Next generation sequencing reveals packaging of host RNAs by brome mosaic virus.

Although RNA viruses evolved the mechanisms of specific encapsidation, miss-packaging of cellular RNAs has been reported in such RNA virus systems as flock house virus or cucumber necrosis virus. To find out if brome mosaic virus (BMV), a tripartite RNA virus, can package cellular RNAs, BMV was propagated in barley and in Nicotiana benthamiana hosts, purified by cesium chloride (CsCl) gradient ultracentrifugation followed by nuclease treatment to remove any contaminating cellular (host) RNAs. The extracted virion RNA was then sequenced by using next-generation sequencing (NGS RNA-Seq) with the Illumina protocol. Bioinformatic analysis revealed the content of host RNAs ranging from 0.07% for BMV extracted from barley to 0.10% for the virus extracted from N. benthamiana. The viruses from two sources appeared to co-encapsidate different patterns of host-RNAs, including ribosomal RNAs (rRNAs), messenger RNAs (mRNAs) but also mitochondrial and plastid RNAs and, interestingly, transposable elements, both transposons and retrotransposons. Our data reveal that BMV virions can carry host RNAs, having a potential to mediate horizontal gene transfer (HGT) in plants.

RNA recombination in brome mosaic virus: effects of strand-specific stem-loop inserts.

A model system of a single-stranded trisegment Brome mosaic bromovirus (BMV) was used to analyze the mechanism of homologous RNA recombination. Elements capable of forming strand-specific stem-loop structures were inserted at the modified 3' noncoding regions of BMV RNA3 and RNA2 in either positive or negative orientations, and various combinations of parental RNAs were tested for patterns of the accumulating recombinant RNA3 components. The structured negative-strand stem-loops that were inserted in both RNA3 and RNA2 reduced the accumulation of RNA3-RNA2 recombinants to a much higher extent than those in positive strands or the unstructured stem-loop inserts in either positive or negative strands. The use of only one parental RNA carrying the stem-loop insert reduced the accumulation of RNA3-RNA2 recombinants even further, but only when the stem-loops were in negative strands of RNA2. We assume that the presence of a stable stem-loop downstream of the landing site on the acceptor strand (negative RNA2) hampers the reattachment and reinitiation processes. Besides RNA3-RNA2 recombinants, the accumulation of nontargeted RNA3-RNA1 and RNA3-RNA3 recombinants were observed. Our results provide experimental evidence that homologous recombination between BMV RNAs more likely occurs during positive- rather than negative-strand synthesis.

Studies on functional interaction between brome mosaic virus replicase proteins during RNA recombination, using combined mutants in vivo and in vitro.

Two viral proteins, 1a and 2a, direct replication of brome mosaic bromovirus (BMV) RNAs as well as they participate in BMV RNA recombination. To study the relationship between replication and recombination, double BMV variants that carried mutations in 1a and 2a genes were tested. The observed effects revealed that the 1a helicase and 2a N-terminal or core domains were functionally linked during both processes in vivo. The use of a series of mutant BMV replicase (RdRp) preparations demonstrated in vitro the participation of the 1a and 2a domains in BMV RNA copying and in template switching during minus-strand synthesis. The observed effects support previous observations that the characteristics of homologous and nonhomologous recombination can be modified separately by mutations at different sites on BMV replicase proteins.

Frequent homologous recombination events between molecules of one RNA component in a multipartite RNA virus.

Brome mosaic bromovirus (BMV), a tripartite plus-sense RNA virus, has been used as a model system to study homologous RNA recombination among molecules of the same RNA component. Pairs of BMV RNA3 variants carrying marker mutations at different locations were coinoculated on a local lesion host, and the progeny RNA3 in a large number of lesions was analyzed. The majority of doubly infected lesions accumulated the RNA3 recombinants. The distribution of the recombinant types was relatively even, indicating that both RNA3 counterparts could serve as donor or as acceptor molecules. The frequency of crossovers between one pair of RNA3 variants, which possessed closely located markers, was similar to that of another pair of RNA3 variants with more distant markers, suggesting the existence of an internal recombination hot spot. The majority of crossovers were precise, but some recombinants had minor sequence modifications, possibly marking the sites of imprecise homologous crossovers. Our results suggest discontinuous RNA replication, with the replicase changing among the homologous RNA templates and generating RNA diversity. This approach can be easily extended to other RNA viruses for identification of homologous recombination hot spots.

Mapping sequences active in homologous RNA recombination in brome mosaic virus: prediction of recombination hot spots.

The mechanism of homologous recombination has been studied previously in brome mosaic virus (BMV), a tricomponent, positive-stranded RNA virus of plants, by using artificial sequences (reviewed by J. J. BujarskiP. D. Nagy (1996). Semin. Virol. 7, 363-372). Here we extend these studies over BMV-derived sequences to obtain clues on prediction of homologous recombination hot spots. First, mismatch mutations, which reduced the AU content, were introduced into the common 60-nt recombination hot-spot sequence, either in the RNA2 or in both RNA2RNA3 components. This decreased the frequency of targeted homologous RNA2/RNA3 recombinationchanged the distribution of junction sites. Second, several short BMV RNA1- or RNA2-derived sequences were introduced into the RNA3 component, homologous recombination activity of these sequences was compared with that observed for previously characterized artificial sequences. Third, sequences at homologous recombinant junctions were compared among a large number of targetednontargeted recombinants. All these studies revealed several factors important for homologous recombination including the length of sequence identity, the extent of sequence identity, the AU content of the common sequences, the relative position of the AU-rich segment vs a GC-rich segment,the presence of GC-rich sequences. Based on this novel model, we suggest that recombination hot spots can be predicted by means of RNA sequence analysis. In addition, we show that recombination can occur between positivenegative strands of BMV RNAs. This provides further clues toward the mechanism of recombination processes in BMV.

Mutations in the N terminus of the brome mosaic virus polymerase affect genetic RNA-RNA recombination.

Previously, we have observed that mutations in proteins 1a and 2a, the two virally encoded components of the brome mosaic virus (BMV) replicase, can affect the frequency of recombination and the locations of RNA recombination sites (P. D. Nagy, A. Dzianott, P. Ahlquist, and J. J. Bujarski, J. Virol. 69:2547-2556, 1995; M. Figlerowicz, P. D. Nagy, and J. J. Bujarski, Proc. Natl. Acad. Sci. USA 94:2073-2078, 1997). Also, it was found before that the N-terminal domain of 2a, the putative RNA polymerase protein, participates in the interactions between 1a and 2a (C. C. Kao, R. Quadt, R. P. Hershberger, and P. Ahlquist, J. Virol. 66:6322-6329, 1992; E. O'Reilly, J. Paul, and C. C. Kao, J. Virol. 71:7526-7532, 1997). In this work, we examine how mutations within the N terminus of 2a influence RNA recombination in BMV. Because of the likely electrostatic character of 1a-2a interactions, five 2a mutants, MF1 to MF5, were generated by replacing clusters of acidic amino acids with their neutral counterparts. MF2 and MF5 retained nearly wild-type levels of 1a-2a interaction and were infectious in Chenopodium quinoa. However, compared to that in wild-type virus, the frequency of nonhomologous recombination in both MF2 and MF5 was markedly decreased. Only in MF2 was the frequency of homologous recombination reduced and the occurrence of imprecise homologous recombination increased. In MF5 there was also a 3' shift in the positions of homologous crossovers. The observed effects of MF2 and MF5 reveal that the 2a N-terminal domain participates in different ways in homologous and in nonhomologous BMV RNA recombination. This work maps specific locations within the N terminus involved in 1a-2a interaction and in recombination and further suggests that the mechanisms of the two types of crossovers in BMV are different.

Silencing homologous RNA recombination hot spots with GC-rich sequences in brome mosaic virus.

It has been observed that AU-rich sequences form homologous recombination hot spots in brome mosaic virus (BMV), a tripartite positive-stranded RNA virus of plants (P. D. Nagy and J. J. Bujarski, J. Virol. 71:3799-3810, 1997). To study the effect of GC-rich sequences on the recombination hot spots, we inserted 30-nucleotide-long GC-rich sequences downstream of AU-rich homologous recombination hot spot regions in parental BMV RNAs (RNA2 and RNA3). Although these insertions doubled the length of sequence identity in RNA2 and RNA3, the incidence of homologous RNA2 and RNA3 recombination was reduced markedly. Four different, both highly structured and nonstructured downstream GC-rich sequences had a similar "homologous recombination silencing" effect on the nearby hot spots. The GC-rich sequence-mediated recombination silencing mapped to RNA2, as it was observed when the GC-rich sequence was inserted at downstream locations in both RNA2 and RNA3 or only in the RNA2 component. On the contrary, when the downstream GC-rich sequence was present only in the RNA3 component, it increased the incidence of homologous recombination. In addition, upstream insertions of similar GC-rich sequences increased the incidence of homologous recombination within downstream hot spot regions. Overall, this study reveals the complex nature of homologous recombination in BMV, where sequences flanking the common hot spot regions affect recombination frequency. A replicase-driven template-switching model is presented to explain recombination silencing by GC-rich sequences.

RNA recombination in brome mosaic virus, a model plus strand RNA virus.

Studies on the molecular mechanism of genetic recombination in RNA viruses have progressed at the time when experimental systems of efficient recombination crossovers were established. The system of brome mosaic virus (BMV) represents one of the most useful and most advanced tools for investigation of the molecular aspects of the mechanism of RNA-RNA recombination events. By using engineered BMV RNA components, the occurrence of both homologous and nonhomologous crosses were demonstrated among the segments of the BMV RNA genome. Studies show that the two types of crossovers require different RNA signal sequences and that both types depend upon the participation of BMV replicase proteins. Mutations in the two BMV-encoded replicase polypeptides (proteins 1a and 2a) reveal that their different regions participate in homologous and in nonhomologous crossovers. Based on all these data, it is most likely that homologous and nonhomologous recombinant crosses do occur via two different types of template switching events (copy-choice mechanism) where viral replicase complex changes RNA templates during viral RNA replication at distinct signal sequences. In this review we discuss various aspects of the mechanism of RNA recombination in BMV and we emphasize future projections of this research.

Engineering of homologous recombination hotspots with AU-rich sequences in brome mosaic virus.

Previously, we observed that crossovers sites of RNA recombinants clustered within or close to AU-rich regions during genetic recombination in brome mosaic bromovirus (BMV) (P. D. Nagy and J. J. Bujarski. J. Virol. 70:415-426, 1996). To test whether AU-rich sequences can facilitate homologous recombination, AU-rich sequences were introduced into parental BMV RNAs (RNA2 and RNA3). These insertions created a homologous RNA2-RNA3 recombination hotspot. Two other AU-rich sequences also supported high-frequency homologous recombination if a common sequence with high or average G/C content was present immediately upstream of the AU-rich element. Homologous RNA recombination did not require any additional sequence motifs or RNA structures and was position nonspecific within the 3' noncoding region. These results suggest that nucleotide content (i.e., the presence of common 5' GC-rich or moderately AU-rich and 3' AU-rich regions) is the important factor that determines the sites of homologous recombination. A mechanism that involves replicase switching during synthesis of positive-sense RNA strands is presented to explain the observed results.

Structure-based rationale for the rescue of systemic movement of brome mosaic virus by spontaneous second-site mutations in the coat protein gene.

We describe spontaneous second-site reversions within the coat protein open reading frame that rescue the systemic-spread phenotype and increase virion stability of a mutant of brome mosaic virus. Based on the crystal structure of the related cowpea chlorotic mottle virus, we show that the modified residues are spatially clustered to affect the formation of hexamers and pentamers and therefore virion stability.

A mutation in the putative RNA polymerase gene inhibits nonhomologous, but not homologous, genetic recombination in an RNA virus.

Brome mosaic bromovirus (BMV), a positive-stranded RNA virus, supports both homologous and nonhomologous RNA recombinations. Two BMV (temperature-sensitive) mutants with alterations in the 2a protein, the putative RNA polymerase component of the viral replicase, were tested for their ability to support both types of recombination. Here we report that one of these mutants with the Leu-486 substituted by Phe did not support nonhomologous recombination. Effect on homologous recombination was mainly on the location and precision of crossover events. The other 2a mutant with Asn-458 substituted by Asp did not negatively affect either type of recombination. Apparently, BMV RNA polymerase participates differently in the two types of recombination events.

Effect of 5' and 3' terminal sequences, overall length, and coding capacity on the accumulation of defective RNAs associated with broad bean mottle bromovirus in planta.

Broad bean mottle bromovirus (BBMV) was shown to accumulate RNA2-derived defective interfering (DI) RNAs [Romero et al., Virology 194, 576-584 (1993); Pogany et al., Virology 212, 574-586 (1995)]. In this work, we utilize three sets of BBMV RNA2-derived artificial DI RNA constructs to determine factors that affect the accumulation of defective RNAs in planta. One set of deletion constructs was used to localize sequences required for efficient accumulation within the 5' 883 nt and the 3' 387 nt of the DI RNAs. The second set had a gradually increasing size of 3' nested deletions to determine the minimal length required for efficient DI RNA accumulation. The smallest DI RNA still accumulating in plants was found to be 1712 nt long. The third set consisted of frameshift mutants which showed that at least 64.4% of BBMV DI RNA sequences must have the 5' portion of the 2a open reading frame to ensure efficient accumulation. The importance of these factors in the selection of DI RNAs is discussed.

Homologous RNA recombination in brome mosaic virus: AU-rich sequences decrease the accuracy of crossovers.

Brome mosaic virus, a tripartite positive-stranded RNA virus of plants, was used for the determination of sequence requirements of imprecise (aberrant) homologous recombination. A 23-nucleotide (nt) region that included a 6-nt UUAAAA sequence (designated the AU sequence) common between wild-type RNA2 and mutant RNA3 supported both precise and imprecise homologous recombination, though the latter occurred with lower frequency. Doubling the length of the 6-nt AU sequence in RNA3 increased the incidence of imprecise crossovers by nearly threefold. Duplication or triplication of the length of the AU sequence in both RNA2 and RNA3 further raised the frequency of imprecise crossovers. The majority of imprecise crosses were located within or close to the extended AU sequence. Imprecise recombinants contained either nucleotide substitutions, nontemplated nucleotides, small deletions, or small sequence duplications within the region of crossovers. Deletion of the AU sequence from the homologous region in RNA3 resulted in the accumulation of only precise homologous recombinants. Our results provide experimental evidence that AU sequences can facilitate the formation of imprecise homologous recombinants. The generation of small additions or deletions can be explained by a misannealing mechanism within the AU sequences, while replicase errors during RNA copying might explain the occurrence of nucleotide substitutions or nontemplated nucleotides.

De novo generation of defective interfering-like RNAs in broad bean mottle bromovirus.

Broad been mottle virus (BBMV) is the only member of the bromoviruses that is known to accumulate defective-interfering (DI) RNAs (Romero et al., Virology 194, 576-584, 1993). De novo generation of DI-like RNAs was demonstrated during serial passages of BBMV in broad bean using either DI RNA-free virion RNA preparations or transcribed genomic RNA inocula. As for previously described DI RNAs, all but one of the characterized de novo generated DI-like RNAs were derived by a single in-frame deletion from the RNA2 component. The sole exception was derived by two shorter in-frame deletions from RNA2. The maintenance of an open reading frame by all DI-like RNAs suggests the importance of coding capacity and/or the shortened 2a protein in the accumulation of these RNAs during infection. The deletion junction sites were between nucleotides 1152 and 2366, suggesting that the retained regions are essential for the efficient accumulation of BBMV DI-like RNAs in planta. Short regions of sequence similarity and/or complementarity were revealed at the 5' and 3' junction borders. We speculate that these regions can facilitate DI (DI-like) RNA formation. In addition to DI-like RNAs, the full-length nucleotide sequences of RNA2 components of the Type and Morocco strains of BBMV are presented.

Mutations in the helicase-like domain of protein 1a alter the sites of RNA-RNA recombination in brome mosaic virus.

A system that uses engineered heteroduplexes to efficiently direct in vivo crossovers between brome mosaic virus (BMV) RNA1 and RNA3 (P. Nagy and J. Bujarski, Proc. Natl. Acad. Sci. USA 90:6390-6394, 1993) has been used to explore the possible involvement of BMV 1a protein, an essential RNA replication factor, in RNA recombination. Relative to wild-type 1a, several viable amino acid insertion mutations in the helicase-like domain of BMV 1a protein affected the nature and distribution of crossover sites in RNA3-RNA1 recombinants. At 24 degrees C, mutants PK19 and PK21 each increased the percentage of asymmetric crossovers, in which the RNA1 and RNA3 sites joined by recombination were not directly opposite each other on the engineered RNA3-RNA1 heteroduplex used to target recombination but rather were separated by 4 to 85 nucleotides. PK21 and another 1a mutant, PK14, also showed increases in the fraction of recombinants containing nontemplated U residues at the recombination junction. At 33 degrees C, the highest temperature that permitted infections with PK19, which is temperature sensitive for RNA replication, the mean location of RNA1-RNA3 crossovers in recombinants recovered from PK19 infections was shifted by nearly 25 bp into the energetically less stable side of the RNA1-RNA3 heteroduplex. Thus, mutations in the putative helicase domain of the 1a protein can influence BMV RNA recombination. The results are discussed in relation to models for recombination by template switching during pausing of RNA replication at a heteroduplexed region in the template.

Foreign complementary sequences facilitate genetic RNA recombination in brome mosaic virus.

We have demonstrated that local antisense sequences can mediate genetic recombination within the 3' noncoding region among brome mosaic virus (BMV) RNAs (P. Nagy and J. J. Bujarski, 1993, Proc. Natl. Acad. Sci. USA 90, 6390-6394). Here we show that foreign complementary inserts can direct crossovers between BMV RNA3 components within an internal region. A 170-nt polynucleotide derived from the cowpea chlorotic mottle virus (CCMV) RNA3 was inserted just upstream of the initiation codon of the BMV coat protein open reading frame in either sense or antisense orientations. The resulting respective mutants, BCC+ and BCC-, maintained unchanged CCMV inserts when inoculated separately on leaves of a local lesion host for BMV. In contrast, when a mixture containing both mutated RNAs3 was inoculated, a significant fraction of lesions accumulated the BMV RNA3 lacking the CCMV insert. The presence of a 3' marker mutation confirmed that the BMV RNA3 progeny arose due to crossovers between BCC+ and BCC- within the complementary sequences. The highest frequency of recombinant appearance was observed when the RNA mixtures were annealed prior to inoculation on the host plants. Our results confirm a concept predicting the general nature of the heteroduplex-mediated recombination functioning in RNA viruses. Examples of possible applications of this approach in recombinant RNA technology are discussed.

Efficient system of homologous RNA recombination in brome mosaic virus: sequence and structure requirements and accuracy of crossovers.

Brome mosaic virus (BMV), a tripartite positive-stranded RNA virus of plants engineered to support intersegment RNA recombination, was used for the determination of sequence and structural requirements of homologous crossovers. A 60-nucleotide (nt) sequence, common between wild-type RNA2 and mutant RNA3, supported efficient repair (90%) of a modified 3' noncoding region in the RNA3 segment by homologous recombination with wild-type RNA2 3' noncoding sequences. Deletions within this sequence in RNA3 demonstrated that a nucleotide identity as short as 15 nt can support efficient homologous recombination events, while shorter (5-nt) sequence identity resulted in reduced recombination frequency (5%) within this region. Three or more mismatches within a downstream portion of the common 60-nt RNA3 sequence affected both the incidence of recombination and the distribution of crossover sites, suggesting that besides the length, the extent of sequence identity between two recombining BMV RNAs is an important factor in homologous recombination. Site-directed mutagenesis of the common sequence in RNA3 did not reveal a clear correlation between the stability of predicted secondary structures and recombination activity. This indicates that homologous recombination does not require similar secondary structures between two recombining RNAs at the sites of crossovers. Nearly 20% of homologous recombinants were imprecise (aberrant), containing either nucleotide mismatches, small deletions, or small insertions within the region of crossovers. This implies that homologous RNA recombination is not as accurate as proposed previously. Our results provide experimental evidence that the requirements and thus the mechanism of homologous recombination in BMV differ from those of previously described heteroduplex-mediated nonhomologous recombination (P. D. Nagy and J. J. Bujarski, Proc. Natl. Acad. Sci. USA 90:6390-6394, 1993).

Mutational analysis of the coat protein gene of brome mosaic virus: effects on replication and movement in barley and in Chenopodium hybridum.

The coat protein (CP) open reading frame (ORF) of brome mosaic virus (BMV) has been mutated to study host-related CP functions in barley, a systemic host, and in Chenopodium hybridum L. which supports both local lesion formation and systemic spread of BMV. To test the role of the N-terminal region of CP, mutants C1 to C3, which synthesized the CP lacking first seven amino acids, and mutant D1, which had Trp 22 and Thr 23 replaced with Phe-Gly-Ser, were generated. C1 to C3 inhibited virus systemic spread in C. hybridum but not in barley while D1 only reduced virus accumulation in noninoculated leaves of C. hybridum. More internal CP regions were tested by mutation of Lys 63 to Leu (mutant SP3) and Lys 129 to Arg (mutant SP1). SP1 behaved similarly to C1 to C3 while SP3 similarly to D1. In addition, SP3 reduced concentrations of RNA3 and RNA4 in both hosts. Apparently, various CP regions differentially affect, either directly or indirectly, virus translocation in different hosts, suggesting both the CP and host factors to be important for virus spread. Larger deletions in the CP ORF (mutants BB4 and SX1) or a decrease of CP production by using a frameshift mutant C, inhibited virus systemic spread in both hosts, and delayed the appearance of smaller local lesions on C. hybridum. Thus, the CP is not required for cell-to-cell movement but is required for systemic translocation of BMV.