Investigating the interactions of endornaviruses with each other and with other viruses in common bean, Phaseolus vulgaris

Endornaviruses accumulate in their hosts as non-enveloped double-stranded RNA molecules, although it is now thought that they evolved from positive-sense RNA viruses [1, 2]. The monopartite RNA genomes of endornaviruses encode a single polyprotein, which contains functional protein domains corresponding to RNA-dependent RNA polymerase and RNA helicase enzymes, as well as a variety of domains that occur in some but not all endornaviruses. Plant endornaviruses are classified, along with certain others infecting fungi and oomycetes, in the genus Alphaendornavirus [2]. Plant endornaviruses are only transmitted vertically via seed and pollen; they are not horizontally transmissible by vectors or wounding [2]. Plant endornaviruses, as well as certain other viruses exclusively transmitted through seed and pollen (partitiviruses, for example) are sometimes referred to as ‘persistent’ viruses to distinguish them from the better understood ‘acute’ viruses, i.e., infectious viruses that cause obvious disease symptoms [3].

Endornaviruses occur in certain lineages of many wild and cultivated plant species including, among others, members of the family Fabaceae (e.g., Vicia faba, Phaseolus vulgaris), species of Capsicum, and cereals including rice and barley [2, 4‐11]. In common bean (P. vulgaris), the endornaviruses Phaseolus vulgaris endornavirus (PvEV) 1, PvEV2 and PvEV3 can occur singly or in combination in various lines, but they are not present in all lineages [4, 8, 11, 12].

It has been proposed that endornaviruses may provide benefits to their hosts [13], giving an example of virus-host mutualism [14]. This may perhaps explain why these inherited viruses are so widespread and persist over many generations. The demonstration that common bean plants of the Black Turtle Soup type carrying PvEV1 and PvEV2 yielded longer pods and produced a greater mass of seeds than non-carriers [15], supports the idea that plant endornaviruses may be mutualistic. However, in broad bean (V. faba), a double-stranded RNA (later confirmed to be an endornavirus) conferred a male sterility phenotype [16], which argues against a beneficial role. A recent survey of common bean varieties, that are popular in East Africa, identified lines of plants containing PvEV1, PvEV1 and PvEV2, and PvEV1, PvEV2 plus PvEV3 [4]. However, it is not known if or how these endornaviruses modify the phenotypes of plants of these common bean lines.

Common bean is an important crop in East and Central Africa, where it is an essential source of dietary protein and carbohydrates. Common bean is also rich in iron and zinc, which are of vital importance in this region, which has a high incidence of anaemia [17‐19]. Mixed cropping systems in East and Central Africa often include common bean or other legumes as intercrops. This is because it enriches the soil with fixed nitrogen to support cultivation of other crops including, among others, maize, banana, and potato [20]. Acute viruses, including the potyviruses bean common mosaic virus (BCMV) and bean common mosaic necrosis virus (BCMNV) as well as cucumber mosaic virus (CMV), a cucumovirus, cause a range of disease symptoms including stunting, developmental abnormalities, and can decreases in crop yield [21‐24]. In contrast to the endornaviruses, which induce no obvious disease symptoms and that are not horizontally transmissible, BCMV, BCMNV and CMV are efficiently transmitted horizontally by aphid vectors and wounding, as well as vertically through seed [21, 24, 25]. In previous work we characterised endornaviruses present in a range of common bean lines that are grown as crops in east Africa [4]. In this work we investigated how these persistent viruses interact with each other, with acute viruses, if endornaviruses affect seed number and weight, and if seedborne transmission of BCMV or CMV is affected by endornaviruses.

PvEV1 and PvEV3 levels were similar in all lines harbouring these endornaviruses, with PvEV2 being variable in its accumulation. The levels of RNA accumulation for PvEV1, PvEV2 and PvEV3 do not appear to be controlled by interactions with each other. Thus, one of our starting hypotheses, i.e., that synergy or interference occurs between endornaviruses appears not to be correct. This was surprising, since the RNA sequences of these viruses have extensive regions of similarity, which we thought might trigger RNA silencing [32, 33], and the similarity of the viruses also suggested that they might compete for the same host factors. Silencing and competition can act as underlying mechanisms for protection of the host from virus infection by viral cross-protection, sometimes referred to as superinfection exclusion [34, 35]. The most likely conclusion is that in common bean, variation in PvEV2 accumulation is regulated by host factors, but it is puzzling why neither PvEV1 nor PvEV3 levels vary between lines.

Synergism occurs in plants infected by two or more dissimilar viruses, leading to increased titres for at least one of the viruses, especially if one of the viruses possesses a strong RNA silencing suppressor [33]. CMV encodes a strong viral suppressor of RNA silencing, the 2b protein [33], as do the potyviruses BCMV and BCMNV, which encode P1/HCPro proteins [24]. However, in most lines acute viruses induced no statistically significant changes in endornavirus RNA accumulation, although in a few there were decreases and, less frequently, increases. For example, in the line RWR2245 CMV and BCMV induced increases in PvEV1 and PvEV2 accumulation, whilst in BCMNV-infected plants these endornaviruses dropped to a hundredth of their normal levels. MCM2001 was the only line in which CMV induced decreases in the accumulation of all three endornaviruses. Measurement of the accumulation of the RNAs of acute viruses in the presence or absence of endornaviruses also yielded complex results, with no consistent relationship to the presence of the endornaviruses. This suggested that differences in CMV, BCMV and BCMNV titres reflected host properties. We conclude that, in common bean, little or no synergy or interference occurs between endornaviruses and acute viruses.

Plant host developmental stage profoundly influences viral infection cycles [36]. Our investigations of endornavirus-endornavirus and endornavirus-acute virus interactions focused on leaves, but it cannot be ruled out that in other tissues, or at different developmental stages, there would have been more virus-virus interactivity. In most mature plant leaf cells, endornaviruses are unlikely to be replicating actively and so may be less likely to compete with or synergise other viruses. A study of Oryza sativa endornavirus (OsEV) in rice (Oryza sativa) showed that this virus accumulated to similar levels (c. 100 copies per cell) in various tissues. However, OsEV RNA levels were far higher (c. 1000 copies per cell) in actively dividing suspension cell cultures derived from these plants [37]. It appears that endornavirus replication is tied to active cell division. Other work with OsEV indicated that endornavirus RNA is not subject to turnover by RNA-degrading mechanisms. Paradoxically, in rice plants engineered to suppress expression of Dicer-Like 2 (DCL2), an endonuclease component of the antiviral RNA silencing system, OsEV accumulation was diminished or abolished [38]. In rice, DCL2 is expressed at its highest levels in egg cells [39], when endornavirus replication is likely to be active. Thus, it seems that endornaviruses somehow exploit RNA silencing to accumulate.

All the viruses included in this investigation can be seed-transmitted. For the endornaviruses the rate is virtually 100% [2], while BCMV and BCMNV seed transmission rates are in the range 10–30% [40], and for CMV reported rates in common bean are highly variable with some studies detecting < 1% and others up to 100% [25]. Using relatively low numbers of seeds, we found no marked effects of endornaviruses on CMV or BCMV transmission. However, using larger numbers of seeds from a wider range of lines might detect subtle effects. Investigation of the effects of CMV and BCMV infection on seed yield revealed variable outcomes in different bean lines. In Wairimu Dwarf plants, which harbour no endornaviruses, neither CMV nor BCMV engendered marked decreases in seed number or seed mass but, in plants of the line RED40 which harbours PvEV1, both viruses caused decreases in seed number (but not seed mass) and CMV had a similar effect in plants of RWR2245, which harbours PvEV1 and PvEV2. The most curious results were obtained for the line RWR2075, which contains PvEV1, PvEV2, and PvEV3. We noted that plants of RWR2075 yielded comparatively few seeds, but it is possible that growth conditions in the UK are not well suited for this African variety. Interestingly, RWR2075 plants infected with CMV yielded more and heavier seeds. It would be interesting to see if such results, in which CMV appears to be beneficial, occur under this line’s normal cultivation conditions, or if this is related to the stress of growing under glasshouse conditions in the UK. Overall, our investigation found no consistent effects of endornaviruses or acute viruses on seed production.

Our data on the effects of endornaviruses on seed yield contrast with the clear-cut results obtained by Khankhum and Valverde [15]. These workers investigated how various traits differ between plants of two nearly isogenic lines belonging to the Black Turtle Soup common bean variety; one line was endornavirus-free and the other harboured both PvEV1 and PvEV2. Plants of both lines produced similar numbers of pods that contained similar numbers of seeds, however, they found that plants harbouring PvEV1 and PvEV2 produced heavier seeds [15]. The Black Turtle Soup variety comprises a wide variety of North American lineages including heirloom breeds [41]. However, careful breeding of certain Black Turtle Soup lineages has produced sets of nearly isogenic lines. These include plants with and without endornaviruses, and lines with genotypic differences, such as pairs of nearly isogenic lines with the partly dominant I locus for BCMV resistance or lacking this locus [42, 43]. We suggest two possible reasons that may explain the contrast between our findings, which suggest that the effects of endornaviruses are variable between common bean lines, and the observations with Black Turtle Soup-derived lines suggesting a beneficial effect of PvEV1 and PvEV2 [15]. Firstly, the benefit of producing larger seeds conferred on plants of Black Turtle Soup by PvEV1 and PvEV2, might be a specific response of this variety to endornaviruses, but may not be conserved in all common bean lineages. Secondly, perhaps the positive effects of endornaviruses on host traits are subtle, and undetectable unless experiments use plant genotypes that are as near identical as possible, as is the case with nearly isogenic lines.

If the second possibility is correct, it may be possible to investigate it further by generating a set of nearly isogenic lines containing various combinations of PvEV1, PvEV2 and PvEV3 for every common bean variety. However, conventional crossing and back-crossing for so many different lines would be time-consuming and impractical. In contrast, by using virus-induced gene silencing it should be feasible to obtain perfectly isogenic lines; for example, by expressing endornavirus-specific RNA silencing sequences using bean pod mottle virus (BPMV) vectors [44]. BPMV would be ideal for producing virus-free seeds since its own seed transmission rate is very low (< 0.1%) [45]. Therefore, a virus-induced gene silencing approach could generate plants and seeds ‘cured’ of endornaviruses to allow phenotypic comparisons between otherwise genetically identical plants with or without endornaviruses present.

In common bean the endornaviruses PvEV1, PvEV2 and PvEV3 appear not to inhibit or synergize each other, and only PvEV2 RNA showed distinct inter-line variation in its titre. It is likely that differences in accumulation of the acute viruses CMV, BCMV and BCMNV between lines were due to host genotype rather than to the presence of endornaviruses, and there seemed to be no direct synergy or competition between acute viruses and endornaviruses. Endornaviruses appeared to have no consistent effects on seed numbers or seed mass, or seedborne transmission of acute viruses.