|Home | About | Journals | Submit | Contact Us | Français|
Filamentous virions of Beet yellows virus contain a long body formed by a major capsid protein and a short tail that is assembled by a minor capsid protein (CPm), an Hsp70-homolog (Hsp70h), a 64-kDa protein (p64), and a 20-kDa protein (p20). Using mutation analysis and newly developed in planta assays, here we investigate the genetic requirements for the tail assembly. We show that the inactivation of CPm dramatically reduces incorporation of both Hsp70h and p64. Furthermore, inactivation of Hsp70h prevents incorporation of p64 into virions and vice versa. Hsp70h and p64 are each required for efficient incorporation of CPm. We also show that the tails possessing normal relative amounts of CPm, Hsp70h, and p64 can be formed in the absence of the major capsid protein and p20. Similar to the tails isolated from the wild type virions, these mutant tails encapsidate the ~700 nt-long, 5’-terminal segments of the viral RNA. Taken together, our results imply that CPm, Hsp70h and p64 act cooperatively to encapsidate a defined region of the closterovirus genome.
There are two principal structural designs of the viral particles: icosahedral, which translates to roughly spherical or ellipsoidal virions, and helical, which results in flexuous filamentous or rigid rod-shaped virions (Harrison, 2001). However, there are many variations of these designs including numerous tailed bacteriophages and morphologically diverse viruses of Archaea that combine icosahedral and helical parts in their particles (Hendrix et al., 2003; Prangishvili et al., 2006). More complex tailed virions are assembled by dozens of proteins responsible for DNA encapsidation, packaging, and delivery to host cells, as well as for particle attachment to c ells and penetration of the cell walls.
A majority of the plant viruses possess simple icosahedral or helical virions that encapsidate positive-strand RNA genomes (Lazarowitz, 2001). An interesting exception to this rule is the family Closteroviridae that is characterized by the large RNA genomes and unusually long and complex filamentous virions (Dolja et al., 2006; Karasev, 2000). The morphology, composition, assembly, and function of the closteroviral virions were studied using two model viruses of a genus Closterovirus, Beet yellows virus (BYV) and Citrus tristeza virus (CTV). It was initially discovered that BYV virions possess two parts, a long, ~1,200 nm virion body made of the major capsid protein (CP) and a short, ~100 nm tail that contained the minor capsid protein (CPm) (Agranovsky et al., 1995). This principle of particle organization was also confirmed for CTV (Febres et al., 1996) and Lettuce infectious yellows virus, genus Crinivirus (Tian et al., 1999).
Later work demonstrated that in addition to CPm, BYV virions contain a virus-coded Hsp70 homolog (Hsp70h) and a 64 kDa protein (p64, see Fig. 1A for BYV genome map) (Napuli et al., 2003; Napuli et al., 2000). Comparative genomic analyses revealed that Hsp70h and p64-like proteins are conserved in all known viruses of the family Closteroviridae (Dolja et al., 2006). It was also established that p64 possess the CP-like C-terminal domain that is embedded into the virion, and the unique N-terminal domain that is exposed at the virion’s surface (Napuli et al., 2003). Interestingly, work with CTV and BYV demonstrated that Hsp70h and p64 or its CTV ortholog are each required for efficient assembly of the tailed closterovirus virions (Alzhanova et al., 2001; Napuli et al., 2003; Satyanarayana et al., 2000). Moreover, it was found that the BYV virions contain yet additional, ~20-kDa protein (p20) that is not essential for the assembly of either virion bodies or tails (Prokhnevsky et al., 2002).
The tandem of recent papers further advanced our understanding of the molecular architecture of closteroviruses. It was found that CTV CPm recognizes the packaging signal near the 5’-end of the viral RNA and is capable of encapsidating RNA regions of variable length in the absence of other structural proteins (Satyanarayana et al., 2004). Moreover, it was shown that if CPm is co-expressed with Hsp70h and p61 (p64 ortholog in CTV), the tails of apparently normal length are formed in the absence of virion bodies suggesting that Hsp70h and p61 cooperatively control the proper tail assembly. The parallel work with BYV reported the isolation, molecular composition, and complex morphology of the virion tails (Peremyslov et al., 2004). It was demonstrated that Hsp70h, p64, and p20 are integral tail components in addition to CPm. It was also revealed that the tails are narrower than the bodies (8 nm vs 12 nm), and that they exhibit a peculiar three-segment architecture with the pointed tip segment likely formed by p20 (Peremyslov et al., 2004). For both CTV and BYV, the tails were shown to encapsidate 600–700 5’-terminal nucleotides of the viral RNA. On the other hand, either CP or CPm alone were capable of encapsidating the (nearly) entire viral genome (Alzhanova et al., 2001; Satyanarayana et al., 2004).
Interestingly, each of the BYV structural proteins was also implicated in virus transport within the infected plants (Alzhanova et al., 2000; Peremyslov et al., 1999; Prokhnevsky et al., 2002). In particular, CP, CPm, Hsp70h, and p64 are each essential, but not sufficient for virus movement between cells (Fig. 1A). P20 is dispensable for local movement, but is required for long-distance transport via the phloem. These data indicated that particle assembly is a prerequisite for BYV transport, and allowed us to advance a concept of the virion tail as a specialized device that evolved to facilitate transport of the large RNA genomes of the closteroviruses (Dolja, 2003).
Further progress in the investigation of the assembly mechanisms was hindered by the requirement of the virion proteins for virus transport. Because inactivation of the structural proteins prevented systemic infection, transport-deficient mutant virions were examined in the transfected protoplasts. This approach provided only low amounts of virions sufficient to detect CP and CPm, but not Hsp70h or p64.
Here we describe the use of a new, agroinfection-based technology that allows isolation of the movement-defective virions from plants for the systematic analysis of the genetic requirements for tail assembly (Chiba et al., 2005). We demonstrate that the efficient tail assembly by CPm requires both Hsp70h and p64, and that the two latter proteins strictly require each other for the incorporation to virions. We further show that the apparently normal tails, which contain CPm, Hsp70h, and p60 and encapsidate ~700 5’-terminal nucleotides of the viral RNA can be formed in the absence of CP and p20.
The Agrobacterium-mediated transient expression of the messenger and other RNAs in plants is an acclaimed technology. Its application for launching viral replicons provided a useful alternative to mechanical or vector-assisted plant inoculation with viruses. It was found, however, that only very few cells become infected following agroinfiltration with either Tobacco mosaic virus (Marillonnet et al., 2005) or BYV (Chiba et al., 2005). The efficiency of agroinfection can be dramatically improved by co-expression of the viral suppressors of RNA silencing (Chiba et al., 2005). Throughout this study, we took advantage of this technology by launching wild type and mutant BYV replicons together with either BYV silencing suppressor p21 or Tobacco etch virus suppressor P1/HC-Pro. The leaves were collected at 8–12 days post infiltration, and the virions were isolated and analyzed using standard techniques.
In accord with CPm being a principal tail component, previous study of BYV revealed that elimination of CPm by mutations results in formation of the tailless virions, in which the genomic RNA is partially or completely encapsidated by CP (Alzhanova et al., 2001). It was not known, however, if these virions contained either Hsp70h or p64. To determine if the two latter proteins can be incorporated to virions in the absence of the functional CPm, we employed two mutant variants of BYV replicon. In the NoCPm variant, expression of CPm was prevented by elimination of the translation initiation codon AUG, while in the CPmR128D, the invariant Arg residue was replaced with Asp resulting in the assembly-incompetent CPm.
The mutant virions were isolated from agroinfiltrated leaves, and their protein composition was compared to that of the wild type virions using immunoblot analysis and the antisera specific to CP, CPm, Hsp70h, and p64 (Fig. 1B). It is important to emphasize that the amounts of the virions in the samples were equalized relative to CP to facilitate comparative analysis of the incorporation of tail proteins (Fig. 1B, row CP). As expected, CPm was undetectable in the virions formed by either NoCPm, or CPmR128D variants (Fig. 1B row CPm; Fig. 1C). Analyses for the Hsp70h and p64 revealed that these proteins were present in the isolated NoCPm virions, but their amounts were only ~ 5% of the levels found in the wild type virions (Fig. 1B, rows Hsp70h and p60). Furthermore, neither Hsp70h, nor p64 was detectable in the CPmR128D virions. These data indicate that CPm is required for the efficient incorporation of both Hsp70h and p64 to virions.
The previously described NoHsp70h mutant in which first AUG codon was inactivated, and Hsp70hΔXho mutant in which most of the Hsp70h ORF was deleted were used to determine Hsp70h function in the tail assembly (Peremyslov et al., 1999). Analyses of both mutants revealed virtually identical phenotypes with neither Hsp70h, nor p64 present in the mutant virions (Fig. 2A, rows Hsp70h and p64, respectively). Furthermore, elimination of Hsp70h affected incorporation of CPm that was reduced to ~25% of the wild type level (Fig. 2A, row CPm; Fig. 1C).
The role of p64 in the virion tail assembly was addressed using Nop64 mutant, in which p64 expression was abolished by virtue of replacing translation initiation codon, and p64R386A mutant, in which the invariant Arg in the CP-like domain was replaced with Ala (Napuli et al., 2003). Similar to analogous CPmR128D mutant, the latter p64 mutation resulted in assembly- and movement-incompetent protein. Strikingly, the phenotypes of Nop64 and p64R386A mutants were very similar not only to each other, but also to those of the Hsp70h mutants. The isolated mutant virions lacked detectable p64 and Hsp70h, while they both exhibited the reduced levels of CPm (Figs. 2B and and1C1C).
The nearly identical phenotypes of the Hsp70h- and p64-deficient mutants indicated that these proteins act in concert to facilitate incorporation of each other and CPm to functional virion tails. This interpretation was further supported by the analysis of a double NoHsp70h/Nop64 variant in which expression of both proteins was eliminated by mutations. It was found that the corresponding virions contained no Hsp70h or p64, and that the level of CPm incorporation was very similar to that of the single NoHsp70h or Nop64 mutants (Figs. 2C and and1C).1C). The requirement of Hsp70h for incorporation of p64 to virions and vice versa suggested that these two proteins physically interact and that the resulting complex assists in assembly of the tails by CPm.
To determine if the encapsidation of the 5’-terminal region of BYV genome by the tail proteins CPm, Hsp70h and p64 requires the body formation by CP or incorporation of p20, we generated a mutant BYVΔ in which CP ORF was replaced with GFP reporter, while p20 and p21 ORFs were deleted (Fig. 1A). The particles presumably formed by the three tail proteins were isolated and their protein composition was determined using immunoblotting. It was found that, similar to the wild type tails, these particles contained CPm, Hsp70h, and p64 (Fig. 3A and B). To compare the levels of Hsp70h incorporated to wild type and mutant particles, the preparations were equalized relative to CPm (Fig. 3A). It was found that the relative levels of Hsp70h in five independent isolates of mutant particles varied from 79% to 111% of the wild type level with the mean of 91%. As seen from Fig. 3B, the ratios of p64 to CPm were also similar in the mutant and wild type particles.
To determine the RNA content of the mutant particles, the hybridization analysis was done using probes specific to 5’- and 3’-terminal BYV genome regions (Fig. 3C). It was found that the mutant virions contained the 5’-terminal RNA segment with the length varying from ~650 to ~800 nts (Fig. 3C, left panel, lanes T1 and T2). Strikingly, size distribution of the encapsidated RNA was very similar to that of the virion tails isolated from wild type virions (Fig. 3C, lane Ts) (Peremyslov et al., 2004). Virtually no signal was detected using 3’-terminal hybridization probe (Fig. 3C, right panel).
The RNA encapsidated by the CP-less BYVΔ mutant was RT-PCR amplified, and the resulting cDNA was cloned and sequenced. The location of 3’-termini of the 18 sequenced cDNA fragments encompassed the area from nt 623 to 858 with the majority of cDNAs (13 out of 18) ending between nts 682 and 724 (Fig. 4).
These experiments demonstrated that the virion tails containing normal ratios of CPm to Hsp70h and p64 can be formed in the absence of CP and p20. Moreover, these mutant tails encapsidated the same, 5’-terminal, ~700 nts-long, expanse of the viral RNA, as the tails in the wild type virions.
In this work, we establish interdependence of the three tail components, CPm, Hsp70h, and p64, for the assembly of the BYV tails of the proper morphology. Furthermore, strict mutual requirement for Hsp70h and p64 for tail assembly strongly suggests that Hsp70h and p64 physically interact with each other. Given that Hsp70h is the member of a large and functionally diverse family of molecular chaperones whose action invariably involves the aid of cochaperones (Bukau et al., 2006), p64 could be considered as a virus-specific cochaperone of Hsp70h. It is well established that the cellular chaperone-cochaperone pairs only transiently interact with their client proteins. In a sharp contrast, the Hsp70h-p64 pair does not only appear to interact with its client CPm, but is incorporated into the mature virion tails as the integral component. Interestingly, we also demonstrate that none of other virion proteins, namely, CP, nor p20, is required for the formation of the tails that contain CPm, Hsp70h and p64 in normal proportions. Moreover, the resulting bodiless tails encapsidate the ~700 nt-long, 5’-terminal region of the viral RNA that is identical to that present in the wild-type virion tails (Peremyslov et al., 2004).
It appears likely that the roles of Hsp70h, p64, and CPm in the tail assembly are conserved in the viruses of a genus Closterovirus because of the largely overlapping requirements found for BYV and CTV. Indeed, CTV tails incorporate ~630 nt-long, 5’-terminal part of the viral genome with Hsp70h and p61 being both required for the formation of the tails of proper length in the absence of CP (Satyanarayana et al., 2004). It should be noted, however, that the tail RNA of CTV was mapped within the area of ~20 nucleotides, while in BYV it spanned over ~200 nucleotides (Fig. 4). This apparent difference in precision with wich the tail length is determined in two closteroviruses may be either intrinsic or simply reflect differential accessibility of the unencapsidated virion RNA to exogenous RNases upon tail isolation. The mechanism by which Hsp70h and p64 regulate the incorporation of CPm and the tail length is yet to be unraveled.
It is not known, if CTV virions incorporate a functional analog of BYV p20. Protein sequence comparisons do not reveal significant similarity between p20 and any of CTV proteins. On the other hand, putative ortholog of BYV p20, p19, is encoded in Grapevine leafroll-associated virus-2 (Zhu et al., 1998). Examination of the virion morphology using atomic force microscopy revealed that the latter virus possesses tails with the pointed tip similar to that of BYV (I.A. Andreeev, M.E. Taliansky, A.I.P., V.V.P., and V.V.D., unpublished data). It should be emphasized that p20 is not required for either encapsidation of the 5’-region of the BYV genome (this work) or the viral cell-to-cell movement (Alzhanova et al., 2000). The only function of p20 identified so far is its requirement for the efficient long-distance transport of BYV via the phloem (Prokhnevsky et al., 2002). As we proposed earlier, p20 could form a tail’s pointed tip and play a role in virion stabilization in the phloem and/or mediate virus exit from the sieve elements (Dolja, 2003).
Recent phylogenetic analyses of the family Closteroviridae suggested an evolutionary scenario according to which Hsp70h- and p64-like proteins were present in a common family ancestor, while CPm appeared later and independently in each of the three recognized genera of this family (Dolja et al., 2006). The corollary of this scenario is that originally, the virion tails could have been formed by only Hsp70h and p64 that co-evolved to interact with each other and later were involved in facilitating CPm incorporation.
Even with the recent progress in our understanding of the virion assembly of BYV and CTV, many challenging questions remain unanswered. What is the exact molecular architecture of the tails? Where are CPm, Hsp70h, and p64 located relative to each other and to tail segments? Where are the packaging signals (if any) recognized by Hsp70h, p64 and CP? Even more vexing is the question regarding the mechanism by which the virion tails facilitate BYV cell-to-cell movement. The merely genome protective role appears unlikely given that CP (Alzhanova et al., 2001) and, in the case of CTV, CPm (Satyanarayana et al., 2004) may encapsidate the entire genome. Furthermore, it is likely that the virion tails participate in the guiding the virions to plasmodesmata. This notion is supported by the autonomous targeting of Hsp70h to plasmodesmata in association with actomyosin motility system (Prokhnevsky et al., 2005). In addition, the lesser diameter of the tails compared with the virion bodies may suggest a role in direct virion insertion to plasmodesmata. However, these hypotheses still await the experimental validation, e.g., the genetic separation of the functions of tail proteins in assembly and cell-to-cell movement.
Recent analysis of the potyviruses using atomic force microscopy revealed the presence of the tail-like structures at one end of the virions corresponding to the 5’-terminus of the RNA genome (Torrance et al., 2005). Furthermore, it was demonstrated that the viral helper component-proteinase is present in potyviral tails. Because helper component-proteinase is required for viral transport within infected plants (Cronin et al., 1995), potyviral tails could be functionally analogous to those of closteroviruses. It seems likely that further investigation of the helical virions of plant viruses that were thought to contain a single capsid protein will uncover more examples of the tail-like structures that evolved to aid a directed, 5’-end-first, virus transport (Atabekov et al., 2000; Lough et al., 2006).
All variants of the BYV cDNAs used in this study were the derivatives of the binary vector p35SBYV-GFP that contained the full-length BYV genome tagged via insertion of the GFP reporter (Prokhnevsky et al., 2002). Generation of the mutations in the BYV genes encoding CPm, Hsp70h, and p64 in the background of pBYV-GFP were described previously (Alzhanova et al., 2000; Napuli et al., 2003; Peremyslov et al., 1999). The corresponding mutant cDNA fragments were transferred to p35SBYV-GFP using Sna BI and Bst EII restriction endonuclease sites. A double NoHsp70h/Nop64 mutation was generated by first combining the individual mutations in the pBYV-GFP and then transferring the Sna BI-Bst EII fragment to p35SBYV-GFP. The BYVΔ mutant (Fig. 1A) lacking the genes encoding CP, p20, and p21 was generated by deletion of the Fse I-Mlu I region in p3’BYV-GFP (Peremyslov et al., 1999) followed by transfer of the Bam HI-Eco RI fragment to p35SBYV-GFP.
The binary vectors were mobilized to Agrobacterium tumefaciens strain C58 by electroporation. The bacterial suspensions at OD600 1.0 were used to infiltrate Nicotiana benthamiana plants (6 to 8-leaf stage). These suspensions contained two agrobacterial strains engineered to launch BYV genome and either BYV or Tobacco etch virus RNA silencing suppressors, p21 and P1/HC-Pro, respectively, in the ratio of 1:5 (Chiba et al., 2005). Plants were propagated in the greenhouse for two weeks, the infiltrated leaves were harvested and screened for GFP expression using Leika MZ6 stereozoom microscope. The leaves that exhibited largest numbers of the green fluorescent cells were used to isolate virions as described (Napuli et al., 2000).
The rabbit polyclonal antisera to the BYV CP (Napuli et al., 2000), CPm (He et al., 1997), Hsp70h (Peremyslov and Dolja, 2002), and p64 (Napuli et al., 2003) were used for the immunoblot analyses in 1:1000 dilution. Protein bands were visualized using ECL western blotting detection kit (Amersham/Pharmacia Biotech). To ensure the uniformity of the virion composition analyses, all tested virion isolates were normalized relative to the CP content, and 1:2 and 1:4 dilutions were used in addition to undiluted samples. In contrast, when the protein composition of the bodiless virions formed by the BYVΔ mutant was determined, the amounts of analyzed virions were normalized relative to a principal tail component, CPm. Quantification of the CPm, Hsp70h, and p64 in all mutant virions relative to that in the wild type control was done using the Personal Densitometer SI (Molecular Dynamics) and the image analysis package ImageQuant provided by the manufacturer.
Hybridization analyzes of the RNA isolated from BYVΔ mutant were done using the conditions, probes, and BYV-derived RNA size markers as described previously (Peremyslov et al., 2004). The RNAs isolated from the wild type full-size virions and virion tails were used as controls (lanes V and Ts, respectively, in Fig. 3C). During the virion isolation, the RNA regions that were not encapsidated by the tail proteins were degraded by the endogenous plant RNases. The RNA isolated from particles formed by BYVΔ mutant was poly(A)-tailed and RT-PCR amplified using the primer specific to the genome 5’-end (5’-GTTTTTAACCATCCTTCTACTAGAC) and the oligo(dT) primer. The amplification product was cloned, and the nucleotide sequences of the 18 resulting cDNA clones were determined.
We thank Tatineni Satyanarayana and William Dawson for stimulating discussions. This work was supported by a grant from the National Institutes of Health (GM053190) to V.V.D.
Publisher's Disclaimer: This is a PDF le of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its nal citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.