Structure and nuclear import function of the C-terminal domain of influenza virus polymerase PB2 subunit.
ABSTRACT The trimeric influenza virus polymerase, comprising subunits PA, PB1 and PB2, is responsible for transcription and replication of the segmented viral RNA genome. Using a novel library-based screening technique called expression of soluble proteins by random incremental truncation (ESPRIT), we identified an independently folded C-terminal domain from PB2 and determined its solution structure by NMR. Using green fluorescent protein fusions, we show that both the domain and the full-length PB2 subunit are efficiently imported into the nucleus dependent on a previously overlooked bipartite nuclear localization sequence (NLS). The crystal structure of the domain complexed with human importin alpha5 shows how the last 20 residues unfold to permit binding to the import factor. The domain contains three surface residues implicated in adaptation from avian to mammalian hosts. One of these tethers the NLS-containing peptide to the core of the domain in the unbound state.
- SourceAvailable from: vir.sgmjournals.org[show abstract] [hide abstract]
ABSTRACT: The influenza A virus RNA-dependent RNA polymerase consists of three subunits PB1, PB2 and PA. The 5' and 3' terminal sequences of the viral RNA (vRNA) form the viral promoter and are bound by the PB1 subunit. The putative promoter-binding sites of the PB1 subunit have been mapped in previous studies but with contradictory results. The aim of the current study was to investigate the function of two evolutionary conserved regions in PB1 - from aa 233 to 249 and 269 to 281, which lie immediately N- and C-terminal, respectively, of a previously proposed binding site for the 3' end of the vRNA promoter. The previously proposed binding site extended from aa 249 to 256 and centred on two phenylalanine residues (F251 and F254). However, the fact that F251 is required for polymerase activity was not confirmed here. Instead, it was proposed that the 233-249 region contains a new 5' vRNA promoter-binding site, and arginine residues crucial for this activity were characterized. However, residues 269-281 were unlikely to be directly involved in promoter binding. These results are discussed in relation to the previous studies and a new model for vRNA promoter binding to the influenza RNA polymerase is presented.Journal of General Virology 04/2006; 87(Pt 3):679-88. · 3.13 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: The capped RNA primers required for the initiation of influenza virus mRNA synthesis are produced by the viral polymerase itself, which consists of three proteins PB1, PB2 and PA. Production of primers is activated only when the 5'- and 3'-terminal sequences of virion RNA (vRNA) bind sequentially to the polymerase, indicating that vRNA molecules function not only as templates for mRNA synthesis but also as essential cofactors which activate catalytic functions. Using thio U-substituted RNA and UV crosslinking, we demonstrate that the 5' and 3' sequences of vRNA bind to different amino acid sequences in the same protein subunit, the PB1 protein. Mutagenesis experiments proved that these two amino acid sequences constitute the functional RNA-binding sites. The 5' sequence of vRNA binds to an amino acid sequence centered around two arginine residues at positions 571 and 572, causing an allosteric alteration which activates two new functions of the polymerase complex. In addition to the PB2 protein subunit acquiring the ability to bind 5'-capped ends of RNAs, the PB1 protein itself acquires the ability to bind the 3' sequence of vRNA, via a ribonucleoprotein 1 (RNP1)-like motif, amino acids 249-256, which contains two phenylalanine residues required for binding. Binding to this site induces a second allosteric alteration which results in the activation of the endonuclease that produces the capped RNA primers needed for mRNA synthesis. Hence, the PB1 protein plays a central role in the catalytic activity of the viral polymerase, not only in the catalysis of RNA-chain elongation but also in the activation of the enzyme activities that produce capped RNA primers.The EMBO Journal 11/1998; 17(19):5844-52. · 9.82 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: The interaction of the PB1 subunit of the influenza virus polymerase with the viral RNA (vRNA) template has been studied in vitro. The experimental approach included the in vitro binding of labeled model vRNA to PB1 protein immobilized as an immunoprecipitate, as well as Northwestern analyses. The binding to model vRNA was specific, and an apparent Kd of about 2 x 10(-8) M was determined. Although interaction with the isolated 3' arm of the panhandle was detectable, interaction with the 5' arm was prominent and the binding was optimal with a panhandle analog structure (5'+3' probe). When presented with a panhandle analog mixed probe, PB1 was able to retain the 3' arm as efficiently as the 5' arm. The sequences of the PB1 protein involved in vRNA binding were identified by in vitro interaction tests with PB1 deletion mutants. Two separate regions of the PB1 protein sequence proved positive for binding: the N-terminal 83 amino acids and the C-proximal sequences located downstream of position 493. All mutants able to interact with model vRNA were capable of binding the 5' arm more efficiently than the 3' arm of the panhandle. Taken together, these results suggest that two separate regions of the PB1 protein constitute a vRNA binding site that interacts preferentially with the 5' arm of the panhandle structure.Journal of Virology 02/1999; 73(1):631-7. · 5.08 Impact Factor
Structure and nuclear import function of the C-terminal
domain of influenza virus polymerase PB2 subunit
Franck Tarendeau1,5, Julien Boudet2,5, Delphine Guilligay1, Philippe J Mas1, Catherine M Bougault2,
Se ´bastien Boulo3, Florence Baudin3, Rob W H Ruigrok3, Nathalie Daigle4, Jan Ellenberg4, Stephen Cusack1,
Jean-Pierre Simorre2& Darren J Hart1
The trimeric influenza virus polymerase, comprising subunits PA, PB1 and PB2, is responsible for transcription and replication
of the segmented viral RNA genome. Using a novel library-based screening technique called expression of soluble proteins by
random incremental truncation (ESPRIT), we identified an independently folded C-terminal domain from PB2 and determined
its solution structure by NMR. Using green fluorescent protein fusions, we show that both the domain and the full-length PB2
subunit are efficiently imported into the nucleus dependent on a previously overlooked bipartite nuclear localization sequence
(NLS). The crystal structure of the domain complexed with human importin a5 shows how the last 20 residues unfold to permit
binding to the import factor. The domain contains three surface residues implicated in adaptation from avian to mammalian
hosts. One of these tethers the NLS-containing peptide to the core of the domain in the unbound state.
The eight segments of the influenza virus RNA genome are packaged
into ribonucleoprotein particles (RNPs) containing the nucleoprotein
(NP) and the trimeric RNA-dependent RNA polymerase complex,
which comprises subunits PA, PB1 and PB2. The polymerase operates
in two distinct modes: in the first, it transcribes virally encoded genes,
using a cap-snatching mechanism to prime transcription and ensure
proper 5¢ capping of viral messenger RNA; in the second, it replicates
full-length viral RNA to produce first positive-strand complementary
RNA and then progeny viral RNA1. The PB1 subunit binds the
conserved 5¢ and 3¢ ends of the viral RNA2–4and carries both the
polymerase active site and the endonuclease activity that cleaves host-
cell mRNA bound by the cap-binding PB2 subunit5. PB1 residues
implicated in the endonuclease and polymerase active sites have been
identified5, although the location of the cap-binding site of PB2
remains controversial5–7. The polymerase is active in the nucleus
and NLSs have been identified on PB1 (ref. 8), PB2 (ref. 9) and
PA10. Recent results suggest that cytoplasmically expressed PB1 and PA
may be imported by the importin RanBP5 as a subcomplex, which
then assembles with separately imported PB2 (refs. 11,12). The host
proteins mediating PB2 nuclear import have not been characterized,
although it seems to be chaperoned by Hsp90 during transport13.
Current concern about the adaptation of highly pathogenic H5N1
avian influenza strains to humans has highlighted the need to under-
stand the genetic and molecular determinants of virulence and
transmissibility of influenza viruses14. Recent studies have focused
attention on polymerase subunit and NP point mutations that
accompany natural or laboratory-selected adaptations from avian to
mammalian hosts, notably in the case of the 1918 pandemic strain15–18.
However, the functional implications of these mutations—for instance,
whether they affect polymerase activity, assembly, stability or interac-
tions with host factors—are poorly understood. This is partly due to
the lack of detailed structural information on the polymerase, although
low-resolution negative-stain electron microscopic reconstructions of a
mini-RNP19and of the polymerase trimer20are available. The intract-
ability of the polymerase to structure determination results partly
from the fact that bacterially expressed subunits are insoluble and
yields of active enzyme produced in insect cells or mammalian cells
are prohibitively low. Additionally, biochemical or bioinformatic
knowledge of the domain structure of the polymerase subunits is
limited. As a new approach to overcome these problems, we devised a
directed evolution–type screen called ESPRIT to isolate bacterial clones
expressing soluble fragments of polymerase subunits from an essen-
tially complete library of terminal truncation variants. Here we present
the structural and functional characterization of a PB2 C-terminal
domain identified by this screen. This is the first successful application
of a new soluble expression screening strategy that should be of general
use in finding soluble domains in otherwise insoluble proteins for both
structural and functional studies.
Received 13 December 2006; accepted 31 January 2007; published online 25 February 2007; doi:10.1038/nsmb1212
1European Molecular Biology Laboratory (EMBL) Grenoble Outstation, 6 rue Jules Horowitz, BP181, 38042 Grenoble Cedex 9, France.2Institut de Biologie Structurale
Jean-Pierre Ebel UMR 5075 CNRS-CEA-UJF, 41 rue Jules Horowitz, 38027 Grenoble Cedex 1, France.3Unit of Virus Host Cell Interactions UMR 5233 UJF-EMBL-
CNRS, 6 rue Jules Horowitz, BP181, 38042 Grenoble Cedex 9, France.4EMBL Gene Expression Programme, Meyerhofstrasse 1, D-69117 Heidelberg, Germany.
5These authors contributed equally to this work. Correspondence should be addressed to D.J.H. (firstname.lastname@example.org) or S.C. (email@example.com).
NATURE STRUCTURAL & MOLECULAR BIOLOGY
VOLUME 14NUMBER 3MARCH 2007229
© 2007 Nature Publishing Group http://www.nature.com/nsmb
Isolation of a soluble domain by random construct screening
An exonuclease protocol21was used to generate a comprehensive,
unbiased set of 5¢ pb2 gene deletions, to which N-terminal
methionine and lysine were genetically fused. A library of approxi-
mately 26,880 constructs with a seven-fold oversample of each clone
was then arrayed robotically onto nitrocellulose membranes and
screened for soluble protein expression in a colony format using a
C-terminal biotin acceptor peptide22. This short peptide is biotiny-
lated in vivo by endogenous BirA and has been extensively exploited as
a tool for affinity purification and immobilization using avidin. Here
we used efficient biotinylation of expressed variants as a quantitative
readout for expression of stable, soluble protein. Colony array blots
were prepared by in situ lysis and expression of biotinylated proteins
was detected with a fluorescent streptavidin conjugate and fluorimager
(Fig. 1a). Intense positive clones were sequenced to identify the
truncation boundaries, revealing that many were predicted to express
C-terminal protein fragments of 80–110 amino acid residues. Two
clones in particular, corresponding to residues 678–759 (designated
DPDE, from the first four residues) and 661–759 (TTKR), were
expressed in soluble form at high levels.
Multimilligram quantities of each protein
were purified and characterized as mono-
disperse by gel filtration.
NMR solution structure of the DPDE
The1H-15N HSQC NMR spectrum of DPDE
shows a good dispersion of the amide reso-
nances at 10 1C, characteristic of a folded
domain (Supplementary Fig. 1 online).
TTKR differs only in the extent of disordered
N-terminal peptide. Measurements on a
sample of uniformly13C,15N-labeled DPDE
led to the assignment of backbone chemical
shifts for all residues except Asp678, Gly693
and Ser741. At 25 1C, the intensities of the
resonances of the N-terminal (678–685) and
C-terminal (744–759) regions were reduced,
as were those of Arg702, Gln728, Gly729 and
relaxation at 10 1C and analysis of data
showed that, with the exception of Lys738,
all temperature-sensitive residues belong to
flexible parts of the molecule (data not
shown). The lack of1H-1H NOEs for residues
678–684 confirms the absence of a stable
conformation for the N-terminal extremity.
Structure calculations were therefore performed for residues 685–759
using NMR restraints measured at 10 1C. Corresponding structural
statistics are presented in Supplementary Table 1 online.
680690 700710720 730740750
Figure 1 Identification and solution NMR structure of PB2 C-terminal domain. (a) Protein expression
screen of 26,880 random deletion constructs of the pb2 gene. Stable expression of soluble protein
results in efficient in vivo labeling of a C-terminal biotin acceptor peptide. Detection is by fluorescent
streptavidin and fluorimaging. (b) Ribbon diagram of the ten lowest-energy NMR structures superimposed
using backbone heavy atoms (r.m.s. deviation 0.94 A˚for 224 atoms). Indicated are Asp701, Arg702
and Ser714, which are implicated in cross-species transmission, and basic regions corresponding to the
minor (purple) and major (gold) sites of the bipartite NLS. (c) Primary sequence alignment comparing
influenza (FLU) A, B and C strains. Indicated are residues implicated in cross-species transmission (blue
triangles), the minor site (purple triangles) and major site (gold triangles) of the bipartite NLS, conserved
buried hydrophobic residues (black ovals) and secondary structure elements.
Control RKRx12KRIR RQRx12KRIR RKRx12NRIQ RQRx12NRIQ
Figure 2 Nuclear import of PB2 C-terminal domain and full-length
PB2 subunit is directed by a bipartite NLS. (a) Steady-state subcellular
localization of PB2 C-terminal domain (DPDE) or full-length PB2 fused to
enhanced GFP in NIH 3T3 cells. Constructs have either the wild-type NLS
(737-RKRx12KRIR-755) or the point mutations shown in red. Images show
2-mm single confocal sections of GFP fluorescence (GFP) and merged DIC
and fluorescence (Merge). Contrast and brightness have been adjusted for
display purposes. Scale bar, 10 mm. (b) Ratio of nuclear to cytoplasmic
concentration of GFP-fusion proteins, as a measure of nuclear import.
Nuclear and cytoplasmic mean fluorescence intensities were background-
subtracted and normalized to differences in total cell fluorescence intensity
for each mutant. Mean nuclear/cytoplasmic fluorescence intensities are
shown with s.d. (n ¼ 12 cells for each mutant). Asterisks mark significant
differences from wild-type construct (P o 0.05, Student’s t-test).
230 VOLUME 14NUMBER 3 MARCH 2007NATURE STRUCTURAL & MOLECULAR BIOLOGY
© 2007 Nature Publishing Group http://www.nature.com/nsmb
The solution structure of the C-terminal domain of PB2 reveals
a compact a-b domain with a hydrophobic core formed by a
three-stranded antiparallel b-sheet, onto which a short amphipathic
a-helix packs (Fig. 1b). Except for the flexible loop 726–730 connect-
ing the b2 and b3 strands and the last 20 residues at the C terminus,
the structure was determined with a high resolution (r.m.s. deviation
of 0.42 ± 0.12 A˚, calculated for the 81 backbone atoms of the
secondary structure elements). Regions of lower structural resolution
are correlated with enhanced dynamics (Supplementary Fig. 2
online). The PB2 C-terminal domain seems to be a novel
structural motif, as no similar protein with a Z-score 42 was found
using Dali23. It is likely that the equivalent domains of PB2 from
influenza B and C strains have similar structures because of
their relatively high sequence conservation, particularly of buried
hydrophobic residues (Fig. 1c).
Importin a5 binding and live-cell localization studies
Efficient nuclear localization of PB2 is believed to depend on an
internal region (residues 449–495) and a classical monopartite basic
motif, 736-KRKR-739 (ref. 9), which is present in DPDE. Mutation9of
the basic motif, deletion24of a C-terminal region containing it or
deletion of the internal domain9all abrogate nuclear localization. In
the NMR structure, 736-KRKR-739 immediately follows strand b3,
and all the basic side chains are surface exposed, except that of Lys736,
which is partly buried. To investigate the nuclear localization proper-
ties of the C-terminal domain, we studied its interaction with human
importin a5 (a importin), the classical eukaryotic nuclear import
adaptor. We first showed that DPDE forms a stable complex with
human importin a5 in vitro (Supplementary Fig. 3 online), suggest-
ing that the domain may be capable of directing nuclear import. We
then performed live-cell localization studies using a green fluorescent
protein (GFP)-DPDE fusion, which was found almost exclusively in
the nucleus (29 ± 8 times more nuclear than cytoplasmic), whereas a
GFP control remained equilibrated between the cytoplasm and
nucleus (Fig. 2). As previous results showed that the monopartite
motif was itself insufficient for nuclear import9, we examined
sequence alignments for other residues that might be implicated
(Fig. 1c). These show that although the monopartite NLS motif is
conserved in influenza B, only two basic residues occur in influenza C
(KRTA). However, a second basic motif, 752-KRIR-755, conserved in
influenza B and C strains, suggested the presence of a classical bipartite
NLS (KRx12–15KRxR, where x is any residue and x12–15is any sequence
12–15 residues long)25. To test this, the GFP-DPDE fusion was
mutated in the putative bipartite NLS (Fig. 2). Mutating the first
(‘minor’) part of the NLS (K738Q) mildly but significantly diminished
nuclear import (19 ± 5 times more nuclear; P ¼ 0.0004 compared
with wild-type), in agreement with studies on the whole PB2 protein9.
Mutating the second (‘major’) part (K752N R755Q) greatly dimin-
ished nuclear import (5 ± 1 times more nuclear; P ¼ 1.6 ? 10–8
compared with wild-type). Combining minor (K738Q) and major
(K752N R755Q) NLS mutations abolished import to levels of GFP
alone (2 ± 1 times more nuclear; P ¼ 9.5 ? 10–8compared with wild-
type and P ¼ 0.86 compared with GFP). Similar data were obtained
with a second mutation in the minor region (R737Q; data not shown)
and with full-length PB2-GFP fusions (Fig. 2). Consistent results were
also obtained when DPDE domains bearing the same NLS mutants
were analyzed for importin a5 complex formation in vitro (Supple-
mentary Fig. 3). Thus, the C-terminal PB2 sequence 738-KRx12KRIR-
755 functions as an efficient, classical, bipartite nuclear import signal,
with a similar response to mutations as that of the retinoblastoma
protein (KRx11KKLR) or phosphoprotein NIN2 (KRx12KKSK)25,26.
Although this NLS is functional in the isolated domain as well as the
full-length PB2 subunit, we cannot exclude the possibility that
other sequences present in PB2 are responsible for the very low
residual nuclear localization seen with the double NLS mutant of
the whole protein.
Cocrystal structure of DPDE with human importin a5
To provide further evidence for this revised mechanism of PB2 nuclear
import, we cocrystallized DPDE with residues 66–512 of human
importin a5 (lacking the autoinhibitory N-terminal region) and
determined the X-ray structure at 2.2-A˚resolution (see Methods
and Supplementary Fig. 4 online). The structure of importin a5
comprises ten armadillo repeats and is similar to that of yeast
karyopherin a (PDB 1BK5), with an r.m.s. deviation of 2 A˚(for 400
Ca positions of 422 residues aligned) and 56% sequence identity. This
is the first structure of a human importin a family member and
the first of a complex of importin with a complete folded domain,
rather than simply an NLS-containing peptide. DPDE interacts with
Figure 3 X-ray structure of the PB2 C-terminal domain complexed with
importin a5. (a) Ribbon diagram showing DPDE (red) bound to human
importin a5 (blue), comprising ten armadillo repeats. The C-terminal helix of
the importin is unpacked and mediates domain-swap dimer formation in the
crystal. The bipartite NLS at the C terminus of DPDE binds classically
within the superhelical groove of importin a5. Basic residues Arg737,
Lys738 and Arg739 from the minor site (purple) interact with the C-terminal
armadillo repeats; Lys752, Arg753 and Arg755 from the major site (gold)
interact with the N-terminal armadillo repeats. Lys736 does not interact
with importin a5 but makes intramolecular hydrogen bonds in DPDE,
perhaps preventing further unfolding of the C terminus. Lys718 makes three
hydrogen bonds with importin a5. (b) Comparison of the PB2 domain
structure in complexed (red) and free solution state (cyan) demonstrates
unfolding of residues 736–759 (purple) upon binding to importin a5.
Residue Asp701, important in host specificity and virulence, forms a salt
bridge with Arg753 of the major NLS motif and tethers the C terminus to
the core of the domain in the unbound state. Residues Arg702 and Ser714
are also implicated in interspecies transmission. Note different orientations
of the N-terminal helix of DPDE in the two structures.
NATURE STRUCTURAL & MOLECULAR BIOLOGY
VOLUME 14NUMBER 3 MARCH 2007231
© 2007 Nature Publishing Group http://www.nature.com/nsmb
importin a5 with its C-terminal residues beyond Lys736 in an
extended conformation, permitting binding of the bipartite NLS to
two distinct regions within the superhelical groove of importin a5
(Fig. 3a). The mode of NLS binding is similar to that previously
observed for bipartite NLS peptides binding yeast and mouse
a importins25,27. Strong interactions of the side chains of three basic
residues in both the minor (737-RKR-739) and major (752-KRIR-755)
sites, within discrete pockets of the importin, are supplemented by
additional hydrogen bonds, notably from the NE1 atoms of importin
tryptophans Trp149, Trp191, Trp234 and Trp360 to the DPDE NLS
peptide main chain carbonyl oxygens (Supplementary Fig. 5 online).
Residues 742–747 of the long (12-residue) linker are not visible in the
electron density, nor are the C-terminal residues 758–759, presumably
owing to flexibility. The side chain of Lys736 of DPDE makes strong
intradomain hydrogen bonds with the main chains of Ser714 and
Leu716 in both the free and bound states of the DPDE domain
(Supplementary Fig. 5), consistent with it not being crucial for
nuclear import9. The folded region of the DPDE domain packs against
importin a5 with a total buried surface area of about 916 A˚2; in
particular, Lys718 makes three hydrogen bonds with Gly284, Asn286
and Thr325 of importin a5 (Fig. 3a).
The structure of the core of the DPDE domain (residues 693–736)
observed in solution by NMR is very similar to the structure in the
crystalline state in complex with importin a5 (r.m.s. deviation of Ca
positions of 1.87 A˚for 41 residues aligned), although the X-ray
structure shows more secondary structure (Fig. 3b). The comparison
reveals how the region 736–759 unfolds to allow binding of the
bipartite NLS to the import factor (Fig. 3b). In the unbound NMR
structure, the carboxyl group of Asp701 forms a salt bridge with
Arg753, an important basic residue of the major site of the bipartite
NLS motif, thus tethering the C-terminal extremity to the core of the
domain. Binding to importin a5 requires that this interaction be
broken, as shown by the structure of the complex (Fig. 3a). A number
of recent studies have shown that PB2 residue 701 is important for
interspecies transmission. The D701N substitution has been found to
contribute considerably to the acquisition of pathogenicity to mice of
both H5N1 (ref. 15) and H7N7 (ref. 17) avian strains and has been
observed in both fatal and nonfatal cases of H5N1 infecting humans in
Vietnam28. We note that the mutation D701N would disrupt the salt
bridge with Arg753 of the NLS, potentially altering the energetics of
importin binding and, in turn, the efficiency of trimeric polymerase
assembly. The adjacent residue 702 is also an important determinant
of host specificity and, with only rare exceptions, is an arginine in
human isolates, including the 1918 H1N1 pandemic virus, and a lysine
in avian strains16. This residue is completely solvent exposed in
both structures, and its function is not yet known. Mutation of a
third residue in this domain, Ser714, in common with position 701,
affects levels of polymerase activity of reconstituted ribonucleoprotein
complexes in mammalian 293T cells; mutants with Asn701 and
Arg714 have the highest activity17. The fact that these three
cross-species-transfer residues15–17are exposed to solvent suggests a
possible role of this domain in intermolecular interactions with other
viral or host proteins; the interaction with importin a5 and the
regulation of nuclear transport of PB2 both could be important in
Identification of soluble constructs. DNA encoding PB2 from strain
A/Victoria/3/75(H3N2)29fused to a C-terminal biotin acceptor peptide was
truncated from the 5¢ end with exonuclease III. An Escherichia coli BL21 strain
was transformed with a pool of 24,036 plasmids, and clones were screened for
protein expression by robotic arraying onto nitrocellulose membranes over
agar. Colonies were lysed in situ and streptavidin conjugates used to detect
biotinylated proteins (Supplementary Methods online).
NMR spectroscopy and structure calculation. Heteronuclear NMR data
for assignment and extraction of structural restraints were collected at
10 1C on a 0.67 mM uniformly13C,15N-labeled DPDE sample in 50 mM
Tris buffer (90% (v/v) H2O, 10% (v/v) D2O) at pH 7.5 with 200 mM NaCl,
1 mM EDTA and 0.01% (w/v) NaN3. Structure calculations including
residues 685–759 were performed using CNS30and refined in explicit water
Crystallography. Cocrystals of DPDE and human importin a5 (NSMB Protein
AAP35605; residues 66–512) were grown by vapor diffusion with 2-methyl-2,
4-pentanediol (MPD) as precipitant. Data to 2.2-A˚resolution were measured
on beamline ID23-1 at the European Synchrotron Radiation Facility, using
X-rays of wavelength 1.072 A˚and a temperature of 100 K, and integrated with
XDS31. Two complexes were located in the asymmetric unit by molecular
replacement using PDB 1Q1S (mouse importin a with NLS peptide) as search
model. The map was greatly improved using RESOLVE32with two-fold
noncrystallographic symmetry averaging, allowing the complete structure
(residues 84–508 of importin a5, and 686–741 and 748–757 of DPDE) to be
constructed and refined using REFMAC33. The final R-factor was 20.6%
(Rfree¼ 24.7%). According to Molprobity (http://molprobity.biochem.duke.
edu/), 98.4% and 99.6% of the total of 966 residues are within the favored
or allowed regions of the Ramachandran plot, respectively, with four outliers
(see Table 1 and Supplementary Methods).
Nuclear localization. Mouse Swiss National Institutes of Health embryonic
fibroblasts (NIH 3T3) were transfected with plasmids encoding the PB2
Table 1 Crystallographic data collection and refinement statistics
Human importin a5 with influenza A
PB2 C-terminal domain
a, b, c (A˚)
I / sI
94.97, 100.55, 151.84
No. reflections (total / free)
Protein (chain A, chain B)
DPDE (chain D, chain E)
Bond lengths (A˚)
Bond angles (1)
74,089 / 2,969
Values in parentheses are for highest-resolution shell.aTwo molecules in the asymmetric unit.
232VOLUME 14NUMBER 3 MARCH 2007 NATURE STRUCTURAL & MOLECULAR BIOLOGY
© 2007 Nature Publishing Group http://www.nature.com/nsmb
domain or full-length PB2 fused C-terminally to enhanced GFP. Cells were
imaged by quantitative confocal microscopy (Supplementary Methods).
Accession codes. BioMagResBank (http://www.bmrb.wisc.edu): The1H,13C
and15N chemical shifts for the C-terminal domain of PB2 with Pro679 in the
trans (major) and cis (minor) conformations have been deposited with
accession number 7056 and linked to the atomic coordinates of the ten
structures of lowest energy in the Protein Data Bank (PDB 2GMO and RCSB
037302). Protein Data Bank: coordinates and structure factors of the PB2
domain–importin a5 complex have been deposited with accession codes 2JDQ
and R2JDQSF, respectively.
Note: Supplementary information is available on the Nature Structural & Molecular
We thank J. Ortin (Centro Nacional de Biotecnologia, CSIC, Madrid) for the pb2
gene, A. Favier (Institut de Biologie Structurale, Grenoble) for NMR scripts and
the EMBL Centre for Molecular and Cellular Imaging for suggestions. Screening
for crystals was done by the Partnership for Structural Biology high-throughput
crystallization facility. We thank the European Synchrotron Radiation Facility
and EMBL Joint Structural Biology group for assistance with the synchrotron
beamtime and T. Crepin (EMBL, Grenoble) for help with data collection. Partial
funding was provided by the European Commission Framework 5 Integrated
Project ‘Structural Proteomics in Europe’ (SPINE, contract QLG-CT-2002-00988).
D.J.H. conceived the ESPRIT method. D.J.H., F.T. and P.J.M. implemented
ESPRIT. D.G. purified wild-type and mutant DPDE for in vitro binding studies
and crystallization and made double-labeled protein for NMR. C.M.B., J.B. and
J.-P.S. performed the NMR measurements and structural analysis. S.B. purified
importin a5 under the supervision of F.B. and cocrystallized it with DPDE.
R.W.H.R. and S.C. initiated the influenza polymerase project, and S.C.
determined the crystallographic structure. F.T. and N.D. performed nuclear
import assays with instrumentation and methodology established by J.E.
S.C. and D.J.H. compiled the text, with contributions from all authors.
COMPETING INTERESTS STATEMENT
The authors declare competing financial interests (see the Nature Structural &
Molecular Biology website for details).
Published online at http://www.nature.com/nsmb/
Reprints and permissions information is available online at http://npg.nature.com/
1. Noah, D.L. & Krug, R.M. Influenza virus virulence and its molecular determinants. Adv.
Virus Res. 65, 121–145 (2005).
2. Jung, T.E. & Brownlee, G.G. A new promoter-binding site in the PB1 subunit of the
influenza A virus polymerase. J. Gen. Virol. 87, 679–688 (2006).
3. Li, M.L., Ramirez, B.C. & Krug, R.M. RNA-dependent activation of primer RNA
production by influenza virus polymerase: different regions of the same protein subunit
constitute the two required RNA-binding sites. EMBO J. 17, 5844–5852 (1998).
4. Gonzalez, S. & Ortin, J. Characterization of influenza virus PB1 protein binding to viral
RNA: two separate regions of the protein contribute to the interaction domain. J. Virol.
73, 631–637 (1999).
5. Li, M.L., Rao, P. & Krug, R.M. The active sites of the influenza cap-dependent
endonuclease are on different polymerase subunits. EMBO J. 20, 2078–2086 (2001).
6. Honda, A., Mizumoto, K. & Ishihama, A. Two separate sequences of PB2 subunit
constitute the RNA cap-binding site of influenza virus RNA polymerase. Genes Cells 4,
7. Fechter, P. et al. Two aromatic residues in the PB2 subunit of influenza A RNA
polymerase are crucial for cap binding. J. Biol. Chem. 278, 20381–20388
8. Nath, S.T. & Nayak, D.P. Function of two discrete regions is required for nuclear
localization of polymerase basic protein 1 of A/WSN/33 influenza virus (H1 N1). Mol.
Cell. Biol. 10, 4139–4145 (1990).
9. Mukaigawa, J. & Nayak, D.P. Two signals mediate nuclear localization of influenza virus
(A/WSN/33) polymerase basic protein 2. J. Virol. 65, 245–253 (1991).
10. Nieto, A. et al. Nuclear transport of influenza virus polymerase PA protein. Virus Res.
24, 65–75 (1992).
11. Fodor, E. & Smith, M. The PA subunit is required for efficient nuclear accumulation of
the PB1 subunit of the influenza A virus RNA polymerase complex. J. Virol. 78,
12. Deng, T. et al. Role of ran binding protein 5 in nuclear import and assembly
of the influenza virus RNA polymerase complex. J. Virol. 80, 11911–11919
13. Naito, T., Momose, F., Kawaguchi, A. & Nagata, K. Involvement of hsp90 in assembly
and nuclear import of influenza virus RNA polymerase subunits. J. Virol. 81,
14. Russell, C.J. & Webster, R.G. The genesis of a pandemic influenza virus. Cell 123,
15. Li, Z. et al. Molecular basis of replication of duck H5N1 influenza viruses in a
mammalian mouse model. J. Virol. 79, 12058–12064 (2005).
16. Taubenberger, J.K. et al. Characterization of the 1918 influenza virus polymerase
genes. Nature 437, 889–893 (2005).
17. Gabriel, G. et al. The viral polymerase mediates adaptation of an avian influenza virus
to a mammalian host. Proc. Natl. Acad. Sci. USA 102, 18590–18595 (2005).
18. Salomon, R. et al. The polymerase complex genes contribute to the high virulence of
the human H5N1 influenza virus isolate A/Vietnam/1203/04. J. Exp. Med. 203,
19. Martin-Benito, J. et al. Three-dimensional reconstruction of a recombinant influenza
virus ribonucleoprotein particle. EMBO Rep. 2, 313–317 (2001).
20. Area, E. et al. 3D structure of the influenza virus polymerase complex: localization of
subunit domains. Proc. Natl. Acad. Sci. USA 101, 308–313 (2004).
21. Ostermeier, M. & Lutz, S. The creation of ITCHY hybrid protein libraries. Methods Mol.
Biol. 231, 129–141 (2003).
22. Beckett, D., Kovaleva, E. & Schatz, P.J. A minimal peptide substrate in biotin
holoenzyme synthetase-catalyzed biotinylation. Protein Sci. 8, 921–929 (1999).
23. Holm, L. & Sander, C. Protein structure comparison by alignment of distance matrices.
J. Mol. Biol. 233, 123–138 (1993).
24. Perales, B., de la Luna, S., Palacios, I. & Ortin, J. Mutational analysis identifies
functional domains in the influenza A virus PB2 polymerase subunit. J. Virol. 70,
25. Fontes, M.R., Teh, T., Jans, D., Brinkworth, R.I. & Kobe, B. Structural basis for the
specificity of bipartite nuclear localization sequence binding by importin-alpha. J. Biol.
Chem. 278, 27981–27987 (2003).
26. Zacksenhaus, E., Bremner, R., Phillips, R.A. & Gallie, B.L. A bipartite nuclear
localization signal in the retinoblastoma gene product and its importance for biological
activity. Mol. Cell. Biol. 13, 4588–4599 (1993).
27. Conti, E. & Kuriyan, J. Crystallographic analysis of the specific yet versatile recognition
of distinct nuclear localization signals by karyopherin alpha. Structure 8, 329–338
28. de Jong, M.D. et al. Fatal outcome of human influenza A (H5N1) is associated with
high viral load and hypercytokinemia. Nat. Med. 12, 1203–1207 (2006).
29. de la Luna, S., Martinez, C. & Ortin, J. Molecular cloning and sequencing of influenza
virus A/Victoria/3/75 polymerase genes: sequence evolution and prediction of possible
functional domains. Virus Res. 13, 143–155 (1989).
30. Brunger, A.T. et al. Crystallography & NMR system: a new software suite for macro-
molecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921
31. Kabsch, W. Automatic processing of rotation diffraction data from crystals of initially
unknown symmetry and cell constants. J. Appl. Cryst. 26, 795–800 (1993).
32. Terwilliger, T.C. Maximum-likelihood density modification. Acta Crystallogr. D Biol.
Crystallogr. 56, 965–972 (2000).
33. Murshudov, G.N., Vagin, A.A. & Dodson, E.J. Refinement of macromolecular structures
by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr. 53, 240–255
NATURE STRUCTURAL & MOLECULAR BIOLOGY
VOLUME 14NUMBER 3 MARCH 2007233
© 2007 Nature Publishing Group http://www.nature.com/nsmb