Journal of General Virology

Visualization by atomic force microscopy of tobacco mosaic virus movement protein–RNA complexes formed in vitro

O. I. Kiselyova,¹ I. V. Yaminsky,¹ E. M. Karger,² O. Yu. Frolova,² Y. L. Dorokhov² and J. G. Atabekov²

Faculty of Physics and Faculty of Chemistry¹, Department of Virology and A. N. Belozersky Institute of Physico-Chemical Biology², Moscow State University, Vorobiovy Gory, Moscow 119899, Russia

The structure of complexes formed in vitro by tobacco mosaic virus (TMV)-coded movement protein (MP) with TMV RNA and short (890 nt) synthetic RNA transcripts was visualized by atomic force microscopy on a mica surface. MP molecules were found to be distributed along the chain of RNA and the structure of MP–RNA complexes depended on the molar MP:RNA ratios at which the complexes were formed. A rise in the molar MP:TMV RNA ratio from 20:1 to 60–100:1 resulted in an increase in the density of the MP packaging on TMV RNA and structural conversion of complexes from RNase-sensitive 'beads-on-a-string' into a 'thick string' form that was partly resistant to RNase. The 'thick string'-type RNase-resistant complexes were also produced by short synthetic RNA transcripts at different MP:RNA ratios. The 'thick string' complexes are suggested to represent clusters of MP molecules cooperatively bound to discrete regions of TMV RNA and separated by protein-free RNA segments.

Intercellular translocation of plant virus genomes within virus-infected plants is mediated by virus-coded movement proteins (MPs). The best-studied viral MP is the 30 kDa protein of tobacco mosaic virus (TMV). The MPs of numerous plant viruses have the ability to bind single-stranded RNA in a sequence-independent manner. Therefore, RNA-binding ability can be regarded as an intrinsic property of MPs (for reviews, see Lucas & Gilbertson, 1994; Ghoshroy et al., 1997; Lazarovitz & Beachy, 1999; Citovsky, 1999; Tzfira et al., 2000). It has been proposed that TMV MP and genomic RNA form extended linear ribonucleoprotein (RNP) complexes that are targeted to and translocated through plasmodesmata in infected cells. The long unfolded 1.5–2.0 nm diameter MP–RNA complexes formed in vitro were visualized by means of electron microscopy (Citovsky et al., 1992). These observations suggested that TMV MP binds RNA cooperatively so that some RNA molecules are fully coated with MP while other RNA molecules remain free of protein. Alternatively, clusters of MP molecules could bind cooperatively to discrete regions of RNA, with the clusters being separated by protein-free segments of TMV RNA. Apparently, this latter type of MP–RNA interaction is favoured by subsaturating MP:RNA ratios (Li & Palukaitis, 1996; Karpova et al., 1997). MP binding to short single-stranded RNA molecules is highly cooperative and is sequence non-specific, with a minimal binding site of 4–7 nt per 30 kDa MP molecule and an interbase separation of 0.53 nm per base in the unfolded MP–RNA complex (Citovsky et al., 1990).

Atomic force microscopy (AFM) (Binnig et al., 1986) is a high-resolution technique which allows visualization of proteins (Kiselyova & Yaminsky, 1999), nucleic acids and nucleoprotein complexes (Drygin et al., 1998; Lyubchenko et al., 1993, 1995; Valle et al., 1996; Shlyakhtenko et al., 1998; Yodh et al., 1999; Jafri et al., 1999). To our knowledge this work presents the first report of examination of the TMV MP–RNA RNP complexes by means of AFM.

Virus and RNA. TMV U1 was isolated from systemically infected Nicotiana tabacum L. cv. Samsun plants as described previously (Karpova et al., 1997).

Expression and purification of (His)₆-tagged TMV MP. Expression in Escherichia coli and purification of the 30 kDa (His)₆-tagged TMV MP was carried out as described by Karpova et al. (1997). The purity of all the MP preparations isolated from E. coli by metal chelate affinity chromatography was verified by SDS–PAGE.

TMV MP–RNA complex formation. MP in twice-distilled water (0.1 µg/µl) was incubated with TMV RNA or short synthetic RNA transcripts on ice for 5 min at different molar MP:RNA ratios. In some cases the complexes were treated with RNase A (20 min at room temperature with 0.5 µg of RNase). RNA transcripts were obtained by in vitro transcription with T7 polymerase (Ribomax kit, Promega) of the linearized plasmid that contained an 890 nt artificial construct. The construct was cloned into pBluescript SK+ vector (Stratagene).

AFM measurements. For AFM measurements, the preparations were deposited onto substrates of freshly cleaved mica, incubated for 5 min, rinsed with distilled water and dried in airflow. Observations were with a Nanoscope III multimode scanning probe microscope (Digital Instruments). Standard silicon 125 µm cantilevers (NanoProbe) with 300–350 kHz resonant frequency were used in all the experiments carried out in tapping mode. For image processing, user-friendly software Femtoscan 001 (Filonov & Yaminsky, 1997) was employed.

Sedimentation analyses. These were done in a Spinco Model E analytical ultracentrifuge equipped with absorption optics.

In preliminary experiments, AFM images of purified MP and protein-free TMV RNA molecules were obtained on mica (not presented). The height of the majority of MP molecules was 1.8–2.2 nm, the apparent diameter measured at half of the height being 10–12 nm. It should be noted that in AFM images horizontal dimensions of objects are always overestimated due to the well-known effect of the geometry of the AFM tip, which has finite dimensions. Using the method described by Stemmer & Engel (1990) the lateral dimensions of MP molecules could be estimated as 2.4–3.3 nm (taking the radius of the AFM tip to be 10 nm). The height of the particles observed in the image was not uniform and varied from 1.8 to 6.0 nm. Presumably, the MP preparations were not homogeneous and the smaller particles of 1.8–2.2 nm in height correspond to monomers, whereas the bigger ones represent dimers and oligomers of the MP. At a concentration of 0.1 µg/µl, about 50–60 % of the preparation was represented by the putative monomers. This is consistent with the results of sedimentation analyses of MP preparations, which revealed a major component with s_20,w = 2.3S, presumably corresponding to monomer, and smaller amounts of MP aggregates (s_20,w = 4.1S).

MP–TMV RNA complexes formed at different molar MP:RNA ratios were visualized by AFM. A mica surface is negatively charged in distilled water, whereas the MP is positively charged under these conditions. Therefore, we used freshly cleaved unmodified mica for imaging the MP–RNA complexes. In the first experiments the complexes produced at a molar MP:TMV RNA ratio of 20:1 (corresponding to 320 nt per MP molecule) were examined. In a previous study of TMV MP–RNA complex formation using a nitrocellulose membrane filter binding assay it was reported that at an MP:RNA molar ratio of 25:1 virtually all the RNA remained unbound, whereas at a ratio of 50:1 about 60 % of the TMV RNA was involved in complex formation (Karpova et al., 1997). However, we show here by AFM that some MP–RNA complexes were formed even at a molar MP:RNA ratio as low as 20:1 (Fig. 1 a, b). The structures presented in Fig. 1 are referred to as 'beads-on-a-string', with the apparent dimensions of the 'beads' identical to those of individual MP molecules in control MP preparations. Fig. 1 shows that the MP molecules are distributed along the RNA chain with detectable minimal distances between the centres of two vicinal resolved globules ('beads') of about 15 nm, although the distance between the neighbour globules varied over a wide range. Apparently, the MP molecules observed as vicinal 'beads' in this type of MP–RNA complex are in fact separated by protein-free RNA segments of about 12 nm (Fig. 1 a, b). The majority of the globules appear to represent individual MP molecules bound to RNA independently of each other and separated by protein-free RNA segments of varying length (Fig. 1 c). Not infrequently these RNA segments could be revealed by AFM (arrows in Fig. 1 a, b). These results imply that at a molar MP:RNA ratio as low as 20:1 the binding of MP to RNA is not cooperative, i.e. the MP molecules do not interact with each other upon RNP complex formation. The larger globules (presumably dimers and/or trimers) made up not more than 30 % of the total amount of RNA-bound MP. Individual TMV RNA molecules free of MP could not be detected in these samples under conditions used to visualize MP–RNA complexes since a special cationic treatment of the mica surface is needed for RNA immobilization. Thus, the presence of MP-free RNA molecules in the MP–RNA mixture is very likely.

Fig. 1. (a, b) AFM images of the 'beads-on-a-string'-type MP–TMV RNA complexes formed at a molar MP:RNA ratio of 20:1 (corresponding to about 300 nt per 1 MP molecule). Arrows indicate the MP-free regions of RNA. (c) Simplified model of the MP–TMV RNA complex structure and the result of deconvolution. Possible protein and RNA positions are indicated by circles and lines, respectively. The distance between the nearest neighbour MP molecules bound to RNA varied over a wide range, the minimum measured distance being about 15 nm. Scale bar in (a) corresponds to 100 nm; in (b) it corresponds to 50 nm; vertical scale bar indicates 4 nm.

According to the model presented in Fig. 1, much of the RNA molecule involved in the MP–RNA complex at an MP:RNA ratio of 20:1 is not in fact coated with MP and should be accessible to RNase attack. In agreement with this hypothesis, only individual globules were observed by AFM after RNase treatment of the MP–RNA complexes produced under the aforesaid conditions (not shown).

Next, AFM was used to characterize MP distribution along TMV RNA within complexes formed at MP:RNA ratios of 60:1 and 100:1 (corresponding to 100 and 60 nt per MP molecule, respectively). Large globular particles (about 40 nm in height and 150–200 nm in diameter) with unresolved structure were revealed in AFM images after the MP–RNA mixture was incubated on ice (data not shown). However, the globules formed at the molar ratio of 60:1 could be unfolded by 40 min incubation at room temperature, producing linear structures with several bends and approximately constant lateral dimensions (Fig. 2 a, b). This type of complex will be referred to as a 'thick string' particle. Unlike the 'beads-on-a-string' formed at a molar ratio of 20:1 the MP monomers are evidently packed more densely in the 'thick string'. The length of 'thick string' complexes could not be determined precisely because of their tendency to form linear aggregates. However, the apparent width of the complexes measured at half height (about 15–18 nm) exceeded that of a single MP molecule (10–12 nm) and the height of the complex (2.5–3.5 nm) was several times that of protein-free RNA (i.e. 0.3–0.5 nm, according to Drygin et al., 1998) and 1.5–2.0 times higher than that of individual MP molecules. In most images quasi-periodic variations of height were observed within the structure of the 'thick string' complexes (Fig. 2 a, b), suggesting that they consist of tightly packed blocks, as clearly seen on the cross-section of the AFM image (Fig. 2 c). It should be emphasized that the distance between the centres of blocks varied greatly (from 22 to 50 nm). Therefore, the period of 25.5 nm revealed in Fig. 2(c) is not a universal parameter of 'thick string' complexes, but characterizes only the dimensions of cross-section made along a particular randomly selected region of the complex. The segments with constant height and width (marked by arrows in Fig. 2 a) might consist of more densely packed molecules which were not resolved by AFM. The presence of bends in the 'thick string' complexes (Fig. 2) can be explained by taking into account the fact that an AFM image is a two-dimensional projection of a three-dimensional conformation of the complex. Presumably, the segments densely covered with MP molecules acquire rigidity and are seen as practically straight lines. Most probably, bending occurs at sites where RNA is free of MP.

Fig. 2. (a, b) AFM images of the 'thick string' type of MP–TMV RNA complexes formed at a molar MP:RNA ratio of 60:1 (about 100 nt per 1 MP molecule). (c) Cross-section made along the marked line in (b) illustrates the height of the complex and periodical height variations. Triangles in (b) indicate the position of two dotted vertical lines drawn in (c). Scale bar in (a) corresponds to 200 nm; in (b) it corresponds to 100 nm; vertical scale bar indicates 6 nm.

As mentioned above, the TMV RNA was accessible to RNase attack in the 'beads-on-a-string'-type complexes. It seemed likely, however, that in the 'thick string' complexes formed at higher MP:RNA ratios the TMV RNA would be more resistant to RNase. Indeed, RNase treatment of such complexes resulted in production of linear segments 150–350 nm long (data not shown). Assuming that the interbase separation in MP–RNA complexes is 0.53 nm (Citovsky et al., 1992), the RNase-resistant TMV RNA fragments in the 'thick string' complexes would contain about 300–650 nt, which is reasonably close to the data of Karpova et al. (1999), who found that RNA fragments of about 500 nt could be isolated after RNase treatment of MP–TMV RNA complexes formed at the molar MP:RNA ratio of 100:1. It is possible that the discrete segments of TMV RNA protected from RNase correspond to the segments of the 'thick string' complexes where the MP molecules were densely packed within discrete clusters along the RNA chain.

In further experiments, short (890 nt) synthetic RNA transcripts were used instead of TMV RNA in order to examine the structure of short MP–RNA complexes produced at non-saturating MP:RNA ratios corresponding to 80 nt, 30 nt and 20 nt per MP molecule (molar MP:RNA ratios of 10:1, 30:1 and 40:1, respectively). According to Citovsky et al. (1990), most RNA transcripts should be fully coated with MP under such conditions, although some molecules remain entirely free of protein and, as was mentioned above, cannot be visualized by AFM under the conditions used. Fig. 3 (a–c) shows the complexes formed by the 890 nt transcripts with TMV MP. The mean length of these complexes was about 350 nm, suggesting that they consist of about 130 individual MP molecules with average lateral dimensions of about 3 nm (see above). According to Citovsky et al. (1990), each MP molecule binds 7 nt in the unfolded MP–RNA complex produced in vitro, which is consistent with a packing density of about 130 MP molecules per 890 nt. It is worth mentioning that the structure of the complexes produced at all three molar MP:RNA ratios was identical and indistinguishable from that of 'thick string'-type complexes described above, i.e. the centre-to-centre distance (22–50 nm) between the blocks, the height (2.5–3.5 nm) and the diameter measured at half height (15–18 nm) corresponded to the parameters of the 'thick strings' formed by the MP and TMV RNA (Fig. 2). It is important to note that these complexes were resistant to RNase attack (not shown). It can thus be suggested that the RNP particles produced by short RNA transcripts and MP structurally mimic the clusters of densely packed MP molecules cooperatively bound to discrete regions of full-length TMV RNA in the 'thick strings'. Interestingly, we did not detect complexes of the 'beads-on-a-string' type in experiments with short RNA transcripts.

Fig. 3. AFM images of complexes formed by TMV MP with the 890 nt RNA transcripts at molar ratios of about: (a) 10:1 (80 nt per 1 MP molecule), (b) 30:1 (30 nt per 1 MP molecule), and (c) 45:1 (20 nt per 1 MP molecule). Arrows in (a)–(c) indicate selected positions of 'thick string' structures in RNP complexes produced at different MP:RNA transcript molar ratios. (d) Possible AFM-based three-dimensional model of the 'thick string' complex structure. The MP molecules are organized in a helix-like array on RNA (vertical solid line); the blocks visualized by AFM represent the pitches with size (L) varying from 22 to 45 nm. Scale bars correspond to 200 nm; vertical scale bars indicate 4 nm.

A possible three-dimensional model of the structural organization of the 'thick string' complexes which matches the AFM data well is presented in Fig. 3(d), in which the MP molecules are arrayed in a helix-like rather than in a linear fashion. According to this model, the blocks visualized by AFM within the structures of the 'thick string' type would represent pitches of a helix of varying size (e.g. 22 nm and 45 nm in Fig. 3 d) consisting of MP molecules tightly packed around RNA. It is apparent that individual vicinal molecules would not be resolved due to the tip-induced broadening effect described above.

It is believed that MP–TMV RNA complexes may represent cell-to-cell intermediates (transport form) of virus infection (for recent review, see Tzfira et al., 2000). It is widely accepted that TMV MP binding to RNA is cooperative, i.e. the MP molecules interact with each other upon RNP formation. The results of AFM analysis taken together with the results of nitrocellulose membrane filter binding assays (Karpova et al., 1999) indicate that at subsaturating MP concentrations at molar MP:RNA ratios of 60–100:1 only clusters of MP molecules can be bound cooperatively to discrete regions of RNA, producing the 'thick string' structures. The MP molecules in 'thick string' particles are packed tightly enough to protect RNA from RNase attack. It is tempting to suggest that the MP–RNA complexes produced in vivo are formed under saturating MP concentrations and therefore represent a continuous extended RNP with the 'thick string' structure.

The real dimensions of unfolded MP–RNA visualized by AFM (height of the complex 2.5–3.5 nm) and by electron microscopy (1.5–2.5 nm in diameter) (Citovsky et al., 1992) are of the same order of magnitude. Taking into account that the permeability of the plasmodesmata is increased by TMV MP from 0.75–1.0 kDa to 20.0 kDa (Wolf et al., 1989; Waigmann et al., 1994), the modified plasmodesmata correspond to a dilated channel diameter of 5–9 nm. In other words, the extended complexes with a diameter between 2.5 and 3.5 nm are potentially suitable for trafficking through the MP-modified plasmodesmata.

Recent studies indicate that TMV MP and/or MP–RNA complexes interact with cytoskeleton elements (Heinlein et al., 1995; McLean et al., 1995) and with the endoplasmic reticulum (Heinlein et al., 1998). Consequently, both intracellular and cell-to-cell translocation of the viral genome may involve the interaction of MP–RNA complexes with different cellular structures, including cytoskeleton elements, endoplasmic reticulum and plasmodesmata.

The authors thank Drs Yu L. Lyubchenko and Yu F. Drygin for helpful discussions. This work was partly supported by the Russian Foundation for Basic Research (grant 00-07-90016, 00-04-55020), Russian Ministry of Science and Technology, Program Surface Atomic Structures (grant N 1.11.99), NATO Science Program (Linkage Grant LST.CLG.975.161) and by the Fogarty International Center grant 1 R03 TW01239-01.

Binnig, G., Quate, C. F. & Gerber, C. (1986). Atomic force microscope. Physical Review Letters 56, 930–933.

Citovsky, V. (1999). Tobacco mosaic virus: a pioneer of cell-to-cell movement. Philosophical Transactions of the Royal Society of London B Biological Sciences 354, 637–647.

Citovsky, V., Knorr, D., Schuster, G. & Zambryski, P. (1990). The P30 movement protein of tobacco mosaic virus is a single strand nucleic acid binding protein. Cell 60, 637–647.

Citovsky, V., Wong, M., Show, A., Prasad, B. V. & Zambryski, P. (1992). Visualization and characterization of tobacco mosaic virus movement protein binding to single-stranded nucleic acids. Plant Cell 4, 397–411.

Drygin, Yu. F., Bordunova, O. A., Gallyamov, M. O. & Yaminsky, I. V. (1998). Atomic force microscopy examination of TMV and virion RNA. FEBS Letters 425, 217–221.

Filonov, A. S. & Yaminsky, I. V. (1997). SPM Control and Image Processing Software. Chichester: Advanced Technologies Center, Moscow.

Ghoshroy, S., Lartey, R., Sheng, J. & Citovsky, V. (1997). Transport of proteins and nucleic acids through plasmodesmata. Annual Review of Plant Physiology and Plant Molecular Biology 48, 27–50.

Heinlein, M., Epel, B. L., Padgett, H. S. & Beachy, R. N. (1995). Interaction of tobamovirus movement proteins with the plant cytoskeleton. Science 270, 1983–1985.

Heinlein, M., Padgett, H. S., Gens, J. S., Pickard, B. G., Casper, S. J., Epel, B. L. & Beachy, R. N. (1998). Changing patterns of localization of tobacco mosaic movement protein and replicase to the endoplasmic reticulum and microtubules during infection. Plant Cell 10, 1107–1120.

Jafri, S., Evoy, S., Cho, K., Craighead, H. G. & Winans, S. C. (1999). An Lpr-type transcriptional regulator from Agrobacterium tumefaciens condenses more than 100 nucleotides of DNA into globular nucleoprotein complexes. Journal of Molecular Biology 288, 811–824.

Karpova, O. V., Ivanov, K. I., Rodionova, N. P., Dorokhov, Yu. L. & Atabekov, J. G. (1997). Nontranslatability and dissimilar behavior in plants and protoplasts of viral RNA and movement protein complexes formed in vitro. Virology 230, 11–21.

Karpova, O. V., Rodionova, N. P., Ivanov, K. I., Kozlovsky, S. V., Dorokhov, Yu. L. & Atabekov, J. G. (1999). Phosphorylation of tobacco mosaic virus movement protein abolishes its translation repressing ability. Virology 261, 20–24.

Kiselyova, O. I. & Yaminsky, I. V. (1999). Proteins and membrane-protein complexes. Colloid Journal 61, 1–19.

Lazarovitz, V. & Beachy, R. N. (1999). Viral proteins as probes for intracellular and intercellular tracking in plants. Plant Cell 11, 535–548.

Li, Q. & Palukaitis, P. (1996). Comparison of the nucleic acid- and NTP-binding properties of the movement protein of cucumber mosaic cucumovirus and tobacco mosaic tobamovirus. Virology 216, 71–79.

Lucas, W. J. & Gilbertson, R. L. (1994). Plasmodesmata in relation to viral movement within leaf tissues. Annual Review of Phytopathology 32, 387–411.

Lyubchenko, Yu. L., Shlyakhtenko, L. S., Harrington, R. E., Oden, P. I. & Lindsay, S. M. (1993). Atomic force microscopy of long DNA: imaging in air and under water. Proceedings of the National Academy of Sciences, USA 90, 2137–2140.

Lyubchenko, Yu. L., Jacobs, B. L., Lindsay, S. M. & Stasiak, A. (1995). Atomic force microscopy of nucleoprotein complexes. Scanning Microscopy 9, 705–727.

McLean, B. G., Zupan, J. & Zambryski, P. (1995). Tobacco mosaic movement protein associates with the cytoskeleton in tobacco cells. Plant Cell 7, 2101–2114.

Shlyakhtenko, L. S., Potaman, V. N., Sinden, R. R. & Lyubchenko, Yu. L. (1998). Structure and dynamics of supercoil-stabilized DNA cruciforms. Journal of Molecular Biology 280, 61–72.

Stemmer, A. & Engel, A. (1990). Imaging biological macromolecules by STM: quantitative interpretation of topographs. Ultramicroscopy 34, 129–140.

Tzfira, T., Rhee, Y., Chen, M.-H., Kunik, T. & Citovsky, V. (2000). Nucleic acid transport in plant–microbe interactions: the molecules that walk through the walls. Annual Review of Microbiology 54, 187–219.

Valle, M., Valpuesta, J. M., Carrascosa, J. L., Tamayo, J. & Garcia, R. (1996). The interaction of DNA with bacteriophage 29 connector: a study by AFM and TEM. Journal of Structural Biology 116, 390–398.

Waigmann, E., Lucas, W., Citovsky, V. & Zambrysky, P. (1994). Direct functional assay for tobacco mosaic virus cell-to-cell movement protein and identification of a domain involved in increasing plasmodesmal permeability. Proceedings of the National Academy of Sciences, USA 91, 1433–1437.

Wolf, S., Deom, C. M., Beachy, R. N. & Lucas, W. Y. (1989). Movement protein of tobacco mosaic virus modifies plasmodesmal size exclusion limit. Science 246, 377–379.

Yodh, J. G., Lyubchenko, Y. L., Shlyakhtenko, L. S., Woodbury, N. & Lohr, D. (1999). Evidence for nonrandom behavior in 208–12 subsaturated nucleosomal array populations analyzed by AFM. Biochemistry 38, 15756–15763.

JGV Direct table of contents

© 2001 SGM

This article is now available in the June 2001 print issue of JGV (vol. 82, 1503–1508). The complete issue of the journal may be seen in electronic form on JGV Online.