|Home | About | Journals | Submit | Contact Us | Français|
The Hsp70 homolog (Hsp70h) of Beet yellows virus (BYV) functions in virion assembly and cell-to-cell movement and is autonomously targeted to plasmodesmata in association with the actomyosin motility system (A. I. Prokhnevsky, V. V. Peremyslov, and V. V. Dolja, J. Virol. 79:14421-14428, 2005). Myosins are a diverse category of molecular motors that possess a motor domain and a tail domain involved in cargo binding. Plants have two classes of myosins, VIII and XI, whose specific functions are poorly understood. We used dominant negative inhibition to identify myosins required for Hsp70h localization to plasmodesmata. Six full-length myosin cDNAs from the BYV host plant Nicotiana benthamiana were sequenced and shown to encode apparent orthologs of the Arabidopsis thaliana myosins VIII-1, VIII-2, VIII-B, XI-2, XI-F, and XI-K. We found that the ectopic expression of the tail domains of each of the class VIII, but not the class XI, myosins inhibited the plasmodesmatal localization of Hsp70h. In contrast, the overexpression of the motor domains or the entire molecules of the class VIII myosins did not affect Hsp70h targeting. Further mapping revealed that the minimal cargo-binding part of the myosin VIII tails was both essential and sufficient for the inhibition of the proper Hsp70h localization. Interestingly, plasmodesmatal localization of the Tobacco mosaic virus movement protein and Arabidopsis protein RGP2 was not affected by myosin VIII tail overexpression. Collectively, our data implicate class VIII myosins in protein delivery to plasmodesmata and suggest that more than one mechanism of such delivery exist in plants.
It is well established that plant virus movement from cell to cell occurs via plasmodesmata, the plant-specific organelles that mediate the translocation of proteins, RNAs, and viruses between adjacent cells (3, 11, 21). All nondefective plant viruses encode one or more proteins that are required for the cell-to-cell movement. Many of these movement-associated proteins share an ability for plasmodesmatal localization in the absence of other virus-encoded products, indicating that their targeting relies on endogenous cellular pathways (5, 21). Despite extensive effort, the mechanisms of such autonomous plasmodesmatal targeting are poorly understood. One recurring theme in studies of plasmodesmatal targeting is the potential role of the endomembrane trafficking pathway, including the endoplasmic reticulum (ER) and secretory or endocytic vesicles (12, 15, 17). Another, perhaps complementary theme is the involvement of microtubular and actomyosin motility systems in the delivery of plasmodesma-associated proteins to their destination (6, 12, 15, 17, 42). Most data supporting the role of cytoskeletal pathways in protein targeting to plasmodesmata are obtained using drugs that affect the integrity of microtubules and actin microfilaments or inhibit the cytoskeleton-associated molecular motors. We also used these techniques to reveal that inhibitors of actomyosin, but not microtubular transport machineries, abolish the plasmodesmatal targeting of the Hsp70 homolog (Hsp70h), a movement-associated protein encoded by Beet yellows virus (BYV) (30).
The two principal components of actomyosin machinery are actin, which forms long actin microfilaments, and myosin, a molecular motor that is capable of moving along the microfilaments (40). The energy for the physical translocation of myosin and its cargo is provided by ATP hydrolysis catalyzed by the N-terminal motor domain that is also responsible for binding microfilaments. This domain is conserved among all known myosins (10). The C-terminal parts of myosins are more variable and contain (sub)domains required for dimerization, the binding of regulatory proteins (myosin light chains), and the movement of cargo by myosins. Myosins are paneukaryotic proteins; at least 24 distinct myosin classes have been identified, with functions varying from muscle contraction to organelle transport to cell motility to host cell invasion by parasites (10). Higher plants typically possess 10 to 20 individual myosins that belong to classes VIII and XI, each of which is evolutionarily related to animal and fungal class V myosins (10, 13, 33). Although the biophysical properties of the class XI myosins suggest their involvement in extremely rapid and unceasing intracellular trafficking of plant organelles (37, 39), the specific functions of individual myosin genes have not been established. Likewise, subcellular localization studies have implicated class VIII myosins in cell plate formation and plasmodesmatal function (32, 34), but no experimental support of this hypothesis has been provided so far. It is also not known which, if any, myosins are required for the plasmodesmatal targeting of viruses or their movement-associated proteins.
BYV is a filamentous virus with a 15.5-kb, positive-strand RNA genome that has emerged recently as an attractive model for investigating the mechanisms of viral transport (9). The cell-to-cell movement machinery of BYV involves five proteins, among which only one, an ER-localized transmembrane protein, p6, fits the definition of a dedicated movement protein (MP) (1, 28, 29). The remaining four movement-associated proteins are integral virion components. Three of them, including Hsp70h, form a narrow tail at the virion end that encapsidates the 5′ extremity of the genome (27). Yet an additional tail protein, p20, is specifically required for viral long-distance transport via the plant vascular system (31). Because tail formation is dispensable for genome protection but is essential for virus transport between cells and organs of the infected plant, the tail can be considered a specialized transport device (2, 27, 31). Interestingly, Hsp70h is the only BYV movement-associated protein that was found in plasmodesmata upon ectopic expression or virus infection (22, 30). As indicated above, autonomous plasmodesmatal targeting of Hsp70h requires the actomyosin motility system (30).
In this work, we show that interference with the cargo-binding activity of the class VIII but not the class XI myosins abolishes the plasmodesmatal localization of Hsp70h. Our results suggest that class VIII myosins function in protein delivery to plasmodesmata and can be used by viruses for the needs of their intracellular and perhaps intercellular translocation.
A conserved region in myosin mRNAs was reverse transcribed and PCR amplified using degenerate primers as described previously (4). The resulting PCR products were cloned into pGEM-T vector (Promega). Nucleotide sequencing of the 300 individual clones yielded six distinct variants with more than 10 identical clones for each of these variants (2a). The cDNA cloning of the 5′- and 3′-terminal regions of the corresponding mRNAs to obtain complete myosin cDNAs was done using the FirstChoice RNA ligase-mediated rapid amplification of cDNA ends kit (Ambion). Once again, at least 10 individual clones were sequenced to obtain the consensus nucleotide sequence for each fragment. The nucleotide and amino acid sequences corresponding to each of the six complete myosins of N. benthamiana were used as a query to find the highest-scoring BLASTN and BLASTP hits among Arabidopsis thaliana myosins to determine likely orthologous relationships. Myosin domain maps (see Fig. Fig.1)1) were obtained using the program SMART 5 (18).
The cDNAs encoding complete or partial variants of each myosin were amplified by reverse transcription-PCR and cloned into binary vector pCB302, modified to accommodate a protein expression cassette (26). A sequence encoding a triple hemagglutinin (HA) tag has been inserted into this cassette using NcoI and AvrII restriction endonuclease sites, while the myosin-encoding cDNAs were inserted downstream from the HA coding sequence using AvrII and XbaI sites (primer sequences are available upon request). The resulting clones were sequenced to confirm the lack of mutations introduced during the cloning. The exact borders of the coding sequences for each of the 18 myosin variants used in this work are shown in Table Table11.
Agrobacterium tumefaciens strain C58 GV2260 cells were transformed with each of the binary expression vectors, and the resulting bacteria at an optical density at 600 nm of 0.2 to 0.5 were used for N. benthamiana leaf infiltrations (30). The expression of the full-size and truncated myosin variants was assayed by immunoblotting using a rat anti-HA monoclonal antibody (Roche) ~24 h postinfiltration. A mouse monoclonal green fluorescent protein (GFP)-specific antibody (Roche) was used for immunoblot detection of the GFP-tagged proteins. The binary vectors for the ectopic expression of GFP- or monomeric red fluorescent protein (mRFP)-tagged variants of BYV Hsp70h, as well as Hsc70 of A. thaliana and Tobacco mosaic virus (TMV) MP-GFP variants, were generated as previously described (30). The GFP-tagged A. thaliana protein RGP2 was expressed using binary vector pBinAtRGP2GFP, described previously (35). To coexpress these flurophore-tagged proteins with myosin variants, the corresponding bacterial strains were mixed prior to agroinfiltration.
Confocal laser scanning microscopy was done using a Zeiss LSM 510 META microscope fitted with the following configurations of excitation and emission filters, respectively: 488 and 508 nm for GFP, 558 and 583 nm for mRFP, and 513 and 527 nm for yellow fluorescent protein.
The GenBank accession numbers for the sequences encoding the N. benthamiana myosins described herein are as follows: XI-2, DQ875135; XI-F, DQ875136; XI-K, DQ875137; VIII-1, DQ875138; VIII-2, DQ875139; and VIII-B, DQ875140.
Because N. benthamiana is a convenient experimental host for BYV (8) that also provides a facile system for transient protein expression, we set out to clone and sequence myosin cDNAs from this plant species. Altogether, six complete myosin cDNAs were characterized (2a). Amino acid sequence comparisons revealed that the isolated cDNAs encoded three class XI and three class VIII myosins. Database searches identified apparently orthologous relationships between N. benthamiana myosins and A. thaliana class XI myosins previously named MYA2, XI-F, and XI-K, as well as class VIII myosins designated ATM1, ATM2, and VIII-B (33). In an attempt to make the myosin nomenclature more systematic yet related to that for A. thaliana myosins, we named the N. benthamiana myosins XI-2, XI-F, XI-K, VIII-1, VIII-2, and VIII-B, respectively. As shown in Fig. Fig.1,1, the domain architecture of these myosins is typical for plant myosins and includes a large, N-terminal motor or head domain with ATPase activity, several IQ motifs that bind regulatory proteins, a dimerization coiled-coil motif, and a C-terminal globular tail domain (GTD) that, in class XI myosins, also includes a “dilute” motif that is conserved between class V and XI myosins (10, 20, 25).
Because the recognition and attachment of the cargo are the functions of myosin tails (20, 25), the overexpression of the tails should saturate the tail-binding capacities of the matching cargoes and inhibit their translocation by the endogenous myosin motors. In addition, free tails may interact with the heads of cognate myosins, thus reducing the motor activity (19). This tail overexpression-based dominant negative inhibition strategy was employed to determine the potential roles of each of the six N. benthamiana myosins in the delivery of BYV Hsp70h to plasmodesmata. The myosin tails that encompassed IQ motifs, coiled-coil motifs, and GTDs (Fig. (Fig.1)1) were tagged with HA epitopes and ectopically expressed at similar levels in the leaves (Fig. (Fig.2C2C).
The localization patterns of Hsp70h in the cells overexpressing myosin tails were monitored using the Hsp70h-GFP fusion and confocal laser scanning microscopy. As described previously (30), in the absence of myosin tails, Hsp70h-GFP was localized to paired punctate bodies identified as plasmodesma-rich pit fields (Fig. (Fig.2A,2A, top panels). Similar localization was observed when the 70-kDa heat shock cognate protein (Hsc70) from A. thaliana was coexpressed with Hsp70h-GFP (Fig. (Fig.2B,2B, top panels), indicating that the ectopic expression of the recombinant protein per se does not perturb Hsp70h-GFP distribution. Furthermore, the overexpression of each of the class XI myosin tails (Fig. (Fig.2A,2A, three bottom rows) did not interfere with the plasmodesmatal targeting of Hsp70h-GFP.
Strikingly, the ectopic expression of the tails of myosins VIII-1, VIII-2, and VIII-B resulted in a pattern of Hsp70h-GFP distribution dramatically different from that occurring in the absence of myosin tails. Instead of localized punctate bodies, Hsp70h exhibited virtually uniform distribution at the cell peripheries (Fig. (Fig.2B,2B, three bottom rows). It should be stressed that such a pattern was observed in 100% of the cells that showed microscopically detectable levels of Hsp70h-GFP expression. In fact, an identical Hsp70h-GFP distribution pattern in the presence of the microfilament-disassembling drugs latrunculin and cytochalasin D and the generic myosin inhibitor 2,3-butanedione monoxime was described previously (30).
To ensure that the observed inhibition of the proper Hsp70h-GFP targeting was due to the ectopic expression of the myosin VIII tails as proteins rather than to RNA interference triggered by corresponding mRNA (3), Hsp70h-GFP was coexpressed with myosin VIII-B tails and helper component protease (HC-Pro), a potent potyviral suppressor of RNA interference that acts via the nonspecific binding of small interfering RNAs (14, 16). Because HC-Pro did not restore the plasmodesmatal localization of Hsp70h-GFP (data not shown), we concluded that the inhibitory effect of the myosin VIII tails was protein mediated. It is also important to emphasize that the ectopic expression of the myosin VIII tails did not affect the architecture of the actin cytoskeleton in any detectable way (Fig. (Fig.33).
These results were interpreted to indicate that the class VIII myosins are required for either delivering or anchoring Hsp70h to plasmodesmata. Because each of the three tested class VIII myosin tails had similar effects, it can be concluded that these myosins are at least in part functionally redundant.
To determine if the inhibitory effect of the class VIII myosin tails on Hsp70h is indeed dependent on the interference with cargo binding rather than on some fortuitous disturbance of myosin function, we systematically tested the effects of a series of six distinct domain combinations (Fig. (Fig.11).
As expected, the ectopic expression of the full-size myosin VIII-1 (Fig. (Fig.4C,4C, lane F) did not affect Hsp70h-GFP targeting to plasmodesmata (Fig. (Fig.4A,4A, top panels). Likewise, the overexpression of the motor domain (Fig. (Fig.4C,4C, lane M) did not interfere with Hsp70h-GFP localization (Fig. (Fig.4A,4A, middle panels). Furthermore, the extended configuration that included the motor domain along with IQ and coiled-coil motifs (Fig. (Fig.4C,4C, lane MC) also had no detectable effect on Hsp70h-GFP targeting to plasmodesmata (Fig. (Fig.4A,4A, bottom panels).
In addition to the entire myosin tails encompassing IQ and coiled-coil motifs and the GTD (Fig. (Fig.4D,4D, lane IQC), we examined the effects of shorter configurations that included the coiled-coil motif and the GTD (Fig. (Fig.4D,4D, lane CCC) and the GTD only (Fig. (Fig.4D,4D, lane GTD). The ectopic expression of each of these three domain configurations resulted in apparently complete abolishment of the plasmodesmatal localization of Hsp70h-GFP (Fig. (Fig.4B).4B). It should be emphasized that mapping experiments were also done using full-size myosin VIII-2, as well as its five truncated variants corresponding to those described above for myosin VIII-1. Once again, only the three GTD-containing tail variants, but not the full-size myosin VIII-2 or its motor domain, interfered with Hsp70h localization (Fig. (Fig.55).
Taken together, these results demonstrated that the GTD of the class VIII myosins is both essential and sufficient for the inhibition of Hsp70h-GFP delivery to plasmodesmata. Because the principal function of the GTD is recognizing and binding the cognate cargo, we concluded that the observed mislocalization of Hsp70h-GFP in the presence of the class VIII myosin GTD was likely due to the inhibition of the direct or indirect association of viral protein with the proper molecular motor.
To determine if myosin VIII-mediated targeting is a common mechanism whereby diverse plasmodesmatal proteins are delivered to their subcellular destination, we employed GFP fusion forms of two well-characterized proteins capable of plasmodesmatal localization, TMV MP (24, 38) and A. thaliana RGP2 (35). As shown in Fig. 6 A and C, the coexpression of either of these proteins with the myosin VIII-2 tails resulted in the proper plasmodesmatal localization. In fact, this result was not unexpected since we have previously found that none of the microfilament-disrupting drugs affect TMV MP localization to plasmodesmata (30).
To ensure that the same myosin VIII tail-expressing N. benthamiana cells were competent for the plasmodesmatal localization of MP-GFP or RGP2-GFP but incompetent for the delivery of Hsp70h-GFP to plasmodesmata, we coexpressed the Hsp70h-mRFP fusion with myosin VIII-2 tails and either MP-GFP or RGP2-GFP. Although both MP-GFP and RGP2-GFP retained their characteristic plasmodesmatal localization, Hsp70h-mRFP was uniformly distributed at the cell peripheries (Fig. 6B and D). These coexpression experiments were also done using tails of myosins VIII-1 and VIII-B (data not shown); the corresponding protein distribution patterns were identical to those shown in Fig. Fig.66 for myosin VIII-2 tails.
Collectively, these data suggest that the myosin VIII-mediated pathway of plasmodesmatal protein targeting is not universal and that at least one pathway that does not involve these myosins exists in the plant cells. It should also be emphasized that the ability of MP-GFP and RGP2-GFP to reach plasmodesmata in these experiments indicates that the expression of myosin VIII tails does not disrupt plasmodesmatal structure in a way that would inhibit an alternative pathway of protein targeting to these organelles.
The evolutionary conservation and proliferation of two plant-specific classes of myosin genes in flowering plants (2a, 10, 33) points to the functional importance of the myosin molecular motors in plant physiology. It is also well accepted that the actomyosin motility system plays a paramount role in the interior dynamics of the plant cells. Based primarily on indirect data, it was proposed previously that class XI myosins are responsible for rapid trafficking of plant organelles such as Golgi stacks, mitochondria, and peroxisomes and slower movements of the nucleus and chloroplasts (37, 41). However, the experimental data on class XI myosin functions are limited so far to the identification of the role of myosin XI-K in root hair elongation (23). Virtually no information on class VIII myosins is available, although the functional association of both actin and class VIII myosin with plasmodesmata was suggested previously (7, 34). In addition, the actomyosin system was implicated in the targeting of certain viral MPs to plasmodesmata (12, 42), as well as in the cell-to-cell movement of TMV (15).
Here, we employed dominant negative inhibition of myosin function based on the overexpression of the myosin cargo-binding domains to identify the myosins involved in the localization of the BYV Hsp70h to plasmodesmata. Our data demonstrate unequivocally that it is class VIII, but not class XI, myosin tails that interfere with proper Hsp70h targeting. Furthermore, we showed that the minimal, very C-terminal GTD of class VIII myosins was able to abolish Hsp70h localization. Because the primary function of the GTD is binding the cognate cargo, we assume that the inhibitory effects of GTDs were due to the interference of the GTDs with the direct or indirect association between class VIII myosins and Hsp70h-GFP. Our pilot experiments aimed at the coimmunoprecipitation of the Hsp70h and the GTD suggested that their interaction may be indirect and/or transient.
Interestingly, class VIII myosin-mediated targeting appears not to be the only mechanism for protein delivery to plasmodesmata. Indeed, the overexpression of the class VIII myosin tails did not disrupt the plasmodesmatal localization of either TMV MP or the A. thaliana protein RGP2. Based on the tight association of TMV MP with the ER (15, 42), it seems possible that this protein can migrate along ER tubules and reach the desmotubule, a modified outlet of the continuous ER network (21). Because RGP2 is also found in the Golgi apparatus (35), it can be assumed that this protein can arrive at the plasmodesmata via a specialized branch of a secretion route.
The mechanism by which class VIII myosins mediate Hsp70h localization to plasmodesmata is yet to be determined. One possibility is that class VIII myosin motors are specifically associated with and move along the subpopulation of the actin microfilaments anchored in the plasmodesmatal vicinity. An alternative scenario is the diffusion of Hsp70h into the cytosol, followed by class VIII myosin-assisted anchoring to plasmodesmata. This scenario would be in line with the previous suggestion of a tight association of the class VIII myosins with plasmodesmata (34). A daunting extension of this possibility would be the actual movement of myosin VIII motors along the actin microfilaments that transit through plasmodesmata and interconnect adjacent cells.
In addition to unraveling the mechanism by which class VIII myosins act to assist Hsp70h localization to plasmodesmata, the potential role of these myosins in the cell-to-cell movement of BYV and other viruses and also plant proteins and RNAs that traffic between cells via plasmodesmata needs to be investigated. However, the data presented herein support the involvement of plant-specific, class VIII myosins in processes of intercellular transport and communication that occur via plant-specific organelles, the plasmodesmata.
We thank Bernard Epel for kindly providing the RGP2-GFP binary vector. We acknowledge the Confocal Microscopy Facility of the Center for Genome Research and Biocomputing and the Environmental and Health Sciences Center at Oregon State University.
This publication was made possible in part by grant number 1S10RR107903-01 from the National Institutes of Health. The research was supported by a grant from the National Institutes of Health (GM053190) to V.V.D. D.A. was supported by Vaadia-BARD postdoctoral fellowship award no. F1-354-2004 from BARD, the United States-Israel Binational Agricultural Research and Development Fund.
Published ahead of print on 16 January 2008.