evolve. Thus, the wild-type SL1 sequence may
replicate more efficiently in the long run than the mutant
constructs that replicate in protoplasts at 24 days p.i., as
long as the stem can form.
Roles of tetraloops
The one conserved sequence (between PAV6 and PAV129) that is not in the conserved pocket is the tetraloop
at the distal end of SL3 that fits the highly stable GNRA
consensus. GNRA and UNCG tetraloops have intraloop
interactions that greatly stabilize them and the helix.
Tetraloops have also been implicated in protein binding.
Thus, this element could also be a potential binding site
for replication proteins (reviewed in Lai, 1998). A third
interaction for GNRA tetraloops involves interaction with
a “receptor” in a specific RNA helix, as in the group I
intron of Tetrahymena thermophilia and the td intron of
the bacteriophage T4 (Butcher et al., 1997; Costa and
Michel, 1997; Pley et al., 1994). Such interactions have
not been reported in viral RNAs.
Interestingly, sequences of the SL1 and the SL2 terminal tetraloops were not important for replication in our
assay. The lack of a requirement for specific sequences
in the loops of the essential 3⬘-proximal stem-loops has
been demonstrated in vitro in TCV (Song and Simon,
1995) and CymRSV, which also contains a 3⬘-terminal
GNRA tetraloop (Havelda and Burgyan, 1995). On the
other hand, RNA2 of RCNMV, which is closely related to
BYDV, was unable to replicate with an alteration in the
three base loop of the 3⬘-terminal stem-loop structure
122
KOEV ET AL.
(Turner and Buck, 1999). Perhaps the more stable tetraloops provide an advantage only in more competitive or
longer term replication conditions. Such was the case for
a GNRA tetraloop in the 5⬘-proximal stem-loop of Potato
virus X (PVX) (Miller et al., 1998). A preference for a GNRA
tetraloop was obvious in SL3 and SL4 of PAV6, but the
PAV-129 3⬘ end permits replication with a much larger
loop 4. This may explain the reduced accumulation of
chimeric RNA or perhaps the longer stem 4 of PAV-129
compensates for loss of the stabilizing tetraloop.
Possible roles of upstream sequence elements
The 3⬘-terminal 109 nts that contain the four stemloops were not sufficient for the wild-type level of RNA
accumulation in protoplasts. Sequence upstream (bases
5503–5567) appear to play a supporting role. They may
be involved in functions other than (⫺)-strand synthesis,
such as (⫹)-strand synthesis. In TCV and RCNMV genomes, elements located hundreds of nucleotides upstream of the 3⬘ end are needed for efficient replication
(Carpenter et al., 1995; Turner and Buck, 1999). Elements
located upstream of the 3⬘-terminal TLS stimulate replication of BMV and TMV RNAs (Lahser et al., 1993; Sullivan and Ahlquist, 1999; Takamatsu et al., 1990). In the
bacteriophage Q␤ RNA, an internal region 1.5 kb upstream of the 3⬘ end is the actual replicase binding site
(Brown and Gold, 1996). Long-distance base-pairing delivers the replicase complex to the 3⬘ end (Klovins et al.,
1998). Thus, it is possible that other upstream elements
of the BYDV genome may play a role in the (⫺)-strand
synthesis. We can rule out bases 2789–4515, 5016–5045,
and 5183–5205 because they can be deleted with little
effect on BYDV RNA replication (Mohan et al., 1995; Paul
et al., 2001).
Regulation of replication by an embedded 3⬘ end?
The computer modeling and phylogenetic comparisons predict that the 3⬘ end of BYDV RNA is embedded
in a helix that is coaxially stacked with stem 3 (Fig. 1). All
changes that disrupted and restored the GGGU 5619–5622:
ACCC 5674–5677 helix destroyed RNA replication (except for
the natural U/A covariation in PAV-129 which supports a
role for this helix). Given that all Tombusviridae and all
viruses with related polymerases terminate in essential
CCC or CCA sequences, it is highly likely that the alterations to CCC 5675–5677 prevented (⫺)-strand synthesis,
even if the base-pairing were restored. The combination
of the covariation mutagenesis data that support functional roles for SL1, SL3, SL4, and the short, four-base
SL2, with the probing data that indicate GGGU 5619–5622 is
base-paired but not to ACUC 5638–5641, leave little choice but
for the GGGU 5619–5622 to base pair to the 3⬘-terminal ACCC.
The embedded nature of the 3⬘-terminus is intriguing,
given that a free CCA or CCC 3⬘ end is required for
(⫺)-strand initiation in other viruses. Base-paired 3⬘-termini was suggested to be a feature peculiar to RNA
bacteriophages (Klovins et al., 1998). It is tempting to
speculate that the putatively base-paired 3⬘ end of BYDV
RNA negatively regulates minus-strand synthesis. Basepairing has been implicated in preventing initiation of
minus-strand RNA synthesis from promoter-like elements within the TLS in TYMV (Singh and Dreher, 1998).
Possible regulatory conformational changes in viral 3⬘UTRs have been suggested for AMV (Olsthoorn et al.,
1999) and flaviviruses (Khromykh et al., 2001). Conceivably, a conformational switch between the embedded
3⬘-end structure, and the alternative structure with the
free ACCC 5674–5677 mediate a switch between negative
(free) and positive (embedded) strand synthesis in BYDV.
The increase in (⫺)-strand accumulation resulting from
deletion of SL1, and to a lesser extent from deletion of
SL2 and SL4, and the concomitant loss of (⫹)-strand
accumulation is consistent with a role of these structures
in inhibiting (⫺)-strand synthesis. (⫺)-Strand synthesis
could be inhibited directly, by the structures themselves,
or indirectly by their recruitment of a protein factor.
Translation factor EF-1␣ has been proposed to negatively regulate (⫺)-strand synthesis of TYMV RNA by
binding the TLS (Dreher, 1999). Blocking (⫺)-strand promoter activity also could facilitate translation of genomic
RNA, since translation and replication are incompatible
on the same RNA molecule (Gamarnik and Andino,
1998).
Such regulation would predict that the (⫹)-strand promoters would be different from the (⫺)-strand promoter.
Indeed the 5⬘ ends of BYDV genomic (⫹) strand and
subgenomic RNAs 1 and 2 begin with an essential GUGAAG sequence that is absent in the 5⬘ end of the (⫺)
strand, and they show no obvious similarity to the (⫺)strand promoter.
Implications for BYDV taxonomy
Isolate PAV-129 differs substantially in sequence from
other PAV and MAV isolates in ORFs 1, 2, and in the long
3⬘-untranslated region. It also has very different symptomatology (Chay et al., 1996). Thus, it might be considered a different virus than BYDV. However, the fact that a
chimera containing the replicase from PAV6 and the 3⬘
origin of replication from PAV129 can replicate in protoplasts (Fig. 6) and in plants (S. Liu, unpublished observations) leads us to conclude that they are merely divergent isolates of one virus species.
The polymerase, translational control signals, and 3⬘terminal cis-acting replication structures of BYDV (genus
Luteovirus, family Luteoviridae) and those of family Tombusviridae are very similar. In contrast, the polymerase
genes and 3⬘-termini of the genus Polerovirus (ending in
GU), family Luteoviridae, bear no resemblance to those
BYDV RNA 3⬘ END STRUCTURE
of BYDV (Koonin and Dolja, 1993; Miller et al., 1995). Thus
genus Luteovirus should perhaps be reassigned to the
Tombusviridae.
MATERIALS AND METHODS
Plasmids
The mutant constructs were derived from the fulllength infectious clone of BYDV, pPAV6 (Di et al., 1993), by
PCR-mediated site-directed mutagenesis using Vent
DNA polymerase (New England Biolabs). The mutants
p3⬘SL1D, p3⬘SL1UUCG, p3⬘SL1GAGA, p3⬘SL1GACA,
p3⬘SL11, p3⬘SL12, p3⬘SL13, p3⬘SL14, and p3⬘SL2D were
constructed by PCR amplification of the 3⬘ portion of the
BYDV genome with the downstream mutagenic primers
and the upstream primer CB0416 (GGTCTAGATAATACGACTCACTATAGGGTACACAAACAAGCGAAT). The product was digested with the restriction endonucleases
KpnI and SmaI, purified by 0.8% low-melt agarose gel
electrophoresis, and inserted into pPAV6 cut with the
same enzymes. The remaining mutants were generated
by two-step PCR (Landt et al., 1990). In the first step,
upstream mutagenic primers containing the mutant
bases flanked by appropriate homologous sequence
were used with the downstream primer 3⬘wt (ATACCCCGGGTTGCCGAA). The product was gel-purified and
used as a downstream primer in the second-step PCR
with the upstream primer CB0416. The second-step PCR
product was cut with KpnI and SmaI, gel-purified, and
placed in pPAV6 cut with the same enzymes. The construct p3⬘ used for RNA structural probing was generated by subcloning a BglII–SalI fragment of pgl016, containing the 3⬘ end of BYDV, into pSS1 (Miller and Silver,
1991) cut with BamHI and SalI. Plasmid preparations
were performed using the QuantumPrep DNA purification kit (Bio-Rad). All mutants were confirmed by sequencing.
pPAV6-129 was derived from pPAV129-CPRT, a plasmid
derived from pPAV6, in which ORFs 3, 4, and 5 of PAV6
were replaced by the counterparts from PAV129. To replace the 3⬘-end sequence of PAV6 in pPAV129-CPRT, a
KpnI–SmaI fragment (nts 4156–5677) was amplified from
two subclones (p129–10 and p129–12) of PAV129 and
joined by three PCR reactions. The first PCR amplified an
832-bp fragment from pPAV129–10, corresponding to positions 4054–4885. The primers used for PCR were 5⬘CTGACGCCTGTTTCTACCTTGC-3⬘ and 5⬘-CCAATTACCCGAGCTTACGAACC-3⬘ for the 5⬘ and 3⬘ ends, respectively. The second PCR generated a fragment
corresponding to bases 4863–5677 in PAV129, using
primers 5⬘-GGTTCGTAAGCTCGGGTAATTGG-3⬘ (complementary to the above 3⬘ primer) and 5⬘-GGGATGTGCCGGGCTTCTCTTTC-3⬘ and p129-12 as template. The
two PCR products were agarose gel purified and about
50 ng of each used as templates for the third PCR using
123
the 5⬘ primer of the first PCR with the 3⬘ primer of the
second PCR to generate a single product of 1522 bases.
This product was digested with KpnI and ligated into the
corresponding sites (KpnI and SmaI) of pPAV129CPRT.
The resulting plasmid was confirmed by partial sequencing and restriction mapping.
Genome-length PCR
To allow direct in vitro transcription of correctly terminated, full-length viral RNA from uncloned PCR products,
full-length cDNA templates were generated by PCR using VENT DNA polymerase (New England Biolabs). An
upstream primer, annealing to the 5⬘ end of the viral
genome, 5⬘-ACGCGGCCGCTAATACGACTCACTATAGAGTGAAGATTGACCATCTCACAAAAGC-3⬘ (T7 promoter in
italics, transcription initiation site underlined, viral sequence in bold), was paired with either of two 3⬘-end
primers, 5⬘-GGGTTGCCGAACTGCTCTTTCGAGTGTCAGCGTTAAGAGTACCC-3⬘ (wild-type) or 5⬘-CTCATGCCGAACTGCTCTTTCGAGTGTCAGCGTTAAGAGTACCC-3⬘
(for mutants 3⬘SL24 and 3⬘SL25). The mutant 3⬘ primer
was used to amplify 3⬘SL22 template to generate mutant
3⬘SL24. The same primer was used with PAV6 template
to generate 3⬘SL25. As controls, the wild-type primer was
used on the same templates to allow transcription of
RNA from PCR products rather than linearized plasmids
for comparison purposes. Full-length transcripts were
obtained using Ambion’s T7 Megascript kit.
Protoplast infection and Northern blot analysis
Wild-type and mutant infectious transcripts were generated by in vitro transcription of SmaI-linearized plasmids using Megascript T7 RNA polymerase system (Ambion, Austin, TX). We used 10–15 ␮g RNA for electroporation of oat protoplasts prepared as described in
Dinesh-Kumar and Miller, 1993. Total RNA was extracted
from protoplasts ⬃24 hpi using RNeasy plant RNA isolation kit (Qiagen, Los Angeles, CA). For minus-strand
detection, RNA was isolated following the procedure
described in Seeley et al., 1992, using aurin tricarboxylic
acid as RNase inhibitor. RNA (5–10 mg) was analyzed by
Northern blot hybridization essentially as described in
Seeley et al., 1992. Equal amounts of total RNA were
loaded in each lane as verified by ethidium bromide
staining. A 32P-labeled riboprobe complementary to the
3⬘-terminus of BYDV was used to detect viral gRNA and
sgRNAs. For positive-strand detection, plasmid pSP10
(Dinesh-Kumar et al., 1992) was linearized with HindIII
and transcribed in vitro with T7 RNA polymerase (New
England Biolabs). For negative-strand detection, the
same plasmid was linearized with SmaI and transcribed
in vitro with SP6 RNA polymerase (Promega, Madison,
WI). GeneScreen nylon membranes (DuPont) were hybridized with the probes and exposed to phosphorim-
124
KOEV ET AL.
ager screens (Molecular Dynamics, Sunnyvale, CA) for
5–24 h (positive-strand detection) or 2–4 days (negativestrand detection).
Transcripts of the stem-loop deletion mutants were
tested for translation in wheat germ extract translation
system (Promega) following manufacturer’s instructions,
using 35S-labeled methionine, as described in Wang et
al., 1999. Translation products were analyzed by SDS–
PAGE (10% acrylamide) as described in Wang et al., 1999.
All experiments were performed at least twice.
RNA sequence and structural analysis
Sequence alignments of BYDV isolates were performed using the GCG software package (Madison, WI).
RNA secondary structure was predicted using MFOLD,
version 3.0, at the MFOLD web site (http://bioinfo.math.
rpi.edu/⬃mfold/) (Mathews et al., 1999; Zuker et al.,
1999). RNA secondary structure probing was performed
with imidazole (Vlassov et al., 1995) and with T1 nuclease
(Miller and Silver, 1991) on in vitro transcripts 5⬘ endlabeled with 32P essentially as described in Koev et al.,
1999. Imidazole cleavage was performed at 25°C for 1–2
h. The transcripts were derived from p3⬘ linearized with
SmaI.
ACKNOWLEDGMENTS
This work was supported by Grant 2001-35319-10011 from USDANRI. We thank Kay Scheets for advice on RNA extraction for detection
of minus-strand viral RNA, Mark Young and Sergei Filichkin for providing unpublished sequence of the 3⬘ end of BYDV-MAV, Theo Dreher for
useful discussion, and Jacquelyn Jackson for advice on full-length
RT-PCR. This is paper J-18711 of the Iowa Agriculture and Home
Economics Experiment Station, Project 3545.
REFERENCES
Bringloe, D. H., Pleij, C. W., and Coutts, R. H. (1999). Mutation analysis
of cis-elements in the 3⬘- and 5⬘-untranslated regions of satellite
tobacco necrosis virus strain C RNA. Virology 264, 76–84.
Brown, D., and Gold, L. (1996). RNA replication by Q␤ replicase: A
working model. Proc. Natl. Acad. Sci. USA 93, 11558–11562.
Buck, K. W. (1996). Comparison of the replication of positive-stranded
RNA viruses of plants and animals. Adv. Virus Res. 47, 159–251.
Butcher, S. E., Dieckmann, T., and Feigon, J. (1997). Solution structure of
a GAAA tetraloop receptor RNA. EMBO J. 16, 7490–7499.
Carpenter, C. D., Oh, J.-W., Zhang, C., and Simon, A. E. (1995). Involvement of a stem-loop structure in the location of junction sites in viral
RNA recombination. J. Mol. Biol. 245, 608–622.
Carpenter, C. D., and Simon, A. E. (1998). Analysis of sequences and
predicted structures required for viral satellite RNA accumulation by
in vivo genetic selection. Nucleic Acids Res. 26, 2426–2432.
Chapman, M. R., and Kao, C. C. (1999). A minimal RNA promoter for
minus-strand RNA synthesis by the brome mosaic virus polymerase
complex. J. Mol. Biol. 286, 709–720.
Chay, C., Smith, D. M., Vaughan, R., and Gray, S. M. (1996). Diversity
among isolates within the PAV serotype of barley yellow dwarf virus.
Phytopathology 86, 370–377.
Chen, M. H., and Frey, T. K. (1999). Mutagenic analysis of the 3⬘
cis-acting elements of the rubella virus genome. J. Virol. 73, 3386–
8403.
Cheong, C., Varani, G., and Tinoco, I., Jr. (1990). Solution structure of an
unusually stable RNA hairpin, 5⬘GGAC(UUCG)GUCC. Nature 346,
680–682.
Costa, M., and Michel, F. (1997). Rules for RNA recognition of gnra
tetraloops deduced by in vitro selection: Comparison with in vivo
evolution. EMBO J. 16, 3289–3302.
D’Arcy, C. J., Domier, L. L., and Mayo, M. A. (2000). Family Luteoviridae.
In “Virus Taxonomy: Seventh Report of the International Committee
on the Taxonomy of Viruses” (M. H. V. van Regenmortel, C. M.
Fauquest, D. H. L. Bishop, E. B. Carstens, M. K. Estes, S. M. Lemon,
J. Maniloff, M. A. Mayo, D. J. McGeoch, C. R. Pringle, and R. B.
Wickner, Eds.), pp. 775–784. Academic Press, San Diego, CA.
Di, R., Dinesh-Kumar, S. P., and Miller, W. A. (1993). Translational
frameshifting by barley yellow dwarf virus RNA (PAV serotype) in
Escherichia coli and in eukaryotic cell-free extracts. Mol. Plant Microbe Interact. 6, 444–452.
Dinesh-Kumar, S. P., Brault, V., and Miller, W. A. (1992). Precise mapping
and in vitro translation of a trifunctional subgenomic RNA of barley
yellow dwarf virus. Virology 187, 711–722.
Dinesh-Kumar, S. P., and Miller, W. A. (1993). Control of start codon
choice on a plant viral RNA encoding overlapping genes. Plant Cell
5, 679–692.
Dreher, T. W. (1999). Functions of the 3⬘-untranslated regions of positive
strand RNA viral genomes. Annu. Rev. Phytopathol. 37, 151–174.
Dreher, T. W., and Hall, T. C. (1988). Mutational analysis of the sequence
and structural requirements in brome mosaic virus RNA for (⫺)
strand promoter activity. J. Mol. Biol. 201, 31–40.
Gallie, D. R., and Walbot, V. (1990). RNA pseudoknot domain of tobacco
mosaic virus can functionally substitute for a poly(A) tail in plant and
animal cells. Genes Dev. 4, 1149–1157.
Gamarnik, A. V., and Andino, R. (1998). Switch from translation to RNA
replication in a positive-stranded RNA virus. Genes Dev. 12, 2293–
2304.
Guo, L., Allen, E., and Miller, W. A. (2000). Structure and function of a
cap-independent translation element that functions in either the 3⬘ or
the 5⬘ untranslated region. RNA 6, 1808–1820.
Haldeman-Cahill, R., Daros, J. A., and Carrington, J. C. (1998). Secondary structures in the capsid protein coding sequence and 3⬘ nontranslated region involved in amplification of the tobacco etch virus
genome. J. Virol. 72, 4072–4079.
Havelda, Z., and Burgyan, J. (1995). 3⬘ terminal putative stem-loop
structure required for the accumulation of cymbidium ringspot viral
RNA. Virology 214, 269–272.
Heus, H. A., and Pardi, A. (1991). Structural features that give rise to the
unusual stability of RNA hairpins containing GNRA loops. Science
253, 191–194.
Hsue, B., and Masters, P. S. (1997). A bulged stem-loop structure in the
3⬘ untranslated region of the genome of the coronavirus mouse
hepatitis virus is essential for replication. J. Virol. 71, 7567–7578.
Huang, C. Y., Huang, Y. L., Meng, M., Hsu, Y. H., and Tsai, C. H. (2001).
Sequences at the 3⬘ untranslated region of bamboo mosaic potexvirus RNA interact with the viral RNA-dependent RNA polymerase.
J. Virol. 75, 2818–2824.
Jaeger, L., Michel, F., and Westhof, E. (1994). Involvement of a GNRA
tetraloop in long-range RNA tertiary interactions. J. Mol. Biol. 236,
1271–1276.
Khromykh, A. A., Meka, H., Guyatt, K. J., and Westaway, E. G. (2001).
Essential role of cyclization sequences in flavivirus RNA replication.
J. Virol. 75, 6719–6728.
Kim, C. H., Kao, C. C., and Tinoco, I., Jr. (2000). RNA motifs that
determine specificity between a viral replicase and its promoter. Nat.
Struct. Biol. 7, 415–423.
BYDV RNA 3⬘ END STRUCTURE
Klovins, J., Berzins, V., and van Duin, J. (1998). A long-range interaction
in ␤ RNA that bridges the thousand nucleotides between the m-site
and the 3⬘ end is required for replication. RNA 4, 948–957.
Koev, G., and Miller, W. A. (2000). A positive strand RNA virus with three
very different subgenomic RNA promoters. J. Virol. 74, 5988–5996.
Koev, G., Mohan, B. R., and Miller, W. A. (1999). Primary and secondary
structural elements required for synthesis of barley yellow dwarf
virus subgenomic RNA1. J. Virol. 73, 2876–2885.
Koonin, E. V., and Dolja, V. V. (1993). Evolution and taxonomy of positivestrand RNA viruses: Implications of comparative analysis of amino
acid sequences. Crit. Rev. Biochem. Mol. Biol. 28, 375–430.
Lahser, F. C., Marsh, L. E., and Hall, T. C. (1993). Contributions of the
brome mosaic virus RNA-3 3⬘-nontranslated region to replication and
translation. J. Virol. 67, 3295–3303.
Lai, M. M. C. (1998). Cellular factors in the transcription and replication
of viral RNA genomes: A parallel to DNA-dependent RNA transcription. Virology 244, 1–12.
Landt, O., Grunert, H. P., and Hahn, U. (1990). A general method for
rapid site-directed mutagenesis using the polymerase chain reaction. Gene 96, 125–128.
Lauber, E., Guilley, H., Richards, K., Jonard, G., and Gilmer, D. (1997).
Conformation of the 3⬘-end of beet necrotic yellow vein benevirus
RNA 3 analysed by chemical and enzymatic probing and mutagenesis. Nucleic Acids Res. 25, 4723–4729.
Mathews, D. H., Sabina, J., Zuker, M., and Turner, D. H. (1999). Expanded sequence dependence of thermodynamic parameters provides robust prediction of RNA secondary structure. J. Mol. Biol. 288,
911–940.
Melchers, W. J., Hoenderop, J. G., Bruins Slot, H. J., Pleij, C. W., Pilipenko, E. V., Agol, V. I., and Galama, J. M. (1997). Kissing of the two
predominant hairpin loops in the Coxsackie B virus 3⬘ untranslated
region is the essential structural feature of the origin of replication
required for negative-strand RNA synthesis. J. Virol. 71, 686–696.
Miller, E. D., Kim, K. H., and Hemenway, C. (1999). Restoration of a
stem-loop structure required for potato virus X RNA accumulation
indicates selection for a mismatch and a GNRA tetraloop. Virology
260, 342–353.
Miller, E. D., Plante, C. A., Kim, K. H., Brown, J. W., and Hemenway, C. L.
(1998). Stem-loop structure in the 5⬘ region of potato virus X genome
required for plus-strand RNA accumulation. J. Mol. Biol. 284, 591–
608.
Miller, W. A., Brown, C. M., and Wang, S. (1997). New punctuation for the
genetic code: Luteovirus gene expression. Semin. Virol. 8, 3–13.
Miller, W. A., Dinesh-Kumar, S. P., and Paul, C. P. (1995). Luteovirus gene
expression. Crit. Rev. Plant Sci. 14, 179–211.
Miller, W. A., and Koev, G. (2000). Synthesis of subgenomic RNAs by
positive-strand RNA viruses. Virology 273, 1–8.
Miller, W. A., and Silver, S. L. (1991). Alternative tertiary structure attenuates self-cleavage of the ribozyme in the satellite RNA of barley
yellow dwarf virus. Nucleic Acids Res. 19, 5313–5320.
Mirmomeni, M. H., Hughes, P. J., and Stanway, G. (1997). An RNA
tertiary structure in the 3⬘ untranslated region of enteroviruses is
necessary for efficient replication. J. Virol. 71, 2363–2370.
Mohan, B. R., Dinesh-Kumar, S. P., and Miller, W. A. (1995). Genes and
cis-acting sequences involved in replication of barley yellow dwarf
virus-pav RNA. Virology 212, 186–195.
Murphy, F. L., and Cech, T. R. (1994). Gaaa tetraloop and conserved
bulge stabilize tertiary structure of a group I intron domain. J. Mol.
Biol. 236, 49–63.
Olsthoorn, R. C. L., Mertens, S., Brederode, F. T., and Bol, J. F. (1999). A
conformational switch at the 3⬘ end of a plant virus RNA regulates
viral replication. EMBO J. 18, 4856–4864.
Paul, A. V., Rieder, E., Kim, D. W., van Boom, J. H., and Wimmer, E. (2000).
Identification of an RNA hairpin in poliovirus RNA that serves as the
125
primary template in the in vitro uridylylation of VPg. J. Virol. 74,
10359–10370.
Paul, C. P., Barry, J. K., Dinesh-Kumar, S. P., Brault, V., and Miller, W. A.
(2001). A sequence required for ⫺1 ribosomal frameshifting located
four kilobases downstream of the frameshift site. J. Mol. Biol. 310,
987–999.
Pilipenko, E. V., Poperechny, K. V., Maslova, S. V., Melchers, W. J., Slot,
H. J., and Agol, V. I. (1996). Cis-element, orir, involved in the initiation
of (⫺) strand poliovirus RNA: A quasi-globular multi-domain RNA
structure maintained by tertiary (“kissing”) interactions. EMBO J. 15,
5428–5436.
Pley, H. W., Flaherty, K. M., and McKay, D. B. (1994). Model for an
RNA tertiary interaction from the structure of an intermolecular
complex between a GAAA tetraloop and an RNA helix. Nature 372,
111–113.
Rieder, E., Paul, A. V., Kim, D. W., van Boom, J. H., and Wimmer, E.
(2000). Genetic and biochemical studies of poliovirus cis-acting
replication element cre in relation to VPg uridylylation. J. Virol. 74,
10371–10380.
Sachs, A. B. (2000). Physical and functional interactions between the
mrna cap structure and the poly(A) tail. In “Translational Control of
Gene Expression” (N. Sonenberg, J. W. B. Hershey, and M. B.
Mathews, Eds.). Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, NY.
Seeley, K. A., Byrne, D. H., and Colbert, J. T. (1992). Red lightindependent instability of oat phytochrome mRNA in vivo. Plant
Cell 4, 29–38.
Siegel, R. W., Adkins, S., and Kao, C. C. (1997). Sequence-specific
recognition of a subgenomic RNA promoter by a viral RNA polymerase. Proc. Natl. Acad. Sci. USA 94, 11238–11243.
Singh, R. N., and Dreher, T. W. (1998). Specific site selection in RNA
resulting from a combination of nonspecific secondary structure and
-CCR- boxes: Initiation of minus strand synthesis by turnip yellow
mosaic virus RNA-dependent RNA polymerase. RNA 4, 1083–1095.
Sit, T. L., Vaewhongs, A. A., and Lommel, S. A. (1998). RNA-mediated
transactivation of transcription from a viral RNA. Science 281, 829–
832.
Song, C. Z., and Simon, A. E. (1995). Requirement of a 3⬘-terminal
stem-loop in in vitro transcription by an RNA-dependent RNA polymerase. J. Mol. Biol. 254, 6–14.
Sullivan, M. L., and Ahlquist, P. (1999). A brome mosaic virus intergenic
RNA3 replication signal functions with viral replication protein 1a to
dramatically stabilize RNA in vivo. J. Virol. 73, 2622–2632.
Takamatsu, N., Watanabe, Y., Meshi, T., and Okada, Y. (1990). Mutational
analysis of the pseudoknot region in the 3⬘ noncoding region of
tobacco mosaic virus RNA. J. Virol. 64, 3686–3693.
Todd, S., Towner, J. S., Brown, D. M., and Semler, B. L. (1997). Replication-competent picornaviruses with complete genomic RNA 3⬘ noncoding region deletions. J. Virol. 71, 8868–8874.
Turner, R. L., and Buck, K. W. (1999). Mutational analysis of cis-acting
sequences in the 3⬘- and 5⬘-untranslated regions of RNA2 of red
clover necrotic mosaic virus. Virology 253, 115–124.
van Rossum, C. M. A., Brederode, F. T., Neeleman, L., and Bol, J. F.
(1997). Functional equivalence of common and unique sequences in
the 3⬘ untranslated regions of alfalfa mosaic virus RNAs 1, 2, and 3.
J. Virol. 71, 3811–3816.
Vlassov, V. V., Zuber, G., Felden, B., Behr, J. P., and Giege, R. (1995).
Cleavage of tRNA with imidazole and spermine imidazole constructs:
A new approach for probing RNA structure. Nucleic Acids Res. 23,
3161–3167.
Wang, S., Guo, L., Allen, E., and Miller, W. A. (1999). A potential mechanism for selective control of cap-independent translation by a viral
RNA sequence in cis and in trans. RNA 5, 728–738.
Wang, X., Ullah, Z., and Grumet, R. (2000). Interaction between zucchini
126
KOEV ET AL.
yellow mosaic potyvirus RNA-dependent RNA polymerase and host
poly-(A) binding protein. Virology 275, 433–443.
Williams, G. D., Chang, R. Y., and Brian, D. A. (1999). A phylogenetically
conserved hairpin-type 3⬘ untranslated region pseudoknot functions
in coronavirus RNA replication. J. Virol. 73, 8349–8355.
Yoshinari, S., Nagy, P. D., Simon, A. E., and Dreher, T. W. (2000). CCA
initiation boxes without unique promoter elements support in vitro
transcription by three viral RNA-dependent RNA polymerases. RNA
6, 698–707.
Yu, H., Grassmann, C. W., and Behrens, S. E. (1999). Sequence and
structural elements at the 3⬘ terminus of bovine viral diarrhea virus
genomic RNA: Functional role during RNA replication. J. Virol. 73,
3638–3648.
Zuker, M., Mathews, D. H., and Turner, D. H. (1999). Algorithms and
thermodynamics for RNA secondary structure prediction: A practical
guide. In “RNA Biochemistry and Bio/Technology” (J. Barciszewski,
and B. F. C. Clark, Eds.), pp. 11–43. Kluwer Academic Publishers,
Dordrecht/Norwell, MA.