Interviral Recombination between Plant, Insect, and Fungal RNA Viruses: Role of the Intracellular Ca²⁺/Mn²⁺ Pump

Viruses with RNA genomes are abundant, and their genomic sequences show astonishing variation. Genetic recombination in RNA viruses is a major force behind their rapid evolution, enhanced pathogenesis, and adaptation to their hosts. We utilized a previously identified intracellular Ca²⁺/Mn²⁺ pump-deficient yeast to search for interviral recombinants. Noninfectious viral replication systems were used to avoid generating unwanted infectious interviral recombinants. Altogether, interviral RNA recombinants were observed between plant and insect viruses, and between a fungal double-stranded RNA (dsRNA) virus and an insect virus, in the yeast host. In addition, interviral recombinants between two plant virus replicon RNAs were identified in N. benthamiana plants, in which the intracellular Ca²⁺/Mn²⁺ pump was depleted. These findings underline the crucial role of the host in promoting RNA recombination among unrelated viruses.

ABSTRACT

Recombination is one of the driving forces of viral evolution. RNA recombination events among similar RNA viruses are frequent, although RNA recombination could also take place among unrelated viruses. In this paper, we have established efficient interviral recombination systems based on yeast and plants. We show that diverse RNA viruses, including the plant viruses tomato bushy stunt virus, carnation Italian ringspot virus, and turnip crinkle virus-associated RNA; the insect plus-strand RNA [(+)RNA] viruses Flock House virus and Nodamura virus; and the double-stranded L-A virus of yeast, are involved in interviral recombination events. Most interviral recombinants are minus-strand recombinant RNAs, and the junction sites are not randomly distributed, but there are certain hot spot regions. Formation of interviral recombinants in yeast and plants is accelerated by depletion of the cellular SERCA-like Pmr1 ATPase-driven Ca²⁺/Mn²⁺ pump, regulating intracellular Ca²⁺ and Mn²⁺ influx into the Golgi apparatus from the cytosol. The interviral recombinants are generated by a template-switching mechanism during RNA replication by the viral replicase. Replication studies revealed that a group of interviral recombinants is replication competent in cell-free extracts, in yeast, and in the plant Nicotiana benthamiana. We propose that there are major differences among the viral replicases to generate and maintain interviral recombinants. Altogether, the obtained data promote the model that host factors greatly contribute to the formation of recombinants among related and unrelated viruses. This is the first time that a host factor’s role in affecting interviral recombination is established.

IMPORTANCE Viruses with RNA genomes are abundant, and their genomic sequences show astonishing variation. Genetic recombination in RNA viruses is a major force behind their rapid evolution, enhanced pathogenesis, and adaptation to their hosts. We utilized a previously identified intracellular Ca²⁺/Mn²⁺ pump-deficient yeast to search for interviral recombinants. Noninfectious viral replication systems were used to avoid generating unwanted infectious interviral recombinants. Altogether, interviral RNA recombinants were observed between plant and insect viruses, and between a fungal double-stranded RNA (dsRNA) virus and an insect virus, in the yeast host. In addition, interviral recombinants between two plant virus replicon RNAs were identified in N. benthamiana plants, in which the intracellular Ca²⁺/Mn²⁺ pump was depleted. These findings underline the crucial role of the host in promoting RNA recombination among unrelated viruses.

INTRODUCTION

High-frequency mutations and RNA recombination contribute to the rapid evolution of plus-strand RNA [(+)RNA] viruses, including the emergence of new strains or new viruses (1–7). Genetic recombination in RNA viruses can rapidly lead to small as well as dramatic changes in virus genomes. Accordingly, new traits and novel cis-acting elements or coding sequences could be acquired by viruses through recombining or rearranging evolutionarily successful RNA sequences. RNA recombination also contributes to repairing truncated or mutated viral RNA genomes, thus increasing viral fitness (5). A major role for RNA recombination in the emergence of new viruses or virus strains is well documented for numerous human, plant, animal, bacterial, insect, and fungal viruses (8–18). Recombination is a mechanism to generate defective interfering (DI) RNAs that play roles in viral pathogenesis (19). RNA recombination can also occur between different viruses (termed interviral or interspecies recombination) or between viral and host RNAs (20–26).

Tomato bushy stunt virus (TBSV), a tombusvirus, is intensively used to study RNA recombination based on a yeast (Saccharomyces cerevisiae) model host. Accordingly, genome-wide screens in yeast have led to the identification of many cellular factors affecting TBSV recombination (8, 10, 27–29). The advantages of tombusviruses include high-frequency RNA recombination and development of in vitro and in vivo methods, facilitating the accumulation of various recombinants (8, 10, 30, 31).

The TBSV-encoded p33 and p92^pol replication proteins are directly translated from the genomic RNA (gRNA), including translational readthrough of the p33 stop codon to obtain p92^pol (32–34). Viral and host components, such as heat shock protein 70 and phospholipids, regulate the RNA-dependent RNA polymerase (RdRp) function of p92^pol (35–38). The abundant p33 RNA chaperone is the master regulator of virus replication, with key roles in the recruitment of the viral RNA template for replication and in the assembly of the membrane-bound viral replicase complexes (VRCs) and rewiring of several cellular pathways (37, 39–43). The two replication proteins are essential VRC components (36, 44).

The template-switching-type recombination driven by viral replicase during RNA synthesis is the major mechanism, although RNA ligation has also been documented (1, 12, 45–48). In addition to playing essential roles in viral replication, host factors also play critical roles in viral genomic recombination (2, 29, 49, 50). High-throughput genomic screens with tombusviruses in the yeast model host have identified more than 50 host genes affecting viral RNA recombination (8, 10, 51). Subsequent studies with several host factors firmly established that various cellular pathways and factors are involved in tombusviral recombination in yeast and plants (9, 31, 49, 51–55). For example, we have previously shown that there is an increase in TBSV RNA recombination in pmr1Δ yeast (31). Pmr1 (for “plasma membrane ATPase related”) is a conserved ATPase-driven Ca²⁺/Mn²⁺ pump, which regulates intracellular Ca²⁺ and Mn²⁺ influx into the Golgi apparatus from the cytosol (56). SERCA (sarcoplasmic-endoplasmic reticulum Ca²⁺-ATPase) and hSPCA1 (human secretory pathway Ca²⁺-ATPase 1), the mammalian orthologs of yeast Pmr1p, also regulate intracellular Ca²⁺/Mn²⁺ homeostasis (57).

Interviral recombination between different viral species has been observed in many different groups of RNA viruses (58), but the viral and host factors involved have not yet been established. In this paper, we exploited a yeast mutant strain which had its PMR1 gene deleted to study interviral recombination. We find rapid accumulation of RNA recombinants between a TBSV replicon RNA (repRNA) or the closely related gRNA of carnation Italian ringspot virus (CIRV) and unrelated insect viruses, namely, Flock House virus (FHV) and Nodamura virus (NoV) RNAs, in pmr1Δ yeast. The recombinant minus-strand RNAs [(−)recRNAs] were copurified with the tombusvirus replicase from membranes, suggesting that they are generated by the viral replicase via a template-switching mechanism during (−)RNA synthesis. Interviral recombinants were also detected when FHV replicated in pmr1Δ yeast harboring the L-A double-stranded RNA (dsRNA) virus of yeast. To confirm that Mn²⁺ homeostasis is also important for interviral recombination in plants, we isolated recombinants formed between the repRNA of TBSV and a satellite RNA of turnip crinkle virus (TCV) in Nicotiana benthamiana plants, in which the orthologous genes of yeast PMR1 were silenced. Replication studies revealed that some of the interviral recombinants are replication competent in cell-free extracts (CFEs), in yeast, and in N. benthamiana. Altogether, the obtained data promote the model that host factors could greatly contribute to the formation of recombinants among related and unrelated viruses.

RESULTS

Efficient interviral RNA recombination between a plant virus-derived replicon RNA and an insect virus RNA.

To identify putative interviral recombination events between unrelated (+)RNA viruses, we took advantage of yeast as a common surrogate host for these viruses. Yeast pmr1Δ and wild-type (wt) strains were transformed with plasmids that can simultaneously induce the replication of TBSV repRNA and FHV RNA1 in the same cells. Northern blot analysis of total nucleic acid extracts from wt and pmr1Δ yeasts revealed the efficient replication of both TBSV repRNA and FHV RNA1 in the same yeast (Fig. 1A). This Northern blot analysis, however, was not sensitive enough to detect the accumulation of interviral recombinants during a 48-h replication cycle. Therefore, the emergence of putative interviral recombinants was monitored with reverse transcription-PCR (RT-PCR) using selective primer combinations, which could amplify interviral RNA recombinants among (+)- and (−)RNA strands of these viruses (see Table S1 in the supplemental material). Interestingly, two primer combinations in RT-PCR led to the amplification of interviral recombinants in pmr1Δ yeast and, to a lesser extent, in wt yeast (Fig. 1B, lanes 1 and 2 versus lanes 3 and 4, and Fig. 1E). The most abundant interviral recombinants from both pmr1Δ and wt strains were detected by RT-PCR from the (−)RNA pool, with TBSV sequences representing the 5′ segment, whereas FHV RNA1 sequences were located within the 3′ segment of the recombinants (schematically shown for the most frequent recombinant in Fig. 1C). We also observed interviral (−)RNA recombinants between FHV RNA3, which is a subgenomic RNA made from RNA1 during replication, and TBSV repRNA in pmr1Δ yeast (Table S2). The RNA sequences around the recombination junction sites are presented in Table S2.

FIG 1.

Characterization of interviral recombinants formed between TBSV repRNA and FHV RNA1 in yeast. (A) Replication of TBSV repRNA and FHV RNA1 in the presence of tombusvirus replicase in yeast. TBSV (+)repRNA and FHV (+)RNA1 were expressed separately or together from plasmids in wt and pmr1Δ yeasts. Accumulation of the progeny (+)RNAs was measured by Northern blotting 24 h after initiation of replication using either TBSV-specific or FHV-specific probes. Note that the tombusvirus p33 and p92^pol replication proteins were expressed from plasmids to support the replication of cognate repRNAs. (B) RT-PCR analysis of the occurrence of (−)recRNA recombinants in wt and pmr1Δ yeast strains. TBSV repRNA, p33 and p92^pol replication proteins, and FHV RNA1 were coexpressed in yeast from plasmids, as shown. Samples in lanes 7 and 8 were derived from pmr1Δ yeast coexpressing TBSV repRNA, p33 replication protein, and FHV RNA1 without the tombusvirus p92 RdRp. Lanes 9 and 10 represent samples where TBSV repRNA and FHV RNA1 were replicated separately in pmr1Δ yeast, followed by total RNA isolation and in vitro mixing of the samples in a 1:1 ratio. Next, RT-PCR was performed to detect putative artifactual recombinants. Lanes 11 and 12 contain the same total RNA samples as those in lanes 1 and 2, but only PCR was used, while the RT step was omitted. (C) Schematic representation of the most frequent interviral (−)RNA recombinant containing the FHV sequence at the 3′ region and the TBSV sequence at the 5′ region (based on the coordinates in TBSV DI-72), as shown. The number of isolations of different hot spot regions in the interviral recombinants is shown. Note that this is not the recombination frequency. The actual sequences of the recRNAs are presented in Table S2 in the supplemental material. (D) RT-PCR analysis of Flag affinity-purified tombusvirus replicase from the detergent-solubilized membrane fraction of pmr1Δ yeast to identify (−)recRNAs from yeast replicating TBSV repRNA and FHV RNA1. Lanes 2 and 4 represent yeast samples with Flag-p33 and Flag-p92^pol, whereas lanes 1 and 3 show yeast expressing 6×His-tagged p33 and p92^pol as negative controls. (E) RT-PCR analysis of the occurrence of (+)recRNA recombinants in wt and pmr1Δ yeast strains. See above (B) for further details. (F) The number of isolations of different hot spot regions in the interviral recombinants from panel E. See above (C) for further details. (G) Proposed tombusvirus replicase-driven template-switching model of the formation of interviral (−)recRNAs between TBSV repRNA and FHV RNA1.

Replication of both TBSV repRNA and FHV RNA1 in pmr1Δ yeast was required to detect (−)recRNAs (Fig. 1B, lanes 7 and 8). This suggests that the viral replicase(s) is likely involved in the generation of interviral recombinants. Mixing the isolated total RNA extracts from the pmr1Δ strain separately replicating TBSV repRNA and FHV RNA1 did not result in interviral (−)recRNA detection, excluding the artifactual generation of recRNAs during the RT-PCR step (Fig. 1B, lanes 9 and 10). Moreover, omitting the RT step during RT-PCR amplification of interviral recombinants from pmr1Δ yeast did not result in (−)recRNA detection (Fig. 1B, lanes 11 and 12), suggesting that the interviral (−)recRNAs were not generated via plasmid-based DNA recombination in pmr1Δ yeast, but rather, they were generated by the involvement of RNA substrates.

We also used a second TBSV repRNA, the highly recombinogenic DI-AU, which contains an AU-rich stretch (between region I and region II of the DI-72 repRNA) (59). The DI-AU repRNA participated in interviral recombination with FHV RNA1 in a similar manner to the above-described DI-72 repRNA, since interviral (−)RNA recombinants were detected in pmr1Δ and, to a lesser extent, in wt yeast strains (Table S2).

The location of recombination sites in TBSV repRNAs and FHV RNA1 included many different positions, yet they were not completely random. Among the 123 identified interviral recombinants between the TBSV repRNAs and FHV RNA1 obtained from pmr1Δ yeast, the hot spot region in FHV RNA1 represented position 302 and sequences around position 432. The hot spot regions in the TBSV repRNA were around positions 237 and 111 (Fig. 1C). Also, the junction sites frequently contained one or more nontemplated nucleotides, which are characteristic features of RNA recombination events with tombusviruses (11, 49). In most recRNAs, sequences around the junction sites lacked regions with high sequence similarity in the parental TBSV repRNAs and FHV RNAs. We could identify only one type of TBSV repRNA-FHV RNA1 interviral recombinant from wt yeast (Fig. 1C). This was a (−)recRNA, and the TBSV sequence represented the 5′ segment, whereas the FHV RNA1 sequence was located within the 3′ segment of the recombinant and three nontemplated nucleotides at the recombination junction, similar to the interviral recombinants obtained from the pmr1Δ strain.

To demonstrate if the interviral recombinants are associated with the TBSV replicase complex in yeast cells, we purified the Flag-tagged TBSV replicase from the detergent-solubilized membrane fraction of the pmr1Δ strain replicating both TBSV repRNA and FHV RNA1. We were able to detect (−)recRNAs formed between the TBSV repRNA and FHV RNA1 in the purified tombusvirus replicase preparation (Fig. 1D, lane 4), whereas the (−)recRNA was missing in the control samples purified from the pmr1Δ strain replicating both the TBSV repRNA and FHV RNA1, and the TBSV replicase was lacking a Flag tag (Fig. 1D, lane 3). Altogether, these data support the model that the TBSV replicase likely generates the (−)recRNAs between the TBSV repRNA and FHV RNA1 (Fig. 1D).

In contrast to the abundant (−)RNA recombinants between the TBSV repRNAs and FHV RNA1, the (+)RNA recombinants were less frequent (Fig. 1E and F). Analysis of the recombination sites revealed that unique (+)RNA recombinants (i.e., those with unique recombination sites) were identified more frequently than those (+)RNA recombinants that shared junction site sequences with the pool of (−)RNA recombinants. Therefore, it is likely that the largest pool of (−)recRNAs could not replicate, and only a fraction (25%) of interviral FHV-TBSV recombinants with the shared junction sites between (+)- and (−)recRNAs might replicate in yeast. In contrast, the unique interviral (+)RNA recombinants could be generated during plus-strand synthesis. Another interesting feature of (+)RNA recombinants is that only 43% of them contained extra (nontemplated) nucleotides at the junction sites, whereas ∼70% of (−)RNA recombinants between TBSV repRNA and FHV RNA1 contained extra nucleotides (Table S3). In addition, ∼80% of (−)RNA recombinants between TBSV DI-AU repRNA and FHV RNA1 contained extra nucleotides (Table S2). These observations indicate that interviral RNA recombination occurring during (−)RNA synthesis is likely more prone to incorporating extra (nontemplated) nucleotides at the junction sites than those events happening during plus-strand synthesis in pmr1Δ yeast. The dominant occurrence of (−)RNA recombinants and the frequent presence of nontemplated nucleotides at the junction sites in comparison with (+)RNA recombinants suggest that most of the interviral recombination events are driven by the tombusviral replicase via a template-switching mechanism during (−)RNA synthesis (Fig. 1G). It is also possible that the cellular RNase could more easily access the (+)recRNAs, which are released from the VRCs, whereas the (−)recRNAs are likely present within the VRCs in a protected environment, as we have shown for the TBSV dsRNA replication intermediates (60).

Since FHV is a two-component virus, we also tested interviral recombination between FHV RNA2-derived repRNA (named DI634) and TBSV repRNA in pmr1Δ and wt yeast strains. Similar to the picture observed with FHV RNA1, the DI634 repRNA also supported (−)recRNAs with TBSV repRNA in pmr1Δ yeast (Fig. 2A, lanes 1 to 3) and not at a detectable level in the wt strain (lanes 4 to 6). Also, replication of both TBSV repRNA and FHV repRNA was required to detect (−)recRNAs in pmr1Δ yeast (Fig. 2A, lanes 9 and 10). The recombination hot spot was around positions 240 and 340 in the TBSV repRNA (Fig. 2B), which are also recombination hot spots in intraviral recombination events (49). The most frequent recombination sites were within the 5′ 42-nucleotide (nt) segment of the FHV repRNA, although less frequent junction sites were spread across the entire FHV repRNA. Many of the recombination junctions contained one or more nontemplated nucleotides (Table S4). Sequences around the junction sites in most recRNAs lacked homologous regions in the parental TBSV and FHV repRNAs.

FIG 2.

Characterization of interviral recombinants formed between TBSV repRNA and FHV DI634 RNA in yeast. (A) RT-PCR analysis of the occurrence of (−)recRNA recombinants in wt and pmr1Δ yeast strains. TBSV repRNA, p33 and p92^pol replication proteins, and FHV RNA1 were coexpressed from plasmids, and DI634 (derived from FHV RNA2) was expressed from the yeast chromosome. Samples in lanes 7 and 8 were derived from pmr1Δ yeast coexpressing TBSV repRNA, p33 replication protein, and FHV DI634 without the tombusvirus p92 RdRp. Lanes 9 and 10 represent samples where TBSV repRNA and FHV RNA1 and DI634 were replicated separately in pmr1Δ yeast, followed by total RNA isolation and in vitro mixing of the samples in a 1:1 ratio. Next, RT-PCR was performed to detect putative artifactual recombinants generated during RT-PCR. Lanes 11 and 12 contain the same total RNA samples as the ones in lanes 1 and 2, but only PCR was used, while the RT step was omitted. (B) Schematic representation of the most frequent interviral (−)RNA recombinant containing the FHV DI634 sequence at the 3′ region and the TBSV sequence at the 5′ region, as shown. The number of isolations of different hot spot regions in the interviral recombinants is shown. The actual sequences of the recRNAs are presented in Table S4 in the supplemental material. (C) Proposed tombusvirus replicase-driven template-switching model of the formation of interviral recRNAs between TBSV repRNA and FHV DI634.

We obtained two TBSV repRNA-FHV DI634 repRNA interviral recombinants from wt yeast (Fig. 2B). These were (−)recRNAs similar to the interviral recombinants obtained from the pmr1Δ strain.

Based on the requirement of p92 RdRp expression for recombinant accumulation (Fig. 2, lanes 9 and 10), the presence of the dominant recombination sites from the TBSV repRNA, and the 5′ location of TBSV sequences in the (−)recRNAs (schematically shown for the most frequent recombinant in Fig. 2B), we propose that the TBSV replicase drove most of the interviral recombination events between TBSV and FHV RNAs during (−)RNA synthesis in pmr1Δ yeast (Fig. 2C).

Interviral RNA recombination between CIRV and FHV in pmr1Δ yeast.

We also tested recombination between the genomic RNA (gRNA) of CIRV, a tombusvirus closely related to TBSV, and FHV RNA1 in pmr1Δ and wt yeasts. RT-PCR analysis of total RNA samples from pmr1Δ yeast detected interviral (−)recRNAs formed between CIRV gRNA and FHV RNA1 (Fig. 3A, lanes 4 to 6), whereas (−)recRNA was detected to a lesser extent in wt yeast (lanes 1 to 3). Sequencing of (−)recRNAs revealed that CIRV sequences represented the 5′ segment, whereas FHV RNA1 sequences were found within the 3′ segment in the (−)recRNAs (Fig. 3B). The hot spot region was at position 1 (the 5′ end of the template) in CIRV and at position 302 in FHV RNA1 (Table S5), which was also found as a hot spot in interviral (−)recRNAs formed with the TBSV repRNA (Fig. 1). The end of the template has frequently been observed in tombusvirus intraviral recombinants (9, 12, 31, 55, 59). Altogether, these data support the model that interviral recombinants between CIRV gRNA and FHV RNA1 accumulate, with the tombusvirus replicase likely performing most of the recombination events during (−)RNA synthesis. Thus, two genomic viral RNAs (Fig. 3) behave similarly to repRNAs (Fig. 1) in supporting interviral recombination events.

FIG 3.

Characterization of interviral recombinants formed between CIRV genomic RNA and FHV RNA1 in yeast. (A) RT-PCR analysis was used to detect (−)recRNA recombinants in wt and pmr1Δ yeast strains. CIRV gRNA, the p33 and p92^pol replication proteins, and FHV RNA1 were coexpressed in yeast from plasmids, as shown. Samples in lanes 7 and 8 were derived from pmr1Δ yeast coexpressing CIRV gRNA, p33 replication protein, and FHV RNA1 without the tombusvirus RdRp. Lanes 9 and 10 represent samples where CIRV gRNA and FHV RNA1 were replicated separately in pmr1Δ yeast, followed by total RNA isolation and in vitro mixing of the samples in a 1:1 ratio. Next, RT-PCR was performed to detect putative artifactual recombinants generated during RT-PCR. Lanes 11 and 12 contain the same total RNA samples as the ones in lanes 1 and 2, but only PCR was used, while the RT step was omitted. (B) Schematic representation of the most frequent interviral (−)RNA recombinant containing the FHV RNA1 sequence at the 3′ region and the CIRV sequence at the 5′ region, as shown. The number of isolations of different hot spot regions in the interviral recombinants is shown. Note that the most frequent (−)recRNAs contained the entire CIRV gRNA sequence. The actual sequences of the recRNAs are presented in Table S5 in the supplemental material. (C) Proposed tombusvirus replicase-driven template-switching model of the formation of interviral (−)recRNAs between CIRV gRNA and FHV RNA1.

Efficient interviral RNA recombination between the plant virus TBSV and the insect virus NoV in pmr1Δ yeast.

To gain additional insights into interviral RNA recombination, we also launched simultaneous replication of TBSV repRNA with Nodamura virus (NoV) RNA1 in pmr1Δ and wt yeasts. Similar to FHV, NoV is also an alfanodavirus infecting insects. The occurrence of putative interviral recombinants was monitored with RT-PCR using selective primer combinations to amplify interviral RNA recombinants among plus and minus RNA strands (Table S1). However, interviral recombinants were detected only by RT-PCR amplifying the minus strands (Fig. 4A, lanes 1 and 2) in pmr1Δ yeast. TBSV sequences represented the 5′ segment, whereas NoV RNA1 sequences were located within the 3′ segment of the (−)recRNAs (Fig. 4B). The recombination junctions in TBSV repRNA were most frequently found around the hot spot region at positions 239 and 240, indicating that most of the interviral recombinants might be generated by the TBSV replicase (Fig. 4C), similar to the interviral recombinants with TBSV and FHV RNAs (Fig. 1 and 2). Accordingly, we did not detect (−)recRNAs when NoV RNA1 was coexpressed with the TBSV repRNA but without the tombusvirus p92 RdRp in yeast (Fig. 4A, lanes 7 and 8). Also, the frequent occurrence of nontemplated nucleotides at the junction sites was observed (Table S6).

FIG 4.

Characterization of interviral recombinants generated between TBSV repRNA and NoV RNA1 in yeast. (A) RT-PCR analysis of the occurrence of (−)recRNA recombinants in wt and pmr1Δ yeast strains. TBSV repRNA, p33 and p92^pol replication proteins, and NoV RNA1 were coexpressed in yeast from plasmids, as shown. See further details in the legend of Fig. 1B. (B) Schematic representation of the most frequent interviral (−)RNA recombinant containing the NoV RNA1 sequence at the 3′ region and the TBSV sequence at the 5′ region, as shown. The number of isolations of different hot spot regions in the interviral recombinants is shown. The actual sequences of the recRNAs are presented in Table S6 in the supplemental material. (C) Proposed tombusvirus replicase-driven template-switching model of the formation of interviral (−)recRNAs between TBSV repRNA and NoV RNA1.

Interviral RNA recombination between the yeast L-A dsRNA virus and FHV RNA1 in pmr1Δ yeast.

To test if deletion of PMR1 could lead to interviral recombination between other RNA viruses, we took advantage of the presence of the L-A dsRNA virus of yeast (61, 62) in pmr1Δ and wt yeasts expressing FHV RNA1 from a plasmid. The occurrence of interviral recombinants was detected with RT-PCR in pmr1Δ and, less frequently, in wt yeasts, among minus strands and, surprisingly, among plus and minus RNA strands too (Fig. 5A and C). Sequencing of the interviral recombinants revealed one type of interviral recRNA in which the L-A virus plus-strand sequence was fused with the FHV RNA1 plus-strand sequence that was located at the 3′ end of (+)recRNA (Fig. 5B). The abundant and diverse second type of interviral recRNAs (Fig. 5C) contained the L-A virus plus-strand sequences at the 5′ position joined to FHV RNA1 minus-strand sequences, which were present at the 3′ end of the recRNAs (Table S7). Half of the recRNAs identified contained a short stretch of shared nucleotides, which could be derived from either FHV or L-A virus. The FHV–L-A interviral recombinants might emerge through a template-switching mechanism (Fig. 5D). Altogether, these data highlight the possibility that Pmr1 might be a major host factor affecting RNA virus recombination between diverse viruses.

FIG 5.

Characterization of interviral recombinants isolated from yeast replicating L-A virus genomic dsRNA and FHV RNA1. (A) RT-PCR analysis was used to detect (+)recRNA recombinants in wt and pmr1Δ yeast strains. FHV RNA1 was expressed in yeast from a plasmid, whereas L-A virus was present in yeast strains, except in the WT* (wild-type yeast [free of L-A virus]). Samples in lanes 7 and 8 were derived from WT* yeast (free of L-A virus), which expressed FHV RNA1. Lanes 9 and 10 represent samples from pmr1Δ yeast replicating L-A virus only. Lanes 11 and 12 were samples where L-A virus and FHV RNA1 were replicated separately in pmr1Δ and WT* yeast, followed by total RNA isolation. We then mixed the samples in a 1:1 ratio, followed by RT-PCR to detect putative artifactual recombinants generated during RT-PCR. (B) Schematic representation of an interviral (+)RNA recombinant containing the FHV RNA1 sequence in the 5′ region and the L-A virus sequence in the 3′ region, as shown. (C) RT-PCR analysis was used to detect (+)/(−)recRNAs in wt and pmr1Δ yeast strains. FHV RNA1 was expressed in yeast from a plasmid, whereas L-A virus was present in yeast strains, except in the WT*. Lane 13 contains the same total RNA samples as the ones in lane 4, but only PCR was used, while the RT step was omitted. See above (A) for further details. (D) Proposed replicase-driven template-switching model of the formation of interviral recRNAs between L-A virus and FHV RNA1. We also show schematically the most frequent interviral (+)/(−)RNA recombinant containing the FHV (−)RNA1 sequence in the 3′ region and the L-A (+)RNA sequence in the 5′ region, as shown. The number of isolations of different hot spot regions in the interviral recombinants is shown. The actual sequences of the recRNAs are presented in Table S7 in the supplemental material.

Silencing of PMR1 orthologs in plants promotes interviral recombination among plant virus-associated RNAs.

To analyze the role of PMR1 orthologs in interviral recombination in plants, we knocked down the expression levels of ECA3 and LCA1 intracellular Ca²⁺/Mn²⁺ pumps via virus-induced gene silencing (VIGS) in N. benthamiana (31). Next, we coinoculated the plants with cucumber necrosis virus (CNV) gRNA (as a helper virus very closely related to TBSV) and the TBSV (+)repRNA (i.e., DI-72), as well as the distantly related TCV, as a helper virus, and the associated satellite RNA C (satC) (+)RNA (19). The emergence of putative interviral recombinants was monitored with RT-PCR and various primer combinations. We targeted the TBSV repRNA and satC RNA for recombination since they can accumulate to high levels in plants. As found previously with intravirus recombination (31), the addition of Mn²⁺ to the soil greatly increased the generation of interviral recombinants (Fig. 6A and D, lanes 1 and 2 versus lanes 3 and 4) in ECA3 and LCA1 knockdown plants. We found two types of (−)recRNAs, depending on the locations of TBSV-derived sequences: either in the 5′ segment (Fig. 6B and C) or in the 3′ segment (Fig. 6E) of (−)recRNAs. Interestingly, the most frequent recombinants contained double-recombination junctions: one between two satC sequences (intraviral recombination) and one between satC and TBSV sequences (schematically shown in Fig. 6C; see also Table S8 for the junction sequences). We did not find (+)RNA recombinants, suggesting that most interviral recombination events occur during minus-strand synthesis in plants. The locations of the TBSV-derived versus TCV satC-derived sequences in (−)recRNAs suggest that the TCV replicase and possibly also the TBSV replicase participated in interviral recombination events. However, the interviral recombination hot spots did not include the characteristic positions 239 and 240 in TBSV repRNA (Table S8), suggesting that intervirus recombination might be driven dominantly by the TCV replicase. Alternatively, the abundant satC RNA, functioning as an RNA decoy, might affect the activity of host RNases on the TBSV repRNA, resulting in new recombination hot spots. The host RNases play a role in TBSV recombination by creating incompletely degraded TBSV RNAs, which are especially active in RNA recombination (54). The observed difference in TBSV recombination hot spots between Ca²⁺/Mn²⁺ pump-deficient yeast and plants is unlikely due to these hosts, because we found similar recombination hot spots in TBSV RNAs when we studied intravirus RNA recombination with TBSV in yeast and plants (31).

FIG 6.

Isolation of interviral recombinants formed between TBSV repRNA and TCV-associated satC RNA in N. benthamiana. (A) RT-PCR analysis to detect (−)recRNAs in wt (not silenced) and ECA3- and LCA1-silenced N. benthamiana. ECA3 and LCA1 expression was knocked down (KD) via VIGS in N. benthamiana leaves, whereas these genes were not silenced in control plants (indicated as WT). On the 9th day of the VIGS treatment, leaves were agroinfiltrated to coexpress TBSV (+)repRNA, CNV (to provide p33 and p92^pol replication proteins), satC (+)RNA, and its helper TCV from plasmids. Samples in lanes 1 and 2 and in lanes 7 and 8 were derived from ECA3- and LCA1-silenced N. benthamiana grown in soil supplemented with MnCl₂ (indicated as Mn⁺⁺). Lanes 7 to 12 represent samples where TBSV repRNA and satC were replicated separately in ECA3- and LCA1-silenced or not-silenced N. benthamiana, followed by total RNA isolation and in vitro mixing of the samples in a 1:1 ratio. Next, RT-PCR was performed to detect putative artifactual recombinants generated during RT-PCR. (B and C) Schematic representation of the most frequent interviral (−)RNA recombinant containing the satC sequence at the 3′ region and the TBSV sequence at the 5′ region, as shown. The number of isolations of different hot spot regions in the interviral recombinants is shown. The actual sequences of the recRNAs are presented in Table S8 in the supplemental material. Note that in panel C, the satC RNA participated as a head-to-tail dimer in recombination with TBSV repRNA, as drawn. (D) A second RT-PCR analysis with a different primer combination from panel A to detect (−)recRNAs in wt (not silenced) and ECA3- and LCA1-silenced N. benthamiana. The same RNA samples as the ones in panel A were used. (E) Schematic representation of the most frequent interviral (−)RNA recombinant containing the TBSV sequence at the 3′ region and the satC sequence at the 5′ region, as shown. The number of isolations of different hot spot regions in the interviral recombinants is shown.

The interviral recombinants between TBSV and FHV are replication competent in a cell-free system based on reconstituted tombusvirus replicase.

We produced four characteristic TBSV repRNA-FHV RNA1 interviral recombinants as full-length (+)recRNAs (Fig. 7A and D), since the cell-free system can be programmed only with TBSV (+)RNAs and not with (−)RNAs (43). Interestingly, two recRNAs were replication competent in the presence of TBSV p33 and p92^pol replication proteins. These recRNAs produced both the dsRNA replication intermediate and the new (+)recRNA progeny (Fig. 7B, lanes 4 and 14). However, the efficiency of replication of these recRNAs was altered in comparison with the parental TBSV repRNA. We observed an ∼2-fold increase in dsRNA replication intermediate production, and ∼10- to 20-fold less (+)RNA progeny for the recRNAs than the TBSV repRNA. These differences are likely due to the presence of FHV-derived sequences at the 5′ region and the lack of TBSV-derived 5′ sequences, including the plus-strand initiation promoter, in the recRNAs (63). The other two recRNAs tested in the CFE assay showed replication at a low level (Fig. 7D and E).

FIG 7.

Testing replication of FHV-TBSV recRNAs with reconstituted TBSV replicase in yeast cell-free extracts. (A) Schematic representation of the two recRNAs used in the CFE-based assay. (B) Nondenaturing PAGE analysis of the ³²P-labeled recRNA products obtained in the CFE-based assay programmed with in vitro-transcribed (+)rec1 RNA or (+)rec2 RNA and purified recombinant MBP-p33 (each sample) and MBP-p92^pol (only those samples that are marked with “+”) replication proteins of TBSV. TBSV (+)repRNA (DI-72) was used as a control. The CFEs were prepared from the BY4741 yeast strain. Single-strand-specific RNase 1 and heat treatment to denature dsRNA were used to confirm that the recRNAs could be used as the templates to produce the single-stranded RNA (ssRNA) and dsRNA forms. Note that we adjusted the sample loading from CFEs obtained with recRNAs to facilitate the visualization of the less abundant RNA products. The heat-treated sample (lane 10) is taken as 100%. The molecular markers (M) are the in vitro transcripts of each recRNA. Each experiment was repeated three times, and the data were used to calculate standard deviations. (C) RNase H digestion experiments with various primers were used to confirm the presence of FHV- and TBSV-derived sequences in the progeny recRNA products. The hybridization positions of the primers on the recRNAs are shown schematically in panel A. (D) Schematic representation of three recRNAs used in the CFE-based assay. The nontemplated nucleotides at the recombination junctions are shown. Note that rec1 RNA was tested in panel B. (E) Denaturing PAGE analysis of the ³²P-labeled recRNA products obtained in the CFE-based assay programmed with in vitro-transcribed (+)rec1, (+)rec5, or (+)rec6 RNAs and purified recombinant MBP-p33 (each sample) and MBP-p92^pol replication proteins of TBSV. The value for rec1 replication is from panel B (in comparison with the TBSV repRNA, which is 100%). Each experiment was repeated three times.

To confirm that the FHV sequences in the recRNAs participated in replication in the cell-free system, we characterized the progeny recRNAs by conducting RNase H profiling with a set of FHV-specific primers (Fig. 7C). The specific cleavage of both recRNA progeny but not the TBSV repRNA (lacking FHV sequences) confirmed that the FHV sequences were maintained during recRNA replication in the cell-free system. Thus, the in vitro-reconstituted TBSV replicase can support the replication of selected interviral recombinants.

Interviral RNA recombinants between the plant virus TBSV and the insect virus FHV can replicate in yeast.

To further test if the interviral recombinants between TBSV and FHV could replicate in yeast, we cloned three recombinants into yeast expression plasmids, followed by launching their replication in the presence of tombusvirus replication proteins (Fig. 8A). Interestingly, the three interviral recombinants replicated in yeast lacking the Xrn1p exoribonuclease, whereas their detection was more difficult in wt yeast, likely due to their rapid degradation by Xrn1 (Fig. 8C). The coexpression of p33 and p92^pol replication proteins was needed to support the replication of recRNAs (Fig. 8B). The expression of the p33 replication protein alone (in the absence of p92 RdRp) did not provide enough protection to the recRNAs transcribed from plasmids (Fig. 8B). Northern blot analysis with TBSV- and FHV-specific probes revealed that the chimeric sequences were present in the recRNA progeny, which replicated to higher levels during 24 h of incubation in a medium that suppressed recRNA expression from plasmids (Fig. 8C). The true replication of these interviral recombinants was also confirmed by the increased accumulation of (−)recRNA in the presence of tombusvirus replicase (Fig. 8D). In addition to the replication of full-length recRNAs, we also observed the emergence of new (shorter) recombinants, suggesting that the interviral recRNAs are further evolved in yeast.

FIG 8.

Replication of selected recRNAs in the presence of tombusvirus replicase in yeast. (A) Schematic representation of the three recRNAs tested in yeast. The TBSV-FHV recRNAs were expressed as plus-strand recRNAs from plasmids in xrn1Δ yeast. (B) Replication of the recRNAs was measured by Northern blotting 0, 10, and 24 h after initiation of recRNA replication. We used either TBSV-specific or FHV-specific probes to detect the accumulation level of the (+)recRNAs. Note that the tombusvirus p33 and p92^pol replication proteins were expressed from plasmids, as shown. (C) The accumulation level of the (+)recRNAs in wt (BY4741) and xrn1Δ yeasts was measured by Northern blotting 0, 10, and 24 h after initiation of recRNA replication. Note that the tombusvirus p33 and p92^pol replication proteins were expressed in each sample. The molecular markers are the in vitro transcripts of each recRNA. See above (A) for further details. (D) Three cloned TBSV-FHV recRNAs (see above [A]) were expressed as (+)recRNAs from plasmids in xrn1Δ yeast. Accumulation of the (−)recRNA progeny was measured by Northern blotting 0, 10, and 24 h after initiation of (+)recRNA replication. We used either TBSV-specific or FHV-specific probes to detect the accumulation level of the (−)recRNAs. Note that the tombusvirus p33 and p92^pol replication proteins were expressed from plasmids. (E) The accumulation level of the (+)recRNAs in xrn1Δ yeasts was measured by Northern blotting 0, 10, and 24 h after initiation of recRNA replication. Note that the FHV RNA1 helper (to provide the FHV RdRp) was expressed in each sample. Also note the rapid emergence of a new RNA species in lanes 5 and 6, which is detected with the FHV (−)RNA2 probe and not with the TBSV (−)RNA probe. See above (B) for further details.

To test if the replication of the selected interviral recRNAs is also supported by the FHV replicase, we launched their replication in the presence of FHV RNA1, which supplies the protein A RdRp (64). In spite of the expression of the FHV protein A RdRp, we observed rapid degradation of interviral recRNAs (Fig. 8E). The only exception was rec4 RNA, which contains the entire FHV DI RNA (in addition to the 3′ segment from TBSV repRNA). rec4 efficiently generated the cognate FHV DI RNA, which was then replicated by the FHV protein A RdRp (Fig. 8E). This FHV DI RNA progeny derived from rec4 was not detected when the tombusvirus replicase was expressed in yeast (Fig. 8B). Based on these results, we suggest that the selected interviral recombinants are replicated by the tombusvirus replicase in yeast, whereas the FHV RdRp could replicate only the FHV-derived full-length DI RNA and not the interviral recombinants. Therefore, we propose that there are major differences among the viral replicases to generate and maintain interviral recombinants.

Interviral RNA recombinants between the plant virus TBSV and the insect virus FHV can replicate in plants.

One of the advantages of this recombination system is that the interviral recombinants could also be tested in a plant host (for tombusviruses). Accordingly, the expression of the three interviral FHV-TBSV recombinants from plasmids in N. benthamiana plants in the presence of TBSV replication proteins revealed the efficient accumulation of the interviral recombinants (Fig. 9A). The accumulation of the interviral recombinants reached rRNA levels when the AtRH20 DEAD box helicase, a coopted host factor for TBSV, is expressed (Fig. 9B) (65). Importantly, the expression of the functional TBSV replicase (p33 plus p92^pol) is required for the replication of the interviral recombinants (Fig. 9A). We also observed the further evolution of the interviral recombinants, albeit the newly generated interviral recombinants preserved some FHV-derived sequences based on Northern blot analysis (Fig. 9A, bottom).

FIG 9.

Replication of selected recRNAs in the presence of tombusvirus replicase in N. benthamiana. (A) The three cloned TBSV-FHV recRNAs (Fig. 8A) were expressed as plus-strand recRNAs from plasmids introduced into N. benthamiana via agroinfiltration. Replication of the recRNA was measured by Northern blotting 2 days after agroinfiltration. Note that the tombusvirus p33 and p92^pol replication proteins were expressed from plasmids, as shown. The molecular markers are the in vitro transcripts of each recRNA. See the legend of Fig. 8B for further details. The middle panel shows an ethidium bromide-stained agarose gel in which the recRNAs are visible (arrowheads). (B) Efficient replication of selected recRNAs in the presence of tombusvirus replicase in N. benthamiana. Three cloned TBSV-FHV recRNAs were coexpressed with the tombusvirus p33 and p92^pol replication proteins and the Arabidopsis RH20 host factor from plasmids introduced into N. benthamiana via agroinfiltration. Shown is an ethidium bromide-stained agarose gel in which the recRNAs are visible (arrowheads). Also note that rec4 produces a shorter abundant RNA species (marked with an asterisk). See above (A) for further details.

DISCUSSION

Viruses with RNA genomes are abundant, and variation in their sequences is astonishing. Recent viral metagenomic analyses have also revealed rampant genetic exchanges among different viruses (66, 67). RNA recombination is a major force behind these intraviral and interviral events. Based on current models, the most frequent RNA recombination events are driven by the viral replicase, which is proposed to “jump” from one site to a different site in the same RNA (intramolecular) or from one viral RNA to another viral RNA (intermolecular) during RNA synthesis (1, 11, 12, 47, 68). Next, the newly made viral recombinants are placed instantly under selection pressure to compete with the parental viruses, frequently leading to their disappearance. However, a few recombinants could become successful, contributing to the never-ending cycle of virus evolution, adaptation to various hosts, and altered viral pathogenesis.

In addition to the major roles of viral components in RNA recombination (1, 7, 49), host factors are also critical drivers of viral RNA recombination. In this paper, we utilized a previously identified intracellular Ca²⁺/Mn²⁺ pump-deficient yeast to search for interviral recombinants. We exploited pmr1Δ yeast because previous works demonstrated that the increased Mn²⁺ concentration in the cytosol affects the activity of tombusvirus replicase and enhances the recombination frequency (31). We used noninfectious viral replication systems to avoid generating unwanted infectious viral recombinants.

Interestingly, various interviral recombination events were observed between plant viruses, including tombusviruses, and a carmovirus, insect nodaviruses, and a fungal virus. Based on the characterization of interviral recombinants, the role of the intracellular environment in these recombination events can be significant in yeast and plant. Deletion of the Pmr1 Ca²⁺/Mn²⁺ pump in a yeast surrogate host or knockdown of orthologous LCA1 and ECA3 expressions in plants (when Mn²⁺ was also supplied) led to an increase in interviral recombinants with all six viruses and replicons tested here. Thus, it is likely that the activity of the cellular Ca²⁺/Mn²⁺ pump in different hosts and with different viruses could be a general factor in suppressing RNA recombination.

The occurrence and abundance of interviral recombinants in wt yeast were much lower than in the pmr1Δ strain. However, we detected interviral (−)RNA recombinants in wt yeast replicating TBSV and FHV RNAs, which also included the recombination hot spot. wt and pmr1Δ yeast strains differ in one gene, and comparable differences are expected to occur in nature within a given host species. Accordingly, there is high genetic and genomic variation among individuals within a species, as revealed by recent large-scale sequencing data. This variation within a host species could be best described as the pangenome (69–71). The tomato reference genome is missing ∼5,000 genes present in the tomato’s pangenome (72). Moreover, mutations were also observed in human gene orthologs of yeast PMR1. For example, mutations within hSPCA1 cause Hailey-Hailey disease in humans, an autosomal-dominant skin disorder (73). Loss of function of the SERCA2 pump leads to another human disease, called Darier disease, which is a rare autosomal-dominant genodermatosis (74). Because comparable mutations in the plant orthologs of yeast PMR1 likely occur in nature, we propose that those individuals within a host species could be especially active in promoting viral (intra- and intervirus) recombination. Alternatively, high Mn²⁺ concentrations in some soils may also increase the probability of viral recombination in those individual plants growing there. Altogether, genetic and environmental conditions might greatly influence the occurrence of interviral RNA recombination.

We find that the interviral recombinants are generated by a replicase-driven template-switching mechanism. The supporting observations are as follows: (i) the most abundant and diverse sets of interviral recombinants are generated as (−)RNA recombinants in all virus combinations tested here, although these viruses were expressed in host cells as (+)RNA templates (except that L-A virus is a dsRNA virus); (ii) the (−)RNA recombinants derived from FHV RNA1 and TBSV repRNA were copurified with the tombusvirus replicase from the membrane fraction, suggesting that most recombination events are driven by the tombusvirus replicase during (−)RNA synthesis; (iii) the high abundance of nontemplated nucleotides at the recombination junction sites in (−)RNA recombinants indicates putative pausing in combination with adding extra nucleotides by the replicase during recombination events; and (iv) expression of the tombusvirus p92 RdRp was required for detection of interviral recombinants derived from FHV RNAs and TBSV repRNA. The plasmid-based expression of viral (+)RNAs in the absence of tombusvirus p92 RdRp failed to lead to the detection of interviral recombinants derived from FHV RNAs and TBSV repRNA, indicating that the RNA ligation-based (nonreplicative end-joining) mechanism is not efficient in these systems.

The recombination junction sites in the TBSV repRNA were frequently located in regions which were previously shown to be intraviral recombination “hot spots” created by endoribonuclease- and exoribonuclease-based cleavages (54), suggesting that a similar mechanism exists during intra- and intervirus recombination. Sequences around the junction sites lacked sequences with high similarity in the parental TBSV and FHV RNAs in most recRNAs, suggesting that a similarity-nonessential class of recombinants (1) is generated between these viruses.

Most of the interviral recombinants between tombusviruses (TBSV and CIRV) and the insect alfanodaviruses (FHV and NoV) are likely generated by the tombusvirus replicase. This model is supported by the following observations: (i) the TBSV and CIRV sequences were located at the 5′ end in (−)recRNAs, suggesting that the replicase jumped from TBSV or CIRV (+)RNA to the FHV (+)RNA template during (−)RNA synthesis (Fig. 1C); (ii) the TBSV recombination hot spots (including the most frequent region at or around positions 239 and 240) occurred in both intraviral and interviral TBSV recRNAs (31); (iii) the (−)recRNAs were associated with the tombusvirus replicase purified from yeast (Fig. 1D); (iv) there were unique, extra, nontemplated nucleotides present at the recombination junctions, which is a characteristic feature of tombusvirus recombination events (46, 54); (v) expression of the FHV RdRp alone in the presence of FHV or NoV RNAs and TBSV repRNA (in the absence of the tombusvirus p92 RdRp protein) did not lead to the accumulation of interviral recRNAs at detectable levels in yeast; (vi) the in vitro-reconstituted tombusvirus replicase in the yeast CFE, which depends on cis-acting elements present in TBSV (+)RNA (37) for activation of the RdRp function, was able to replicate selected recRNAs (Fig. 7); and (vii) the cloned recRNAs were replication competent with the tombusvirus replicase but not with the FHV RdRp in yeast (Fig. 8). All these observations are more compatible with the model that the tombusvirus replicase is mostly responsible for driving the interviral recombination events between tombusviruses and nodaviruses.

In contrast, interviral recombination between TBSV repRNA and TCV satC is mostly generated by the TCV RdRp, based on the following observations: (i) the TCV RdRp can use TBSV cis-acting elements, such as replication enhancers, efficiently in vitro (45); (ii) the recombination hot spots in TBSV RNA generated by cellular endonuclease-driven cleavages on (+)RNA templates are infrequently present in the TBSV-TCV interviral recombinants; (iii) the TCV satC sequences participating in interviral RNA recombination with the TBSV repRNA are mostly derived from satC dimers, suggesting that the TCV replicase is highly active in RNA recombination; and (iv) the junction sites more frequently contained a 4- to 8-nucleotide stretch, which was identical in the two parental RNAs, in TBSV-TCV interviral recombinants than in those observed with TBSV intraviral (46, 54) or interviral (this work) recombinants. Altogether, it seems that the interviral recombinants between TCV satC and TBSV repRNA are rather diverse, likely due to the involvement of both TCV replicase and, to a lesser extent, TBSV replicase in the generation of interviral recombinants.

The surprising finding of interviral recombinants between a dsRNA virus of yeast and an insect (+)RNA virus (i.e., L-A virus and FHV) indicates that RNA viruses from different classes could exchange genetic materials if they coinfect the same cells. Many of the L-A virus–FHV recRNAs are unusual in a sense that the recRNAs contained plus-strand sequences from L-A and minus-strand sequences from FHV. How these recRNAs form is currently unknown since the basic recombination strategies supported by FHV or L-A virus replicases have not yet been revealed. L-A virus replicates its RNA inside the virion in the yeast cytosol (61, 62), whereas FHV replicates in vesicle-like structures in the outer membrane of mitochondria (75). If the L-A virus replicase creates these recRNAs, then the FHV RNA1 likely needs to be encapsidated into L-A virions with some frequency. The other possibility is that the FHV replicase uses the endonuclease-cleaved (+)RNA of L-A virus as a “primer” to start (−)RNA synthesis internally on FHV (+)RNA1. We were unable to detect recombinants between TBSV repRNA and yeast L-A virus in pmr1Δ or wt yeasts after 48 h of replication, suggesting that interviral recombination occurrence is different in various combinations of RNA viruses.

Interestingly, the subcellular localization of different RNA viruses might have played a limited role in interviral recombination. For example, interviral recombinants were identified between the peroxisomal TBSV and mitochondrial FHV and NoV in yeast and the TCV satC in plants. Moreover, interviral recombinants were also observed between the cytosolic L-A virus and the membrane-bound FHV (Fig. 5). Overall, the occurrence of these interviral recombinants is likely due to rare chance events, because Northern blot analysis did not detect their accumulation, likely due to low frequency and the short replication period during these experiments.

Comparison of (+)- and (−)RNA interviral recombinants seems to promote the idea that most of the interviral recombinants are not replication competent and mostly accumulate only as (−)RNA interviral recombinants. This is likely due to the fact that the interviral recombinants contain heterologous cis-acting elements from different viruses, which are likely not adapted to the cognate replicases. However, we still observed replication of the interviral recombinants, especially in plants (Fig. 9). Intriguingly, not only could the interviral recombinants accumulate, but they also are involved in further generation of new RNA recombinants. Thus, the originally formed interviral recombinants might contribute to further evolution and adaptation of the recombinants to their hosts. Much longer-lasting evolution and adaptation studies will be needed to find out if interviral recombinants could be better adapted to some hosts or environmental conditions than the original wt viruses. There are certainly ample examples in the scientific literature of the occurrence of interviral recombinants and the putative role of interviral recombination in increasing genomic sequence variability of RNA viruses.

MATERIALS AND METHODS

Recombination between TBSV DI-72 RNA and FHV RNA1 and RNA3 in yeast.

Yeast strain BY4741 or pmr1Δ was transformed with plasmids pGBK(HIS)-Cup-His33/Gal-DI-72 (31), expressing His₆-tagged p33 of cucumber necrosis virus (CNV) and the TBSV DI-72 repRNA (76); pGAD-Cup-His92 (31), expressing His₆-tagged p92^pol of CNV; and pESC(Ura)-Gal-FHV RNA1, expressing FHV RNA1. The transformed yeasts were selected on SC-ULH⁻ (synthetic complete [uracil, leucine, and histidine minus]) plates and then grown for 48 h at 23°C in selective medium supplemented with 2% galactose and 50 μM CuSO₄. Pelleted yeasts were used for total RNA extraction. Isolated RNA samples were analyzed with reverse transcription-PCR (RT-PCR) for the presence of interviral recombinants. The following primers were used for the detection of recombinant RNAs (recRNAs) formed between TBSV DI-72 and FHV RNA1: primer 3633 was used in the RT reaction, and 3633 and 719 were used in PCR (35 cycles, 52°C annealing and 68°C polymerization steps).

To detect recombinants formed between TBSV DI-72 RNA and the FHV subgenomic RNA3, the following primers were used: primer 5432 was used in the RT reaction, and 5432 and 719 were used in PCR (35 cycles, 52°C annealing and 68°C polymerization steps).

For the detection of recRNAs associated with Flag-tagged TBSV replicase, the following primers were used: primer 5331 was used in the RT reaction, whereas 719 and 5331 were used in the first PCR (30 cycles, 52°C annealing and 68°C polymerization), and 5406 and 5407 were used in the second (i.e., nested) PCR (25 cycles, 52°C annealing and 68°C polymerization steps).

To differentiate between recombination of DI-72 repRNA with RNA3 or the 3′-terminal part of RNA1, the following primers were used: primer 5571 (corresponding to the region spanning positions 2701 to 2721 to the very 3′ fragment of a solely RNA1 part of FHV RNA1) was used in the RT reaction, and 5571 and 719 were used in PCR (35 cycles, 52°C annealing and 68°C polymerization).

To detect plus-strand recombinants formed between TBSV DI-72 repRNA and FHV RNA1, the following primers were used: primer 719 was used in the RT reaction, whereas 719 and 3633 were used in the 1st PCR (30 cycles, 52°C annealing and 68°C polymerization), and 5406 and 5407 were used in the nested PCR (25 cycles, 52°C annealing and 68°C polymerization steps).

The control PCR with RNA samples directly (omitting the RT reaction) included RNA samples isolated from (i) yeasts with the pESC(Ura)-Gal-FHV RNA1 plasmid omitted, (ii) yeasts with the pGAD-Cup-His92 plasmid omitted, or (iii) RNA samples obtained by mixing the first and second samples. PCR products containing a mixture of interviral recombinants were digested with EcoRI and HindIII and inserted into pUC19, digested with the same restriction enzymes. Ligation products, each isolated from a single bacterial colony, were purified and sequenced.

The level of accumulation of rRNA and TBSV DI-72 repRNA was detected with Northern blotting as previously reported (36). For FHV RNA1 detection, the first PCR amplification was performed using pESC(Ura)-Gal-FHV RNA1 as a substrate with the 5361A and 3633 primers. The 300-nt product of the T7-based transcription reaction with [³²P]UTP was used as a probe in Northern blotting for the detection of FHV RNA1.

In vitro TBSV replication assay in the cell-free yeast extract.

Preparation of the cell-free extract (CFE) from the BY4741 yeast strain was performed as described previously (43). The CFE assays were performed as reported previously (43). Briefly, a 20-μl total volume contained 200 ng purified maltose-binding protein (MBP)-p33, 200 ng purified MBP-p92^pol, 2 μl of the CFE, and 0.15 μg unlabeled DI-72 (+)repRNA or rec1 or rec2 (+)RNA transcripts.

To probe for the presence of both FHV and TBSV sequences or fragments in the ³²P-labeled products of CFE replication, an RNase H digestion method was used, as described previously (77). Briefly, after phenol-chloroform extraction and isopropanol precipitation, ³²P-labeled CFE reaction products were dissolved in 1× STE buffer (10 mM Tris [pH 8.0], 2 mM EDTA [pH 8.0], and 50 mM NaCl). After 100 pmol of the corresponding oligonucleotide (or water) was added to each tube, samples were heated to 94°C in a PCR machine and gradually cooled to room temperature in 15 min. The RNase H digestion was carried out in a 100-μl final volume in the presence of a solution containing 20 mM Tris (pH 8.0), 50 mM NaCl, and 10 mM MgCl₂ with 1 U of RNase H at 30°C for 15 min. Each sample was then phenol-chloroform extracted and precipitated with isopropanol.

Tombusvirus replicase purification from yeast.

Yeast strain pmr1Δ was transformed with plasmids pGBK-HIS-Cup-Flag33/Gal-DI-72 (78), expressing Flag-tagged p33 of CNV and the TBSV DI-72 repRNA, and pGAD-Cup-Flag92 (79), expressing Flag-tagged CNV p92^pol and pESC(Ura)-Gal-FHV RNA1. The transformed yeasts were pregrown in SC-ULH⁻ medium supplemented with 2% glucose at 29°C. After that, yeast cells were centrifuged at 2,000 rpm for 3 min, washed with SC-ULH⁻ medium supplemented with 2% galactose, and resuspended in SC-ULH⁻ medium supplemented with 2% galactose and 50 μM CuSO₄. After growing for 24 h at 23°C, yeasts were pelleted, and the Flag-tagged tombusvirus replicase was purified according to a previously reported procedure (80). The total membrane fraction and the purified fraction, which was eluted from an anti-Flag M2-agarose resin column, was analyzed for the presence of recombinants by RT-PCR. The following primers were used for the detection of recombinants of TBSV DI-72 with FHV RNA1: primer 5331 was used in the RT reaction, 5331 and 719 were used in the 1st PCR (35 cycles, 52°C annealing and 68°C polymerization steps), and 5406 and 5407 were used in the nested PCR (25 cycles, 52°C annealing and 68°C polymerization steps). As a negative control, the same replicase purification procedures were performed with the yeast pmr1Δ strain transformed with plasmids pGBK(HIS)-Cup-His33/Gal-DI-72, pGAD-Cup-His92, and pESC(Ura)-Gal-FHV RNA1.

Replication of FHV-TBSV interviral recombinants in yeast.

To measure the level of replication of several recRNAs by the tombusvirus replicase in yeast, yeast strains BY4741 and Δxrn1 were transformed with plasmids pGBK(HIS)-Cup-His33/Gal-DI-72, expressing His₆-tagged p33 of CNV and the TBSV DI-72 repRNA; pESC(Ura)-Cup-His-p92 (81), expressing His₆-tagged p92^pol of CNV; and pESC(Leu)-Gal-Rec, expressing interviral recRNAs formed between FHV and TBSV. To measure the level of replication of several recRNAs by FHV replicase in yeast, yeast strains BY4741 and Δxrn1 were transformed with plasmids pESC(Ura)-Gal-FHV RNA1, expressing FHV RNA1; pESC(Leu)-Gal-Rec, expressing interviral recRNAs formed between FHV and TBSV; and pGBK(HIS). The transformed yeasts were selected on SC-ULH⁻ plates and pregrown for 18 h at 29°C in ULH⁻ medium supplemented with 2% galactose. After centrifugation and washing of the yeast pellet with ULH⁻ medium supplemented with 2% glucose, yeasts were grown for 24 h at 23°C in ULH⁻ medium supplemented with 2% glucose and 50 μM CuSO₄. Pelleted yeasts were used for total RNA extraction. Isolated RNA samples were analyzed with agarose gel electrophoresis followed by Northern blot hybridization as described above.

Replication of FHV-TBSV interviral recombinants in plants.

To measure the level of replication of several recRNAs supported by the CNV replicase in plants, cultures of Agrobacterium tumefaciens strain C58C1 carrying pGD-p33, pGD-p92 (82), pGD-p19 (83), pGD-RH20 (84), and pGD-Rec2, pGD-Rec3, or pGD-Rec4 were grown individually, and the mixture of these strains was then infiltrated into young leaves of N. benthamiana plants. After waiting for 2 to 3 days, plant samples were taken, and total RNA was isolated. Isolated RNA samples were analyzed with agarose gel electrophoresis followed by Northern blot hybridization.

Characterization of the progeny of FHV-TBSV interviral recombinants in Δxrn1 yeast and plants.

RNA samples isolated from yeast or plant cells which replicated rec2, rec3, or rec4 interviral recombinants were treated with 1 μl of T4 RNA ligase to transform RNA molecules into the circular form (85). This was followed by the RT reaction with a primer selective for the FHV-TBSV interviral recombinants: primers 7274, 7275, and 7276 in the case of rec4, rec2, and rec3 recombinants, respectively. DNA products obtained from RT reactions were PCR amplified separately using primers 7274, 7275, and 7276 as reverse primers for rec4, rec2, and rec3 recombinants, in combination with 1000 as the forward primer. The resulting PCR products were purified, and sequencing was performed by ACGT, Inc. (Wheeling, IL, USA).