Spectroscopy on the assembly of cowpea chlorotic mottle virus

This thesis describes the characterization of cowpea chlorotic mottle virus (CCMV) by using spectroscopic techniques. In chapter one and two the main properties of CCMV, which belongs to the bromoviruses, are summarized. The application of spectroscopic techniques in the study of other viruses is reviewed in chapter two.The changes in the nucleoprotein particles, induced by variation of the pH, the ionic strength or the divalent cation concentration, may be important in the process of infection of the plant. After the penetration of the virus particles into the cell, the RNA must be released from its protective protein coat (dissociation) to be transcribed, to start the production of protein and to influence the metabolism of the cell. Subsequently coat protein subunits and RNA are assembled to form new spherical nucleoprotein particles, that can be transported to other parts of the plant for new infection.Dissociation and assembly can be induced in vitro by changing the composition of the buffer solution. The accompanying changes in the structure of the coat protein subunits and in the interaction between the different constituents of the nucleoprotein particles were investigated by using a number of different spectroscopic techniques to be described subsequently.With fluorescence and phosphorescence spectroscopy, the characteristics of tryptophanyl and tyrosyl residues in the protein subunits were observed. From fluorescence quantum yield and lifetime measurements (chapter 3) it can be concluded that the presence of RNA in the protein particles results in a static quenching of part of the tryptophanyl residues. These residues become fluorescent and exposed to the solvent, when the interaction between protein and RNA is diminished by an increase in ionic strength or is completely absent as in empty protein shells or in dissociated virus at pH 7.5 and high ionic strength. A rise in pH from 5.0 to 7.5 results in a similar effect on the fluorescence, probably because of a different location of the RNA in the swollen particles, compared to the stable virus particles at pH 5.0.Because the phosphorescence spectra of proteins show more fine structure than the fluorescence spectra, the effects of conformation on the phosphorescence properties of the virus particles at 77 K were investigated. In contrast to the expectations, no influence of the conformational changes on the position of the phosphorescence maxima was observed (chapter 4). Contrary to the interpretation obtained from fluorescence experiments. the location of the peaks in the phosphorescence spectra (413 and 440 run) is more representative for tryptophanyl residues buried in the interior of the protein subunits.A more refined characterization of the phosphorescent triplet state is made by the determination at 1.2 K of the zero field splitting parameters, using optical detection of magnetic resonance. The observed |D| - |E| and 2|E| transitions of the tryptophan moieties showed a large bandwidth in the virus, but no fine structure. The large bandwidth must be ascribed to statistical and structural heterogeneity in the environment of the tryptophanyl residues of the coat protein subunits. It also masks the possible shifts in the zero field transitions, when the protein-RNA interaction is reduced. From the slight changes in the ODMR parameters one can conclude that a decrease in the protein-RNA interactions causes an increase in the exposure of some of the tryptophanyl residues.At this stage a quantitative interpretation of the fluorescence and phosphorescence data cannot be given. However in contrast to what has been observed with BMV, the intrinsic spectroscopic properties of (CCMV) can very well be applied to a qualitative study of the assembly of this virus and it can be utilized to check the correct encapsulation of the RNA in the protein coat. In this respect the application of techniques like optical rotatory dispersion (ORD), circular dichroism (CD) and light absorption does lead to a more complex interpretation because of the large contribution of RNA to these spectra. For rodshaped viruses like TMV with an RNA content of 5% in contrast to the 24% of (CCMV) ORD and CD are more useful techniques. In experiments on M these techniques were applied very successfully to obtain information about the involvement of tryptophanyl residues in protein-RNA interactions.It is difficult to relate the observations of the tryptophan fluorescence in CCMV directly to specific residues in each of the protein subunits in the nucleoprotein particles. This is a consequence of the differences in location of the subunits in the nucleoprotein particles. The protein subunits of CCMV are arranged quasi-symmetrically in the coat of the virus particles. As shown by Caspar and Klug the interactions between the subunits may differ in the vicinity of a quasi sixfold and of a fivefold symmetry axis of the protein subunit arrangement. In addition, it is not very likely that the interactions between the
nucleotides and each of the protein subunits are identical. Thus instead of three classes of tryptophanyl residues in the nucleoprotein particles of (CCMV) this number is possibly higher. The heterogeneity may be caused by structural differences within the subunits. This effect is observed in tomato bushy stunt virus, where two different conformations for the protein subunits are present in the virus particle.In the case of TMV one major source of heterogeneity, the interaction between the subunits, is absent, when the 1.5% of the protein subunits on the ends of the rod is neglected. It is also observed that the contact between RNA and each of the TMV subunits in the rod is very similar.Apart from the probes that are already present in the protein molecules, like tryptophan and tyrosine, fluorescent and paramagnetic probes that can be attached covalently to the protein have been employed. These labels provide data about the mobility of certain regions in the protein subunits, about the role of RNA-protein interactions in this process and about the influence of small concentrations of mono- and divalent cations.As a fluorescent probe pyridoxal-5'-phosphate (PLP) was used (chapter 5). The advantage of this label is that it can be introduced into the stable virus particles at pH 5.0. The mobility of the labelled part was estimated from fluorescence polarization studies. Changes in ionic strength or divalent cation concentration at pH 5.0 had no effect on the mobility of the label. Upon swelling of the particles the label became more mobile. Immobilization occurred in the presence of divalent cations and to a smaller extent when monovalent cations were present. Measurements of radiationless energy transfer from tryptophanyl and from tyrosyl residues to the label indicated distinct changes in transfer efficiency when the conditions were varied.A second type of label used in the study of CCMV is a paramagnetic nitroxide spin label (chapter 6). By using electron paramagnetic resonance (EPR) the mobility of the spin label could be monitored from the shape of the EPR spectra. The application of the recently developed technique of saturation transfer EPR (ST-EPR) allows the determination of the slow motions of the virus particles. From the spin label experiments, it was observed that the mobility of the label is sensitive to conditions that influence the protein-RNA interactions and the RNA conformation. The mobility of the spin label at pH 5.0 is increased markedly upon raising the ionic strength. Also a swelling of the particles enhances the mobility, while addition of MgCl ₂ , which reduces the swelling, lowers the flexibility of the labelled region of the protein.The effect of ionic strength on the mobility of the spin label at pH 5.0 is in contrast to the results obtained for PLP-modified virus. However a principal difference between the two probes attached to CCMV is formed by the labelling conditions. PLP is attached to the virus under conditions, that keep the particles in a stable conformation. The spin label is bound to CCMV when it is in the swollen state and where the RNA is sensitive to ribonuclease. Attempts to perform the labelling reaction at pH 5.0 were unsuccessful. Apart from the different sites that are labelled, the differences at pH 5.0 may reflect the incomplete reversibility of the swelling or the incorrect refolding of CCMV after the labelling at pH 7.0.It is likely that the observed local flexibility in the protein subunits, when the nucleoprotein particles are in the swollen or dissociated form, plays an important role in the mechanism of assembly and dissociation. For TMV it has been noticed, that a change in the quarternary structure is accompanied by an increase in the flexibility of a certain section in the protein subunits. This was concluded from the disappearance of a detailed X-ray structure in parts of the double disc of TMV, while this structure could clearly be observed in the case of M nucleoprotein particles. The flexibility within the TMV protein subunits was confirmed by M spectroscopy by De Wit et al. and by Jardetzky et al. .In tomato bushy stunt virus (TBSV) it was not possible to observe a considerable number of amino acid residues at the N-terminus in the X-ray analysis. It is not yet understood, whether this must be explained by a local mobility of the N-terminal arms or by an irregular folding of these arms.The effect of mono- and divalent cations on the conformation of (CCMV) which was also noticed in the experiments with labelled virus, was investigated quantitatively in titration experiments. Mn ²⁺-ions were used as a paramagnetic probe and the binding of these ions was detected by means of EPR spectrscopy. Competition for binding with Mn ²⁺-ions made it possible to determine the binding properties of other cations. In the bromoviruses approximately two to three ions are bound per protein subunit at pH ≥6.5, while at pH 5.0 approximately one binding site was observed. In isolated RNA relatively more binding sites were detected. Since protein without RNA is not stable at low ionic strength no meaningful binding data could be obtained in this case.The binding properties were also considerably influenced by changes in the concentration of Tris-HCl. The binding of Mn ²⁺to the nucleoprotein might reflect a negative cooperative behaviour or non-identical sites. Monovalent cations can effectively displace the divalent manganese ion from the nucleoprotein, whereby the binding constant was approximately a factor 100 smaller than found for divalent cations. Pfeiffer and Durham determined the binding of Ca ²⁺and Mg ²⁺to BMV by measuring the displacement of protons from the nucleoprotein after addition of certain amounts of Ca ²⁺and Mg ²⁺. This method can only be applied at pH values below the pK of the titrating group. However the binding of metal ions to the virus is especially important at pH values above the pK of the titrating group. The more direct method described in this thesis, is more appropriate in this respect. In this way from our binding studies it could be observed that at pH 5.0 one binding site was present, while at pH 7.5 between two and three binding sites were observed. At pH 7.5 the binding of cations is more important in controlling the swelling of the nucleoprotein.In the interpretation of the binding properties similar problems may arise as mentioned for the classification of the tryptophan fluorescence properties. In the investigation of specific binding sites, formed by a protein subunit and the RNA, it seems reasonable to express the binding stoichiometry in sites per subunit. However, since the environments of the subunits in the protein coat are not identical, the observed binding may be an average value from the different sites. Especially for the binding properties, it can be imagined that the stoichiometry is less determined by the individual subunits and more by the clustering of the subunits in hexamers and pentamers. From the binding experiments it can be concluded that RNA is important in the formation of the cation binding sites. The arrangement of RNA between the protein subunits can results in values of stoichiometry not related directly to a single subunit.When the binding of cations is seen in terms of neutralizing charges on the nucleoprotein particles, it will be more important to know the concentration dependence of the binding, which should not necessarily be specific, than the specific binding constants and the stoichiometry. Apart from the direct determination of the binding, it is also possible to observe indirect effects of mono- and divalent cations on native and modified nucleoprotein particles. The presence of these cations can be manifested in structural effects, like flexibility of certain regions in the protein. These effects can be employed to follow the binding. In CCMV modified with PLP (chapter 5) the addition of Mg ²⁺and Na ⁺to the solution influences the mobility of the fluorescent label. A more detailed study in which also other cations are used and their concentration dependence on the mobility of the label is followed, may be a useful means in the interpretation of these effects.In this thesis the characterization of CCMV, its assembly and dissociation is described, using a number of spectroscopic, techniques. In this way information was obtained about the virus conformation and its dynamic behaviour. The application of fluorescence and phosphorescence of intrinsic probes is restricted to tryptophanyl and to a smaller extent to tyrosyl residues. The interpretation of the spectral properties is often very complicated. because of the character of the residues and the large number of environmental aspects, that play a role. Labelling with fluorescent or paramagnetic probes can yield useful information about intramolecular distances, mobility and hydrophobicity in the virus particles. One should however be careful to use only modified nucleoprotein in the investigation of the structureal transitions in the virus particles. Firstly, it is difficult to estimate to what extent the effects on the label reflect the intrinsic effects and secondly, it is uncertain what the contribution is, due to the modification.When information is available about the amino acid sequence and the X-ray structure, the application of spectral techniques becomes more useful. For this reason and because of the use of several mutants. rather detailed information is obtained on tryptophan and tyrosine in the subunits of tobacco mosaic virus. For the study of protein association and virus assembly, this knowledge may yield good perspectives.In recent years the technique of laser-Raman-scattering has become a powerfull technique in the characterization of protein structure. The studies are facilitated by extensive knowledge about the interpretation of Raman spectra of biomolecules. This method also seems to be very promising for the investigation of virus structure. It provides information about several amino acid residues, the folding of the backbone and the structure of the RNA. Also information about the RNA-protein interaction may be deduced from the spectra.The role of membranes in virus assembly and dissociation can be investigated by monitoring the spectroscopic behaviour of viral protein, when nucleoprotein particles and artificial membranes are brought together. By the study of the hydrophobicity of the probes and their mobility in the protein upon penetration in the membrane it is possible to obtain information on this process. The sensitivity of fluorescence to protein-RNA interactions can be used to follow the possible dissociation of the nucleoprotein particles, when they cross the membranes.