162060/FULLTEXT01.pdf · causes translational repression. Structures of multiple RNA segments with altered sequence at some positions which are required for binding to the capsid

Control of Quasi-equivalence in Virus Capsids

BY

CHARLOTTE HELGSTRAND

Dissertation for the Degree of Doctor of Philosophy in Molecular Biotechnologypresented at Uppsala University in 2002

ABSTRACT

Helgstrand, C. 2002. Control of quasi-equivalence in virus capsids. ActaUniversitatis Upsaliensis. Comprehensive Summaries of Uppsala Dissertations fromthe Faculty of Science and Technology 760. 62 pp. Uppsala. ISBN 91-554-5431-3.

The geometric principles underlying the construction of spherical virus capsid donot allow more than 60 protein monomers to from a capsid while maintaining anidentical chemical environment. Most virus capsid, however, contain many moreprotein subunits. Quasi-equivalence explains how the capsid proteins can haveslightly different interactions in the virus shell. Quasi-equivalence requires thecapsids to be constructed from multiples of 60 subunits, where the T numberdenotes the multiplicity. Many T=3 plant and insect viruses use a molecular switch in form oforder/disorder of a segment of the polypeptide chain to regulate the quasi-equivalentcontacts. The structure of a mutant of the T=3 capsid of bacteriophage fr confirmsthat this virus and other members of the Leviviridae family lack a switchmechanism. The structure of the T=4 Nudaurelia capensis Virus shows a molecular switchin form of a C-terminal helix inserted in some contacts between protein dimers. Thisvirus is very similar in structure to the T=3 nodaviruses. In the nodaviruses a five-membered helix bundle, formed by cleaved peptides around the five-fold axes onthe inside of the shell, are suggested to aid in membrane translocation of thegenomic RNA. In Nudaurelia capensis Virus the helix bundle is formed by 10helices, of which 5 are still covalently attached to the capsid proteins. Bacteriophage HK97 has T=7 quasi-symmetry. A domain that is degraded duringmaturation and is not present in the structure of the mature virion controls the quasi-equivalence. During maturation covalent bonds are formed between the proteinsubunits, producing a set of interlocking covalently bound rings, resemblingchainmail. Structural studies of complexes between the bacteriophage MS2 and variants ofits translational operator are also included in this work. A dimer of the MS2 coatprotein binds with sequence specificity to an operator in its genomic RNA, andcauses translational repression. Structures of multiple RNA segments with alteredsequence at some positions which are required for binding to the capsid protein, hasbeen determined.

Charlotte Helgstrand, Department of Cell and Molecular Biology, Box 596, BMC,75124, Uppsala, Sweden

Vi börjar ana att vår vilsegångär ännu djupare än först vi trottatt kunskap är en blå naivitetsom ur ett tillmätt mått av tankesynfått den idén att Gåtan har struktur.

Harry Martinsonur Aniara

PAPERS INCLUDED IN THE THESIS

This thesis is based on the following papers, which are referred to in the textby their Roman numerals:

I Axblom Charlotte, Tars Kaspars, Fridborg Kerstin, Orna Ligita,Bundule Maija, Liljas Lars. (1998). Structure of phage fr capsids with adeletion in the FG loop: Implications for viral assembly. Virology 249:80-88

II Grahn Elin, Moss Timothy, Helgstrand Charlotte, Fridborg Kerstin,Sundaram Mallikarjun, Tars Kaspars, Lago Hugo, Stonehouse Nicola J.,Davies Darrel R., Stockley Peter G., Liljas Lars. (2001). Structural basisof pyrimidine specificity in the MS2 RNA hairpin-coat-protein complex.RNA 7: 1616-1627

III Helgstrand Charlotte, Grahn Elin, Moss Timothy, Stonehouse Nicola J.,Tars Kaspars, Stockley Peter G., Liljas Lars. (2002). Investigating thestructural basis of purine specificity in the structures of MS2 coat proteinRNA translational operator hairpins. Nucleic Acids Research 30: 2678-2685

IV Helgstrand Charlotte, Wikoff William, Duda Robert, Hendrix Roger,Johnson John E., Liljas Lars. The refined structure of a molecular ballon;the HK97 bacteriophage capsid at 3.4Å resolution. (Manuscript)

V Helgstrand Charlotte, Munshi Sanjeev, Johnson John E., Liljas Lars. Therefined structure of the insect virus Nudaurelia capensis Virus.(Manuscript)

Reprints were made with permission from the copyright holders.

Contents

1 Introduction ..............................................................................................91.1 Viruses – general ................................................................................91.2 Classification of viruses .....................................................................91.3 Viral architecture ..............................................................................10

2 Background.............................................................................................112.1 Multifunctionality of viral structural proteins ..................................112.2 Protein-protein interactions within the virus shell............................11

2.2.1 Quasi-symmetry ......................................................................112.2.2 Deviations from quasi-symmetry ............................................13

2.3 Assembly control..............................................................................152.4 Maturation ........................................................................................152.5 Interactions with the host cell...........................................................16

2.5.1 Antibody interactions ..............................................................162.5.2 Receptor interactions ...............................................................172.5.3 Translocation of RNA .............................................................18

2.6 Interactions with the nucleic acid .....................................................192.7 Enzymatic activity ............................................................................20

3 Structure determination methods............................................................213.1 Phasing .............................................................................................213.2 Phase improvement by density averaging ........................................213.3 Methods used in this work................................................................22

3.3.1 The structure of frs5 (Paper I) .................................................223.3.2 Complexes between MS2 and variants of its translationaloperator (Papers II-III) .....................................................................223.3.3 Rebuilding and refinement of the HK97 and Nudaureliacapensis Virus structures...............................................................23

4 Assembly control in the small RNA coliphages (Paper I) ......................244.1 Introduction ......................................................................................244.2 Results ..............................................................................................264.3 Overall conclusions - phage assembly..............................................28

5 Interactions of phage MS2 with its translational operator (Paper II-III) 295.1 Introduction ......................................................................................295.2 Results, Paper II................................................................................315.3 Results, Paper III ..............................................................................355.4 Overall conclusions - RNA binding .................................................38

6 The structure of HK97 - a covalently linked virus shell (Paper IV).......406.1 Introduction ......................................................................................406.2 Results ..............................................................................................41

7 The refined structure of Nudaurelia capensis Virus (Paper V) ..........457.1 Introduction ......................................................................................457.2 Results ..............................................................................................47

8 Acknowledgements.................................................................................52

9 References ..............................................................................................54

Abbreviations

DNA deoxyribonucleic acidRNA ribonucleic acidTBSV Tomato Bushy Stunt VirusSBMV Southern Bean Mosaic VirusRYMV Rice Yellow Mottle VirusSV40 Simian Virus 40BTV Bluetongue VirusICAM-1 intracellular adhesion molecule 1VP4 viral protein 4VP1 viral protein 1FHV Flockhouse VirusBBV Black Beetle VirusTR translational operatorMIR multiple isomorphous replacementMR molecular replacementEM electron microscopyE. coli Echerishia coliUV ultravioletPEG polyethylene glycolHPLC high performance liquid chromatographyA adenineU uracilC cytosineG guaninePy pyrimidinePu purine2one pyrimidine-2-oneBrU bromouracile4one pyridine-4-oneSU 2-thiouracilI inosine2ap 2'-deoxy-2-aminopurine

N V Nudaurelia capensis VirusIg Immunoglobulin

9

1 Introduction

1.1 Viruses – general

We all have daily contact with viruses. Some may cause disease and evendeath. Others are mostly harmless. Viruses that infect livestock and cropscause huge economic losses. There are so far few cures for viral infections.Vaccination is an efficient method, but very costly since a separate vaccinefor each virus has to be developed and administered. My reason to studyviruses is to learn more about them, in every aspect of their life cycle. Thisdoes not necessarily need to be done for viruses that cause human disease.By studying many different types and exploring the variability of theseorganisms we gain a deeper understanding of how viruses function.

1.2 Classification of viruses

Viruses can be classified according to their host specificity as animalviruses, plant viruses or bacterial viruses (phages). Since there is generally awide variety of viruses infecting each host type this classification is notsufficient for a detailed analysis. The most common way to classify virusesis by the type of nucleic acid their genome is composed of. Viruses mayhave either DNA or RNA as genetic material, and both DNA and RNA maybe either single stranded or double stranded. By comparing morphology,infection mechanism, protein composition and genetic similarities theviruses can be further subdivided into orders, families and genera. Theconcept of a species is not really the same in viruses as for other organisms,since viruses evolve not only by mutational divergence, but also byexchanging large parts of their genomes with each other. It is not unusual forviruses in the same family to have varying host specificities. An extremeexample is the reoviruses; different viruses in this family infect plants,animals and possibly also yeast.

10

1.3 Viral architecture

Viruses cannot replicate by themselves; they need a host cell to provide thebuilding blocks and sometimes the necessary enzymes for the replication andprotein synthesis. The key to replication is the nucleic acid, but that may bevulnerable to degradation before infection into a host cell can occur. Thegenome is therefore protected by a protein shell that surrounds it. This shellis commonly one of two types, either rod-shaped or spherical. In additionmany viruses have a lipid membrane, envelope, derived from the host cell.Inside the protein shell there may also be additional proteins required forinfection, replication and protein synthesis. In this thesis non-envelopedspherical viruses have been studied.

11

2 Background

2.1 Multifunctionality of viral structural proteins

The proteins that build up the virus shell, or capsid, are called structuralproteins or capsid proteins. Their primary role is to protect the nucleic acid,but the nature of the viral lifecycle also requires several other interactions.The aim of this chapter is to give some understanding of the variousfunctions performed by viral structural proteins and the complexity that canbe encoded into a single stretch of polypeptide. The viruses discussed hereare only a few examples; detailed information is available for many more.

2.2 Protein-protein interactions within the virus shell

2.2.1 Quasi-symmetry

Since a nucleic acid molecule can not encode a single protein that is largeenough to enclose it, virus capsids must be constructed by multiple copies ofone or a few proteins. The lowest-energy assemblies are those conforming toa geometrical shape. The largest closed body to which identical units can bearranged is an icosahedron (Figure 1). An icosahedron has axes of 2-, 3- and5-fold symmetry. The icosahedron has 60 identical positions, which meansthat 60 would be the maximum possible number of proteins with identicalinteractions building up a virus capsid. However, many viruses contain manymore than 60 capsid protein subunits, often several hundred. The theory ofquasi-equivalence (1) was presented as a way of explaining how more than60 protein subunits could be included into the shell, by allowing the protein

12

subunits to have different interactions and different chemical environment.The theory allows for certain multiples of 60 subunits, indicated by thetriangulation (T) number, to form quasi-equivalent particles. There are onlycertain T numbers that are allowed, following the rule T=H²+HK+K², whereH and K are integers. High-resolution structures of virus particles withtriangulation numbers of 1, 3, 4, 7 and 13 are available. Many viruses withhigher T numbers, up to T=169 (2), are known. In particles with T>1 anumber of Quasi-symmetry axes are also introduced, which may or may notcoincide with the icosahedral symmetry axes.

Figure 1. Left: An icosahedron with the position of 5-, 3- and 2 fold axes indicated.Right: An icosahedron with the location of quasi-symmetric subunits A, B and C ina T=3 virus capsid. Figures 1b, 3a and 3b are from VIPER(http://mmtsb.scripps.edu/viper/viper.html).

An example of quasi-symmetry is the T=3 RNA plant viruses (TBSV,SBMV, etc.) (3,4). The Rice Yellow Mottle Virus (RYMV) capsid proteinhas a 238-residue chain with a typical viral jellyroll fold (5). The shell isconstructed by 180 capsid proteins. The protein exists in three quasi-equivalent conformations, denoted A, B and C (Figure 2). Together one ofeach A, B and C chain forms one icosahedral asymmetric unit. The maindifference between the quasi-equivalent conformations is the order/disorderof an N-terminal segment of the capsid protein in some protein interfaces. Atthe icosahedral fivefold axis, five A subunits interact in a "bent"conformation. At the quasi-sixfold axis (icosahedral threefold axis), B and Csubunits interact. In the B/C interface, an N-terminal segment from the Csubunit is inserted in a cleft between the two proteins, forcing them tointeract in a "flat" conformation (Figure 2). In this particular plant virus, the

13

N-terminal arm from the C subunit is not inserted in the interface of the Csubunit it is part of, instead it inserted in the interface of the neighbouring Csubunit. This "domain swapping" leads to the formation of an internalnetwork of the N-terminal strands (Figure 2).

Figure 2. Top left: A bent subunit contact between two A subunits in RYMV. Topright: A flat subunit contact between the B and the C subunits in RYMV. Bottom:The network of N-terminal strands on the inside of the capsid in RYMV.

2.2.2 Deviations from quasi-symmetry

There are examples of virus shells that do not conform to the quasi-equivalence theory. The capsids of polyoma virus and Simian Virus 40(SV40) with their 360 protein monomers would have a triangulation numberof 6. The structures shows (6,7) that the capsids are formed by pentamersonly, occupying the positions of both pentamers and hexamers in a capsid

14

corresponding to T=7 symmetry (Figure 3). The inner shell of the core ofBluetonuge Virus have 120 protein monomers (8), made possible by the highplasticity of the protein monomer, which allows it to form two differentlyshaped triangles (Figure 4). It is evident that protein molecules are flexibleenough to result in low-energy structures even if the monomers havesignificantly different interactions.

Figure 3. Left: The arrangement of SV40 pentamers in a pseudo T=7 organisation.Right: The T=7 arrangement of HK97 hexamers and pentamers.

Figure 4. The two different conformations adopted by the protein that builds up theBTV core inner shell.

15

2.3 Assembly control

The problem of assembling hundreds of protein subunits into a shell ofpredefined size puts special demands on viral capsid proteins. The proteinsmust form a capsid of the correct size to accommodate the genome and,ideally, the capsid formation should be fast. The shell may assemble fromprotein monomers, but higher order assembly intermediates, like dimers,trimers, pentamers and hexamers, are common. In viruses with T numbersgreater then one, the protein subunits must form different interactions whenadded to the capsid, to be able to define the correct curvature. There arevarious solutions to this problem. Small viruses tend to have specialsegments that become either ordered/disordered or in discreteconformational states during the assembly of the capsid (3,4,9-11). Thesesegments act as switches to produce different quasi-equivalent states of theproteins in the growing capsid (Figure 2). If these segments are altered ordeleted, aberrant assembly forms are often produced (12,13). In this case allnecessary information is provided in the amino acid sequence of the capsidprotein.

Some large viruses, like dsDNA phages, herpesvirus or adenovirus, needthe extra help of special scaffolding proteins to produce the correct sizedparticles. There are examples of both internal and external scaffoldingproteins (14). Internal scaffolding proteins do not need to form a symmetricstructure or have specific interactions with the capsid protein; providing acore of an appropriate size onto which the capsid protein can assemble isapparently sufficient for formation of the correct shell (14). In contrast,external protein scaffolds display icosahedral symmetry (15-17).

2.4 Maturation

Several types of viruses undergo a process called maturation, in which theassembled virus capsid has to be altered to produce an infectious virion.Maturation may involve proteolytic cleavage, ordering of protein segments,release of scaffolding proteins and size changes of the particles (swelling orshrinking). Classic examples of maturation are the large dsDNAbacteriophages (P22, T4, T7, , etc.) and herpesvirus. In some virusesmaturation involves large rearrangements of the protein components in the

16

capsid (18-23), which implies that the capsid proteins must have severalalternative interaction surfaces.

The large dsDNA bacteriophages in the siphoviridae family are complexstructures, with a large icosahedral head consisting of several hundred capsidprotein molecules. They have a connector/portal replacing one pentamer inthe head through which DNA is packed, and a flexible tail used to injectDNA into the host cell. The typical assembly and maturation pathway forthese phages starts with the formation of a procapsid (24). In addition to thecapsid proteins, the procapsid contain hundreds of scaffolding proteins (25),which are responsible for guiding the assembly into a correctly formed head.Additional minor proteins may also be present, and there is a portal structurereplacing one of the vertices in the head. The next step in maturation is therelease of the scaffolding proteins, a process which involves proteolyticcleavage since the scaffolding proteins are positioned inside the head andcannot escape without cleavage (25-27). DNA is packed through the portaland simultaneously results in irreversible expansion of the particle. The tailis assembled separately and is attached to the portal complex after DNApacking.

2.5 Interactions with the host cell

To be able to replicate, the viral nucleic acid must enter the host cell. Inmany cases it is the capsid proteins that contacts the cell and mediates theentry of the genetic material. Animal viruses, as well as somebacteriophages, use cell receptors for attachment to the host cell. Anotherentry pathway for phages is the attachment to the bacterial flagella or pili,after which the particles are drawn into the cell. Plant viruses do not appearto have any direct interactions with their host cells; instead they usuallyinfect by "passive entry" through a wound in the cell wall. Animals, on theother hand, have the immune system as a defence against foreign particles,including viruses.

2.5.1 Antibody interactions

Antibodies react with the most exposed parts of the virions, which in non-enveloped viruses are the capsid proteins. Escaping antibodies, whilemaintaining the ability to bind to the receptor, is therefore an important

17

property of capsid proteins. The primary means by which these virusesescape antibodies is by having a large sequence variability of the exposedparts - hypervariable regions.

2.5.2 Receptor interactions

Animal viruses, as well as some bacteriophages, use cell receptors forattachment to the host cell. The receptor can be almost any kind of moleculedisplayed on the cell surface, proteins, glycolipids and carbohydrates. Thedetailed interactions of a few picornaviruses (poliovirus, coxsackievirus andrhinovirus), with their receptors have been thoroughly studied (28-34). Thehuman rhinoviruses can be subdivided into three groups according to theirreceptor specificity, with the major group using the intracellular adhesionmolecule-1 (ICAM-1). There are several structures of rhinoviruses incomplex with their receptor fragments (28,29,32). All structures of differentrhinovirus types have showed a groove, or canyon, around the 5-fold axis.This region has unusually conserved amino acid sequence, possibly becauseit is too narrow for antibodies to penetrate. The ICAM-1 N-terminalimmunoglobulin domain has been shown to bind in this canyon, an elegantway of providing a conserved receptor-binding site that is protected fromantibody neutralisation (Figure 5).

Figure 5. The ICAM-1 receptor binds in a shallow groove on the HRV surface.Antibodies are too large to penetrate the canyon. Adapted from (110).

18

2.5.3 Translocation of RNA

Receptor interactions often lead to endocytosis of the virus particle. Theparticles can become destabilised by the low pH in the endosome. In order toenter the cytoplasm, the genetic material has to be translocated across theendosomal membrane (or the cell membrane if there is no endocytosis). Inenveloped viruses, fusion of the viral and cellular membranes allows entry ofthe particle or its genetic material into the cytoplasm. Membrane-integratedviral proteins mediate the membrane fusion. Non-enveloped viruses mustfind another way to cross the membrane. In picornaviruses the internalprotein VP4 is released upon endocytosis and the N-terminal part of VP1 isexposed, mediating translocation through the membrane (35). In thestructures of two T=3 nodaviruses, Flockhouse Virus (FHV) and BlackBeetle Virus (BBV), another mode of RNA translocation have been found(36). During particle assembly, an autoproteolytic cleavage detaches apeptide from each monomer in the capsid. At the five-fold axes, just insidethe viral surface, these peptides form a five-membered helix bundle (Figure6) (9,10). There are indications that this bundle becomes externalised andassociated with the membrane during endocytosis and can aid in transferringthe RNA across the membrane (36,37).

Figure 6. The arrangement of C-terminal helices in a cylinder shape around thefivefold axis in BBV. All 5 C-terminal helices are shown (black). For clarity onlytwo full A subunits are shown (grey).

19

2.6 Interactions with the nucleic acid

In order to form infectious particles, many virus capsid proteins have toassemble around their genome. This means that some kind of recognitionmust take place between the nucleic acid and the capsid protein to ensurethat the correct DNA or RNA is packed.

The most well studied interaction between a capsid protein and itsgenome is that of the T=3 coliphage MS2. Binding of a capsid protein dimerto a copy of the RNA genome acts as the nucleation point for capsidassembly (38-41). The protein dimer binds specifically to a 19-nucleotidehairpin structure located at the beginning of the viral replicase gene. Thebinding causes translational repression of the replicase, and therefore thehairpin has been named the translational operator (TR) (42). Binding assayshas provided a consensus for the sequence-specific interactions (43).Numerous structures are now available of complexes of the MS2 capsid andshort RNA fragments (44-51), showing how the nucleic acid interacts withthe capsid protein (Figure 7).

Figure 7. Left: The sequence and secondary structure of MS2 TR. The numbering isrelative to the start of the replicase gene. Right: The MS2 TR bound to a dimer ofMS2 coat protein (side view).

-5 U U A A G-C G-C-10 A G-C U-A +1 A-U C-G A-U

5' 3'

20

2.7 Enzymatic activity

There are several examples of virus capsid proteins catalysing chemicalreactions. The most spectacular is the capsid protein of Sindbis Virus andSemliki Forest Virus, which are chymotrypsin-like serine proteases (52-54).The capsid protein acts in cis to cleave itself from the translated polyprotein.Autoproteolytic cleavage of the capsid protein during maturation of virusparticles are common, examples are picornaviruses, nodaviruses andNudaurelia capensis Virus. There are also examples of reactions otherthan proteolysis. The capsid proteins of the -like bacteriophage HK97 formcovalent bonds throughout the virus shell as the final step of maturation(55,56). The reaction is catalysed by the capsid protein itself.

21

3 Structure determination methods

All the structures in this thesis were determined by x-ray crystallography.There are a few things about structure determination of viruses that differsfrom ordinary protein crystallography. The size of the unit cell is often verylarge, since it may contain one or more whole virus capsid. If the virus shellis "empty", the solvent content of the crystal is usually high. This leads toweak data and sometimes to low resolution.

3.1 Phasing

The first high-resolution virus structures were solved with MultipleIsomorphous Replacement (MIR). Recently though, Molecular Replacement(MR) has become the most favoured choice for phasing virus data. If a high-resolution structure of a related virus is available, the molecular replacementmay work at high resolution. An alternative starting model for molecularreplacement is cryo-EM reconstructions at relatively low resolution (57-60)or even hollow shells of electron density - ab initio phasing (61). The phasesobtained at low resolution must then be extended to high resolution by phaseextension.

3.2 Phase improvement by density averaging

Virus particles have 60 identical units in its shell, imposed by the icosahedralsymmetry. This knowledge allows virus crystallographers to use electrondensity averaging over the non-crystallographically related asymmetric unitsto improve the phases. The procedure is performed at a defined resolutionand in real space. An average is calculated for the electron densities in all thenon-crystallographic asymmetric units. This averaged density is then usedfor Fourier back-transformation to yield improved phases that can be

22

combined with the observed structure factor amplitudes. The procedure isrepeated cyclically until no further improvement is seen.

3.3 Methods used in this work

3.3.1 The structure of frs5 (Paper I)

The deletion mutant frs5 was recombinantly overexpressed in E. coli and thecapsids were purified by size-exclusion chromatography. Temperaturestabilities of frs5 and fr wildtype capsids were determined by heatingaliquots of the capsids to various temperatures in 20 mM Tris-HCl, pH 8.0,for 15 minutes. The aliquots were run on an Ethidium Bromide-containingagarose gel and the RNA associated with the capsids visualised by UV-radiation.Crystals of the purified capsid protein were grown using hanging dropvapour diffusion with 0.1 M NaCl, 0.1 M Bicine and 30% w/v polyethyleneglycol (PEG) monomethylether 550 (Hampton Crystal Screen II, solution46). Crystals of 0.5-1.0 mm grew in one week at room temperature. Datawere collected using synchrotron radiation to 3.5 Å. The data were processedusing DENZO (62) and scaled using Scalepack (62). The crystals hadspacegroup R32 with cell dimensions a=b=264.1 Å and c=654.2 Å in thehexagonal setting.

The structure was determined by molecular replacement using XPLOR(63) with the refined structure of recombinant wild type fr capsid (64) as astarting model. Electron density averaging using 10-fold non-crystallographic symmetry was performed using AVE (65) to produce thefinal 2Fo-Fc maps. The fr wildtype model was rebuilt according to the frs5sequence. Positional refinement and temperature factor refinement wasperformed using XPLOR (63).

3.3.2 Complexes between MS2 and variants of its translationaloperator (Papers II-III)

MS2 capsid protein was recombinantly overexpressed in E. coli and thecapsids were purified by size-exclusion chromatography. Crystals of thepurified capsid protein were grown using hanging drop vapour diffusionwith 0.3-0.4 M Na-phosphate buffer, pH 7.4, as well solution. The initialdrops were 10 µl and contained 3.7 mg/ml virus capsids and 0.5-2.0% PEG

23

with a molecular weight of 8000 in 0.05 M NaPO4 buffer, pH 7.4. Crystalsgrew after several weeks in 30° or 37°. The crystals had spacegroup R32with cell dimensions a=b=288.0 Å and c=654.0 Å in the hexagonal setting.

The RNA molecules were produced by solid-phase synthesis and purifiedto homogeneity by HPLC. Some RNA molecules were purchased from DNATechnology A/S, Denmark. Crystals larger than 0.7 Å were used for soakingwith RNA. After transfer to room temperature the crystals were washed with5% PEG in 0.4 M Na-phosphate buffer, pH 7.4. RNA molecules in solutionwere added to a final concentration of 4.5 mg/ml. After about one week ofsoaking, the crystals were mounted in glass capillaries and used for datacollection.

All diffraction data were collected using synchrotron radiation at 5° C.Several positions on each crystal and several crystals were used for eachMS2-RNA complex. The data were processed using DENZO (62) and scaledusing Scalepack (62).

The structures were determined by molecular replacement, using thecoordinates for an empty MS2 particle as initial model. The MR was doneusing XPLOR (63) or CNS (66). Electron density averaging using 10-foldnon-crystallographic symmetry was performed using AVE (65) to producethe final 2Fo-Fc maps.

The models were built in O (67) using coordinates of an existing MS2-RNA complex (44). Positional refinement and temperature factor refinementwas performed using XPLOR (63) or CNS (66).

3.3.3 Rebuilding and refinement of the HK97 and Nudaureliacapensis Virus structures

The starting models were rebuilt in O (67) and positional and temperaturefactor refinement was performed using CNS (66).

24

4 Assembly control in the small RNAcoliphages (Paper I)

4.1 Introduction

Bacteriophages are viruses that infect bacteria. Phages that specifically infectE. coli cells are called coliphages. The small RNA phages belonging to thefamily of Leviviridae were first characterised by Loeb and Zinder (68), whopurified them from sewage in New York. The coliphage MS2 and its closerelatives have since been used as model systems for studying virus assemblyand protein-RNA interactions (41,43-51,69-81). The protein shell structureof MS2 and the related phages fr, Q , GA and PP7 are known (64,82-86).Bacteriophage fr has a T=3 icosahedral protein shell surrounding a singlestranded RNA genome of 3575 nucleotides. The outer diameter of theparticle is 280 Å. These phages infect cells by attachment to bacterial piliand the assembled particle also contains one copy of the A protein(maturation protein) which is believed to contact the host cell pilus duringinfection (87,88). The genetics of phage fr and other Leviviruses are verysimple, with only four proteins encoded: the coat protein, the A protein, asubunit of the viral replicase and a lysis protein. fr has a high genomicidentity to MS2, 77% (89), which for the coat protein means that only 17 outof 129 amino acids are different. The crystallographic structure of fr is verysimilar to that of MS2 (64).

The protein shell is composed of 180 copies of a 129 amino acid capsidprotein. The monomers associate to dimers immediately after synthesis andthese spontaneously assemble into capsids even at low concentrations. Themonomer fold is a simple 5-stranded antiparallel -sheet, with 2 helices atthe C-terminus and an additional -hairpin at the N-terminus (Figure 8).When dimers are formed, the C-terminal helices from each subunit becomeinterlocked and a continuous 10-stranded -sheet is formed. The T=3packing means that there will be 3 slightly different, but distinct quasi-

25

equivalent conformations with slightly different interactions in the shell.These are denoted A, B and C (Figure 8) and are arranged in the shell as A/Bdimers and C/C dimers. The three different conformations are remarkablysimilar, the only difference being a bent conformation of the loop betweenstrand F and G (FG loop) in the B conformation.

Figure 8. A trimer of dimers of MS2 coat protein. The three chains in theasymmetric unit, A, B and C are shown, together with their dimer partners B', A'and C'.

Assembly of a virus particle with more than 60 proteins in the shell requiresthe proteins to be added to the shell in the correct conformation, or the shellwill not get the correct curvature. fr/MS2 coat protein lacks any obvious"switch" region like the one seen for a number of plant and insect viruses(TBSV, SBMV, BBV etc.), where a segment of the N-terminus is ordered insome subunits and disordered on other, producing distinct differences in theinterfaces between the protein subunits (3,4,9-11). The FG loop interacts atthe five- and quasi-sixfold axes, in an extended conformation at the A and Csubunits, while the B subunit FG loop is bent. When the structure was firstsolved, it was revealed that the FG loop had a cis-proline at residue 68 insubunit B, but not in A and C, and it was suggested that this isomerisationwas used as a switch mechanism (84). Further mutational and structuralstudies of MS2, however, revealed that the proline could be replaced with anAsn (90) and correct capsids still formed. The structure of this mutant

26

showed that AsnB78 had a normal trans conformation (91), but a full-lengthclone of with this mutation failed to produce viable phages (92).

4.2 Results

The mutant frs5 (del 70-73) was one in a series of deletion and substitutionmutants made to explore the importance of various regions for capsidassembly (93). In frs5 the putative switch region had been deleted, butcapsids of the correct size still formed, albeit with a much larger fraction ofunassembled dimers present. Purified frs5 capsids were crystallised and thestructure was solved at 3.5 Å resolution. The fold is very similar to wild-typefr and MS2 and the fr wildtype model was easily rebuilt and refined. In thewild-type structure the FG loops at both symmetry axes are held together byhydrogen bonds. In the mutant structure, the FG loops at the q6-fold axishave no hydrogen bonds since the distance between the chains is too long(Figure 9). This also causes high temperature factors. At the 5-fold axis theFG loops were disordered and could not be modelled, probably because thelack of hydrogen bonds (Figure 9). We were able to conclude that the FGloop could not be considered a switch for assembly regulation and, as far asthere are no collisions, they may have certain flexibility. The absence ofhydrogen bonds between the FG loops leads to a lower temperature stability,as was shown by heating both fr wildtype and frs5 capsids to differenttemperatures and detecting the "leakage" of RNA from the particles on anagarose gel.

27

Figure 9. Top: The structure of wildtype fr coat protein, with interactions aroundthe five- and quasi-sixfold axes. Bottom: The structure of the frs5 mutant. The loopsare disordered around the fivefold axes and are considerably shorter than wildtypearound the quasi-sixfold axes.

28

4.3 Overall conclusions - phage assembly

Since there is no obvious molecular switching mechanism in the coatproteins of the small RNA coliphages, the assembly must be regulated bymore subtle mechanisms. It was proposed (85) that the assembly mightinstead be regulated by the interaction surfaces of the assembling dimers.Only one polar contact in the dimer-dimer interface is conserved when allknown structures of the small coliphages are compared. This is a hydrogenbond between the mainchain oxygen of residue 39 and the mainchainnitrogen of residue 94. All three quasi-equivalent subunits form thisinteraction and it might be enough to define the correct curvature of thegrowing shell. If the assembly is initiated by a trimer of dimers(AB+AB+CC) ((93), Vijay Reddy, personal communication), the O39 - N94interaction would be in the centre of the complex, and all interactions of theT=3 capsid would be defined. Additional dimers could then be added to thepreformed quasi-equivalent surfaces.

29

5 Interactions of phage MS2 with itstranslational operator (Paper II-III)

5.1 Introduction

In addition to its role as shell protein, MS2 coat protein also acts atranslational repressor of the replicase gene (42). A 19-nucleotide longhairpin, named the translational operator (TR), in the beginning of thereplicase gene has been identified as the sequence binding to the coatprotein. The hairpin consists of a seven-basepair stem, with a bulgednucleotide between basepairs five and six, and is closed by a four-nucleotideloop. Multiple binding studies of MS2 coat protein and variant RNAsegments in solution (43,69-81,94) have produced a consensus for thebinding requirements (Figure 10). The bases that interact with the coatprotein in a sequence specific way are -10, -7, -5 and -4, where position -4seems to be the most strict; no binding can be detected for bases other thanAdenine (A) in this position.

Figure 10. The translational operator (TR) (left) and the consensus for binding toMS2 coat protein (right).

-5 -5 U U N Py A A -7 A A -4 G-C N-N' G-C N-N'-10 A -10 Pu G-C N-N' U-A +1 N-N' A-U N-N' C-G N-N' A-U 5' 3'

5' 3'

30

The starting point for virus assembly in vivo is believed to be a coat proteindimer bound to one molecule of genomic RNA (38-41). Recombinantexpression of coat protein results in the quick assembly into "empty"capsids, but a significant amount of cellular RNA is also incorporated. Thepresence of viral RNA, or the TR, stimulates capsid formation at lowerprotein concentrations (41), suggesting that there are specific interactionsbetween the TR and the coat protein that stimulates assembly. It has beenproposed that the binding of RNA to the coat protein could induce aconformational change in the dimer that could act as an assembly initiationcomplex (95). The most likely change that could affect assembly is aconformational change of the FG loop. The x-ray structure of the MS2 TR incomplex with wildtype MS2 coat protein does not show any obviousmechanism of linking the binding of TR to a conformational change, sincethe TR binding site is distant from the FG loop.

X-ray crystallography has also been used to study the detailedinteractions of MS2 and the TR (44-51) (Figure 11). Recombinant MS2capsids, containing no genomic RNA, can be crystallised, and we have beenable to soak in short synthetic fragments of RNA, containing variations ofthe MS2 TR. The RNA fragments diffuse into the MS2 capsids, probablythrough pores at the quasi-sixfold axes, and bind to the dimers in the capsid.There are thus 90 RNA binding sites in the capsid. This allows for the use ofnon-crystallographic symmetry in the structure determination. The RNAmolecules bind both to A/B dimers and to C/C dimers, but at the symmetricC/C dimer the RNA can bind either way and appear twofold disordered.Papers II-III are devoted to structural studies of RNA variants with differentbases at positions -5, -7 and -10.

31

Figure 11. The binding of the MS2 TR to the coat protein.

5.2 Results, Paper II

Biochemical work has pointed out the base at position -5 in the TR loop asone of the critical positions for TR binding. According to extensive RNA-coat protein affinity measurements, pyrimidines are strongly preferred at thisposition, with purine affinities at least a 100-fold lower than for purines (69).Four factors appear to influence the binding affinity of the -5 base. The baseinteracts with the coat protein by stacking to the sidechain of Tyr 85, thesubstituent at ring position 2 forms a hydrogen bond to the Asp 87 sidechain,there is a hydrogen bond between the tyrosine hydroxyl group and the -6phosphate, and the substituent at ring position 4 may contact the -6phosphate. The wild-type base at position -5 is uracil (U), but several

32

substitutions have been tested in binding experiments and the structures ofsome have been determined (45,47). Cytosine (C) at -5 gives a significantlyhigher binding affinity than wild-type U (74,79,96), due to a hydrogen bondformed within the RNA molecule (phosphate -6 to the ring position 4 N inthe -5 base) that stabilises the protein in the bound conformation (45). Apartfrom that, the conformation is very similar to the wildtype TR.

In paper II, we report the structures of six different complexes withsubstitutions at position -5, two of them purines. The variant bases testedwere 5-bromouracil-5 (BrU-5), pyrimidine-2-one-5 (2one-5), 2-thiouracil-5(SU-5), 2-thiouracil-5, 6 (SU-5, 6), adenine -5 (A-5) and guanine-5 (G-5).The structures had good resolution, varying from 2.2 to 2.8 Å. One of thecomplexes, BrU-5, had a resolution of 2.2 Å, the highest obtained yet for anMS2 crystal. This allowed improvement of the protein model at some places.The structures of the RNA molecules were all very similar to each other andto the wild-type TR, although with small changes in the detailed interactions(Figure 12).

The BrU complex. Although in a conformation very similar to wildtype,the RNA backbone had moved slightly between nucleotides -7 and -4, in adirection away from the protein. The bromine atom was visible as a strongpeak in the electron density map. To accommodate the bulky bromine, theBrU base had to be somewhat rotated in the plane of the base compared tothe wild-type, which allows for formation of 3 hydrogen bonds. The BrUhairpin has a binding affinity for coat protein that is significantly higher thanthe wild-type TR (96). The three hydrogen bonds (compared to one for thewildtype) and the entropy effect of burying the bulky, hydrophobic bromineatom probably leads to the higher affinity measured in solution.

The 2-one-5 complex. The pyrimidine-2-one-5 RNA was constructed toexplore the relative importance of base stacking vs. polar interactions at the4 position of the base ring. The 2one base can be regarded as a pyrimidinemissing substituents at the 4 position. The structure shows that 2-one-5 baseis both rotated 20° in the plane of the base and moved approximately 0.3 Åcompared to the wildtype TR. The position of the O2 atom is still very closeto the corresponding atom in the wildtype and makes the same hydrogenbond. The ring of TyrA85 has moved a little, apparently to optimise stackinginteraction with the base. The 2-one-5 RNA has a slightly lower bindingaffinity than the wildtype TR, probably caused by the loss of one hydrogenbond.

33

Figure 12. The detailed interactions of some RNA fragments from Paper II. Top left:U-5 (wildtype). Top right: BrU-5. Bottom left: 2one-5. Bottom right: G-5.

The SU-5 and SU-5-6 complexes. The 2-thiouracil substituted SU-5 RNAwas constructed to investigate the effect of weakening the hydrogen bond toAsnA87. The less electronegative sulphur atom at ring position 2 should be aweak hydrogen bond acceptor. The SU-5-6 RNA, with an additional2-thiouracil at base -6 was tested to see if a conformational change could beprovoked, as seen for a previously determined structure with a pyridine-4-one at the -5 position (47). The complexes had structures very similar to

34

each other and to the wildtype TR. The 2-thiouracil base and the RNAbackbone of bases -6 to -4 has moved significantly away from the protein,probably as an effect of the longer and weaker hydrogen bonds between thesulphur atoms and the protein. SU-5 has a binding affinity that is 10 timesweaker than the wild-type TR, consistent with longer hydrogen bondsbetween the TyrA85 OH and the -5 phosphate and between the sulphur atomand the asparagine sidechain.

The A-5 and G-5 complexes. Binding experiments with purines atposition -5 have indicated only very low affinity for these complexes (97).Nevertheless, we were able to determine the structures of the variant RNAhairpins with A and G in the -5 position, respectively. The two complexesare similar to each other and resemble the wildtype TR. The backbonebetween riboses -6 and -4 has moved away from the protein to accommodatethe larger purine at -5. The A and G bases are in the same position as thewildtype TR U, but because of their larger size they are protruding bothtowards and away from the protein. There is no obvious strain in thesecomplexes, but the steric changes involved in fitting these bases mayaccount for the reduced affinity. In the G-5 complex, the phosphate -5 hasmoved 1.8 Å compared to the wildtype TR, disrupting the hydrogen bondbetween TyrA85 OH and this phosphate. G-5 also has the lowest affinity ofthe tested complexes.

From our experiments, it seems like many types of bases can be acceptedat the -5 position, albeit with reduction of binding affinity. There is,however, one example of a base that cannot be bound to the coat protein instandard manner. The structure of a complex containing a pyridine-4one(4one) at position -5 has been determined (47), and it revealed that the -5base is pointing away from the protein and not forming any interaction.Instead, the -6 base is stacking to the TyrA85 side chain (Figure 13). Thisswitching of bases is probably caused by the more electronegative potentialaround the 4 oxygen in the 4one base, which is spatially close to one of theoxygens in phosphate -6. We conclude that the stacking to TyrA85 is thesingle most important interaction of this base, since each of the polarinteractions can be removed or weakened with intact binding to the protein.

35

Figure 13. The detailed interactions of the 4one-5 RNA fragment.

5.3 Results, Paper III

In addition to the -5 base, the bases at -10, -7 and -4 are important forsequence specific binding to the MS2 coat protein. No other base thanadenine (A) at position -4 gives detectable binding (97). The bases atposition -4 and -10 bind in the corresponding binding pockets on each sideof the protein dimer, but the orientation of the -10 base in the pocket isdifferent. At the -10 position both purines are acceptable for binding (98).The -7 base does not make any direct interaction with the protein, instead itis stacked to base -5, which in turn is stacked to TyrA85 of the protein.There is a strong preference for purines at position -7 (76,78,80). MS2 RNAcomplexes with hairpins substituted at positions -10 and -7 are presented inpaper III. The base substitutions tested were guanosine-10 (G-10_1),2'-deoxy-2aminopurine-10 (2ap-10), inosine-10 (I-10), cytidine-10 (C-10),cytidine -7 (C-7) and another guanosine-10 with additional mutations in theRNA stem (G-10_2) (Figure 14). Some of the RNAs also contained U to Csubstitution at position -5, since that mutation increases binding affinity andthus makes binding in the capsids more likely (74,78). Only for three of thecomplexes, C-10, C-7 and G-10_2, could the electron density of the RNA beinterpreted into models (Figure 15).

36

Figure 14. The sequences and secondary structures of the RNA fragments used inPaper III.

The G-10_1 and G-10_2 complexes. If a G replaces the wildtype A-10, theresulting RNA hairpin has low binding affinity (98). This is probably causedby multiple alternative conformations of the RNA fragment, some of whichare unable to bind the coat protein (Figure 14) (98). The G-10_1 RNA hadvery weak density and only the loop bases could be seen. We havepreviously been successful in producing complexes with low affinity RNAhairpins, but in this case possibly only a fraction of the molecules are in thecorrect conformation for binding. To counteract the problem we constructeda hairpin, G-10_2, with two of the G-C basepairs replaced by C-G basepairs(Figure 14). This fragment has a binding affinity similar to wildtype TR (98)and the electron density showed clear binding of the hairpin. Theconformation of the RNA is very similar to wildtype, with the G-10essentially in the same place as the wildtype A-10, but slightly rotated toaccommodate the amino group at position 2 in the base (Figure 15). The

U U -5 U U U C U C U U A A A A A A A A A A G-C G-C G-C G-C G-C G-C C-G G-C G-C G-C-10 A G G C 2ap G-C C-G G-C G-C G-C U-A +1 U-A U-A U-A U-A A-U A-U A-U A-U A-U C-G C-G C-G C-G C-G A-U A-U A-U A-U A-U

wildtype G-10_2 G-10_1 C-10 2ap-10

U U U C U C U C U C A A C A A A A A A A G-C G-C G-C G G-C G-C G-C G G-C G-C I A G-C G-C G-C G-C G-C G-C G-C G U-A U-A U-A U-A U-A A-U A-U A-U A-U A-U C-G C-G C-G C-G C-G A-U A-U A-U A-U A-U

I-10 C-7 alternative basepairing for G-10_1

37

Figure 15. Superpositions of Left: G-10_2 (grey), C-10 (black) and the wildtype A-10 (white). Right: C-7 (black) and the wildtype A-7 (white).

guanosine base forms the same hydrogen bonds to the protein as thewildtype adenosine.

The 2ap-10 and I-10 complexes. The RNA fragments containing2-aminopurine and inosine at position -5 were constructed to explore theimportance of the substituents at the 2 and 6 position in the base,respectively. They could both be regarded as truncated versions of guanine.The 2ap-10 RNA has good density for the upper part of the hairpin, and welldefined density for bases -11 and -9. There is, however, no visible densityfor the -10 base. Phosphates -9 and -10 have moved up to 1.5 Å, indicatingthat the -10 base might be oriented differently than in the wildtype TR. The2ap-10 RNA fragment has been found to have the highest binding affinitymeasured to date in this system (80).

The I-10 map showed no evidence of RNA binding, only isolated densitysimilar to that found in recombinant capsids that have not been soaked withRNA. Both the inosine and the 2-aminopurine bases should be able to formthe same hydrogen bonds to the protein as the guanosine in the G-10_2hairpin, but the electron density indicates that this has not been the case. Aprevious study (98) indicated that both 2ap-10 and I-10 could have a similarproblem with alternative basepairing as the G-10_1 hairpin. That wouldsuffice to explain the poor binding of I-10, but in the 2ap-10 case there is aclear binding of the RNA and indications of flexibility for the -10 base. Theabsence of binding of 2-aminopurine in the binding pocket is even morepuzzling in view of its similarity to guanosine, which in the G-10_2 complexis neatly inserted into the pocket.

The C-10 complex. The structure of the C-10 operator hairpin provesthat even an RNA fragment with very low binding affinity can bind well

38

enough for structure determination. This fragment has a 1000-fold lowerbinding affinity than the wildtype TR (98). The conformation of the RNA isidentical to that of the wildtype, except for base -10. The -10 cytosine is onlypartly inserted into the binding pocket and does not form any hydrogenbonds to the protein, which would explain the poor binding affinity of thismolecule. In a solution structure of the wildtype TR, the -10 base isintercalated into the stem. It is possible that interaction of the base leads to astem conformation which is incompatible with binding to the protein, andthus causes the cytosine to bulge out and insert itself into the binding pocket.

The C-7 complex. The C-7 complex was produced in search of anexplanation for the purine specificity of the -7 base. Various purines givebinding affinities similar to wildtype, while pyrimidines give at least a 100-fold reduction (76). The conformation of this complex is the same as forwildtype TR, with the cytosine -7 base occupying the same position as thewildtype adenosine. A uracil at position -7 would be expected to give verylow affinity since it would form a basepair with A-4, which is absolutelycrucial for the binding. While a cytosine might not form a basepair, it maystill allow a loop conformation where A-4 is unavailable for binding to theprotein. A purine could, by its larger size, force the A-4 to be exposed.Alternatively, the difference in binding affinity could be caused by the morefavourable stacking of a purine between U-5 and G-8.

5.4 Overall conclusions - RNA binding

An interesting observation is that so many RNA hairpins bind to the MS2coat protein in a very similar manner, even if they have significantlydifferent binding affinity. Generally a wide range of bases can beaccommodated in the structure, but at the price of weaker binding. For thebinding of a base at -5, the effect of the interactions on binding affinityseems straightforward. The most important interaction is the stacking toTyrosine A85, followed by the hydrogen bond between Tyrosine A85 OHand the -5 phosphate. The two remaining hydrogen bonding atoms, thesubstituents at ring position 2 and 4 in the base, have additional effects. Theenhanced affinity of a cytidine at -5 is due to an intra-RNA hydrogen bondstabilising the RNA in the bound conformation. For the bases at position -10and -7, the relationships are more complex. The relative affinity of basevariants at these positions seems to not only depend on the boundconformation, but also on how the molecules behave in solution. The purine

39

specificity at the -10 is caused by the ability of these bases to form hydrogenbonds to the protein, since pyrimidines are too small. But the 2-aminopurineRNA, with its strong binding and possibly different conformation obscuresthe scenario. The most probable explanation for a purine specificity at the -7position is the need to have a large base to force the -4 adenine into aposition favourable for binding, rather than any specific properties of thatbase.

40

6 The structure of HK97 - a covalently linkedvirus shell (Paper IV)

6.1 Introduction

The virus capsids of the -like coliphages are large and complex structures.They have a T=7 icosahedral head surrounding the single-stranded DNAgenome, a portal through which the DNA is inserted into the preformedcapsid, and a long tail which is attached after DNA packaging and used toinject the genetic material into the host cell. As the genome is packed, thehead undergoes an expansion and rearrangement of the protein monomerscalled maturation. The size and shape are drastically changed from small andspherical to large and icosahedral. The head is constructed from 415 copiesof the coat protein and a portal complex that takes the place of one pentamerin the T=7 lattice. Many phages in the siphoviridae family (includingcoliphage ) assemble their heads utilising special scaffolding proteins.These proteins help define the shell curvature and size, but are discardedduring maturation. Additional stabilising proteins hold the capsid together.

Bacteriophage HK97 is different from phage in that it does not requirescaffolding or stabilising proteins. Instead, about 100 residues in N-terminusof the coat protein, which is cleaved off prior to DNA packing, might beacting as a guide for head assembly. To stabilise the head, covalent bondsbetween the protein monomers are formed throughout the capsid in such away that the linked monomers form interlocking rings (55,56,99). HK97 isso far the only member of the siphoviridae family which has been studied indetailed by both electron microscopy and x-ray crystallography(18,56,100,101). The work presented in paper IV includes extensiverefinement of an x-ray model of phage HK97, analysis of the interactionswithin the shell.

41

6.2 Results

The structure of the mature head from phage HK97, formed fromrecombinant expression of the capsid protein and chemically expanded, haspreviously been determined (59,102,103). Despite extensive rebuilding,positional refinement, and refinement of various other factors such as cellparameters, particle centre and particle orientation, the R-factor could onlybe lowered from 47,2 % to 36,5 %. While disappointing, we have not beenable to devise any improvement to the model that could lower the R-value.The crystals have a very large unit cell with a solvent content of 87%, whichleads to weak data. Since the R-factor as a function of resolution isreasonably well correlated with average intensity, we suspect that this mightbe the major cause of the high R-factor. Analysis has also showedanisotropic diffraction in different directions of the crystal.

The structure. The HK97 capsid protein monomer has a completely newfold. There are two domains, the A domain which is a mixed alpha-betastructure and the P domain which is unusually elongated and mainly consistsof a helix and a long, kinked -sheet. Together they form an L-shapedprotein with two protrusions, the N-terminal, and a part of the P-domain -sheet named the E-loop (Figure 16). The monomers are organised intopentamers and hexamers, which are also the assembly intermediates (Figure17). The pentamers are responsible for most of the curvature in the shell,while the hexamers are relatively flat. The pentamers are at positions oficosahedral symmetry, but the hexamers do not have any symmetry restraintsand are slightly different in conformation and interactions. In the hexamer,most of the contacts between the monomers are preserved between thedifferent conformations. The protein interactions in the pentamer are adaptedby a more bent conformation of the N-arm and the E-loop, and most of thecontacts between monomers are the same as in the hexamer.

42

Figure 16. The HK97 monomer (A subunit).

Figure 17. The arrangement of HK97 subunits in the T=7 shell. Eight hexamers(grey) and one pentamer (black) are shown, a total of 53 subunits.

43

Formation of a covalent bonds throughout the capsid. The long N-terminus and the E-loop weaves through the shell making stabilisinginteractions with neighbouring monomers. Each monomer contacts no lessthan nine of the surrounding proteins. But the cause of the unusual stabilityof the HK97 particle is the cross-linking of all monomers into one covalentlybound entity. The result, interlocking rings covering the whole capsid andresembling chainmail (Figure 18), has never before been observed forproteins. The residues forming the covalent bond are lysine 169 near the tipof the E-loop and asparagine 356 in the P-domain. The reaction is suggestedto be catalysed by glutamic acid 363 from a third subunit (56). Theevolutionary origin of this crosslink is uncertain, but two related phages withthe ability to form covalent bonds have been identified (104-106). There arealso related phages, with enough sequence similarity to be reasonably wellaligned that does not posses the crosslinking residues.

A comparison with the model of a previous stage in phagematuration. Bacteriophage HK97 undergoes several distinct, temporallyseparated maturation steps, which has been studied in detail by electronmicroscopy (EM) (18,100). Recently, a 12 Å resolution EM structure of anearlier maturation intermediate, the Prohead II, allowed adjustment andrefinement of the subunits from the Head II structure in the density (101). InProhead II, the capsid protein is cleaved, but the particle has not expandedand no crosslinks are formed. Except for the N-terminal arm and the E-loop,the mature head monomers could be fitted as rigid bodies into the EMdensity. The E-loop had to be remodelled as projecting radially outwardsfrom the shell and the N-terminal could not be fit in. The differences inmonomer arrangement are evident. Both pentamers and hexamers havebecome more "closed" with a cuplike structure. The hexamers are much lesssymmetric than in the mature head, and could rather be viewed as a dimer oftrimers. A comparison on atomic level shows that virtually no contactsbetween the monomers are the same as in the mature head. Evidently, thiscapsid protein has at least two entirely different sets of interaction surfaces.

44

Figure 18. The interlocking covalent rings formed by the crosslinking. The sameview as in figure 17, but both hexamers and pentamer are grey, and the E-loops areblack.

45

7 The refined structure of Nudaureliacapensis Virus (Paper V)

7.1 Introduction

Nudaurelia capensis Virus (N V) belong to the tetravirus family, andinfects insects of the lepidopteran order (butterflies and moths). N V is apositive-stranded RNA virus and the genome consists of two separate RNAsegments. The virus capsid is composed of 240 protein subunits of 641amino acids each. The x-ray structure was first published in 1996 (107). Thecapsid proteins are arranged with T=4 quasi-symmetry in the shell, regulatedby parts of the C-termini. These segments are found in two discrete orderedstates, producing either flat or bent contacts similar to T=3 plant viruses andnodaviruses. The monomer fold is an elaboration of the classic viral jellyroll,consisting of 8 antiparallel strands (Figure 20). Insertions at the N and C-termini form a helical domain on the inside of the capsid surface. A largeinsertion in the EF loop forms an immunoglobulin-like (Ig-like) domain onthe particle surface. Other insertions add two strands to the viral jelly-rolland two helices and a small sheet are positioned between the jellyroll andthe Ig-like domain (Figure 19). The protein chain undergoes autoproteolyticcleavage during maturation, cutting off the peptides (residues 571 to 641)from the main portion of the protein.

46

Figure 19. Left: The fold of the N V monomer (C subunit). Right: The monomerfold of BBV (C subunit). The strands that are present in the standard viral jellyrollare coloured black.

The structure is similar to that of four viruses in the nodavirus family, afamily of T=3 viruses that infect insects, fish and some mammals. Theseviruses, Black Beetle Virus (BBV), Flockhouse Virus (FHV), NodamuraVirus and Pariacoto Virus, also produce peptides by autoproteolyticcleavage (9,10,108,109). Inside the 5-fold axis the five helices formed by the peptides from a bundle. For FHV a model has been proposed where the

helix bundle is involved in membrane translocation of the viral RNA (36).Non-enveloped viruses must find a way to transport their genetic materialthrough the cell membrane. Receptor interactions lead to endocytosis of thevirus particles and the low pH of the endosome triggers structural changes inthe virus capsid. For FHV, the helix bundle is proposed to be externalisedupon exposure to low pH, and become attached to the membrane. The innerpart of the helix bundle is associated with the viral RNA and the helixbundle would then transfer parts of the RNA to the cytoplasmic side of thecell membrane. Attachment of ribosomes and subsequent translation of theRNA could then pull the rest of the RNA into the cell. With the structure ofN V it was discovered that five peptides close to the fivefold axis formeda large pore-like structure that could possibly act as a channel fortransporting the RNA through the membrane. The work presented in Paper

47

V consists of rebuilding and refinement of the N V structure, includingsome previously unmodelled segments of polypeptide with implications forthe membrane translocation.

7.2 Results

The preliminary model of N V (107) was rebuilt and refined and severalnew stretches of polypeptide could be added. The R-factor was lowered from27.5% to 22.3%.

The T=4 symmetry. The 240 proteins are arranged with T=4 quasi-symmetry (Figure 20). The different conformations are named A, B, C andD. The virus shell can be viewed as built up by two different kinds oftrimers, ABC and DDD. The T=4 interactions is regulated by either flat orbent contacts of subunits interacting at the five-fold, quasi-sixfold and quasi-twofold axes (Figure 21). C-terminal helices from subunits C and D areinserted as wedges in the contacts between the C and D and between D andB. The part of the chain forming these helices is disordered in the A and Bsubunits. In the capsid, the effect is that DDD trimers are surrounded by theC and D helices, making the interface between DDD trimers and ABCtrimers flat (Figure 20). The interactions between different ABD trimers arebent, since no helices are present in the interface. The Ig-like domain ispositioned on the outside of the virus capsid, and clusters together aroundthe threefold and quasi-threefold axes.

48

Figure 20. Top: The T=4 arrangement of capsid proteins in the N V shell. Asubunits are white, B subunits are light grey, C subunits are dark grey and Dsubunits are black. Five-, three- and twofold axes are indicated. Bottom: theordered C-terminal helices that make protein interfaces flat.

49

Figure 21. The flat and bent contacts between N V subunits. Top left: Bent contactbetween two A subunits close to the five-fold axis. Top right: Bent contact betweensubunits B and C around the quasi-sixfold axis (icosahedral twofold). Bottom left:Flat contact between subunits C and D around the quasi-sixfold axis. Bottom right:Flat contact between subunits D and B around the quasi-sixfold axis. A mageisiumion is shown as a black sphere.

50

The arrangement of peptides around the fivefold axis. The addition ofan N-terminal helix in the B subunits revealed that what was previouslythought to be a cylinder-shaped arrangement of five helices around thefivefold axis was in fact a 10-helix cylinder, consisting of the peptides andfive N-terminal helices (Figure 23). Since the N-terminal helices arecovalently attached to the main portion of the capsid protein, externalisationof this cylinder would be problematic. There is a long stretch of polypeptidewithout secondary structure attached to the N-terminal helices that couldallow some flexibility of the helix cylinder (Figure 23), just long enough forthe cylinder to become externalised. The linker would be lining the outsideof the helix cylinder, however, and could interfere with membrane contacts.In addition, and opposed to FHV, the inside of the helix cylinder ishydrophobic and the outside is polar. It is clear that this arrangement of peptides and N-terminal helices cannot serve as a pore for RNA transportacross the membrane in the form it is present in the virus capsid.

51

Figure 22. The cylinder of 10 helices formed around the fivefold axis in N V. Allfive peptides are shown (dark grey), as well as all five N1 helices and linkers(black). For clarity, only two A and B subunits out of five are shown (grey). Top:View from inside the capsid. Bottom: View from the side.

52

8 Acknowledgements

I want to thank all past and present members in the structure labs of theDepartment of Cell- and Molecular Biology, Uppsala University, theDepartment of Biochemistry, Uppsala University and the Department ofMolecular Biology, Swedish University of Agricultural Sciences for all thenice times I have had and for providing such a supportive and inspiring placeto work.

Lars, my supervisor. Thank you for all your help, especially during theselast months. And thank you for making all the pictures for the manuscripts!

Special thanks goes to my long-time office-mates Elin and Kaspars,work would have been dull without you!

I also want to thank all other members of the virus group, Bror, Roshan,Kerstin, Torsten, Seved, Nina, Hans, Fariborz, Jimmy, Lena and Anettefor help and for fun Friday lunches.

Thank you Alwyn, Sherry, Jill and Alex for getting me hooked oncrystallography in the first place!

To the coffee club members, thanks for all the good laughs, and thediscussions about life, politics, philosophy, underwear and sometimes evenwork during our daily gatherings at 10.15 and 15.15. Margareta, Inger,Ulla, Emma, Patrik, Malin, Karin, Fredrik, Isabella and all other faithful"fikare". I'm sorry I've been absent the last weeks!

Saeid, thank you for our endless talks about children, day-care and work.Thank you Erling and Remco for taking care of the computers and

Solveig and Ingrid, for handling all the bureaucracy.I want to thank Mamma, Pappa, Lina, Lars, Birgitta, Åke, Anna,

Karin, Andreas and other relatives for believing I could make it. And a bigthanks to Mamma and Birgitta for babysitting!

My dearest friends and cooking companions Kicki, Eva, Dennis, Per,Camilla and Pontus. Thank you for all the good times we have had over theyears.

And to my beloved family, Magnus and Axel. Thank you for letting mework all those weekends. I love you and I promise I will spend more timewith you from now on!

53

54

9 References

1. Caspar, D. L. D. and Klug, A. (1962) Physical principles in theconstruction of regular viruses. Cold Spring Harbor Symp. Quant.Biol., 27, 1-24.

2. Yan, X., Olson, N. H., Van Etten, J. L., Bergoin, M., Rossmann, M.G. and Baker, T. S. (2000) Structure and assembly of large lipid-containing dsDNA viruses. Nat Struct Biol, 7, 101-3.

3. Harrison, S. C., Olson, A. J., Schutt, C. E., Winkler, F. K. andBricogne, G. (1978) Tomato bushy stunt virus at 2.9 Å resolution.Nature, 276, 368-373.

4. Abad-Zapatero, C., Abdel-Meguid, S. S., Johnson, J. E., Leslie, A.G. W., Rayment, I., Rossmann, M. G., Suck, D. and Tsukihara, T.(1980) Structure of southern bean mosaic virus at 2.8 Å resolution.Nature, 286, 33-39.

5. Qu, C., Liljas, L., Opalka, N., Brugidou, C., Yeager, M., Beachy, R.N., Fauquet, C. M., Johnson, J. E. and Lin, T. (2000) 3D domainswapping of a molecular switch for quasi-equivalent symmetrymodulates the stability of an icosahedral virus. Structure, 8, 1095-1103.

6. Liddington, R., Yan, Y., Moulai, J., Sahli, R., Benjamin, T. L. andHarrison, S. C. (1991) Structure of simian virus 40 at 3.8 Åresolution. Nature, 354, 278-284.

7. Stehle, T., Yan, Y., Benjamin, T. L. and Harrison, S. C. (1994)Structure of murine polyomavirus complexed with anoligosaccharide receptor fragment. Nature, 369, 160-163.

8. Grimes, J. M., Burroughs, J. N., Gouet, P., Diprose, J. M., Malby, R.,Zientara, S., Mertens, P. P. and Stuart, D. I. (1998) The atomicstructure of the bluetongue virus core. Nature, 395, 470-478.

9. Hosur, M. V., Schmidt, T., Tucker, R. C., Johnson, J. E., Gallagher,T. M., Selling, B. H. and Rueckert, R. R. (1987) Structure of aninsect virus at 3.0 Å resolution. Proteins: Struct. Funct. Gen., 2, 167-176.

10. Fisher, A. J. and Johnson, J. E. (1993) Ordered duplex RNA controlscapsid architecture in an icosahedral animal virus. Nature, 361, 176-179.

55

11. Speir, J. A., Munshi, S., Wang, G., Baker, T. S. and Johnson, J. E.(1995) Structures of the native and swollen forms of cowpeachlorotic mottle virus determined by X-ray crystallography and cryo-electron microscopy. Structure, 3, 63-78.

12. Erickson, J. W., Silva, A. M., Murthy, M. R. N., Fita, I. andRossmann, M. G. (1985) The structure of a T=1 icosashedral emptyparticle from southern bean mosaic virus. Science, 229, 625-629.

13. Dong, X. F., Natarajan, P., Tihova, M., Johnson, J. E. andSchneemann, A. (1998) Particle polymorphism caused by deletion ofa peptide molecular switch in a quasiequivalent icosahedral virus. JVirol, 72(7), 6024-33.

14. Dokland, T. (1999) Scaffolding proteins and their role in viralassembly. Cell. Mol. Life Sci., 56, 580-603.

15. Dokland, T., McKenna, R., Ilag, L. L., Bowman, B. R., Incardona,N. L., Fane, B. A. and Rossmann, M. G. (1997) Structure of a viralprocaspid with molecular scaffolding. Nature, 389, 308-313.

16. Dokland, T., Bernal, R. A., Burch, A., Pletnev, S., Fane, B. A. andRossmann, M. G. (1999) The role of scaffolding proteins in theassembly of the small, single- stranded DNA virus phiX174. J MolBiol, 288, 595-608.

17. Marvik, O. J., Dokland, T., Nøkling, R. H., Jacobsen, E., Larsen, T.and Lindqvist, B. H. (1995) The capsid size-determining protein Sidforms an external scaffold on phage P4 procapsids. J. Mol. Biol.,251, 59-75.

18. Conway, J. F., Duda, R. L., Cheng, N., Hendrix, R. W. and Steven,A. C. (1995) Proteolytic and conformational control of virus capsidmaturation: the bacteriophage HK97 system. J. Mol. Biol., 253, 86-99.

19. Dokland, T. and Murialdo, H. (1993) Structural transitions duringmaturation of bacteriophage lambda capsids. J. Mol. Biol., 233, 682-694.

20. Prasad, B. V., Prevelige, P. E., Marietta, E., Chen, R. O., Thomas,D., King, J. and Chiu, W. (1993) Three-dimensional transformationof capsids associated with genome packaging in a bacterial virus. JMol Biol., 231, 65-74.

21. Zhang, Z., Greene, B., Thuman-Commike, P. A., Jakana, J.,Prevelige, P. E., Jr., King, J. and Chiu, W. (2000) Visualization ofthe maturation transition in bacteriophage P22 by electroncryomicroscopy. J Mol Biol, 297, 615-26.

22. Canady, M. A., Tihova, M., Hanzlik, T. N., Johnson, J. E. andYeager, M. (2000) Large Conformational Changes in the Maturation

56

of a Simple RNA Virus, Nudaurelia capensis ω Virus (NωV). J MolBiol, 299, 573-584.

23. Trus, B. L., Booy, F. P., Newcomb, W. W., Brown, J. C., Homa, F.L., Thomsen, D. R. and Steven, A. C. (1996) The herpes simplexvirus procapsid: structure, conformational changes upon maturation,and roles of the triplex proteins V19c and VP23 in assembly. J. Mol.Biol., 263, 447-462.

24. Casjens, S. and Hendrix, R. (1998) Control mechanisms in dsDNAbacteriophage assembly. In The bacteriophages (Calender. R. ed.), 1,15-91.

25. King, J. and Casjens, S. (1974) Catalytic head assembling protein invirus morphogenesis. Nature, 251(5471), 112-119.

26. Onorato, L. and Showe, M. K. (1975) Gene 21 protein-dependentproteolysis in vitro of purified gene 22 product of bacteriophage T4.J Mol Biol., 92(3), 395-412.

27. Ray, P. and Murialdo, H. (1975) The role of gene Nu3 inbacteriophage lambda head morphogenesis. Virology, 64(1), 247-263.

28. Olson, N. H., Kolatkar, P. R., Oliveira, M. A., Cheng, R. H., Greve,J. M., McCleeland, A., Baker, T. and Rossmann, M. G. (1993)Structure of a human rhinovirus complexed with its receptormolecule. Proc. Natl. Acad. Sci. USA, 90, 507-511.

29. Kolatkar, P. R., Bella, J., Olson, N. H., Bator, C. M., Baker, T. S.and Rossmann, M. G. (1999) Structural studies of two rhinovirusserotypes complexed with fragments of their cellular receptor. EmboJ, 18, 6249-6259.

30. He, Y., Bowman, V. D., Mueller, S., Bator, C. M., Bella, J., Peng,X., Baker, T. S., Wimmer, E., Kuhn, R. J. and Rossmann, M. G.(2000) Interaction of the poliovirus receptor with poliovirus. ProcNatl Acad Sci U S A, 97, 79-84.

31. He, Y., Chipman, P. R., Howitt, J., Bator, C. M., Whitt, M. A.,Baker, T. S., Kuhn, R. J., Anderson, C. W., Freimuth, P. andRossmann, M. G. (2001) Interaction of coxsackievirus B3 with thefull length coxsackievirus- adenovirus receptor. Nat Struct Biol,8(10), 874-8.

32. Hewat, E. A., Neumann, E., Conway, J. F., Moser, R., Ronacher, B.,Marlovits, T. C. and Blaas, D. (2000) The cellular receptor to humanrhinovirus 2 binds around the 5-fold axis and not in the canyon: astructural view. EMBO J., 19(23), 6317-6325.

33. Belnap, D. M., McDermott, B. M., Jr., Filman, D. J., Cheng, N.,Trus, B. L., Zuccola, H. J., Racaniello, V. R., Hogle, J. M. and

57

Steven, A. C. (2000) Three-dimensional structure of poliovirusreceptor bound to poliovirus. Proc Natl Acad Sci U S A, 97(1), 73-8.

34. Xing, L., Tjarnlund, K., Lindqvist, B., Kaplan, G. G., Feigelstock,D., Cheng, R. H. and Casasnovas, J. M. (2000) Distinct cellularreceptor interactions in poliovirus and rhinoviruses. Embo J, 19(6),1207-16.

35. Fricks, C. E. and Hogle, J. M. (1990) Cell-induced conformationalchange in poliovirus: Externalization of the amino terminus of VP1is responsible for liposome binding. J. Virol., 64, 1934-1945.

36. Cheng, R. H., Reddy, V. S., Olson, N. H., Fisher, A. J., Baker, T. S.and Johnson, J. E. (1994) Functional implications of quasi-equivalence in a T=3 icosahedral animal virus established by cryo-electron microscopy and X-ray crystallography. Structure, 2, 271-282.

37. Gallagher, T. M. and Rueckert, R. R. (1988) Assembly-dependentmaturation cleavage in provirions of a small icosahedral insectribovirus. J. of Virol, Sept 1988, 3399-3406.

38. Hohn, T. (1969) The assembly of protein particles of the RNAbacteriophage fr in the absence of RNA. Biochem. Biophys. Res.Comm., 36, 7-17.

39. Hung, P. P. and Overby, L. R. (1969) The reconstitution of infectivebacteriophage Qβ. Biochemistry, 8, 820-828.

40. Ling, C. M., Hung, P. P. and Overby, L. R. (1970) Independentassembly of Qβ and MS2 phages in doubly infected Escherichia coli.Virology, 40, 920-929.

41. Beckett, D., Wu, H.-N. and Uhlenbeck, O. C. (1988) Roles ofoperator and non-operator RNA sequences in bacteriophage R17capsid assembly. J. Mol. Biol., 204, 939-947.

42. van Duin, J. (1988) Single-stranded RNA bacteriophages. In TheBacteriophages. Calendar, R. (ed.), Plenum Press, New York, Vol. 1,pp. 117-167.

43. Witherell, G. W., Gott, J. M. and Uhlenbeck, O. C. (1991) Specificinteraction between RNA phage coat proteins and RNA. Progr. Nucl.Acid Res. Mol. Biol., 40, 185-220.

44. Valegård, K., Murray, J. B., Stockley, P. G., Stonehouse, N. J. andLiljas, L. (1994) Crystal structure of an RNA bacteriophage coatprotein-operator complex. Nature, 371, 623-626.

45. Valegård, K., Murray, J. B., Stonehouse, N. J., van den Worm, S.,Stockley, P. G. and Liljas, L. (1997) The three-dimensionalstructures of two complexes between recombinant MS2 capsids andRNA operator fragments reveal sequence-specific protein-RNAinteractions. J. Mol. Biol., 270, 724-738.

58

46. Grahn, E., Stonehouse, N. J., Murray, J. B., van den Worm, S.,Valegård, K., Fridborg, K., Stockley, P. G. and Liljas, L. (1999)Crystallographic studies of RNA hairpins in complexes withrecombinant MS2 capsids and implications for binding requirements.RNA, 5, 131-138.

47. Grahn, E., Stonehouse, N. J., Adams, C. J., Fridborg, K., Beigelman,L., Matulic-Adamic, J., Warriner, S. L., Stockley, P. G. and Liljas, L.(2000) Deletion of a single hydrogen bonding atom from the MS2RNA operator leads to dramatic rearrangements at the RNA-coatprotein interface. Nucl. Acids Res., 28, 4611-4616.

48. Grahn, E., Moss, T., Helgstrand, C., Fridborg, K., Sundaram, M.,Tars, K., Lago, H., Stonehouse, N. J., Davis, D. R., Stockley, P. G.and Liljas, L. (2001) Structural basis of pyrimidine specificity in theMS2 RNA hairpin-coat protein complex. RNA, 7, 1616-1627.

49. Helgstrand, C., Grahn, E., Moss, T., Stonehouse, N. J., Tars, K.,Stockley, P. G. and Liljas, L. (2002) Investigating the structural basisof purine specificity in the structures of MS2 coat protein RNAtranslational operator hairpins. Nucl. Acids Res., 30, 2678-2685.

50. Convery, M. A., Rowsell, S., Stonehouse, N. J., Ellington, A. D.,Hirao, I., Murray, J. B., Peabody, D. S., Phillips, S. E. V. andStockley, P. G. (1998) Crystal structure of an RNA aptamer-proteincomplex at 2.8 Å resolution. Nature Struct. Biol., 5, 133-139.

51. Rowsell, S., Stonehouse, N. J., Convery, M. A., Adams, C. J.,Ellington, A. D., Hirao, I., Peabody, D. S., Stockley, P. G. andPhillips, S. E. V. (1998) Crystal structure of a series of RNAaptamers complexed to the same protein target. Nature Struct. Biol.,5, 970-975.

52. Tong, L., Wengler, G. and Rossmann, M. G. (1993) Refinedstructure of sindbis virus core protein and comparison with otherchymotrypsin-like serine proteinase structures. J. Mol. Biol., 230,228-247.

53. Choi, H.-K., Tong, L., Minor, W., Dumas, P., Boege, U. andRossmann, M. G. (1991) Structure of Sindbis virus core proteinreveals a chymotrypsin-like serine proteinase and the organization ofthe virion. Nature, 354, 37-43.

54. Choi, H.-K., Lu, G., Lee, S., Wengler, G. and Rossmann, M. G.(1997) Structure of Semliki Forest virus core protein. Prot. Struct.Funct. Gen., 27, 345-359.

55. Duda, R. L., Martincic, K. and Hendrix, R. W. (1995) Genetic basisof bacteriophage HK97 prohead assembly. J. Mol. Biol., 247, 636-647.

59

56. Wikoff, W. R., Liljas, L., Duda, R. L., Tsuruta, H., Hendrix, R. W.and Johnson, J. E. (2000) Topologically linked protein rings in thebacteriophage HK97 capsid. Science, 289, 2129-2133.

57. Simpson, A. A., Chipman, P. R., Baker, T. S., Tijssen, P. andRossmann, M. G. (1998) The structure of an insect parvovirus(Galleria mellonella densovirus) at 3.7 Å resolution. Structure, 6,1355-1367.

58. Wynne, S. A., Crowther, R. A. and Leslie, A. G. (1999) The crystalstructure of the human hepatitis B virus capsid. Mol Cell, 3(6), 771-80.

59. Wikoff, W. R., Duda, R. L., Hendrix, R. H. and Johnson, J. E. (1999)Crystallographic analysis of the dsDNA bacteriophage HK97 matureempty capsid. Acta Cryst., D55, 763-771.

60. Prasad, B. V. V., Hardy, M. E., Dokland, T., Bella, J., Rossmann, M.G. and Estes, M. K. (1999) X-ray crystallographic structure of theNorwalk virus capsid. Science, 286(5438), 287-290.

61. Munshi, S., Liljas, L. and Johnson, J. E. (1998) The structuredetermination of Nudaurelia capensis ω virus. Acta Cryst. D, D54,1295-1305.

62. Otwinowski, Z. and Minor, W. (1996) In Carter, C. W., Jr., andSweet, R. M. (eds.), Methods in Enzymology. Academic Press, NewYork, Vol. 276, pp. 307-326.

63. Brünger, A. T. (1990) XPLOR Manual, Yale University, NewHaven, CT.

64. Liljas, L., Fridborg, K., Valegård, K., Bundule, M. and Pumpens, P.(1994) Crystal structure of bacteriophage fr capsids at 3.5 Åresolution. J. Mol. Biol., 244, 279-290.

65. Jones, T. A. (1992) In Dodson, E. J., Gover, S., and Wolf, W. (eds.),CCP4 Study Weekend 1992: Molecular Replacement. DaresburyLaboratory, Daresbury, England, pp. 91-105.

66. Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros,P., Grosse-Kunstleve, R. W., Jiang, J.-S., Kuszewski, J., Nilges, M.,Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T. and Warren, G.L. (1998) Crystallography & NMR System: A New Software Suitefor Macromolecular Structure Determination. Acta Cryst., D 54, 905-921.

67. Jones, T. A., Bergdoll, M. and Kjeldgaard, M. (1990) In Bugg, C.,and Ealick, S. (eds.), Crystallographic and Modeling Methods inMolecular Design. Springer-Verlag, New York, pp. 189-199.

68. Loeb, T. and Zinder, N. (1961) A bacteriophage containing RNA.Proc. Natl. Acad. Sci. USA, 47, 282.

60

69. Carey, J., Lowary, P. and Uhlenbeck, O. C. (1983) Interaction ofR17 coat protein with synthetic variants of its ribonucleic acidbinding site. Biochemistry, 22, 4723-4730.

70. Carey, J., Cameron, V., de Haseth, P. L. and Uhlenbeck, O. C.(1983) Sequence-specific interaction of R17 coat protein with itsRNA binding site. Biochemistry, 22, 2601-2610.

71. Carey, J. and Uhlenbeck, O. C. (1983) Kinetic and thermodynamiccharacterization of the R17 coat protein-RNA interaction.Biochemistry, 22, 2610-2615.

72. Gott, J. M., Wilhelm, L. J. and Uhlenbeck, O. C. (1991) RNAbinding properties of the coat protein from bacteriophage GA. Nucl.Acids Res., 19, 6499-6503.

73. Johansson, H. E., Liljas, L. and Uhlenbeck, O. C. (1997) RNArecognition by the MS2 phage coat protein. Sem. Virol., 8, 176-185.

74. Lowary, P. T. and Uhlenbeck, O. C. (1987) An RNA mutation thatincreases the affinity of an RNA-protein interaction. Nucl. AcidsRes., 15, 10483-10493.

75. LeCuyer, K. A., Behlen, L. S. and Uhlenbeck, O. C. (1995) Mutantsof the bacteriophage MS2 coat protein that alter its cooperativebinding to RNA. Biochemistry, 34, 10600-10606.

76. Romaniuk, P. J., Lowary, P. T., Wu, H. N., Stormo, G. andUhlenbeck, O. C. (1987) RNA binding site of R17 coat protein.Biochemistry, 26, 1563-1568.

77. Uhlenbeck, O. C., Carey, J., Romaniuk, P., Lowary, P. T. andBeckett, D. (1983) Interaction of R17 coat protein with its RNAbinding site for translational repression. J. Biomol.Struct. Dynam., 1,539-552.

78. Stockley, P. G., Stonehouse, N. J., Murray, J. B., Goodman, S. T. S.,Talbot, S. J., Adams, C. J., Liljas, L. and Valegård, K. (1995)Probing sequence-specific RNA recognition by the bacteriophageMS2 coat protein. Nucl. Acids Res., 23, 2512-2518.

79. Lago, H., Fonseca, S. A., Murray, J. B., Stonehouse, N. J. andStockley, P. G. (1998) Dissecting the key recognition features of theMS2 bacteriophage translational repression complex. Nucleic AcidsRes., 26, 1337-1344.

80. Parrott, A. M., Lago, H., Adams, C. J., Ashcroft, A. E., Stonehouse,N. J. and Stockley, P. G. (2000) RNA aptamers for the MS2bacteriophage coat protein and the wild-type RNA operator havesimilar solution behaviour. Nucl. Acids Res., 28, 489-497.

81. Lago, H., Parrott, A. M., Moss, T., Stonehouse, N. J. and Stockley,P. G. (2001) Probing the kinetics of formation of the bacteriophage

61

MS2 translational operator complex: Identification of a proteinconformer unable to bind RNA. J. Mol. Biol., 305(5), 1131-1144.

82. Valegård, K., Liljas, L., Fridborg, K. and Unge, T. (1990) The three-dimensional structure of the bacterial virus MS2. Nature, 345, 36-41.

83. Golmohammadi, R., Valegård, K., Fridborg, K. and Liljas, L. (1993)The refined structure of bacteriophage MS2 at 2.8 Å resolution. J.Mol. Biol., 234, 620-639.

84. Golmohammadi, R., Fridborg, K., Bundule, M., Valegård, K. andLiljas, L. (1996) The crystal structure of bacteriophage Qβ at 3.5 Åresolution. Structure, 4, 543-554.

85. Tars, K., Bundule, M., Fridborg, K. and Liljas, L. (1997) The crystalstructure of bacteriophage GA and a comparison of phages belongingto the major groups of Escherichia coli leviviruses. J. Mol. Biol.,271, 759-773.

86. Tars, K., Fridborg, K., Bundule, M. and Liljas, L. (2000) Crystalstructure of phage PP7 from Pseudomonas aeruginosa at 3.7 Åresolution. Virology, 272, 331-337.

87. Roberts, J. W. and Steitz, J. A. (1967) The reconstitution of infectivebacteriophage R17. Proc. Natl. Acad. Sci. USA, 58, 1416-1421.

88. Paranchych, W., Ainsworth, S. K., Dick, A. J. and Krahn, P. M.(1971) Stages in phage R17 infection. V. Phage eclipse and the roleof F pili. Virology, 45, 615-628.

89. Adhin, M. R., Avots, A., Berzin, V., Overbeek, G. P. and van Duin,J. (1990) Complete nucleotide sequence of the group I RNAbacteriophage fr. Biochim. Biophys. Acta, 1050, 104-109.

90. Stonehouse, N. J. and Stockley, P. G. (1993) Effects of amino acidsubstitutions on the thermal stability of MS2 capsids lackinggenomic RNA. FEBS Lett., 334, 355-359.

91. Stonehouse, N. J., Valegård, K., Golmohammadi, R., van den Worm,S., Walton, C., Stockley, P. G. and Liljas, L. (1996) Crystal structureof MS2 capsids with mutations in the subunit FG loop. J. Mol. Biol.,256, 330-339.

92. Hill, H. R., Stonehouse, N. J., Fonseca, S. A. and Stockley, P. G.(1997) Analysis of phage MS2 coat protein mutants expressed froma reconstituted phagemid reveals that proline 78 is essential for viralinfectivity. J. Mol. Biol., 266, 1-7.

93. Tars, K. (1999) Structure of small RNA phages. PhD thesis. UppsalaUniversity.

94. Beckett, D. and Uhlenbeck, O. C. (1988) Ribonucleoproteincomplexes of R17 coat protein and a translation operator analog. J.Mol. Biol., 204, 927-938.

62

95. Stockley, P. G., Stonehouse, N. J., Walton, C., Walters, D. A.,Medina, G., Macedo, M. B., Hill, H. R., Goodman, S. T. S., Talbot,S. J., Tewary, H. K., Golmohammadi, R., Liljas, L. and Valegård, K.(1993) Molecular mechanism of RNA-phage morphogenesis.Biochem. Soc. Trans., 21, 627-633.

96. Talbot, S. J., Goodman, S., Bates, S. R. E., Fishwick, C. W. G. andStockley, P. G. (1990) Use of synthetic oligoribonucleotides to probeRNA- protein interactions in the MS2 translational operatorcomplex. Nucl. Acids Res., 18, 3521-3528.

97. Carey, J., Lowary, P. and Uhlenbeck, O. C. (1983) Interaction ofR17 coat protein with synthetic variants of its ribonucleic acidbinding site. Biochemistry, 22, 4723-4730.

98. Wu, H. N. and Uhlenbeck, O. C. (1987) Role of a bulged A residuein a specific RNA-protein interaction. Biochemistry, 26, 8221-8227.

99. Johnson, J. E. and Wikoff, W. R. (1998) Macromolecular assembly:chainmail stabilization of a viral capsid. Curr Biol., 8, R914-R917.

100. Lata, R., Conway, J. F., Cheng, N., Duda, R. L., Hendrix, R. W.,Wikoff, W. R., Johnson, J. E., Tsuruta, H. and Steven, A. C. (2000)Maturation dynamics of a viral capsid: visualization of transitionalintermediate states. Cell, 100, 253-63.

101. Conway, J. F., Wikoff, W. R., Cheng, N., Duda, R. L., Hendrix, R.W., Johnson, J. E. and Steven, A. C. (2001) Virus maturationinvolving large subunit rotations and local refolding. Science, 292,744-8.

102. Wikoff, W. R., Duda, R. L., Hendrix, R. W. and Johnson, J. E.(1998) Crystallization and preliminary X-ray analysis of the dsDNAbacteriophage HK97 mature empty capsid. Virology, 243, 113-118.

103. Wikoff, W. R. and Johnson, J. E. (1999) Virus assembly: Imaging amolecular machine. Curr Biol, 9, R296-300.

104. Koptides, M., Ugorcakova, J., Baloghova, E., Bukovska, G. andTimko, J. (1994) Characterization and sequence analysis of the F2promoter from corynephage BFK20. Acta Virol, 38(4), 223-8.

105. Kropinski, A. M. (2000) Sequence of the genome of the temperate,serotype-converting, Pseudomonas aeruginosa bacteriophage D3. JBacteriol, 182(21), 6066-74.

106. Gilakjan, Z. A. and Kropinski, A. M. (1999) Cloning and analysis ofthe capsid morphogenesis genes of Pseudomonas aeruginosabacteriophage D3: another example of protein chain mail? JBacteriol, 181, 7221-7.

107. Munshi, S., Liljas, L., Cavarelli, J., Bomu, W., McKinney, B.,Reddy, V. and Johnson, J. E. (1996) The 2.8 Å resolution structure

63

of a T=4 animal virus and its implications for membranetranslocation of RNA. J. Mol. Biol., 261, 1-10.

108. Zlotnick, A., Natarajan, P., Munshi, S. and Johnson, J. E. (1997)Resolution of space-group ambiguity and the structure determinationof nodamura virus to 3.3 angstrom resolution from pseudo-R32(monoclinic) crystals. Acta Cryst. D, D53, 738.

109. Tang, L., Johnson, K. N., Ball, L. A., Lin, T., Yeager, M. andJohnson, J. E. (2001) The structure of pariacoto virus reveals adodecahedral cage of duplex RNA. Nat Struct Biol, 8, 77-83.

110. Luo, M., Vriend, G., Kamer, K., Minor, I., Arnold, E., Rossmann, M.G., Boege, U., Scraba, D. G., Duke, G. M. and Palmenberg, A. C.(1987) The atomic structure of mengo virus at 3.0 Å resolution.Science, 235, 182-191.