ABSTRACT
The double-stranded panhandle structure of the influenza virus RNA is important for
replication, transcription and packaging into the virion of the virion RNA. The
solution structure of a 34 nt RNA which contains the conserved panhandle
sequences has been investigated by one- and two-dimensional NMR spectroscopy. The partially complementary 5
'
- and 3
'-ends of the RNA form a double helical structure which is, on average,
close to A-form. The stem contains bulges at nucleotides A10, A12 and C26. In between
these bulges, C11 and G25 form a Watson-Crick base pair. The structural features of the panhandle provide a
framework for the explanation of mutational analysis and for a better
understanding of RNA-polymerase interactions.
The influenza A virus genome is composed of eight segments of single-stranded RNAs of negative polarity (
1
). During influenza virus infection, the negative sense virion RNA (vRNA) serves
as a template for both mRNA synthesis and cRNA synthesis (
2
). Although control of transcription and replication of viral RNA is not well
understood, both processes of RNA replication and RNA transcription are
catalyzed by the same RNA-dependent RNA polymerase (
2
). Since these two modes of RNA synthesis occur at the same promoter in
competition, there must be a kind of regulatory mechanism to ensure the control
of both mRNA and cRNA for optimal growth of the virus.
Sequencing of influenza A virus RNA segments revealed that all the segments
contain 13 and 12 conserved nt at their 5'- and 3'-ends respectively (
3
,
4
). It has been shown that the vRNA promoter is located at the 3'-end of the vRNA segment (
5
,
6
). A model for polymerase promoter recognition has been proposed in which four
contiguous internal nucleotides are crucial for binding, located at or near the
initiation site of the 3'-terminus (
7
). More recently, however, it has been shown that the 5'-end of the vRNA plays an important role in regulation of
transcription as well. The sequences of the 5'- and 3'-ends of the vRNA are partially complementary to each
other (
3
,
4
) and it has been suggested that a panhandle structure can form (
8
; Fig.
1
A). The double-stranded panhandle structure appears to be required for polyadenylation
and for transcription termination (
9
,
10
).
In vitro
studies using recombinant influenza virus polymerase have demonstrated that
both viral 5'- and 3'-ends are required for polymerase activity and that the
polymerase binds sequence-specifically to the partially double-stranded panhandle structure (
11
,
12
). Both an RNA polymerase binding assay and an
in vitro
transcription assay have shown that the panhandle structure is involved in
initiation of transcription (
13
). Nuclease S1 mapping of the influenza virus RNA has indicated that the termini
of the RNA segments form a 15 bp panhandle structure (
8
; Fig.
1
A). Another alignment, which contains an internal loop, has been proposed as a
model structure of the panhandle of the influenza virus RNA (
14
; Fig.
1
B). Recently, however, by using different nucleases and chemical probes, Baudin
et al.
(
15
) have determined a somewhat different panhandle structure, in which the
panhandle contains an internal loop and a bulge (Fig.
1
C).
In this paper, we report the structure of this panhandle RNA determined in
solution by NMR spectroscopy. On average, the influenza panhandle structure is
close to double helical A-form geometry, although it contains a G[middot]U base pair and internal bulges. From our studies we conclude that
C11 and G25 ( Fig.
1
D for nomenclature), which were considered part of an internal loop in the
previously proposed panhandle structures, are base paired. We have changed the
residue U at position 5 to C to increase the transcription yield. We do not
think that this single base substitution can alter the overall secondary
structure of the panhandle RNA.
An RNA oligonucleotide, 5'-GGAGCAGAAACAAGGCUUCGGCCUGCUUUUGCUC-3', was synthesized using T7 RNA polymerase and
synthetic DNA templates (
16
). DNA templates were synthesized chemically on an Applied Biosystems 391 and
were purified by 15% polyacrylamide gel electrophoresis under denaturing
conditions (7 M urea). Synthesis and purification of the RNA oligonucleotide
were done as previously reported (
17
,
18
).
The purified sample was dialyzed extensively against 10 mM sodium phosphate,
0.01 mM EDTA, pH 6.5. It was then lyophilized and dissolved in 0.25 ml 90% H
2
O/10% D
2
O for exchangeable proton experiments. For non-exchangeable proton experiments, the sample was lyophilized several times
from 99.9% D
2
O and dissolved in 0.25 ml 99.96% D
2
O (Aldrich). A microtube from Shigemi was used to decrease the sample volume.
All NMR experiments were done on a Bruker DMX-600 NMR spectrometer operating at 600 MHz proton frequency. Exchangeable
proton spectra were obtained using the 1-1 (jump-return) method (
19
). One-dimensional NOE experiments were performed with a pre-irradiation time of 600 ms. All two-dimensional NMR spectra were acquired in phase-sensitive mode using the TPPI method (
20
). The two-dimensional spectra were acquired with 400 or 512 FIDs of 2048 complex
data points, using spectral widths of 3800 Hz for spectra in D
2
O and 12000 Hz for spectra in H
2
O. The repetition delay was set to ~2 s and 64 scans were averaged for each FID; the total acquisition time was
<24 h. Data were zero-filled to 1 K real points in
t
1
and apodized by using 45-60o phase-shifted squared sine bells in both dimensions.
NOESY spectra in H
2
O with mixing times of 100 and 400 ms were obtained using a jump-return pulse for solvent suppression (
21
). NOESY spectra in D
2
O were obtained at a mixing time of 400 ms. The residual HDO resonance was
presaturated during the relaxation delay. DQF-COSY spectra were obtained using the standard pulse sequence (
22
).
The imino proton spectra were assigned by standard methods using one- and two-dimensional NNR spectroscopy. Uracil imino protons were identified
by their strong NOEs to the H2 protons of base paired adenines. The amino
protons of the cytosines in the stem were assigned by NOEs to their own H5
protons and to the cross-strand guanine imino protons. We have introduced an RNA hairpin loop, 5'-CUUCGG-3', which possesses a unique conformation (
23
,
24
), to increase the stability of the panhandle and to help the assignment. Figure
2
shows the imino proton spectrum of the panhandle RNA studied here (Fig.
1
D). The upfield shifted resonance at 9.90 p.p.m., which shows high thermal
stability, was assigned to the G20 imino proton in the loop from the previously
studied results (
24
). Assignments of the imino protons of G14, G15 and G21 were by sequential imino-imino NOE connectivities (Fig.
2
). A strong NOE was observed between the two imino protons of the G7[middot]U29 base pair. Sequential imino-imino NOEs were observed from the G2[middot]C34 to A9[middot]U27 base pairs. The sharp resonance at 12.64
p.p.m. was assigned to the imino proton of G25, since it gave NOEs to amino
protons of cross-strand C11. A NOESY crosspeak from the G25 imino to C11 H5 proton was also
observed at long mixing times (400 ms), providing further support once the non-exchangeable protons had been assigned.
In this paper we have described the panhandle structure of the influenza virus
RNA. The influenza virus RNA contains 13 and 12 nt conserved sequences at the 5'- and 3'-ends respectively, which are partially complementary
to each other (
3
,
4
). These sequences are known to form a unique panhandle structure in the RNP
complex in virions as well as in infected cells, but not in the naked vRNA (
8
). Baudin
et al.
(
15
) have shown an opposite result, where the 5'- and 3'-ends of the RNA form a base paired panhandle structure
in the naked state, whereas in the RNP complex they are single-stranded. Our results clearly show that in the naked state the RNA forms a
base paired panhandle structure which contains a G[middot]U base pair and internal bulges.
Many detailed mutational analyses of various sequences of the panhandle RNAs
have been performed to investigate the functional importance of specific
sequences for both replication and transcription. The NMR data provide an additional framework for interpreting the mutation studies and for suggesting further biochemical
and genetic studies. Modification by diethyl pyrocarbonate or hydrazine
interference analyses have shown that the influenza virus RNA polymerase binds
sequence-specifically to four residues, ACAA
10-13
, which are located in a proposed bulge region in the 5'-side of the panhandle (Fig.
1
; ref
12
). Mutations of either C11 or G25 have been found to result in large decreases
in the initiation of transcription (
27
). Our NMR data indicate that C11 and G25 form a Watson-Crick base pair that may be important for recognition and binding of the
virus RNA polymerase. Although mutation of A12 did not interfere with
transcriptional activity (
27
), imidazole ring opening by diethyl pyrocarbonate of the residue decreased
polymerase binding (
12
). Mutation of A10 inhibited transcriptional activity (
27
). Systematic mutational analyses have been carried out showing that three
contiguous internal residues, GCU
25-27
, are crucial for polymerase binding (
28
) and transcriptional activity (
7
). We observe that residues A10, A12 and C26 form bulges and all the residues
are stacked into the helix. It is very likely that the backbone might be bent
or kinked in this region, although the overall geometry of the stem conserves
the A-form. Unlike other nucleotides in the stem, the riboses of A12, A13 and
G25 adopt ~40-60% N-type sugar conformation (
26
). The presence of bulges and dynamic sugar conformations suggest that this
region is more flexible than the other parts of the panhandle and this
flexibility might be important for protein-RNA interaction as well as base-specific interactions.
An RNA fork model for the initiation of influenza virus RNA has been proposed in
which the RNA forked structure is partly double-stranded and partly single-stranded (
13
). It was suggested that polymerase binds to the single-stranded 5'-end of the vRNA and the RNA-polymerase complex binds to the single-stranded 3'-end to form an RNA fork. Although
Fodor
et al.
(
27
) assumed that the 5'- and 3'-ends of the vRNA would not form the panhandle
structure, Baudin
et al.
(
15
) and we have shown that the 5'- and 3'-ends are base paired in the naked state. Thus, we
suggest that initially the polymerase recognizes and binds to the proposed
bulge region of the partially double-stranded vRNA termini, followed by melting of the termini, generating the
RNA fork.
The authors are indebted to Dr Baik L. Seong (Hanhyo Institutes of Technology) for suggesting this project and for useful discussions. This work was supported by the Center for Molecular Catalysis and the Korea
Science and Engineering Foundation (to B.-S.C.) and by the Ministry of Science and Technology, the Republic of Korea
(to C.C.).
REFERENCES
Return