Structural studies on rotavirus

The biochemistry and molecular biology of rotaviruses are well characterized (Estes 1996). Rotaviruses belong to the family of viruses called Reoviridae. The rotavirus particle contains multiple protein layers enclosing the genome that consists of 11 segments of double-stranded (ds) RNA. Each segment codes for one protein with the exception of segment 11 that codes for two proteins. Out of the 12 proteins, six are structural and six are non-structural. The outer capsid layer is made of VP7 and VP4. The intermediate layer is composed of VP6. Enclosed within these two layers is the core consisting of VP2, VP1 and VP3. VP1 and VP3 are minor internal proteins present in small quantities. The non-structural proteins participate in various functions during the life cycle of the virus.

To provide a biological perspective to our structural studies, a brief outline of the virus life cycle is given below. The first event in the life cycle of the virus is its interaction with the host cell. It remains unclear as to how rotaviruses enter the cell (see Arias et al 2001, this volume). It is either through an endocytotic pathway or through direct penetration of the cell membrane. The choice of pathway seems to depend upon whether or not the virus is trypsinized. Non-trypsinized viruses may enter the host cell through the rather inefficient endocytotic pathway. On the other hand, the trypsinized viruses may enter by direct penetration (Kaljot et al 1988). During cell entry, mature virions, which are triple-layered particles (TLPs), lose their outer shell and the resulting double-layered particles (DLPs) become transcriptionally active. Endogenous transcription, a common feature of dsRNA viruses, is a dynamic and fascinating process. During this process, the 11 dsRNA segments are transcribed inside the intact DLPs and the mRNA molecules are released. These mRNA molecules act as templates for the progeny RNA and code for all the structural and non-structural proteins (see Patton 2001, this volume). Out of the 12 proteins, 10 are synthesized on the free ribosomes. The other two — VP7, the major outer layer glycoprotein and a non-structural protein, NSP4 — are synthesized on the ribosomes associated with the rough endoplasmic reticulum (RER) where they are co-translationally inserted into the ER membrane. The assembly of progeny DLPs takes place in a specialized compartment called the viroplasm. Newly formed DLPs bud through the ER membrane. NSP4, the non-structural protein localized in the ER membrane, acts as an intracellular receptor and facilitates the budding ofthe DLPs through the ER membrane. Recently, it has been shown that NSP4 is a viral enterotoxin and by itself can cause diarrhoea in mice (see Estes et al 2001, this volume). One of the unique steps in the morphogenesis of rotavirus is the formation of transiently

Rotavirus Dlp

FIG. 1. Structures of rotavirus TLP (a) and DLP shown in the context of cell entry and removal of the outer layer (b), and transcription (c) inside the cell. Various structural features in the outer layer of the TLP — three types of channels, VP4 spikes, and VP7 trimers are indicated by dashed arrows in (a). In (b) a dashed arrow shows one of the 260 VP6 trimers in the DLP. The nascent mRNA molecules, shown as curvy cylinders in (c) exit through type I channels in the DLP (Lawton et al 1997a). All the structures are shown as viewed along their icosahedral threefold axis. The structures of TLP, DLP and the transcribing particle were determined using cryo-EM techniques.

FIG. 1. Structures of rotavirus TLP (a) and DLP shown in the context of cell entry and removal of the outer layer (b), and transcription (c) inside the cell. Various structural features in the outer layer of the TLP — three types of channels, VP4 spikes, and VP7 trimers are indicated by dashed arrows in (a). In (b) a dashed arrow shows one of the 260 VP6 trimers in the DLP. The nascent mRNA molecules, shown as curvy cylinders in (c) exit through type I channels in the DLP (Lawton et al 1997a). All the structures are shown as viewed along their icosahedral threefold axis. The structures of TLP, DLP and the transcribing particle were determined using cryo-EM techniques.

enveloped particles. Rotavirus particles exit the cells by lysing the cell, or by a non-classical vesicular transport from polarized cells.

We have used the cryo-EM technique to understand the structural basis of trypsin-enhanced infectivity, endogenous transcription, replication, assembly processes, and the budding of the DLPs through the ER membrane (Prasad & Estes 1997, 2000). We review here recent progress on the structure—function relationships in rotavirus.

Overall organisation of rotavirus

Rotavirus is an icosahedral particle with an overall diameter of 1000 A (Fig. 1a). The capsid architecture is built on an icosahedral T = 13 lattice (Prasad et al 1988). Two distinct surface features of the structure are spikes, and channels. The channels are located at all the 5- and 6-coordinated positions of the T = 13 lattice. Three types of channels are present, depending on their locations with respect to the icosahedral symmetry axes. The type I channels are at the fivefold axes, type II channels surround the icosahedral 5-fold axes, and type III channels neighbour the threefold axes. Each particle has a total of 132 channels— 12 type I, 60 type II and 60 type III channels. VP7, which constitutes the outer layer, clusters into trimers surrounding the channels. Each particle has 260 trimers or 780 molecules of VP7. There are 60 spikes, located at one edge of the type II channels, in each particle.

Our structural analysis using anti-VP4 monoclonal antibodies (MAb) has shown that the spikes are made up of VP4 (Prasad et al 1990). Two molecules of the Fab bind on either side of the bi-lobed head of each spike. VP4 has been implicated in cell entry, neutralization, virulence, and in some animal viruses it is a haemagglutinin (reviewed in Estes 1996). The first MAb that was used in this study is a neutralizing antibody that affects virus binding to cells and prevents internalization of the virus. This MAb binds to the side of the bilobed head of the spike, indicating that this region on the spike is important for neutralization. In addition to interacting with VP7, the outer layer protein, VP4 also interacts with VP6 through a distinct globular domain that is buried inside the type II channels underneath the VP7 layer (Shaw et al 1993, Yeager et al 1994).

During trypsinization, which is known to cause a significant increase in the infectivity of rotavirus (Estes et al 1981), VP4 is cleaved into two fragments VP5* and VP8* that remain associated with the virion (Arias et al 1996). What is the molecular mechanism of trypsin-enhanced infectivity? One answer to this question could come from locating the two proteolytic fragments in the spike structure. Our recent biochemical and structural studies on the TLPs grown in the presence of trypsin and in the absence of trypsin provided some insights into trypsin-enhanced infectivity in rotavirus (S. E. Crawford, M. K. Estes, S. Mukherjee, A. Shaw, J. A. Lawton, R. F. Ramig & B. V. V. Prasad 2000, unpublished results). Based on these studies we have hypothesized that trypsin, by a mechanism that is not yet clear, imparts stability and order to the spike, and such an ordering may be critical not only during the cell entry but also during virus exit, from infected cells.

During the process ofcell entry, the mature infectious particle loses its outer layer and becomes a DLP (Fig. 1b). Removal of the outer layer exposes the intermediate layer composed of VP6. This layer also has the same T = 13 icosahedral symmetry as the outer layer and consists of 260 trimers of VP6. These trimers are in register radially with the VP7 trimers such that the channels in the VP7 layer continue into this layer. Recently, the atomic resolution structure of the VP6 protein of rotavirus has been determined by X-ray crystallography (F. Rey, personal communication). The VP6 protein has two domains, the distal domain that interacts with VP7 has an eight-stranded antiparallel ^-barrel structure, and the lower domain that interacts with inner VP2 layer is predominantly a-helical. The structure of VP6 bears significant resemblance to the homologous protein (VP7) of blue-tongue virus

(BTV; Grimes et al 1995). Fitting of the X-ray structure of VP6 into the cryo-EM map of the DLP has allowed examination of how this protein adapts to various quasi-equivalent positions of the T = 13 lattice and delineation of the residues that are critical for interactions with VP7, VP4 and VP2.

The outer virion layer consisting of VP7 and VP4 can be removed in vitro by treating the TLPs with 10 mM EDTA. The DLPs thus obtained are transcriptionally active and the dsRNA segments are transcribed continuously in vitro as long as the precursors necessary for transcription are available in the reaction mixture (Cohen 1977, Cohen & Dobos 1979). Several interesting questions can be asked in connection with the endogenous transcription in rotaviruses. Where are the internal proteins VP1 (the RNA-dependent RNA polymerase; Valenzuela et al 1991) and VP3 (the guanylyl transferase; Liu et al 1992) located? Knowing that the DLPs are capable of repeated cycles of transcription, the question is how are the genome segments organized? Where do the mRNA molecules exit from the intact DLPs? Can blocking the RNA exit channels using ligands arrest transcription?

The strategy that we have used to study the internal organization is as follows (Prasad et al 1996). First, the structures of various recombinant virus-like particles (VLPs) (Crawford et al 1994) were compared between themselves and with the native DLPs to deduce the topographical locations of all the internal structural proteins. Then, higher resolution structural analysis was carried out using a medium high-voltage electron microscope to delineate the internal features, particularly of the RNA, in greater detail. Underlying the VP6 layer, VP2 forms the innermost icosahedral layer between the radii 230 Â and 260 Â (Fig. 2). Apart from VP2, no other rotavirus structural protein has the ability to form native-like icosahedral structures. Thus, VP2 performs the function of the size determinant in rotavirus and provides a platform for the correct assembly of VP6. The diameter and the morphological features of the 2/6 and 1/3/2/6-VLPs match very well with that of the native DLPs indicating that the recombinant particles are in all respects native-like. When the structure of the 1/2/3/6-VLP is compared with that of the 2/6-VLP, flower-shaped structural features attached to the inner surface of the VP2 layer are seen at the fivefold vertices. These flower-shaped structures represent VP1/3 complexes. We think that they represent complexes of VP1 and VP3, because these structures are not seen in either 1/2/6-VLPs or 3/2/6-VLPs. Comparison of the native DLP structure with the structure of 1/2/3/6-VLPs indicates that the genomic RNA forms concentric shells of density surrounding the flower-shaped VP1/3 complexes immediately below the VP2 layer (Fig. 2, bottom panel, right).

Rotavirus Dsrna Structure

dsRNA VP11 VP3

-mRNA transcript

FIG. 2. General architectural features of rotavirus. Top panel: (Left) A cutaway representation of the TLP structure. The outer VP7 layer and the intermediate VP6 layer are partially removed to expose the inner VP2 layer. (Middle) The VP2 layer showing the organization of the 120 molecules as 60 dimers (Lawton et al 1997b). (Right) The RNA layers inside the VP2 layer. The VP6 and VP2 layers are 'opened' to expose the RNA core (Prasad et al 1996). Bottom panel: (Left) A side view of an 'isolated' type I channel showing the VP1/VP3 flower-shaped complex attached to the inside tip of the VP2 layer at the fivefold axis. (Right) Proposed exit pathway of the mRNA molecule through the type I channel (Lawton et al 1997a).

In the higher-resolution structure of the DLP, at a slightly higher threshold the RNA density splits into strand-like features, which are about 20 A in diameter and separated by 28 A (Fig. 2, top panel, right). The first layer of RNA immediately below the VP2 layer is icosahedrally ordered. This layer represents about 25% of the genome, or about 5000 base pairs. The remainder of the genome is further inside. The density at the lower radius is not highly reproducible between the independent reconstructions and may not be as well ordered as the first layer of RNA. However, in the near atomic resolution structure of BTV, in which similar features are seen, even the inner layers of RNA are interpreted as being ordered (Grimes et al 1998). The first layer of RNA is well ordered because of its close interactions with VP2, which is icosahedrally ordered and is known to have RNA binding properties (Labbe et al 1994). Since our publication in 1996 (Prasad et al 1996), similar internal structural features including the flower-shaped transcriptase complex and the concentric layers of RNA have been found in several members of Reoviridae such as BTV (Grimes et al 1998), aquareovirus

(Shaw et al 1996, Nason et al 2000), and orthoreovirus (Dryden et al 1998, Reinisch et al 2000).

The structural organization of the VP2 layer is unique (Lawton et al 1997a). The thin shell of VP2 is formed by 60 dimers of VP2 (Fig. 2, top panel, middle). Such an organization, which has been described as a T = 2 structure, appears to be a common feature in the dsRNA viruses whose structures we know to date. Although the atomic resolution structure of VP2 is not yet available, the structure of homologous VP3 in BTV is known from the X-ray crystallographic studies of the BTV cores (Grimes et al 1998). It is quite likely that the structure of the VP2 will be very similar to that of VP3 in BTV with three domains. It appears that this kind of unique organization in all the dsRNA viruses has evolved because of the requirement for endogenous transcription. Such an organization may be critical for the (1) proper positioning of the transcription enzyme complexes, (2) proper organization of the genomic RNA, and (3) proper organization of the outer VP6 layer.

Release of mRNA moleculesfrom intact DLPs

We have used the cryo-EM technique to visualize the exit pathway of the transcripts in actively transcribing DLPs (Lawton et al 1997b). The cryo-EM pictures of the DLPs in the act of transcription clearly show that the structural integrity of the particles is maintained and as many as four mRNA strands are associated with some of the particles. Based on the 3D structural analysis of the actively transcribing particles (Fig. 1c), a model for the exit pathway of the transcript has been proposed (Fig. 2, bottom panel, right). The mRNA molecule that is being synthesized by the transcription complex (flower-shaped structure), exits through the pores in the VP2 layer that are slightly offset from the fivefold axis and gets into the type 1 channel. Similar cryo-EM analysis on transcribing orthoreovirus particles has shown that channels through the turrets at the icosahedral fivefold axes are used for the extrusion of nascent transcripts (M. Yeager, personal communication). It is quite likely that the exit pathway for the mRNA involves fivefold vertices in other members of the Reoviridae.

Although the structural organization of the genome in dsRNA viruses is far from being clear, the existing data appear to support a model in which each genome segment is spooled around a transcription complex. In such a model, consistent with the biochemical data, the segments can be simultaneously and repeatedly transcribed. With such a genomic organization, each dsRNA segment has to move around the transcription-enzyme complex that is anchored to the VP2 layer. What drives such a movement of the template? Are the ends of the dsRNA molecule somehow tethered? What are the interactions that are required in the genomic organization that facilitate such a coordinated movement of the template around the enzyme complexes? Further studies are required to answer these questions. Our recent structural studies to probe further the genomic organization by varying pH and ionic concentrations have shown that hydrogen bond interactions between internal proteins and RNA, and the electrostatic interactions between the RNA strands, are important for proper genome organization (B. Pesavento, J. Lawton, M. K. Estes, B. V. V. Prasad 2000, unpublished results).

Inhibition of transcription by ligands that interact with VP6

Is it possible to arrest transcription by blocking the mRNA exit channels, and can this be exploited to become an antiviral strategy? This question is particularly relevant because it has been shown that 2/6-VLPs can confer protective immunity in mice (O'Neal et al 1998), and also that anti-VP6 IgA molecules can protect mice from rotavirus infection (Burns et al 1996). Can anti-VP6 MAbs inhibit transcription? Earlier biochemical studies with anti-VP6 MAbs showed that some MAbs completely inhibited transcription, while others did not (Ginn et al 1992). A natural ligand to VP6 is VP7. Binding of VP7 to VP6 makes the DLPs transcriptionally incompetent. Why are the TLPs transcriptionally inactive?

Our structural studies have shown that certain ligands (2A11 Fab and VP7) narrow the mRNA exit channels and also cause slight conformational changes in VP6 constricting the exit channels in the interior regions (Lawton et al 1999). A relevant question is whether the narrowing ofthe channel opening and/or the small but consistent conformational changes in the interior of the exit channels are the reasons for transcriptional inhibition. The process oftranscription can be described as having three stages: initiation, elongation and translocation. To determine at which stage transcription is inhibited we have analysed high-resolution RNA gels of transcription reaction mixtures containing transcriptionally competent (DLPs and 8H2 Fab-bound DLPs) and transcriptionally incompetent (TLPs and 2A11 Fab-bound DLPs) particles. These studies have shown that the initiation of transcription is not affected in any of these particles. A consistent feature in all the cases is the presence of a band that corresponds to transcripts of about 6—7 nucleotides in length. Our recent studies have shown that this represents a pause site, and in the case of DLPs the transcription continues past this pause site in an efficient manner to yield full-length transcripts (J. A. Lawton, M. K. Estes & B. V. V Prasad 2000, unpublished results). In the case of 2A11-bound DLPs, only a small number of transcripts of about 70—80 nucleotides in length are made. In TLPs no other transcripts are made past this pause site. It is quite possible that the narrowing of the channel opening is the reason for the transcriptional inhibition in the case of 2A11-bound DLPs. However that is certainly not the reason in the case of TLPs. It is possible that there is a subtle conformational switch that is on when the VP7 layer is removed and off when it is intact. Higher resolution structural analysis may provide some insight into why TLPs are transcriptionally incompetent.

How To Cure Yeast Infection

How To Cure Yeast Infection

Now if this is what you want, you’ve made a great decision to get and read this book. “How To Cure Yeast Infection” is a practical book that will open your eyes to the facts about yeast infection and educate you on how you can calmly test (diagnose) and treat yeast infection at home.

Get My Free Ebook

Post a comment