Previous Article | Next Article ![]()
Journal of Virology, September 2007, p. 9519-9524, Vol. 81, No. 17
0022-538X/07/$08.00+0 doi:10.1128/JVI.00526-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
,
Colin Loney,
R. Paul Yeo,
and
David Bhella*
Medical Research Council Virology Unit, Institute of Virology, Church Street, Glasgow G11 5JR, United Kingdom
Received 13 March 2007/ Accepted 5 June 2007
|
|
|---|
|
|
|---|
Domains of N responsible for interactions with associated cofactors have been mapped for several members of Mononegavirales, including the rhabdovirus rabies virus (RV) (17), Sendai virus (SeV) (23), measles virus (MeV), (both members of the Paramyxovirinae [3]), and RSV (13, 16, 22). Although differences exist among the viruses, there is an overall homology in modular components of the N proteins: the amino terminus of N is largely responsible for the assembly of the nucleocapsid, while the carboxyl terminus contains elements that are necessary for the interaction with the polymerase and M protein (22).
Nucleocapsid formation proceeds concurrently with genome or antigenome synthesis, and N is thought to be delivered to the nascent RNA strand in complex with P (NoP). The NoP complex maintains N in a soluble state and promotes specific binding to both viral genomic and antigenomic RNA (10). For a number of viruses within Mononegavirales, expression of N in heterologous systems results in formation of nucleocapsid-like structures by nonspecific encapsidation of cellular RNAs (4, 12, 30); coexpression of P protein partially inhibits this effect (29).
Paramyxovirus nucleocapsids have a characteristic herringbone morphology when imaged in a transmission electron microscope (4). Structural investigations have proven a challenging area of research, owing to the presence of conformational flexibility in nucleocapsids as well as natively disordered regions within the C-terminal regions of N, which are thought to be critical in mediating cofactor interactions (5, 11, 20, 28). Three-dimensional reconstruction analysis of SeV nucleocapsids from negative stain images revealed a filament approximately 20 nm in diameter with a hollow core of about 5-nm diameter. Several discrete helical conformations were identified, revealing tightly wound helices with pitches (the distance between successive helix turns) of 53 Å and 68 Å and twist values of 13.07 and possibly 12.8 or 13.2 subunits per turn, as well as a longer-pitch helix visible only by metal shadowing (11). Analysis of MeV nucleocapsids produced by recombinant expression of full-length N in insect cells revealed a broad range of helix conformations ranging from 52 to 66 Å in pitch and at least 13 to 13.5 subunits per turn (5). Cryomicroscopy and three-dimensional reconstruction of trypsin-digested MeV nucleocapsids demonstrated improved helical order, allowing higher-resolution analysis and revealing that MeV N comprises two globular domains, one proximal to the helix axis that interacts with neighboring subunits to form the helix core and another distal domain that is at the end of an elongated spoke region that radiates away from the helix axis. Labeling of the RNA component within this structure indicated that it is most likely located close to the helix axis (28). Structural investigations of RSV nucleocapsids demonstrated a somewhat different nucleocapsid morphology, revealing a narrower diameter (14 to 16 nm) and longer pitch of between 68 and 74 Å (4). Pneumovirus nucleocapsids appear less well ordered than those of the Paramyxovirinae, rendering helical reconstruction considerably more difficult. In addition to production of helices, recombinant expression of RSV N also leads to the assembly of decameric and undecameric rings (4, 32), suggesting that the helical nucleocapsid may have 10 to 11 subunits per helix turn. Heterologous expression of rhabdovirus N has also been shown to produce decameric and undecameric rings, and these have recently been the subject of crystallographic analysis to 2.9 Å for vesicular stomatitis virus (VSV) and 3.5 Å for RV (2, 14).
We have conducted a structural investigation of RSV N-RNA purified from insect cells following expression of N from a recombinant baculovirus. We show that N is able to form nucleocapsid-like structures that are consistent with those seen previously; both helical structures and N-RNA rings were identified (4, 32). An additional helical form was observed in cryomicroscopy images with a significantly longer pitch than has been previously described. A three-dimensional reconstruction was calculated from images of N-RNA rings, revealing features strikingly reminiscent of N-RNA decameric and undecameric rings formed by recombinant expression of N proteins from other Mononegavirales. Prediction of disorder in the RSV N protein highlights similarities with other Mononegavirales N proteins. These predicted disordered regions might have functional importance, for example, initiating viral transcription by mediating interactions with the viral polymerase.
|
|
|---|
20% ammonium molybdate before blotting and plunge-freezing in ethane slush according to the method of Adrian et al. (1). Frozen grids were transferred to the JEOL 1200 EX in an Oxford Instruments CT3500 cryo-stage and imaged as described above. Preliminary image processing. Micrographs were digitized on a Dunvegan Hi-scan drum scanner at a raster step size of 10 µm per pixel, corresponding to 3.4 Å in the specimen. Digitized micrographs were transferred to an SGI Octane workstation running Irix 6.5 and converted to Purdue image format using the BSOFT package (15). A total of 12,484 particles were manually selected from 29 micrographs, and defocus values were calculated for each micrograph using X3d, Sumps, and CTFzeros (9). Partial contrast transfer function (CTF) correction was performed by inverting the phase of successive oscillations of the CTF, using CTFmix (9).
Image processing: N-RNA ring reconstruction. Calculation of the three-dimensional reconstruction was performed on a dual processor SGI Octane workstation running Irix 6.5 using EMAN (21). Particles were centered by iterative translational alignments against an average image, calculated from the whole data set. Class averages were generated by multireference alignment and multivariate statistical analysis, and an initial model was then calculated from unique top and side view class averages. Top view class averages revealed a C10 symmetry, which was assumed in all subsequent steps. Iterative classification and alignment of the data set, using reprojections of consecutive models and multireference alignment, resulted in successive improvement in resolution, up to 24 Å. The resolution of the final reconstruction was estimated by measurement of the Fourier shell correlation (FSC) between two separate maps calculated from equal and randomly distributed subsets of the data. The final reconstruction was visualized in Chimera. To compare this structure with crystallographic data from VSV N-RNA (PDB 2GIC) (14), the "fit model in map" option in Chimera was used to dock the decameric ring into our reconstruction (24). The fit was then refined using the Situs program colacor (33).
PONDR analysis. The sequence of RSV N (strain A2; NCBI accession number NC_001803 [31]) was submitted to the PONDR server (http://www.pondr.com/) using the default integrated predictor VL-XT (19, 25, 26). Access to PONDR was provided by Molecular Kinetics (Indianapolis, IN). VL-XT is copyright 1999 by the WSU Research Foundation, all rights reserved. PONDR is copyright 2004 by Molecular Kinetics, all rights reserved.
|
|
|---|
![]() View larger version (132K): [in a new window] |
FIG. 1. RSV nucleocapsids derived from lysed virions were visualized by electron microscopy by negative staining with 2% ammonium molybdate (a). Recombinant N protein nucleocapsids purified by CsCl buoyant density gradients were imaged in negative stain (b) or by cryo-negative stain (c). Images were recorded on a JEOL 1200 EX II TEM at a nominal magnification of x30,000 with Kodak SO 163 film. N-RNA rings are indicated with white arrows, while compact and extended forms of helical nucleocapsids are highlighted with black arrows.
|
N-RNA ring reconstruction. C10 symmetry was apparent in class averages of top views of the ring structure and confirmed by calculating a series of three-dimensional reconstructions with different symmetries imposed. These were found to have poor resolution and not to match the raw data well (not shown). In subsequent stages of refinement, C10 symmetry was imposed.
The EMAN refinement loop was iterated nine times and was deemed to have stabilized when the FSC curves of successive three-dimensional models converged, providing the first indication of the resolution of the reconstruction. To more accurately determine the resolution, the raw data were divided into two groups from which independent reconstructions were calculated. The Fourier shell correlation plot was calculated and the resolution determined from the point where the FSC curve crossed 0.5 (Fig. 2b).
![]() View larger version (47K): [in a new window] |
FIG. 2. Projection images calculated from the three-dimensional model of N-RNA rings and compared to class averages (a). FSC resolution test of the N-RNA ring reconstruction (b). Raw data were split into two subsets, from which independent reconstructions were calculated. The resolution of the reconstruction was estimated as the point at which the FSC crossed 0.5, giving a final resolution value of 24 Å. Surface representations of the three-dimensional reconstructions of recombinant RSV N-RNA rings are top view (c), tilted by 40° (d), side view (e), and bottom view (f).
|
Docking of a high-resolution structure for VSV N-RNA decameric rings into our map was performed to compare the two structures (Fig. 3). An initial manual fit was generated in Chimera (24) and optimized using the "fit model in map" option. This was subsequently refined using Situs (33), giving a correlation value of 0.84. Alternative fits were explored in which the hand of the RSV N-RNA reconstruction was flipped or the crystallographic data were manually placed in different orientations. These experiments yielded lower correlation values, indicating that the proposed alignment of these two structures is optimal.
![]() View larger version (58K): [in a new window] |
FIG. 3. The structure of RSV N-RNA decameric rings compared to the crystallographic structure of VSV N-RNA. The VSV crystal structure was docked to the RSV structure using UCSF Chimera and refined using Situs. Realizations in (a) top view, (b) 50° view, and (c) side view highlight the structural similarities between these two macromolecular assemblies.
|
![]() View larger version (20K): [in a new window] |
FIG. 4. PONDR prediction of disordered regions for RSV N. The sequence of RSV N was submitted to the PONDR server, which predicts regions of disorder with the VL-XT predictor. Regions of 40 amino acids or greater with PONDR scores above 0.5 are classified as disordered.
|
|
|
|---|
Similar morphologies have been observed for nucleocapsids of other Mononegavirales, although visualization of extended helices is less frequent. Helical recombinant nucleocapsids appear identical to virally derived nucleocapsids and would seem to be a good model for structural investigations. While N-RNA rings are most likely not a functional form of N, they provide a more tractable model for the analysis of N-RNA structure, as demonstrated by the successful crystallographic analysis of rings for VSV and RV N-RNA (2, 4, 14, 32).
RSV N-RNA appears morphologically similar to that of the rhabdoviruses VSV and RV. We have previously shown that RSV N-RNA is capable of forming both decameric and undecameric rings, with decamers being the more common form. In this study we have restricted our analysis to the decameric ring structure. Three-dimensional reconstruction provides a low-resolution view of the N protein. At this resolution we cannot derive any information concerning the arrangement of RNA within the ring or interactions between specific domains of N. We have, however, previously established that RSV N binds to six or possibly seven bases of RNA per monomer (32).
Structural characterization of Pneumovirinae and Paramyxovirinae nucleocapsids previously highlighted morphological differences between these two subfamilies, pneumovirus nucleocapsids being narrower with a longer pitch (4). We also observed significant differences between two-dimensional top view averages of RSV N-RNA rings and those of MeV or simian virus 5. Our current work further supports these observations; the published reconstructions of MeV N-RNA (5, 28) show a rather different density for N than we see for RSV. The structure of recombinant RSV N-RNA rings appears rather more comparable to those observed for RV and VSV N-RNA complexes, both of which form rings composed of 10 or 11 N monomers with slightly narrower internal radii of 30 to 38 Å and external radii of 80 Å (2, 14). Docking experiments in which the crystallographic coordinates of the VSV N-RNA decameric ring (14) were fitted into the reconstruction highlight the similar morphology in these two structures (Fig. 3). There are, however, significant biological differences between these two N proteins, not least that rhabdovirus N binds to nine rather than six or seven bases of RNA. This suggests that the RNA is likely to follow a less convoluted path in the RSV ring structure.
Disorder as a functional requirement of Mononegavirales N proteins.
Structural studies on SeV nucleocapsids derived from virus-infected cells appeared to suggest that a range of discrete pitch and twist conformations is adopted by the nucleocapsid (11). Investigations of recombinantly expressed MeV nucleocapsid structure have, however, indicated that rather than discrete conformations there may be significant structural plasticity, showing that pitch can vary continuously between 52 and 66 Å while twist varies from at least 13 to 13.5 subunits per turn (5). Removal of the C-terminal region of MeV N has been shown to cause changes in both pitch and twist (5, 28). The C terminus of MeV N, termed Ntail, is a natively unfolded domain which undergoes induced folding upon binding to partner proteins, including P, Hsp72, interferon regulatory factor 3, and a cell surface N protein receptor (6, 7, 18, 20, 34). The disorder-to-order transition in the MeV Ntail has been shown to involve
-helix formation in Ntail (20). It has also been established that Ntail binding to Hsp72 causes up-regulation of transcription and translation in MeV (34). Ntail is therefore critically important in both the structure and function of nucleocapsids, and it has been suggested that changes in helix morphology induced by binding of partner proteins to Ntail may modulate nucleocapsid function (5). Disordered regions have also been shown to exist in other Mononegavirales N proteins, including both RV and VSV (2, 14). In these N proteins the disordered regions are not at the C terminus; rather, they are between residues 373 and 398 in RV and 340 and 375 in VSV. Interestingly, these disordered regions include residues that are critical in P binding. In RV, serine 389 is required to be phosphorylated for P binding. Our PONDR predictions suggest that the RSV N may have a disordered C terminus, as residues 372 to 391 score above 0.5 and are flagged as disordered. Shorter regions of predicted disorder are also seen, notably at residues 117 to 125 and 183 to 194. The reliability of PONDR predictions increases for regions with a score of >0.5 that are longer than 40 residues; consequently, we should exercise caution with these data. By analogy with other Mononegavirales, however, these regions may be critical to nucleocapsid function, and the effects of induced folding events upon binding partner proteins may also modulate the structure and function of the nucleocapsid.
Published ahead of print on 13 June 2007. ![]()
Supplemental material for this article may be found at http://jvi.asm.org/. ![]()
Present address: Laboratoire d'Analyse Ultrastructurale, Université de Lausanne, Lausanne 1015, Switzerland. ![]()
Present address: Centre for Infectious Diseases, University of Durham, Queen's Campus, Wolfson Research Institute, University Boulevard, Stockton-on-Tees TS17 6BH, United Kingdom. ![]()
|
|
|---|
RII via its C-terminal domain: role in MV-induced immunosuppression. J. Virol. 77:11332-11346.This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»