Structural elements of the 3
'-terminal coat protein binding site in alfalfa mosaic virus RNAs
Structural elements of the 3 '-terminal coat protein binding site in alfalfa mosaic virus RNAs
Chantal B. E. M.
Reusken
and
John F.
Bol*
Institute of Molecular Plant Sciences, Gorlaeus Laboratories, Leiden University,
PO Box 9502, 2300 RA
Leiden
,
The Netherlands
Received April 16, 1996;
Revised and Accepted May 30, 1996
ABSTRACT
The 3
'
-termini of the three genomic RNAs of alfalfa mosaic virus (AlMV) and
ilarviruses contain a number of AUGC-motifs separated by hairpin structures. Binding of coat protein (CP) to
such elements in the RNAs is required to initiate infection of these viruses.
Determinants for CP binding in the 3
'
-terminal 39 nucleotides (nt) of AlMV RNA 3 were analyzed by band-shift assays. From the 5
'
- to 3'
-end this 39 nt sequence contains AUGC-motif 3, stem-loop structure 2 (STLP2), AUGC-motif 2, stem-loop structure 1 (STLP1) and AUGC-motif 1. A mutational analysis showed that
all three AUGC-motifs were involved in CP binding. Mutation of the A- and U-residues of motifs 1 or 3 had no effect on CP binding but
similar mutations in motif 2 abolished CP binding. A mutational analysis of the
stem of STLP1 and STLP2 confirmed the importance of these hairpins for CP
binding. Randomization of the sequence of the stems and loops of STLP1 and
STLP2 had no effect on CP binding as long as the secondary structure was
maintained. This indicates that the two hairpins are not involved in sequence-specific interactions with CP. They may function in a secondary structure-specific interaction with CP and/or in the assembly of the AUGC-motifs in a configuration required for CP binding.
INTRODUCTION
The family
Bromoviridae
consists of four genera of plant viruses with a similar tripartite genome.
These are the bromoviruses, the cucumoviruses, the ilarviruses and the
alfamoviruses. There are two distinctive features that set the alfamo- and ilarviruses apart from other genera in this family. First, the RNAs
of alfamo- and ilarviruses lack the 3'-terminal tRNA-like structure that is present in the RNAs of bromo- and cucumoviruses. Second, the coat protein (CP)
is required for initiation of alfamo- and ilarvirus infection while a mixture of the three genomic RNAs of
bromo- and cucumoviruses is infectious as such. A mixture of the three genomic
RNAs of alfalfa mosaic virus (AlMV), type member of the alfamoviruses, is
infectious only when CP or its subgenomic messenger RNA 4 is present. This
early function of CP has been termed `genome activation'. Evidence has been
obtained that each of the genomic RNAs in an AlMV inoculum has to be complexed with a few CP molecules to permit initiation of infection. RNAs 1 and 2 of AlMV encode the replicase subunits P1 and P2 respectively, whereas RNA 3 encodes both the
movement protein P3 and CP (for a review see
1
). The four RNAs are separately encapsidated into bacilliform particles that are predominantly stabilized by RNA-CP interactions (
2
).
In addition to genome activation, interactions between AlMV CP and RNAs have
been implicated in RNA stability (
3
), RNA replication (
4
,
5
), a switch from minus-strand RNA synthesis to plus-strand RNA synthesis (
6
-
8
) and cell-to-cell spread (
9
). The evidence for the importance of RNA-CP interactions in the life cycle of AlMV has lead to several studies on
the identification of CP binding determinants in the AlMV RNAs. The observation that AlMV RNAs are able to withdraw CP subunits from virus particles
in vitro
indicated that the RNAs contain specific binding sites with a high affinity for
CP (
10
). In subgenomic RNA 4, which is identical in sequence to the 3'-terminal sequence of 881 nucleotides (nt) of RNA 3, these high-affinity binding sites were predominantly localized in the 3'-terminal untranslated region (UTR) (
11
). The 3'-terminal 145 nt of RNAs 1, 2 and 3 show a sequence similarity of
80% and despite the nucleotide differences in these sequences, all termini can be folded into a similar secondary structure (
12
). The proposed secondary structure, based both on phylogenetic comparisons (
12
,
13
) and enzymatic structure mapping data (
14
), consists of several stem-loop structures interspersed with single-stranded AUGC-motifs. This is illustrated in Figure
2
A for the 3'-UTR of RNA 3. This pattern is conserved in the 3'-terminal structures of ilarviruses (Fig.
1
). Previously, we demonstrated by band-shift assays the presence of a minimum of two specific binding sites for
CP in the 3'-UTR of RNA 3 (
23
). Site 1 consists of the region between nt 11 and 127 and site 2 is located
between nt 133 and 208 from the 3'-end. Both sites contain several hairpins and AUGC-motifs (Fig.
2
A). Deletion studies revealed that the two sites could bind CP independently of
each other and by site-directed mutagenesis it was shown that the AUGC-motifs are important for CP binding to both sites. The importance of
the AUGC-motifs for CP binding was further illustrated by Houser-Scott
et al.
(
13
) who demonstrated that mutation of the second AUGC-motif in the 3'-terminal 39 nt fragment of RNA 3 resulted in a loss of CP
binding. Moreover, these authors showed that poly(AUGC)
10
was unable to bind CP, indicating that the AlMV CP recognizes invariant AUGC
motifs in the context of conserved structural elements.
MATERIALS AND METHODS
RNA synthesis and purification
RNAs corresponding to the 3'-terminal 39 nt of AlMV RNA 3 were transcribed
in vitro
by T7 RNA polymerase from synthetic DNA templates containing a 24-base double-stranded T7 promoter and single-stranded sequences corresponding to the 3'-terminal region of AlMV RNA 3. The partially
double-stranded DNA templates were obtained by annealing a standard T7-oligodeoxyribonucleotide fragment (5'-AATTTAATACGACTCACTATAGGG-3') to a second oligodeoxyribonucleotide
fragment which is in part complementary to this T7-oligodeoxyribonucleotide fragment and in part coding for the RNA molecule of choice. Five [mu]M of each oligodeoxyribonucleotide fragment was incubated in annealing buffer (10 mM Tris-HCl pH 7.8, 5 mM MgCl
2
, 1 mM EDTA) for 15 min at 65oC and 20 min at room temperature prior to transcription. All RNAs contained
an addition of GGG at their 5'-end which promotes efficient transcription (
24
) and were radiolabelled by incorporation of [
32
P]UTP during the transcription reaction (reaction conditions: 200 mM Tris-HCl pH 8.2, 50 mM MgCl
2
, 25 mM DTT, 5 mM spermidine, 40% PEG-6000, 0.05% Triton-X-100, 500 nM DNA, 1 mM NTPs, 0.05 mCi [
32
P]UTP, 28 U RNasin and 17.5 U T7 RNA polymerase in total volume of 100 [mu]l). All RNAs were purified on 15% polyacrylamide-7 M urea gels, eluted from the gels with 10 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.1% SDS by shaking overnight at 37oC and ethanol precipitated. Purified RNA was resuspended in deionized
water and concentrations were determined on the basis of the specific activity
of [
32
P]UTP incorporated into the transcripts. Mutant RNA molecules were obtained by
introducing the desired mutations in the oligodeoxyribonucleotide fragment
containing the RNA 3 sequence.
Isolation of AlMV CP
CP was purified from AlMV particles in the form of a stable dimer as described
previously (
23
). Briefly, virions were dissociated into protein and RNA by mixing one volume
of virus suspension (~20 mg/ml) with one volume of 1 M MgCl
2
at 4oC. The RNA precipitate was pelleted by centrifugation and the protein in
the supernatant was dialyzed extensively against a buffer containing 0.05 M
sodium acetate, pH 5.5.
Gel mobility shift assays
Gel mobility shift assay conditions were adapted from Reusken
et al.
(
23
). One pmol of radiolabelled, gel-purified RNA and 3.75 pmol of purified AlMV CP were incubated in a total volume of 10 [mu]l of 10 mM phosphate, 1 mM EDTA, pH 7.0. The molar concentration of CP
was calculated on the basis of its presence in the form of dimers in the
preparation (
23
,
25
). Prior to use all RNAs were heated for 10 min at 65oC and chilled on ice. The resulting mixture was incubated for 30 min at
room temperature. When indicated, unlabeled competitor RNAs were added at a 10-fold mass excess prior to addition of CP. Purified AlMV RNA 4 was used as
the homologous competitor and RNA of turnip yellow mosaic virus (TYMV) was used
as non-homologous competitor (
23
). The incubation mixtures were analyzed by electrophoresis into a 7.5% polyacrylamide gel (acrylamide-bisacrylamide ratio of 40:1) with 1* Tris-borate-EDTA as the running buffer. After electrophoresis for
2 h at 375 V, the gel was dried and exposed to X-ray film.
RESULTS
Specific binding of CP to 3
'
-terminal fragments of RNA 3
Zuidema
et al
. (
26
) reported that CP was able to rebind a 36 nt 3'-terminal fragment of RNA 3 that was protected by CP against
digestion with ribonuclease T1. Subsequently, Houser-Scott
et al
. (
13
) showed that the 3'-terminal 39 nt of RNA 3 did bind CP with high affinity. The
possibility to transcribe this 3'-terminal RNA fragment with T7 RNA polymerase from synthetic DNA templates greatly facilitated a mutational analysis of this
sequence. The wild-type (wt) transcript of 39 nt with three nonviral G-residues at the 5'-end is shown in Figure
2
B. It consists of two hairpin structures flanked by AUGC-boxes 1, 2 and 3. Binding of CP to wt and mutant fragments was tested by
band-shift assays. An example of the technique is shown in Figure
3
A. Lane 1 of Figure
3
A shows the free radiolabelled wt transcript. Addition of CP results in a shift
of part of the fragment (Fig.
3
A, lane 2) which is completely abolished by addition of a 10-fold excess of the non-labelled homologous competitor (Fig.
3
A, lane 4). The observation that addition of a 10-fold excess of a non-homologous competitor did not affect the band-shift (Fig.
3
A, lane 3) demonstrates the specificity of the complex formation. Experiments with wt and mutant fragments were repeated at least four times with
reproducible results.
Role of AUGC-boxes in CP binding
To analyze the role of AUGC-boxes 1, 2 and 3 in the binding of CP to the 3'-terminal 39 nt of RNA 3, the mutations listed in Figure
2
B were made. The results of band-shift assays with the mutant transcripts are summarized in this figure by
a plus sign (binding similar to wt) or a minus sign (no detectable binding).
When the sequence of boxes 1, 2 or 3 was changed into AAAA, no binding of CP
was observed. This indicates that all three boxes play a role in the interaction of CP with the 39 nt fragment. Mutation of the sequence of box 2 to AGGC abolished CP binding but similar mutations of
boxes 1 or 3 did not affect CP binding under our assay conditions (Fig.
2
B). Apparently, sequence requirements for box 2 are more stringent than those for boxes 1 and 3. This conclusion is
further substantiated by the observation that mutation of box 2 to UAGC or UUGC
abolished CP binding whereas these mutations in box 3 had no effect (Fig.
2
B).
Role of stem-loop structure 2 in CP binding
The 5'-proximal hairpin in the 39 nt fragment of RNA 3 is referred to as
stem-loop structure 2 (STLP2) and the 3'-proximal hairpin is designated stem-loop structure 1 (STLP1). The stem structure of STLP2
contains an AUGC-motif at the position of nt 33 to 30 from the 3'-end of RNA 3. To investigate whether the stem structure of
STLP2 or the primary sequence of this fourth AUGC-motif in the 39 nt fragment were important for CP binding, the mutations
shown in Figure
4
were made (mutated nucleotides are indicated by a dot). Mutagenesis was
targeted to the adjacent U-A and A-U basepairs (nt 32-21 and 33-20) highlighted in the upper right panel of Figure
4
. Mutation of nt 33-32 from A-U into U-A in mutant STLP2-1 changed the AUGC-motif into UAGC and disrupted basepairing.
Figure
3
B shows that mutant STLP2-1 was unable to bind CP; the data are summarized by a minus sign in Figure
4
. In mutant STLP2-2 the A-U sequence of nt 21-20 was changed into U-A. This mutation disrupts basepairing but does not
affect the AUGC-motif. As shown in Figure
3
C and summarized by a minus sign in Figure
4
, mutant STLP2-2 was unable to bind CP. In mutant STLP2-3 the mutations of STLP2-1 and STLP2-2 were combined, resulting in a change of the wt
basepairs U-A and A-U into A-U and U-A respectively. Thus, basepairing is restored but
mutant STLP2-3 contains the same modification of the AUGC-motif at position 33-30 as mutant STLP2-1. The observation that mutant STLP2-3 did bind CP (Fig.
3
D; summarized by a plus sign in Fig.
4
) demonstrates that the secondary structure of the stem of STLP2 is essential
for CP binding rather than the AUGC-motif contained in this stem structure.
Role of stem-loop structure 1 in CP binding
Structural elements of CP binding sites in ilarvirus RNAs
The 3'-termini of AlMV and ilarvirus RNAs contain several AUGC-boxes brought together by two hairpin structures of variable
lengths (Fig.
1
). To see if our assay system permits the analysis of binding of AlMV CP to the
3'-termini of ilarvirus RNAs, we transcribed a synthetic
oligonucleotide
in vitro
into a RNA molecule corresponding to the 3'-terminal 49 nt of tobacco streak virus (TSV) RNA 3. This transcript
did bind AlMV CP in the band-shift assay at the same level as obtained with the wt AlMV 39 nt
transcript (Fig.
6
, TSV-3). Also, a chimeric molecule containing STLP1 of TSV and STLP2 of AlMV did bind CP (Fig.
5
, AlMV-TSV). The notion that the major function of the two hairpins is to arrange
the AUGC boxes in the correct position was confirmed by randomizing the stem
and loop sequences of STLP1 and STLP2 in the AlMV 39 nt fragment in such a way
that the folding pattern and overall [Delta]
G
value was maintained. The resulting non-viral RNA did bind AlMV CP as efficiently as the wt 39 nt fragment in the
band-shift assay (Fig.
6
, random).
DISCUSSION
Although AlMV and ilarvirus CP's show little sequence similarity, they are able to bind the 3'-termini of each others RNAs and are interchangeable in genome
activation (
30
-
35
). One of the 3'-structures recognized by these proteins in the viral RNAs probably
consist of three AUGC-boxes or related motifs, interspersed by two hairpins. We have used short
in vitro
synthesized transcripts to study the determinants in this structure required
for CP binding. Under the conditions of our band-shift assay, CP did bind to mutant transcripts either at a wt level or at
a non-detectable level. We cannot rule out the possibility that some mutant
transcripts had a reduced affinity for CP that would have been detectable at
other conditions. Despite this limitation of the technique, the results
presented in Figure
4
indicated that a stable stem structure of the 5'-hairpin (STLP2) in the 39 nt fragment is required for binding of
AlMV CP
in vitro.
Because randomization of the sequence of the stem and loop of STLP2 had little
effect on CP binding (Fig.
6
), this hairpin is probably not involved in sequence-specific interactions with CP. The AUGC-motif in the stem of STLP2 is not conserved in the 3'-terminal structures of ilarviruses (Fig.
1
) and the results from Figures
4
and
6
indicate that this motif has no role in CP binding other than participating in
the stem structure.
ACKNOWLEDGEMENTS
We thank Dr G. W. Witherell for his helpful advice concerning the T7
transcription procedure and C. M. A. van Rossum for the stimulating
discussions. TYMV RNA was a gift from Mrs C. J. Houwing. This work was
supported in part by the Netherlands Foundation for Chemical Research (SON)
with financial aid from the Netherlands Organization for Scientific research
(NWO).
REFERENCES
1 Jaspars, E.M.J. (1985) In J.W. Davies (ed.) Molecular Plant Virology. CRC Press, Boca Raton, Vol I, pp 155-221.
