A Single Amino Acid Change in Turnip Crinkle Virus Movement Protein p8 Affects RNA Binding and Virulence on Arabidopsis thaliana

Abstract

Comparison of the symptoms caused by turnip crinkle virus strain M (TCV-M) and TCV-B infection of a resistant Arabidopsis thaliana line termed Di-17 demonstrates that TCV-B has a greater ability to spread in planta. This ability is due to a single amino acid change in the viral movement protein p8 and inversely correlates with p8 RNA binding affinity.

The mechanisms by which viruses move in plants are not yet well understood. It has long been known that viral proteins are essential to this process, and in the past few years rapid progress has been made in determining their functions (for recent reviews, see references 3, 9, 10, 11, 25, and 26). Cell-to-cell movement of turnip crinkle virus (TCV), a member of the carmovirus group, requires the presence of two small proteins, p8 and p9 (12), and in some hosts the coat protein is also required (12, 19). Currently, the process by which TCV moves from cell to cell is unknown, and the biochemical functions of the proteins involved have not been elucidated.

An Arabidopsis thaliana line termed Di-17 that uniformly developed necrotic lesions when inoculated with the M strain of TCV has been derived from ecotype Di-0 (7). Viral RNA was restricted to these lesions and, for the most part, the remainder of the Di-17 plants exhibited no further symptoms. Therefore, these lesions appear to be part of a resistance-associated phenomenon known as the hypersensitive response (HR). Despite the development of an HR, a small percentage (0 to 25%) of Di-17 plants developed systemic disease symptoms 1 to 3 weeks postinoculation (p.i.), indicating that the virus had spread to the uninoculated portions of the plant (7). TCV-M is associated with a virulent satellite (Sat C), whose presence intensifies the symptoms developed by turnip and susceptible ecotypes of Arabidopsis (1, 21, 27, 28). This effect appears to be dependent on the TCV coat protein (16, 17).

TCV-B is more virulent than TCV-M on Di-17 plants.

To determine if the spread of TCV-M in Di-17 plants is influenced by the presence of Sat C, the symptoms exhibited by TCV-M-infected plants were compared with those produced by inoculation with TCV-B, a closely related strain of TCV (2, 23) that does not carry Sat C. All inoculations were carried out as previously described (7). In three independent experiments, a minimum of 23 Di-17 plants were inoculated with virions of either TCV-M or TCV-B. All of the plants developed lesions synchronously at 3 days p.i. By 3 weeks after inoculation with TCV-M, 23% of the plants (average of three experiments) showed mild systemic symptoms. However, 69% of the plants inoculated with TCV-B developed systemic disease symptoms, a threefold increase over the percentage of plants inoculated with TCV-M. These symptoms appeared at approximately the same time and were of similar severity as those caused by TCV-M (i.e., curling of the bolt and vein clearing of the cauline leaves; data not shown). The mild symptoms correlated with the presence of viral RNA (data not shown).

Though TCV-B is not associated with Sat C, it contains a defective interfering (DI) RNA, DI RNA G, which increases the disease severity on susceptible cruciferous plants (20). To eliminate effects due to the different small RNAs, we obtained the genomic clones for each virus (2, 23). These were used to produce genomic RNA in vitro (13, 14), which was passaged through turnip to obtain large amounts of highly infectious viral RNA. This RNA was normalized by visualization on ethidium-bromide-stained gels (data not shown); this analysis also demonstrated the absence of small DI and Sat RNAs (data not shown).

Following inoculation with the Sat- and DI-free forms of the TCV RNAs obtained from infected turnip, all of the Di-17 plants developed lesions. As was previously observed with the virions, there was a large disparity in the number of plants that developed systemic disease symptoms. RNA synthesized from TCVms (the TCV-M genomic clone) produced systemic disease symptoms in 16% of the plants, while that from pT1d1 (the TCV-B genomic clone) caused disease in 53%. These disease symptoms have been correlated with the presence of TCV RNA (data not shown). Because this disparity in systemic infection was observed in the absence of symptom-altering Sat and DI RNAs, the high level of virulence of TCV-B appears to be due to genomic differences between the two viruses.

The movement protein domain of TCV-B causes larger lesions and increased virulence.

The genomes of TCV-B and TCV-M have previously been cloned and sequenced and shown to contain 16 nucleotide differences which are scattered throughout the genome (Fig. 1A and references 2, 6, and 23). Of these 16 nucleotide changes, 4 result in changes at the amino acid level, 1 results in changes in the replicase coding sequences, 1 results in changes in p8, and 2 result in changes in the coat protein.

FIG. 1.

Structure of TCV and chimeric genomes and their effects on lesion size and virulence. (A) Genetic organization of TCV. The positions and functions of the TCV open reading frames are diagrammed. The boxes below represent TCV-M and TCV-B sequences. The vertical bars between the boxes denote nucleotide changes that are silent at the protein level, while the asterisks represent nucleotide changes that result in different amino acids in the proteins. The locations of the restriction enzyme sites used to produce the chimeric genomic clones are shown. The sizes of the fragments produced by these enzymes are noted at the bottom of panel A. (B) Sizes of lesions produced by chimeric genomes. The genomes are represented by a three-letter code wherein the first letter represents the origin of the replicase region, TCV-B (B) or TCV-M (M), the second letter represents the origin of the p8 and p9 regions, and the third letter represents the origin of the coat and 3′ end. The lesions produced by all possible combinations of the three TCV regions were measured with verniers at 3 (hatched bars) and 6 (solid bars) days p.i. Each bar on the graph represents the average of 20 lesions produced by each chimeric genome on several plants. The standard deviations (error bars) are noted on the graph. As the lesions are not completely symmetrical, the smallest axis was chosen for measurement. (C) Virulence of different chimeric genomes. Plants inoculated with the chimeric genomes were observed for 3 weeks p.i. and scored for the presence of disease symptoms in uninoculated portions. For each experiment, the genome causing the highest percentage of plants with systemic disease symptoms was given the value of 100%, and the remaining genomes were expressed as relative percentages. This experiment was performed six times, and the total relative percentage was averaged over each of the six independent experiments. This procedure was necessary due to variability from experiment to experiment in the absolute numbers of plants that developed symptoms due to environmental effects. Standard deviations (error bars) are noted on the graph.

To identify the alteration(s) responsible for the difference in virulence on Di-17 plants, a series of chimeric viral genomic clones were produced. The genomic clones of TCV-B and TCV-M were divided into three regions, using internal EcoRI and BglI sites and an XbaI site from the vector at the 3′ end of the genome to produce all possible chimeras. In vitro-synthesized genomes were then passaged through turnip, and the level of viral RNA was quantitated and normalized as before.

All Di-17 plants inoculated with the chimeric RNAs developed lesions on the inoculated leaves. These lesions, however, differed markedly in size at 3 days p.i. and later at 6 days p.i. (Fig. 1B). Based on the sizes of these lesions, the chimeric genomes could be divided into two groups: all viruses containing the TCV-B movement domain produced lesions that were roughly 40% larger than those caused by chimeras containing the TCV-M movement domain.

The prevalence of systemically infected Di-17 plants was also monitored (Fig. 1C). As before, the chimeric clones could be divided into two groups, based on the percentage of inoculated plants that developed systemic symptoms. Because the actual number of plants that developed systemic symptoms varied somewhat from experiment to experiment, all data are expressed as the relative percentages of plants that developed disease symptoms. Averaging the results from six independent experiments, each of which encompassed 15 to 35 plants inoculated with RNA from either chimeric clone, indicated that RNAs containing the movement region from TCV-M induced systemic disease symptoms in 20 to 30% (relative) of the plants, while those containing the movement region from TCV-B induced systemic disease symptoms in 80 to 90% (relative) of the plants. Therefore, the increase in virulence correlates with the development of larger lesions, and both phenotypes map to the 421-nt region encoding the movement proteins (MPs) of TCV-B.

The p8 MPs of TCV-B and TCV-M differ at one amino acid.

Sequence analysis of the MP domains from TCV-B and TCV-M revealed a single nucleotide difference (data not shown; 2, 6, 23). This G→A transition changes amino acid 25 of the p8 MP from lysine in TCV-M to glutamate in TCV-B (Fig. 2). Interestingly, the surrounding peptide sequence (Fig. 2) reveals that this region of p8 is highly basic, containing either eight (TCV-B) or nine (TCV-M) basic residues in a region of 14 amino acids whose overall charge is +7 or +9, respectively. It is significant that the chimeras MBM and BMB are the equivalents of reciprocal site-directed mutants of TCV-M and TCV-B, respectively, since there is only one nucleotide difference between these and their parent clones, MMM and BBB. This has been confirmed by sequence analysis of the MP domains from MBM and BMB (data not shown).

FIG. 2.

Fine map of the region encoding the movement proteins. The 0.42-kb EcoRI-to-BglI movement region is represented by the open box, with the single nucleotide change noted by the asterisk. The open reading frames (ORFs) or portions of open reading frames contained within this region are delineated by the solid boxes and labeled. The amino acid (aa) sequence of p8 immediately surrounding the amino acid difference between strains M and B at residue 25 is diagrammed, along with the corresponding difference in net charge of this region.

The p8 proteins of TCV-M and TCV-B bind RNA differentially.

It has been hypothesized that the p8 protein binds the viral RNA genome and thereby facilitates cell-to-cell movement (12). Moreover, the highly basic nature of the region surrounding amino acid 25 suggests that it might be responsible for forming nonspecific salt bridges with the negatively charged phosphate groups on the RNA backbone. To test this hypothesis and assess the effects of the two different amino acids at position 25, the coding sequences of the TCV-B and TCV-M p8 genes (BamHI-to-HindIII fragment) were cloned into the pET15b vector (BamHI to HindIII; Novagen) and expressed as His-tagged fusions in Escherichia coli. After purification from the soluble fraction according to manufacturer’s specifications (Novagen pET system manual), the His-p8 fusion proteins were tested for their ability to bind to nucleic acid in a gel retardation assay (data not shown). His-p8 proteins appeared to bind in a highly cooperative manner to the 5′ 229 nucleotides of the TCV genomic RNA; the probe existed in only two detectable forms, free RNA and a completely retarded complex. Furthermore, His-p8 also bound RNA generated from the polylinker region of pBluescript KS+ (Stratagene), suggesting that there is no sequence specificity to binding (data not shown). This behavior is similar to that of many other plant virus movement proteins (see, for example, references 4, 5, 8, and 24).

While the overall binding characteristics of the His-p8B and His-p8M fusion proteins were similar with regard to substrate specificity, the relative affinities of these two proteins for RNA were different. The ability of purified p8 proteins to bind ³²P-labeled TCV RNA under increasing NaCl concentrations was measured by cross-linking analysis (4) (Fig. 3).

FIG. 3.

Cross-linking of His-p8M and His-p8B to RNA. A uniformly labeled 5′ 229-nucleotide fragment of the TCV genome was incubated with either His-p8B or His-p8M in the presence of the indicated salt concentrations. The reaction mixtures were then irradiated with UV light, treated with RNase A, and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis on a 15% gel. (A) Autoradiogram. The numbers above each lane indicate the molar NaCl concentration in each reaction mixture. The image was obtained with a digital camera. (B) Quantitation of cross-linking data. The data are the averages of results of four independent experiments. Dashed line, binding by His-p8B; solid line, binding by His-p8M. Error bars denote standard deviations. Quantitation was done by densitometry of the X-ray film.

Both proteins exhibited remarkable salt tolerance, with 60% as much binding at 0.4 M NaCl as was detected at 0.17 M NaCl. The ability of His-p8B to bind RNA, however, was more sensitive to salt concentration, showing an 80% reduction in binding at 0.6 M NaCl, while His-p8M binding was reduced only 50% (Fig. 3B). The greater salt sensitivity exhibited by His-p8B suggests that it binds RNA through an interaction with lower affinity and less stability than that observed for His-p8M.

There are several possible explanations for the differing abilities of the p8 proteins to bind RNA. If the basic region of p8 is indeed important for RNA binding, changing a basic amino acid to an acidic amino acid may weaken the overall stability of the interaction. Alternatively, this amino acid change might significantly affect protein structure. Protein structure predictions, however, show only very minor differences between p8M and p8B (15a). Further studies will obviously be necessary to determine the function of this region of the protein.

The ability of TCV-B to cause larger lesions and systemic disease symptoms in a higher percentage of infected Di-17 plants may be related to the reduced affinity of p8B for RNA. Perhaps the viral RNA genome is translocated through the plasmodesmata as a complex with p8. The rate at which this complex is dissociated might then control how rapidly new cells are infected, since the next step in viral proliferation is translation of the genome, and viral MPs can substantially block this process in vitro (15). Thus, the weaker interaction of p8B with the viral genome might increase the rate of the uncoating process relative to that of the p8M-RNA complex. If all of the subsequent steps in the infection process occurred at the same rate, this could explain the difference in the sizes of the lesions caused by the two viruses.

Alternatively, the decreased RNA binding affinity of p8B may not be responsible for the larger lesions and the greater number of Di-17 plants developing systemic disease symptoms after inoculation with TCV-B. It is possible that p8 acts as an elicitor (or avirulence factor) that induces the resistance response and that p8M is recognized by the plant either more rapidly or more effectively than p8B. This might be expected to result in an earlier induction of the plant’s defense responses, including the HR. However, there was no detectable difference in the time of lesion appearance on TCV-B- or TCV-M-infected plants. Furthermore, it was recently suggested that the TCV coat protein is responsible for elicitation of the resistance response in Di-0 plants (17, 23).

Another possible explanation is that p8B interacts more efficiently than p8M with the plasmodesmata or other cellular components involved in cell-to-cell spread. This could result in increased numbers of cells being infected by the time the HR is initiated, thus leading to larger lesions after TCV-B infection. In addition, if more cells become infected prior to the activation of plant defense responses, there is a greater probability that virions may enter the vasculature and cause a systemic infection.

These arguments suggest that the development of systemic symptoms is due to the outcome of early interactions between plant and virus. After the initial recognition of the virus by the plant, the plant responds by initiating defense responses, one of which is the HR. The combination of these responses typically results in the restriction of viral pathogens to cells within or immediately surrounding the lesion (7, 22). While the plant is initiating defense responses, the virus is replicating and spreading to the surrounding cells. If the virus is very efficient in this process, a larger number of cells will become infected by the time the defense response is effective, thus increasing the chances that a viral genome or virion will escape the defensive barrier. Conversely, if the plant establishes the defensive barrier rapidly enough, viral spread will cease. Thus, the development of systemic disease may rest on the outcome of this initial early race between the plant and the virus, which in this case appears to be influenced by the sequence of the p8 MP.