Substantial portions of the 5' and intercistronic noncoding regions of cowpea chlorotic mottle virus RNA3 are dispensable for systemic infection but influence viral competitiveness and infection pathology

Cowpea chlorotic mottle bromovirus replication proteins support template-selective RNA replication in Saccharomyces cerevisiae

Positive-strand RNA viruses generally assemble RNA replication complexes on rearranged host membranes. Alphaviruses, other members of the alpha-like virus superfamily, and many other positive-strand RNA viruses invaginate host membrane into vesicular RNA replication compartments, known as spherules, whose interior is connected to the cytoplasm. Brome mosaic virus (BMV) and its close relative, cowpea chlorotic mottle virus (CCMV), form spherules along the endoplasmic reticulum. BMV spherule formation and RNA replication can be fully reconstituted in S. cerevisiae, enabling many studies identifying host factors and viral interactions essential for these processes. To better define and understand the conserved, core pathways of bromovirus RNA replication, we tested the ability of CCMV to similarly support spherule formation and RNA replication in yeast. Paralleling BMV, we found that CCMV RNA replication protein 1a was the only viral factor necessary to induce spherule membrane rearrangements and to recruit the viral 2a polymerase (2a^pol) to the endoplasmic reticulum. CCMV 1a and 2a^pol also replicated CCMV and BMV genomic RNA2, demonstrating core functionality of CCMV 1a and 2a^pol in yeast. However, while BMV and CCMV 1a/2a^pol strongly replicate each others’ genomic RNA3 in plants, neither supported detectable CCMV RNA3 replication in yeast. Moreover, in contrast to plant cells, in yeast CCMV 1a/2a^pol supported only limited replication of BMV RNA3 (<5% of that by BMV 1a/2a^pol). In keeping with this, we found that in yeast CCMV 1a was significantly impaired in recruiting BMV or CCMV RNA3 to the replication complex. Overall, we show that many 1a and 2a^pol functions essential for replication complex assembly, and their ability to be reconstituted in yeast, are conserved between BMV and CCMV. However, restrictions of CCMV RNA replication in yeast reveal previously unknown 1a-linked, RNA-selective host contributions to the essential early process of recruiting viral RNA templates to the replication complex.

Melandrium yellow fleck bromovirus infects Arabidopsis thaliana and has genomic RNA sequence characteristics that are unique among bromoviruses

Melandrium yellow fleck bromovirus (MYFV) systemically infected Arabidopsis thaliana, although the susceptibility of several A. thaliana accessions to MYFV differed from their susceptibility to the other two bromoviruses infecting A. thaliana. We constructed full-length cDNA clones of MYFV genomic RNAs 1, 2, and 3 and determined their complete nucleotide sequences. Similar to Broad bean mottle bromovirus, (1) the 5'-terminal nucleotide of the MYFV genomic RNAs was adenine, and (2) the "D-arm" was absent from the tRNA-like structure in the 3' untranslated regions (UTRs) of MYFV RNAs. As unique characteristics, MYFV RNA3 lacked the poly(A) tract in the intercistronic region and contained a directly repeated sequence of about 200 nucleotides and polypyrimidine tracts of heterogeneous lengths in the 5' UTR. Co-infection experiments using RNA3 clones with or without the duplicated sequence demonstrated that the duplication contributed to the competitive fitness of the virus in Nicotiana benthamiana.

The conserved, 5' termini of RNAs 1 and 2 of Tomato aspermy virus are dispensable for infection but affect virulence

The 5' terminus of each of the three genomic RNAs (RNAs 1, 2 and 3) of Tomato aspermy virus (TAV) begins with the sequence 5'-GUUU, which is also shared by a number of other viruses. Mutagenic analyses showed that the 5'-GUUU sequence of RNAs 1 and 2 of TAV was dispensable for viral infection and did not prevent symptom induction. On the other hand, substitution of U at position 5 for G in RNA 1, but not RNA 2, induced veinal necrosis symptoms in Nicotiana glutinosa. The mutants constructed included insertion of UUU into the 5'-GUUU sequence of TAV RNAs 1 and/or 2. All RNA 2 mutants induced more severe symptoms than viral RNAs containing either mutated RNA 1 or most combinations of mutated RNAs 1 and 2. Some combinations of mutated RNAs 1 and 2 also induced veinal necrosis in N. glutinosa. Virulence was unrelated to the levels of viral RNA accumulation. Sequence analysis of progeny viral RNAs showed that only the mutant viral RNAs with a G to U substitution in RNA 1 and the deletion of the 5'-GUUU in both RNAs 1 and 2 were able to maintain the same sequence as the inoculum. The other mutants either reverted to the wildtype sequence or underwent further deletion or insertion. None of the constructed mutants were able to compete for accumulation with the wildtype virus after co-inoculation to the plant species tested.

Formation and amplification of a novel tombusvirus defective RNA which lacks the 5' nontranslated region of the viral genome

Defective interfering (DI) RNAs of tomato bushy stunt virus (TBSV) are small, subgenomic, helper-dependent replicons that are believed to be generated primarily by aberrant events during replication of the plus-sense RNA genome. Prototypical TBSV DI RNAs contain four noncontiguous segments (regions I through IV) derived from the 5' nontranslated region (NTR) (I), an internal section (II), and the 3'-terminal portion (III and IV) of the viral genome. We have studied the formation of these molecules by using engineered precursor DI RNA transcripts and report here the consistent accumulation of a novel defective RNA species, designated RNA B. Northern blot, primer extension, and sequence analyses indicated that, unlike prototypical DI RNAs, RNA B lacks region I. In vitro transcripts corresponding to the region II-III-IV structure of RNA B were amplified when coinoculated with helper, indicating that the 5' NTR of the genome does not harbor cis-acting replication elements essential for viral RNA replication. Region I is, however, important for DI RNA fitness, since molecules lacking it accumulated to significantly lower levels ( approximately 10-fold reduction). Analysis of the minus-strand sequence of region I led to the identification of an RNA undecamer sequence, arranged in tandem, at its very 3' terminus. Additional variants of the undecamer motif were also identified at internal positions in region I and in the negative strands of regions II, III, and IV. Features of the undecamer motif, the consensus of which is (-)3'-CCCAAAGAGAG, are consistent with a role as a cis-acting replication element. It is proposed that the ability of RNA B to be amplified is due, in part, to compensatory effects of a strategically positioned undecamer motif in region II. Possible replicase-mediated mechanisms for the generation of this novel viral RNA are also presented.

Comparison of the role of 5' terminal sequences of alfalfa mosaic virus RNAs 1, 2, and 3 in viral RNA replication

The 5' untranslated regions (UTRs) of the genomic RNAs 1, 2, and 3 of alfalfa mosaic virus (AMV) are 100, 54, and 345 nucleotides (nt) long, respectively, and lack extensive sequence similarity to each other. RNA 3 encodes the movement protein P3 and the coat protein and can be replicated in transgenic tobacco plants expressing the replicase proteins P1 and P2 (P12 plants). 5' Cis-acting sequences involved in RNA 3 replication have been shown to be confined to the 5' UTR. When the 5' UTR of RNA 3 was replaced by the 5' UTRs of RNAs 1 or 2, the recombinant RNA was not infectious to P12 plants. Also, when the P3 gene in RNA 3 was put under the control of a subgenomic promoter and the 5' UTR of this RNA was replaced by 5' terminal RNA 1 sequences of 103 to 860 nt long or RNA 2 sequences of 57 to 612 nt long, no accumulation of the hybrid RNAs was observed. Deletion of the 5' 22 nucleotides of RNA 3 resulted in the accumulation of a major progeny that lacked the 5' 79 nt. However, when the 5' 22 nucleotides of RNA 3 were replaced by the complete 5' UTR of RNA 1 or 5' sequences of RNAs 1, 2, or 3 with a length of 5 to 15 nt, accumulation of the full-length mutant RNAs was observed. The effect of mutations in the 5' viral sequences of 5 to 15 nt was analyzed. It is concluded that although elements within nucleotides 80-345 of the 5' UTR of RNA 3 are sufficient for replication, a specific sequence of 3 to 5 nt is required to target the replicase to an initiation site corresponding to the 5' end of the RNA.

Long sequences in the 5' noncoding region of plum pox virus are not necessary for viral infectivity but contribute to viral competitiveness and pathogenesis

The 5'-terminal 31 nucleotides of the 146-nucleotides-long 5' noncoding region of plum pox potyvirus (PPV) are highly conserved in all the members of the Potyvirus genus. To map the sequences of the 5' noncoding region that are necessary in vivo for infectivity, we have constructed a nested set of substitution and deletion mutants. While we were not able to infect Nicotiana clevelandii plants with full-length PPV transcripts bearing mutations in the 5'-terminal 35 nucleotides of the viral genome, the deletion of long sequences located between nucleotides 39 and 145 did not alter either the rate of infection or viral accumulation. Nevertheless, these mutants were not able to compete with the wild-type strain in coinoculation experiments. Plants infected with a PPV mutant that lacked nucleotides 127 to 145 showed a very mild symptomathology; the wild-type symptom severity was recovered after spontaneous second-site mutations.

The polymerase-like core of brome mosaic virus 2a protein, lacking a region interacting with viral 1a protein in vitro, maintains activity and 1a selectivity in RNA replication

Brome mosaic virus (BMV), a member of the alphavirus-like super-family of positive-strand RNA viruses, encodes two proteins required for viral RNA replication: 1a and 2a. 1a contains m7G methyltransferase- and helicase-like domains, while 2a contains a polymerase (pol)-like core flanked by N- and C-terminal extensions. Genetic studies show that BMV RNA replication requires 1a-2a compatibility implying direct or indirect 1a-2a interaction in vivo. In vitro, la interacts with the N-terminal 125-amino-acid segment of 2a preceding the pol-like core, and prior deletion studies suggested that this 2a segment was essential for RNA replication. We have now used protein fusions and deletions to explore possible parallels between noncovalent 1a-2a interaction and covalent fusion of similar protein domains in tobacco mosaic virus and to see whether the N-terminal 2a-1a interaction was the primary basis for 1a-2a compatibility in vivo. We found that 2a can function as part of a tobacco mosaic virus-like 1a-2a fusion and that a 2a segment (amino acids 162 to 697) comprising the pol-like core was sufficient to provide 2a functions in such a fusion. Unexpectedly, the unfused 2a core segment also supported RNA replication when it and wild-type la were expressed as separate proteins. Moreover, in gene reassortant experiments with the related cowpea chlorotic mottle virus, the unfused 2a core segment showed the same 1a compatibility requirements as did wild-type BMV 2a. Thus, the pol-like core of 2a must interact with la in a way that is selective and essential for RNA synthesis, and 1a-2a interactions are more complex than the single, previously mapped interaction of the N-terminal 2a segment with 1a.

Genetic elements of plant viruses as tools for genetic engineering

Viruses have developed successful strategies for propagation at the expense of their host cells. Efficient gene expression, genome multiplication, and invasion of the host are enabled by virus-encoded genetic elements, many of which are well characterized. Sequences derived from plant DNA and RNA viruses can be used to control expression of other genes in vivo. The main groups of plant virus genetic elements useful in genetic engineering are reviewed, including the signals for DNA-dependent and RNA-dependent RNA synthesis, sequences on the virus mRNAs that enable translational control, and sequences that control processing and intracellular sorting of virus proteins. Use of plant viruses as extrachromosomal expression vectors is also discussed, along with the issue of their stability.

Host-specificity restriction by bromovirus cell-to-cell movement protein occurs after initial cell-to-cell spread of infection in nonhost plants

The nonstructural 3a protein of the positive-strand RNA bromoviruses is required for infection spread in plants and is a crucial determinant of host specificity in systemic infection. To determine the paths of wild-type (wt) bromovirus infection spread, the step at which 3a mutants are arrested, and the nature of the host specificity associated with the 3a gene, we used in situ hybridization to examine infection spread by cowpea chlorotic mottle bromovirus (CCMV) and its derivatives at the level of individual cells in cowpea leaf epidermis. From 1 to 3 days post inoculation (dpi), wt CCMV spread from initially infected cells to adjacent cells, creating expanding infection foci whose radii grew by one additional epidermal cell diameter every 5 hr. By 3 to 4 dpi, vascular elements contacting such foci acted as conduits for further infection spread. By contrast, a 3a frameshift derivative multiplied in initially infected epidermal cells but failed to move into neighboring cells even by 4 dpi, showing that the 3a gene is essential for cell-to-cell spread. Most interestingly, a CCMV derivative with the 3a gene replaced by that of a bromovirus not adapted to cowpea, brome mosaic virus (BMV), initially spread from cell to cell in cowpea plants, but stopped spreading between 1 and 2 dpi, when most infection foci encompassed 40-80 epidermal cells. Thus, the host-specificity restriction imposed by BMV 3a protein did not result from an inability to direct the spread of infection out of initially infected cowpea cells, but from a much later block. The apparent absence of any preexisting anatomical boundary at the limit of infection spread and localized tissue changes at the infection foci suggested that induced host responses might have contributed to this block.

Requirement of the 5'-end genomic sequence as an upstream cis-acting element for coronavirus subgenomic mRNA transcription

We have developed a defective interfering (DI) RNA containing a chloramphenicol acetyltransferase reporter gene, placed behind an intergenic sequence, for studying subgenomic mRNA transcription of mouse hepatitis virus (MHV), a prototype coronavirus. Using this system, we have identified the sequence requirement for MHV subgenomic mRNA transcription. We show that this sequence requirement differs from that for RNA replication. In addition to the previously identified requirement for an intergenic (promoter) sequence, additional sequences from the 5' end of genomic RNA are required for subgenomic mRNA transcription. These upstream sequences include the leader RNA and a spacer sequence between the leader and intergenic sequence, which is derived from the 5' untranslated region and part of gene 1. The spacer sequence requirement is specific, since only the sequence derived from the 5' end of RNA genome, but not from other MHV genomic regions or heterologous sequences, could initiate subgenomic transcription from the intergenic sequence. These results strongly suggest that the wild-type viral subgenomic mRNAs (mRNA2 to mRNA7) and probably their counterpart subgenomic negative-sense RNAs cannot be utilized for mRNA amplification. Furthermore, we have demonstrated that a partial leader sequence present at the 5' end of genome, which lacks the leader-mRNA fusion sequence, could still support subgenomic mRNA transcription. In this case, the leader sequences of the subgenomic transcripts were derived exclusively from the wild-type helper virus, indicating that the MHV leader RNA initiates in trans subgenomic mRNA transcription. Thus, the leader sequence can enhance subgenomic transcription even when it cannot serve as a primer for mRNA synthesis. These results taken together suggest that the 5'-end leader sequence of MHV not only provides a trans-acting primer for mRNA initiation but also serves as a cis-acting element required for the transcription of subgenomic mRNAs. The identification of an upstream cis-acting element for MHV subgenomic mRNA synthesis defines a novel sequence requirement for regulating mRNA synthesis in RNA viruses.

The multigenic nature of RNA virus adaptation to plants

Molecular genetic studies of RNA virus host range suggest that successful systemic infection requires many or all viral genes to have some degree of specific adaptation to the host. In some cases, accommodation to the host appears to be required for viral gene products to fulfil their primary functions in viral multiplication, while in other cases, adaptation seems to be required to avoid eliciting host defenses.

Bromovirus movement protein genes play a crucial role in host specificity

Monocot-adapted brome mosaic virus (BMV) and dicot-adapted cowpea chlorotic mottle virus (CCMV) are closely related bromoviruses with tripartite RNA genomes. Although RNAs 1 and 2 together are sufficient for RNA replication in protoplasts, systemic infection also requires RNA3, which encodes the coat protein and the nonstructural 3a movement protein. We have previously shown with bromoviral reassortants that host specificity determinants in both viruses are encoded by RNA3 as well as by RNA1 and/or RNA2. Here, to test their possible role in host specificity, the 3a movement protein genes were precisely exchanged between BMV and CCMV. The hybrid viruses, but not 3a deletion mutants, systemically infected Nicotiana benthamiana, a permissive host for both parental viruses. The hybrids thus retain basic competence for replication, packaging, cell-to-cell spread, and long-distance (vascular) spread. However, the hybrids failed to systemically infect either barley or cowpea, selective hosts for parental viruses. Thus, the 3a gene and/or its encoded 3a protein contributes to host specificity of both monocot- and dicot-adapted bromoviruses. Tests of inoculated cowpea leaves showed that the spread of the CCMV hybrid containing the BMV 3a gene was blocked at a very early stage of infection. Moreover, the BMV hybrid containing the CCMV 3a gene appeared to spread farther than wt BMV in inoculated cowpea leaves. Several pseudorevertants directing systemic infection in cowpea leaves were obtained from plants inoculated with the CCMV(BMV 3a) hybrid, suggesting that the number of mutations required to adapt the hybrid to dicots is small.

Role of the 5' leader sequence of alfalfa mosaic virus RNA 3 in replication and translation of the viral RNA

RNA 3 of alfalfa mosaic virus (AIMV) encodes the movement protein P3 and the viral coat protein which is translated from the subgenomic RNA 4. The 5'-leader sequences of RNA 3 of AIMV strains S, A, and Y differ in length from 314 to 392 nucleotides and contain a variable number of internal control regions of type 2 (ICR2 motifs) each located in a 27 nt repeat. Infectious cDNA clones were used to exchange the leader sequences of the three strains. This revealed that the leader sequence controls the specific ratio in which RNAs 3 and 4 are synthesized for each strain. In addition, it specifies strain specific differences in the kinetics of P3 accumulation in plants. Subsequent deletion analysis revealed that a 5'-sequence of 112 nt containing one ICR2 motif was sufficient for a 10 to 20% level of RNA 3 accumulation in protoplasts and a delayed accumulation in plants. An additional leader sequence of maximally 114 nt, containing two ICR2 motifs, was required to permit wildtype levels of RNA 3 accumulation. The effect of deletions in the leader sequence on P3 synthesis in vitro and in vivo was investigated.